Ride-hailing Platforms, Algorithmic Management, and Everyday

Resistances: A Case Study of Drivers in Lagos, Nigeria.

A thesis submitted to the University of Manchester for the Degree of Doctor of

Philosophy in the Department of Geography.

2022

Daniel Eduweye Arubayi

School of Environment, Education and Development (SEED), Department of Geography.

2 
 
Contents 
List of Tables ........................................................................................................................................... 7 
List of Figures .......................................................................................................................................... 8 
Abstract .................................................................................................................................................... 9 
Declaration .............................................................................................................................................10 
Copyright Statement ..............................................................................................................................11 
List of Acronyms ...................................................................................................................................12 
Acknowledgement .................................................................................................................................14 
Dedication ..............................................................................................................................................16 
1.  Chapter One: Introduction .........................................................................................................17 
1.1.  Introduction .......................................................................................................................... 17 
1.2.  Problem Statement and Research Context .......................................................................... 18 
1.3.  Research Questions .............................................................................................................. 19 
1.4.  Research Aim ........................................................................................................................ 19 
1.4.1.  Sub Aims ...........................................................................................................................20 
1.5.  Relevant Literature and Theoretical Underpinnings for this Research ................................ 20 
1.5.1.  Definition of Key Concepts...............................................................................................21 
1.6.  An Overview of the Research Design and Methodology ...................................................... 24 
1.7.  Significance of the Study....................................................................................................... 25 
1.8.  Thesis Structure Overview .................................................................................................... 27 
2.  Chapter  Two:  Theoretical  Approaches  for  Understanding  Gig  Work  Management  and 
Everyday Worker Resistances ............................................................................................................32 
2.1.  Introduction .......................................................................................................................... 32 
2.2.  Towards the Emergence of the Gig Economy: The Flexibilisation of Labour Based on 
Neoliberal Policies ............................................................................................................................ 33 
2.3.  From Informal Labour to Gig work: The Limits of Neoliberal Policies in the Global South.. 36 
2.4.  The Global Platformisation of Labour: Situating Gig Work in Employment Relations ......... 40 
2.4.1.  Spatial and Temporal Dynamics of Gig Work ..................................................................45 
2.5.  Algorithmic Management and Surveillance: Weapons of Control ....................................... 48 
2.5.1.  The Value of Surveillance in Society ................................................................................48 
2.6.  Global perspectives on Algorithmic Management and Gig Workers’ Agencies ................... 53 
2.6.1.  The Power of Algorithmic Management..........................................................................53 
2.6.2.  Surveillance as a Component of Algorithmic Management ............................................59 
2.6.3.  Gig Workers' Agencies:  Subverting the Power of Algorithmic Management.................62 
2.7.  Everyday Resistances of Gig Workers to Algorithmic Management: Developing a 
Conceptual Model............................................................................................................................. 65 
2.7.1.  James Scott’s Everyday Resistances: Outlining the Hidden and Public Transcripts ........65 


 
8.  The Limitation of James Scott's Concept in Lagos ................................................................ 74 
9.  Conclusion ............................................................................................................................. 75 
 Chapter Three: Research Design and Methodology ................................................................77 
1.  Introduction .......................................................................................................................... 77 
2.  Research Paradigm: Social Constructivism ........................................................................... 78 
2.1.  Ontology – Relativism ......................................................................................................78 
2.2.  Epistemology – Interpretivism .........................................................................................79 
3.  Research Design: Qualitative Study ...................................................................................... 80 
3.1.  Methodology: A Case Study Approach ............................................................................81 
3.2.  The Study Area: Lagos as a Case Site ...............................................................................82 
3.3.  Study Locations ................................................................................................................85 
3.4.  Study Sample ...................................................................................................................87 
4.  Methods and Data Collection ............................................................................................... 88 
4.1.  Desk Research: Document Review and Searches ............................................................89 
4.2.  Data Collection Process and Participant Recruitment .....................................................90 
4.3.  Criteria for Including Participants ....................................................................................90 
4.4.  Focus Group Discussion (FGD) .........................................................................................97 
4.5.  FGDs for Platform Drivers and Taxi Drivers .....................................................................97 
4.6.  Semi-Structured Interviews (SSIs) ..................................................................................100 
5.  Participant Observation ...................................................................................................... 103 
5.1.  Mobile Observations of Platform Drivers ......................................................................104 
5.2.  Online Observations of Platform Drivers .......................................................................110 
6.  Method of Data Analysis: Thematic Analysis ..................................................................... 115 
7.  Limitations of the Study ...................................................................................................... 128 
8.  Methodological Challenges in the Field ............................................................................. 130 
9.  Reflexivity and Positionality................................................................................................ 131 
10.  Ethical Considerations ........................................................................................................ 133 
11.  Health and Safety Precautions ........................................................................................... 136 
12.  Conclusion ........................................................................................................................... 136 
 Chapter Four: The Emergence of Ride-hailing Platforms: From Global North to Global 
South Cities.........................................................................................................................................138 
1.  Introduction ........................................................................................................................ 138 
2.  Comparative Factors of Taxis: Global North vs Global South ............................................. 139 
2.1.  Regulatory Environment ................................................................................................140 
2.2.  Dynamics of Trade Unionism .........................................................................................142 
2.3.  Existing Digitised Infrastructure .....................................................................................144 


 
3.  The Operation of Ride-hailing Platforms: A Brief Overview of Uber.................................. 146 
4.  The Factors of Platform Emergence in Global North Contexts:  From the United States to 
Africa… ............................................................................................................................................ 150 
4.1.  Innovative Environment: Early Experiments in the United States ................................151 
4.2.  Venture Capitalism: Ease of Access to Funding .............................................................154 
4.3.  Regulatory Strength and Permissibility: The Battles Within .........................................156 
4.4.  Hype and Rhetoric of Ride-hailing Platforms.................................................................159 
5.  Ride-hailing Platforms in the Global South: Opportunities and Challenges in Africa ........ 161 
6.  Taxi Driving as Gig Work in Lagos: From the Kabu-Kabu to Modern Taxis ........................ 169 
7.  First Generation Taxis: Challenges in Managing Taxi Work ............................................... 171 
7.1.  The Role of Unions in Controlling Taxi Labour...............................................................175 
8.  Second-generation Taxis: Attempting to Control and Measure Taxi Work ....................... 179 
9.  The Uberisation of the Taxi Industry in Lagos: Third-generation Taxis .............................. 184 
9.1.  Uber Arrives in Lagos: Introducing the Hype and Rhetoric of a Global Platform ..........184 
9.2.  Bolt (Taxify): Competition for Ride-hailing Platforms....................................................186 
9.3.  Oga-Taxi: Integrating the City Culture ...........................................................................188 
10.  Formalising the Informal: Uberising Gig Work in Lagos ..................................................... 192 
11.  Conclusion ........................................................................................................................... 196 
 Chapter  Five:  Examining  the  Impacts  of  Algorithmic  Management  in  Ride-hailing  Gig 
Work in Lagos ....................................................................................................................................198 
1.  Introduction ........................................................................................................................ 198 
2.  Algorithmic Management as a Weapon of the Dominant: Effective Control of Ride-hailing 
Gig Work in Lagos ........................................................................................................................... 200 
2.1.  Big Data and Surveillance...............................................................................................200 
2.2.  Gig Assignments and Performance Evaluation ..............................................................205 
3.  Characterising Algorithmic Impacts: A Mismatch of Data and Ride-hailing Drivers’ 
Realities… ........................................................................................................................................ 207 
3.1.  Price Mechanism and Precarious Vehicle Arrangements: Proliferators of 
Overworking….. ............................................................................................................................210 
3.2.  Data Misrepresentation: Exposing Drivers to Contextual Risks ....................................214 
3.3.  Arbitrary Discipline and Punishment: Unclear Terms of Platform Deactivations .........217 
3.4.  Motivation Paradox: Manipulating Drivers by Incentivising Ride-hailing Gig Work .....224 
3.5.  Digital Map Limitations for Driving in Lagos ..................................................................230 
4.  Algorithmic Management and the Burdens of Labour. ...................................................... 233 
5.  Conclusion ........................................................................................................................... 237 
 Chapter Six: Everyday Resistance: Exploring the Hidden and Public Practices of Platform 
Drivers in Lagos .................................................................................................................................240 


 
1.  Introduction ........................................................................................................................ 240 
2.  Public Resistance of Platform Drivers in Lagos ................................................................... 241 
2.1.  Platform Drivers’ Solidarity: Online Strikes and Offline Protests ..................................242 
2.2.  Public Dissent through Media Engagements .................................................................248 
3.  Hidden Practices of Everyday Resistance in Lagos: Challenging the Weapons of the 
Dominant ........................................................................................................................................ 251 
3.1.  Algorithmic Literacy: Making Sense of Ride-hailing Gig Work ......................................252 
3.2.  False Compliance of Platform Drivers: Gaming Spaces for Rewards.............................255 
3.3.  Sabotage: Manipulating Algorithms to Regain Control .................................................262 
4.  Conclusion ........................................................................................................................... 268 
 Chapter Seven: Conclusion and Recommendations ...............................................................272 
1.  Introduction ........................................................................................................................ 272 
2.  Research Question 1: How do Ride-hailing Platforms Emerge in a Global South Context like 
Lagos?.. ........................................................................................................................................... 272 
3.  Research Question 2: How is the algorithmic management in the ride-hailing sector of the 
gig economy impacting platform drivers in Lagos? ........................................................................ 275 
4.  Research Question 3: What are platform drivers' resistance strategies against algorithmic 
impacts in Lagos? ............................................................................................................................ 277 
5.  Research Question 4: What are the implications of ride-hailing platforms, algorithmic 
management, and drivers’ resistances for the future of labour in the gig economy? .................. 280 
6.  Methodological and Fieldwork Contributions .................................................................... 283 
7.  Recommendations .............................................................................................................. 286 
7.1.  Recommendations for Platform Companies in GS Cities ..............................................286 
7.2.  Recommendations for State Governments and Transport Authorities ........................287 
7.3.  Recommendations for Platform Unions. .......................................................................288 
8.  Future Research – Gig Work and Algorithmic Management in 30 Years ........................... 288 
References ...........................................................................................................................................290 
 Appendices..................................................................................................................................344 
1.  Appendix A: Ethical Approval for Research ........................................................................ 344 
1.1.  Appendix A1: Data Management Plan Application online ............................................344 
1.2.  Appendix A2: Ethical Approval letter .............................................................................356 
2.  Appendix B: Fieldwork Data collection Approach and Instruments .................................. 360 
2.1.  Appendix B1: Participant Information Sheets ...............................................................360 
2.2.  Appendix B2: Participant Consent Form ........................................................................366 
2.3.  Appendix B3: Detailed Ethical Considerations...............................................................368 
3.  Appendix C: Researcher's guide and Coded Participant Information ................................ 374 
3.1.  Appendix C1: Focus Group Discussion Guide for Ride-hailing Platform drivers ...........374 

8.3.2. Appendix C2: Semistructured Interview Guide for Ride-hailing Platform Drivers ........378

8.3.3. Appendix C3: Coded Thematic Data and Pseudonymised Participant Information. ....385

8.3.4. Appendix C4: Relevant Results from a Failed Survey Experiment .................................392

Word Count: 88,840

7 
 
 
 
List of Tables 
 
Table 1: Platform typologies ................................................................................................................. 41 
Table 2: Summarising Algorithmic Management ................................................................................. 56 
Table 3: Key Locations in Lagos ............................................................................................................ 86 
Table 4: Overview of Semi-Structured Interviews, Focus Groups Discussions and Recruitment 
Strategies .............................................................................................................................................. 92 
Table 5: Focus Group Participants, October and November 2018 ....................................................... 98 
Table 6: NUPEDP FGD ........................................................................................................................... 99 
Table 7: Mobile Observations of Ride-hailing Platform Drivers ......................................................... 106 
Table 8: Overview of Online Observations ......................................................................................... 113 
Table 9: Box 1 showing Socio-technical Transition and Multi-Level Perspective (MLP) .................... 115 
Table 10: Focus Group Discussion Themes and Justification ............................................................. 121 
Table 11: NVivo Themes and Justification for Platform Drivers ......................................................... 123 
Table 12: Overview of NVivo Themes for Policy and Union Representatives .................................... 124 
Table 13: Overview of NVivo Themes for Taxi Labour ....................................................................... 125 
Table 14: NVivo Themes and Justification for Platform and Venture Capitalists .............................. 126 
Table 15: Comparative Factors of Taxis .............................................................................................. 139 
Table 16: Comparing GN Factors of Platform Emergence to GS contexts ......................................... 150 
Table 17: Regions Using Uber in Africa ............................................................................................... 163 
Table 18: Table describing Informal Transport modes in Lagos ......................................................... 169 
Table 19: Summary of Requirements for Uber, Bolt, and Oga-taxi .................................................... 189 
Table 20: The Difference between Traditional Taxis and Ride-hailing Platform Gig Work ................ 194 
Table 21: Summarising the Algorithmic Burdens and Impacts in Lagos ............................................ 235 

8 
 
 
List of Figures 
 
Figure 1: Spatial and Temporal Conception of Gig Work ..................................................................... 46 
Figure 2: An Overview of Surveillance Concepts .................................................................................. 50 
Figure 3: Map of Metropolitan Lagos ................................................................................................... 84 
Figure 4: Overview of Data Collection Methods .................................................................................. 88 
Figure 5: Six Phases of Thematic Analysis .......................................................................................... 119 
Figure 6: Risk Zone Classifications in Nigeria ...................................................................................... 134 
Figure 7: A Practical Example of Surge Pricing ................................................................................... 147 
Figure 8: Nudging Drivers to Keep Driving.......................................................................................... 148 
Figure 9: Snapshot Example of Driver Ratings and Trip Metrics ........................................................ 149 
Figure 10: Uber Product Classification ............................................................................................... 153 
Figure 11: A Map Showing the Growth of Ride-hailing Platforms in Africa  ...................................... 168 
Figure 12: Picture of a Typical Kabu-Kabu Taxi ................................................................................... 172 
Figure 13: Yellow Taxi and Car-hire Parking Arrangement in Eko Hotel Taxi Stand Victoria Island, 
Lagos. .................................................................................................................................................. 175 
Figure  14: Red Cabs in Lagos.............................................................................................................. 180 
Figure  15: Metro Taxi and the Introduction of Taximeters in Lagos ................................................. 182 
Figure  16: Mapping the Emergence of Ride-hailing Platforms in Lagos Since Uber’s Arrival ........... 191 
Figure  17: A Driver Roaming the Streets of Victoria Island Lagos for Trips....................................... 198 
Figure  18: Conceptualising the Surveillant Assemblage and External Modalities of a Lagos Driver  201 
Figure  19: Display of Weekly Earnings on Uber and Hours Worked on Bolt ..................................... 204 
Figure  20: Activity Dashboard and Activity Score Page for Bolt ........................................................ 206 
Figure  21: Screenshot Example of an Arbitrary Deactivation on Bolt ............................................... 219 
Figure  22: Example of Blocked Drivers with High Cancellation Rates on Bolt and Uber .................. 221 
Figure  23: Screenshot Example of Fare Underpayment for Bolt and Uber ....................................... 222 
Figure  24: Screenshot of a Bonus Trip on Bolt .................................................................................. 225 
Figure  25: Screenshots showing Surge Trips for Uber and Bolt in Lagos .......................................... 229 
Figure  26: Traffic Congestion on the Muritala Mohammed Airport Road in Lagos .......................... 232 
Figure  27: The First Significant Protest Against Platforms in Lagos .................................................. 245 
Figure  28: A Union Leader Advocating for Drivers at the Bolt Head Office ...................................... 246 
Figure  29: Protest Placards Against Bolt’s Commission Increment ................................................... 246 
Figure  30: A Driver Soliticiting for Trip Requests from Colleagues (Before and After) ..................... 257 
Figure  31: Fake GPS Trip Data Using Lockito Verus Genuine GPS Trip Data ..................................... 263 
Figure  32: Conceptual model of algorithmic management and the everyday resistances in Lagos.. 271 

Abstract

The global financial crisis of 2007 and 2008 led to high unemployment rates,

underemployment, and job uncertainty. In the same period, the Uber ride-hailing platform

emerged to solve the inefficiencies in the taxi industry in the US. It employed workers based

on an independent contractor model characterised by flexibility and labour autonomy. This

gave rise to the platformisation or uberisation of labour, whereby traditional labour sectors

proliferated the gig or platform economy to reduce the cost for organisations. These, however,

have become increasingly precarious by preventing workers from social protection benefits

and reducing their bargaining power in challenging unfairness.

Since Global North (GN) cities were early adopters and emerging grounds for ride-

hailing platforms, scholarship on their emergence and impacts in Global South (GS) cities is

less common. Therefore, this thesis first aims to identify how ride-hailing platforms such as

Uber emerge and proliferate gig work in GS cities, using Lagos, Nigeria, as a core example. It

examines the mode of managing drivers, controlled by algorithms characterised by opacity,

information asymmetries, and biases which are burdens of labour. Using the algorithmic

management concept and James Scott's everyday resistances, this thesis shows how algorithms

impact the labour process and have facilitated public and hidden resistances from drivers.

The thesis is based on an innovative and robust qualitative methodology comprising

semi-structured interviews (SSI), focus group discussions (FGD), mobile participant and

online observations of platform drivers. Other primary data sources include driver forums,

attending driver training sessions and listening to transport radio programmes. The findings

reveal that the algorithmic burdens are underlying factors proliferating impacts in Lagos and

further facilitating drivers’ resistance practices. Compared to GN cites, Lagos's realities vary

because pre-existing contextual characteristics go beyond code and further exacerbate

burdensome labour processes. This thesis theoretically and empirically contributes to

algorithmic management as a concept based on the contextual realities that differ between GN

and South cities. Further, it contributes to the creative accounts of platform research,

particularly in GS cities. Finally, this thesis contributes to rethinking the everyday resistance

concept, particularly concerning how the public and hidden resistances change with time. More

so, how the weapons of the weak are as much a weapon of the dominant in facilitating hidden

counter resistances in Lagos.

Declaration

I, Daniel Eduweye Arubayi, declare that no portion of the work in this thesis has been

submitted in support of an application for another degree or qualification of this or any other

university or other institute of learning.

The following four notes on copyright and the ownership of intellectual property rights must

be included as written as below:

1. The author of this thesis (including any appendices and/or schedules to this thesis)

owns certain copyright or related rights in it (the “Copyright”), and he has given the

University of Manchester certain rights to use such Copyright, including for

administrative purposes.

2. Copies of this thesis, either in full or in extracts and whether in hard or electronic

copy, may be made only in accordance with the Copyright, Designs and Patents Act

1988 (as amended) and regulations issued under it or, where appropriate, in

accordance with licensing agreements which the University has from time to time.

This page must form part of any such copies made.

3. The ownership of certain Copyright, patents, designs, trademarks, and other

intellectual property (the “Intellectual Property”) and any reproductions of copyright

works in the thesis, for example, graphs and tables “(Reproductions”), which may be

described in this thesis, may not be owned by the author and may be owned by third

parties. Such Intellectual Property and Reproductions cannot and must not be made

available for use without the prior written permission of the owner (s) of the relevant

Intellectual Property and/or Reproductions.

4. Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the Copyright and any Intellectual Property and/or

Reproductions described in it may take place in the University IP Policy (see

http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=24420), in any relevant

Thesis restriction declarations deposited in the University Library, the University

Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/)

and in the University’s policy on Presentation of Theses.

List of Acronyms

AB5 Assembly Bill No.5

AES Amalgamation of E-hailing Stakeholders

APCED Association of Professional and Commercial E-hailing Drivers

CPUC California Public Utilities Commission

FIRS Federal Inland Revenue

FGD Focus Group Discussions

GCTU Ghana Cooperative Transport Union

GIS Geographic Information Systems

GPRTU Ghana Private Road Transport Union

GPS Global Positioning System

GTB Guaranty Trust Bank

GDP Gross Domestic Product

GN Global North

GS Global South

HND Higher National Diplomas

IMF International Monetary Fund

IT Information Technology

ILO International Labour Organization

LGA Local Government Area

LIRS Lagos State Internal Revenue

LAMATA Lagos Metropolitan Area Transport Authority

LASDRI Lagos Drivers Institute

LASEEDS Lagos State Economic Empowerment and Development

LASRRA Lagos State Residents Registration Agency

LSTDCOA Lagos State Taxi Drivers and Cab Operators Association

MOTPA Moving Train Pilots Association

NIPOST Nigerian Postal Service

NURTW National Union of Road Transport Workers

NUPEDP National Union of Professional E-hailing Driver Partners

NBS Nigerian Bureau of Statistics

PHV Private Hire Vehicles

PEDPA Professional E-hailing Driver-partners Association

PROTOA Progressive Transport Owners Association

RDU Rideshare Drivers United

RTEAN Road Transport Employers Association

SFMTA San Francisco Municipal Transport Agency

SAP Structural Adjustment Programmes

SATAWU South African Transport and Allied Workers Union

SECDAN Self-Employed Commercial Drivers Association of Nigeria

SMCN Social Media and Communication Networks

SSI Semi-Structured Interviews

TIN Tax Identification Number

TNC Transport Network Company

TFL Transport for London

UNILAG University of Lagos

ULT Union Leader Council

UK United Kingdom

US United States

VGC Victoria Garden City

Acknowledgement

I want to thank God for giving me the strength and ability to complete this thesis. It was not

easy, but it has been worth it. I also want to thank my parents, Professor Eric and Professor

Diana Arubayi, for the support you both have shown me throughout this journey. Inspired by

both your achievements as academics, I garnered the impetus to keep pressing on. May God

continue to bless and keep you both. To my mum, Professor Mrs Diana Arubayi, thank you for

all your prayers and your sacrifices to ensure I completed this journey. Thank God for

preserving your life; this is for you. Love you both!

I want to thank my supervisors, Dr Martin Dodge and Professor James Evans, for all your

support, guidance, constructive feedback, conference suggestions and job opportunities. I

could not have asked for better supervisors. Thank you both.

To my siblings Erica Ifeka, Eric Arubayi (late), Dr Dereck Arubayi, Mine Arubayi and Bros

Ifeanyi Ifeka. Thank you all for your emotional support, pieces of advice, and overall

unwavering support. To my Big sister Erica, thank you for always cheering me on; my visits

to yours and vice versa were critical for recharging my academic batteries. To my bro Dr

Dereck, the fact you completed your PhD three years before I began at SEED was enough

inspiration to keep going. Thanks for all your encouragement and prayers. To Bros Ifeanyi,

thank you for your encouragement and pieces of advice throughout this journey. I want to thank

Mine Arubayi for being a sister and for all the support, encouragement, and prayers. Thank

you all for blessing me with wonderful nephews and nieces. Thank you all for your support,

prayers, and contribution to my nephews and niece, Annabel, Jotham, Jayden, Aldrick, Alvin,

Arthur, and Armani. I know the future is bright for all of you. Love you all, family!

To my partner, Sylvia, thank you for your unwavering support, prayers, and inspiration to keep

pressing on in this Journey. Congratulation on getting your master’s degree. You are truly a

blessing, and I could not have asked for a better partner. Love you!

I also want to thank the rest of my family members, uncles, aunties, cousins, and friends. To

my cousins and sisters/brothers: Tosan, Huldah, Yele, Ola, Ruth, Weyinmi, Gbubemi, and

Ama, thanks for the support. To my friends: Kevwe, Toke, Sophiri, Udon, Nuella, Ini, Odele,

Foma, Zino, Nita, Michael, Edirin, Dizzy, Lansky, Uche, Saze, Donald, Darlington, Hafis,

Nurudeen, Dr Donald, Lola, MJ, and several others too numerous to mention, thank you all.

To my uncles and aunties: mummy Joyce, mummy Otsema, mummy Bridget, Aunty Ejiro,

Aunty Nkechi, Uncle Ejiro, Dr Uchenna, Bros Jide. Thank you all for your support one way or

another. To my fellow PhD colleagues past and present, Jo, Amby, Sawyer, Amish, Seun, Yong

and Chantal, thank you all for making the experience an enjoyable one. Also, thanks for

keeping me fit and making the PhD interesting to my Tuesday lunchtime football colleagues.

A big thank you to all the interviewees who agreed to take part in this study, though mostly

anonymised; thanks for contributing to this study. Finally, thanks to Mark Graham and the

entire Fairwork team for your support. Looking forward to continuing the push for a fairer gig

economy for all gig workers.

Dedication

I would like to dedicate this thesis to my late brother Eric Arubayi. Next year would make it

five years since you left us. It still feels like yesterday. But I am comforted that you are in a

better place. I still find it weird that I received my PhD admission exactly one week after you

passed, following a couple of previous rejections. That had to be you telling me to go for it,

right? Of course, it was a difficult decision to take and keep going, but I am sure you would be

proud to know I did. It has not been an easy ride, big bro, but we keep moving. Keep resting,

Makanaki; we pin this side! Love you forever.

1. Chapter One: Introduction

1.1. Introduction

The last decade saw a geometric increase in digital platforms across different sectors

such as housing (e.g., Airbnb), delivery and logistics (e.g., Deliveroo, JustEat, UberEats),

remote work or cloudwork (e.g., Upwork, Freelancer.com, Prolific), mobility (e.g., Uber, Bolt,

Ola) and several others (Möhlmann and Zalmanson, 2017; Chen and Foley, 2019; Cant, 2020).

These innovative platforms are scalable across different regions at a time and intending to

simplify labour processes using digital platforms or mobile applications (apps). The business

model for many of these platforms is such that workers operate as independent contractors,

managed by algorithms, thereby reducing platform companies' responsibility. Many of these

digital platforms are conceived in GN contexts and spread to GS contexts. For example, the

first major ride-hailing platform, Uber, emanated from San Francisco, USA, but is currently in

over 10,000 cities globally, including over 30 countries in the GS (Uber, 2020).

This thesis focuses on the labour dynamics of ride-hailing platform drivers in Lagos. It

unravels how ride-hailing platforms emerge in a GS context, the pros and cons of how

algorithms manage drivers, and how these drivers are responding to the impact of code, i.e.,

algorithmic management (Scott, 1985; 1989; Lee et al., 2015). Embedded in a social

constructivist paradigm (Guba and Lincoln, 1994; Young and Colin, 2004), this research takes

on a qualitative approach including semi-structured interviews (SSIs), focus group discussions

(FGDs), and participant observations which are divided into online and mobile observations.

Secondary data sources, such as policy documents, articles, grey literature, and online

YouTube videos, were integral for corroborating or negating data from primary data sources.

Based on the six phases of doing thematic analysis by Braun and Clark (2006), this research

also examines the pre-existing taxi regime or mobile workforce in Lagos to understand the

novelty of ride-hailing platforms, which Rayle et al. (2014) argue are more efficient and

reliable taxis because of the availability of technology.

Following this brief introduction, this chapter outlines the problem and the significance

of this topic, the research questions, objectives, theoretical context, and definition of key

concepts; a broader methodological overview; and the thesis chapter overview. The next

section begins by framing the problem of this study within the context of labour, management

and control, and resistance.

1.2. Problem Statement and Research Context

Since the introduction of the first widely known ride-hailing platform, Uber, in 2009,

the platformisation of labour has spread into different sectors, creating more precarity because

of the misclassification of workers as mostly independent contractors characterised by job

insecurity and lack of safety nets for these gig workers (Dudley et al., 2017; Healy, 2017).

More so, the labour process has become more atomised such that algorithms through big data

and surveillance infrastructures are critical in managing workers in the absence of human

managers (Andrejevic and Gates, 2014). Simply put, algorithmic management (Lee et al.,

2015), which is a core concept in this thesis, is central to the management of ride-hailing

drivers, and it is responsible for assigning trips to drivers, facilitating piece-rate payments,

monitoring and evaluating their performances, and sanctioning deviant behaviours during ride-

hailing gig work (Rosenblat and Stark, 2016; Möhlmann and Zalmanson, 2017). The

platformisation of taxi labour, in the form of ride-hailing platforms, has digitised manual

processes and optimised labour functions which minimise conventional regulatory policies and

reduce the importance of a human intermediary.

In hindsight, while algorithms can manage the labour process of several drivers in quick

succession with an increased sense of autonomy and flexibility, these are exacerbating old

challenges and introducing new ones. The process of algorithmic management is characterised

by a lack of transparency, bias and information asymmetries that are burdensome for drivers,

leading to resistance behaviours that attempt to subvert the power of algorithms. While a

growing number of scholars have started to examine the impacts of platforms across

transportation (Rosenblat and Stark, 2016; Rosenblat, 2018; Möhlmann and Zalmanson, 2017)

(e.g., Uber), housing platforms (e.g. Airbnb), delivery platforms (Veen et al., 2019; Cant,

2020); cloudwork platforms (Anwar and Graham, 2018; Wood et al., 2019) and other such

platforms, few scholars have examined the impacts of digital platforms, specifically

algorithmic management, in GS contexts (Carmody and Fortuin, 2019; Pollio, 2019; Amorim

and Moda, 2020; Kaye-Essien, 2020). This thesis fills the gap in the literature on how

algorithmic management impacts workers and how they resist the power of this hidden weapon

of control in GS contexts. It also adds to the limited account of methodologies that enable

researchers to study gig workers, especially within accounts of the black box society, i.e.,

opaque platform business models in GS contexts.

Using the city of Lagos, Nigeria, as a core GS perspective, this thesis examines the

impacts of ride-hailing platforms and drivers’ resistance strategies to algorithmic processes of

management. The overarching goal is to improve the understanding of how these platforms

first emerged in such contexts, their mode of operationalisation, and the heterogeneity between

the GN and South contexts. Considering that such innovative ideologies emerge from the GN

perspectives (Kwet, 2019; Birhane, 2020), there are bound to be contextual challenges that

limit algorithmic management in GS cities like Lagos. Therefore, Lagos, a megacity with over

21 million inhabitants (Bloch et al., 2015), the innovative, financial, and economic hub of

Nigeria (Lagosstate.gov.ng; Ramachandran, 2019), serves as a unique case site to examine the

impacts and everyday resistances against platform weapons of control, i.e., algorithmic

management. The Lagos case is also interesting in filling this gap because of the high

emergence of ride-hailing platforms even before Uber came into Nigeria in 2014 and a pre-

existing informal labour economy compounded by high unemployment rates, weak regulatory

frameworks and debilitating urban infrastructure (Olajide et al., 2018; Arubayi, 2021a).

Therefore, this chapter will provide an overview of the seven chapters of this thesis through

the lens of the research questions, aims, and objectives in the subsequent section.

1.3. Research Questions

• How do ride-hailing platforms emerge in a GS context like Lagos?

• How is the algorithmic management in the ride-hailing sector of the gig economy

impacting platform drivers in Lagos?

• What are platform drivers' resistance strategies against algorithmic impacts in

Lagos?

• What are the implications of ride-hailing platforms, algorithmic management, and

drivers’ resistances for the future of labour in the gig economy?

1.4. Research Aim

• To understand the extent to which platforms through algorithmic management

impact ride-hailing platform drivers and how these impacts lead to resistance

strategies that subvert the power of algorithms.

1.4.1. Sub Aims

• To understand how ride-hailing platforms emerge in a GS context using Lagos,

Nigeria, as an example.

• To outline and discuss the peculiar impacts surrounding algorithmic management

of ride-hailing drivers in Lagos while identifying similarities and differences

across other GS and GN contexts.

• To critically unpack platform drivers’ everyday hidden resistance and public

strategies in Lagos.

• To imagine how the consequences of algorithmic management and drivers’

resistance practices could shape the future landscape of the gig economy in the

GS, starting with the mobility services like ride-hailing platforms and more

broadly.

1.5. Relevant Literature and Theoretical Underpinnings for this Research

This research is based on theoretical approaches around emerging ideologies of work,

labour management and surveillance, and worker resistance. Observing the trends of non-

standard employment since the beginning of neoliberalism, such as part-time work, temporary

agency, contract employment, including zero contract work, has demonstrated the need for

more control of labour, more flexibility of labour and emerging resistance practices (Karlberg,

2000; De Stefano, 2014). These classifications highlight the existence of gig work which

permeated labour frontiers that existed before the platformisation or uberisation of labour.

More so, the gig economy, as it is widely known, embraces all forms of digitally intermediated

work with precarious classifications that isolate workers such as independent contractors,

partners, or entrepreneurs (Healy, 2017). With the progression of technology and innovation,

gig work is now encompassed within platforms and referred to as the platform or gig economy.

Although early experiments demonstrated using platforms to manage taxi labour in Lagos, the

idea of gig work and the use of technology to manage and control labour was perfected by the

emergence of international ride-hailing platforms such as Uber and Bolt, explored in Chapter

four.

Consequently, this thesis integrates algorithmic management as a foreground that

demonstrates how digitally mediated gig work, specifically ride-hailing platforms, ensures the

control of drivers while enabling the idea of flexibility and autonomy (Lee et al., 2015;

Rosenblat and Stark, 2016; Shapiro, 2018; Wood et al., 2019; Woodcock 2021). Considering

the novelty of the concept and the widely analogous mode of labour management in the

mobility sector, specifically taxi labour in Nigeria, algorithmic management showcases

workers’ experiences in a GS context. Arguably, the process of algorithmically managing

workers is entwined in the surveillance of workers, which demonstrates invisible and

unverifiable power, compelling workers to succumb to control as well as boost labour

productivity (Rosenblat and Stark, 2016; Mölhmann and Zalmanson 2017; Richardson 2019;

Cant, 2020). The process of rating drivers, assigning cancellation and acceptance percentages,

recording the hours worked and weekly payments are relevant dynamics that did not exist in

Lagos compared to GN cities. However, the concept highlights more cons than pros in ride-

hailing gig work in Lagos. Therefore, it is critical to show how workers succumb to control

from platforms in Lagos using algorithmic management as an invisible form of power.

However, while algorithms demonstrate invisibility, they are not unnoticed. Hence, this thesis

also adopts James Scott’s everyday resistance concept to pinpoint the everyday resistances of

drivers, which comprise hidden and public practices critical in subverting algorithmic control

(Scott, 1985; 1989; Möhlmann and Zalmanson, 2017).

This forms a conceptual model that identifies how and when drivers resist algorithmic

management. The power of algorithms inherently acknowledges the prevalence of worker

agencies (Mölhmann and Zalmanson, 2017). The everyday resistance concept progresses the

understanding of worker agencies by distinguishing between the hidden and public driver

practices, emphasising the spatiality and temporality that exist between them. This thesis

develops the idea of gig work through mobile workplaces, such as ride-hailing platforms, to

show how new frontiers of management and control, i.e., algorithmic management, lead to

everyday resistance of workers in a GS context using Lagos, Nigeria, as a core case setting.

The section below defines key concepts highlighted in this section, also used throughout this

thesis.

1.5.1. Definition of Key Concepts

This section defines the recurrent key concepts used throughout this thesis.

Algorithmic Management: As part of contributions to Human-Computer Interaction (HCI)

studies, Min Kyung Lee and colleagues coined the concept of algorithmic management in

2015. The concept attempts to understand how human workers respond to algorithms that have

replaced human managers (Lee et al., 2015), considering these have become central to

allocating, optimising and evaluating work in both platformised settings like Uber, Bolt,

Amazon Mechanical Turk and traditional work settings such as for engineers, warehouse

workers and other such roles. Mateescu and Nguyen (2019, p. 1) succinctly define “algorithmic

management as a diverse set of technological tools and techniques to remotely manage

workforces, relying on data collection and surveillance of workers to enable automated or semi-

automated decision-making”.

Gig work and the gig economy: Gig work is a hybrid model of contingent work or non-

standard employment that embodies aspects of zero-hour contracts with unguaranteed hours;

and temporary agency work characterised by third-party intermediaries (O’Sullivan et al.,

2015; Duggan et al., 2019). Before the invention of platforms, gig workers were identified as

dog walkers, day labourers, janitors, personal care attendants, coffee shop attendants, and other

classifications involving temporary, often low-waged, and uncertain terms of contingent work

(Friedman, 2014). Flexible hours, project-based work, piece-rate payments, and often short-

term relationships are all comparable traits across contingent work, non-standard work or other

kinds of temporary arrangements that are embedded within the confines of gig work (Duggan

et al., 2019). The gig work model utilises independent contractor status or short-term contracts,

which involve flexibility and autonomy as a licence for low wages, lack of safety benefits and

job insecurity while retaining employer-worker property relations extended through constant

surveillance from digital platforms (Muntaner, 2018). The growth of gig work, especially since

the platformisation of labour, is widely referred to as the gig economy. This is often used

interchangeably with the platform economy or sharing economy. However, this thesis widely

uses the term gig economy because it highlights gig work that preceded the introduction of

platforms.

Ride-hailing platforms: In the literature, these have been referred to as ride-sourcing

platforms, ridesharing, Transportation Network Companies (TNCs), and ride-hailing

platforms. Examples include Uber, Bolt, Lyft, Didi Chuxing and other such platforms globally.

This thesis chooses the term ride-hailing platforms because of the action of hailing or flagging

traditional taxis using one’s hand. With these platforms, hailing or flagging down vehicles is

done by pushing a button or touching the screen on a ride-hailing app. According to Rayle et

al. (2014), because drivers also offer rides in exchange for a fare, it is closely related to

traditional taxis, although more efficient and reliable because of real-time matching technology

and algorithmic management. Ride-hailing platforms are clearly defined by Rayle et al. (2014).

Accordingly, riders request rides via a smartphone app to often non-commercially licensed

drivers; the algorithm can facilitate real-time matching using global positioning system (GPS)

technology to the nearest available rider or driver. The app calculates the fare for the trip, which

can be variable or fixed, and processes a certain percentage for the platform while the driver

receives the remaining fare. A core characteristic of platforms is the rating system, where the

algorithm enables riders to rate drivers over a scale of 5 and vice versa for drivers to riders

(Rosenblat and Stark, 2016). Because of the evolving nature of gig work, the characteristics of

ride-hailing platforms remain volatile.

Surveillance: Drawing from the architectural prison designed by Jeremy Bentham, Michel

Foucault developed the notion of panopticism, and since then, ideologies of surveillance, i.e.,

being watched, have emerged over the years. Surveillance as a concept is embedded in

algorithmic management, which facilitates subtle control or invisible power through digital

platforms. Drawing from the literature, Galic et al. (2017) examine three phases of surveillance

concepts from architectural theories and discipline of society, e.g., panopticism to digital

infrastructural theories (e.g., control of society by Deleuze, 1992) and new conceptualisations

from below, which also identify ideologies of surveillance from below such as sousveillance

(Mann, 2004). These were relevant in understanding the history of surveillance in the

workplace which showed the centrality of devices, practices and norms that ensure control of

workers, for example, call centre operators who are constantly monitored on the job (Bain,

2002). Considering its theoretical roots, the notion of surveillance is simply an act of being

watched, monitored, or tracked knowingly or unknowingly. Particularly, this thesis draws on

Haggerty’s surveillant assemblage, which is critical in how algorithmic management utilises

different layers in its design to manage the labour process of workers. Critically, it was also

critical in how platforms initiate counter-resistances, as the findings show. With ride-hailing

platforms, drivers are constantly monitored, tracked, and rated through algorithms, which

makes surveillance integral in examining the algorithmic management of ride-hailing drivers.

Everyday Resistances: As mentioned above, this identifies the unnoticed or hidden practices

of subordinate groups, according to James Scott (1985, 1989). With the hidden managerial

style of ride-hailing platforms, the need to distinguish between hidden practices and more overt

practices has become pertinent in examining gig workers in Lagos. The ‘everyday’, which

highlights time or temporality, is critical in examining when and how drivers resist unfair

practices in Lagos. Most importantly, it highlights how researchers might consider the notion

of everyday resistances in examining labour within digital platforms as gig work evolves.

1.6. An Overview of the Research Design and Methodology

This research responds to the challenges of studying existing and emerging digital

platforms in the GS by implementing creative, innovative, and adaptable methodological

approaches. Considering that research on the plaformisation of labour in African cities is

lacking, the methodologies enclosed within this research contribute to approaches that guide

researchers in studying gig work in these contexts.

The novelty of research in platform studies and gig work creates the need for a social

constructivist research paradigm, which is critical in understanding and interpreting these new

concepts through interactions of ride-hailing drivers in Lagos. Their involvement in ride-

hailing gig work and core understanding of the city facilitates knowledge creation which is

mutually beneficial to the researcher, who is also a resident of the city. Although it is not as

representative as a positivist epistemological stance (Guba and Lincoln, 1994), it critically

produces an in-depth understanding because of the multiple realities of drivers. My position as

a student and researcher in a GN context and as a native of Lagos also facilitated mutual

knowledge creation on the research subject. This thesis adopts a collective case study approach

with Uber and Bolt as core cases (Yin, 2003). The initial plan was to examine one international

platform and one domestic platform, but the reality in the field facilitated core choices between

two international ride-hailing platforms based on more drivers on these platforms. While these

two platforms (Uber and Bolt) are consistent throughout driver stories, the idea behind adopting

a collective or multiple case study approach enabled this thesis also to analyse the emergence

of domestic platforms like Oga-taxi, conventional taxis, and defunct innovative platforms

preceding Uber in Lagos. The multiplicity of this phenomenon identifies a background of

intentional, innovative ideologies to platformise taxi driving in Lagos before the emergence of

Uber. These case studies were only focused on a single case setting, Lagos, due to time,

logistics and political (e.g., election) constraints. In addition, the city of Lagos is the innovative

hub of Nigeria; the emergence of most ride-hailing platforms from international communities

first appeared in the city before progressing to the capital Abuja and other cities in Nigeria.

The sources of data in this thesis are both primary and secondary. In terms of primary

sources, qualitative data collection methods that ensure validity and reliability by triangulating

across different techniques were utilised, for example, the use of FGDs, SSIs, and participant

observations through online and mobile observations (Sheller and Urry, 2006; Hine, 2008;

Caliandro, 2018). A total of 70 participants were involved across SSIs and FGDs. The

observation of participants included taking ride-hailing or taxi trips, monitoring drivers’ posts

on social media groups and attending drivers’ training sessions. Considering ride-hailing

platforms as a mobile gig workplace, developing mobile observations contributes to the

development of mobile methodologies which help capture the evasive nature of these platforms

(Sheller and Urry, 2006; Hein et al., 2008). Participants in this study included ride-hailing

platform drivers, a platform company, taxi drivers, passengers, transport policy representatives,

a start-up venture capitalist and two lecturers who were key informants in the field. Secondary

data sources comprised a review of the Lagos Metropolitan Area Transport Authority

(LAMATA) policies and other policy and scarce research articles in Lagos. Alternative

secondary data sources included YouTube interviews and online news articles such as The

Guardian and Vanguard news sources. The information generated from these resources was

analysed using NVivo by categorising data according to participant groups based on themes

generated from the field and themes across previous research on ride-hailing platforms,

algorithmic management, and resistances. Because of the novelty of the research at the time, it

was difficult to model data collection along with preconceived themes. However, it was critical

to establish knowledge throughout the empirical chapters from the thematic analysis results.

1.7. Significance of the Study

The recent innovative platform trends after the great recession in 2008 have furthered

the ideals of having an increased labour force with lean business models with more control in

the labour process (Srnicek, 2017). The growth of platforms has permeated different aspects of

mainstream businesses, services, and gradually traditional labour environments, especially

since the Covid-19 pandemic. The pandemic has demonstrated that work can be done online,

and labour is an increasingly mobile concept, which further boosts flexibility and autonomy

(Rani and Dahir, 2020). It has also indicated, if anything, the need for less precarious labour

classifications like independent contractors. Ride-hailing platforms are a core facet of the gig

economy and are integral to how these growing ideals are perceived in technology and labour

contexts. However, there is an indication that gig work platforms are creating a geometric race

to the bottom because of how precarious and burdensome the labour process and conditions

are to workers.

This is also disproportionately felt across GN and GS dynamics, based on a stark

variation between regulatory and labour laws in these contexts. Also, the knowledge gap and

digital divide across regions (Kwet, 2019) highlight why most innovative ideologies first

emerge in GN countries before proceeding to the South. Particularly, this thesis outlines and

examines four factors in understanding the political economy of how platforms first emerge in

GN contexts and their motivation for emerging in GS contexts. These include the innovative

environment, venture capitalism based on ease of access to capital, regulatory battles and

permissibility, and hype and rhetoric of platforms. It, therefore, indicates that contexts like

Lagos rely on the GN cities to establish new platforms and frameworks that enable the overall

fairness of workers in the gig economy; if not, they are left behind. The significance here is

that this research facilitates an understanding of how these platforms work in cities like Lagos

compared to developed cities and why indigenous companies should develop fair and less

burdensome labour for ride-hailing drivers.

Accordingly, this research is impactful in a few ways. First, it develops workers’

narratives on their perceptions and experiences of ride-hailing gig work, which are lacking, as

previously mentioned; all of these are comparable and can be scalable to other GS cities. This

reflects the research paradigm based on social constructivism with multiple realities of drivers,

which are unique to Lagos State, Nigeria (Guba and Lincoln, 1994; Creswell et al., 2007).

Secondly, considering the non-disclosure of information to third parties and the opaque

management on Uber and Bolt, the information generated from this research creates some

awareness of what occurs within the app and how algorithms, through invisible or hidden

power, are dictating the labour process even from GN cities. It also shows why using algorithms

as a managerial tool in controlling workers does not immediately solve the problems of

inefficiency, conflicts, transparency, corruption, and other such challenges in the Lagos

context. Instead, it creates the realisation that algorithmic management as a lone tool of control

still has a long way to go in terms of incorporating multiple complexities. These complexities

include the unpredictability of human behaviour, in this case, Lagosians (drivers and riders);

the nuances of the city of Lagos, such as unexpected traffic congestion, road blockages; and

the inequity in labour standards that enable continuous penetration by ride-hailing platforms

without poor rebuttals from the state. Thirdly, it shows that researching platforms like Uber

require creative methodologies and triangulation of information that is transferable to

qualitative researchers. For instance, the hidden nature of ride-hailing gig work extends to the

processes of searching for drivers because of their evasive nature, considering their highly

mobile working environment. This realisation further facilitates the researcher's consciousness

as to how participants are sourced without the possibility of being monitored by these

platforms.

The information presented in this research is relevant to researchers, policymakers,

innovators or start-up companies, riders, and drivers. Researchers in the GS can build on

existing knowledge, spot loopholes, and develop new ideologies that further embolden the

existence of ride-hailing platforms and why they should be taken more seriously. This thesis

enables policymakers in Lagos, such as representatives from the Ministry of Transport and the

Lagos Metropolitan Area Transport Authority (LAMATA), to examine, re-evaluate and create

policies that enable fairness for ride-hailing platform drivers. Similarly, it provides a

background for start-up and innovative companies to develop platforms that treat workers like

human beings. New entrants need to realise that the resistance against algorithmic management

is due to the hidden nature of ride-hailing gig work. Thus, creating a transparent and less

precarious platform for workers is critical. For drivers, this helps enable an external

information symmetry by reading the multiple realities of their work and some of the strategies

of Uber and Bolt in Lagos and other cities globally. Finally, it enables riders to become aware

and ensure that they are accountable for their actions, considering that Uber and Bolt

systematically possess rider biases in the city context.

1.8. Thesis Structure Overview

This is an overview of all the chapters written in this thesis. It guides the reader in terms

of what to expect in each chapter before the appendices and bibliography.

Chapter 1 outlines the thesis' problem with articulated research questions, aims, research gap

and problem statement. It further highlights the relevant literature and theoretical

underpinnings, definition of key concepts, a summary of the methodological approach and the

significance of the study.

Chapter 2 reviews historical and theoretical approaches that precede the emergence of the gig

economy. This chapter argues that neoliberal ideologies that facilitated free market rule,

deregulation and privatisation of labour markets exacerbated the individualisation of labour

with growing non-standard employment classifications reflective of gig work today. In other

words, the gig economy began long before the innovation of digital platforms – digital

platforms reflect the growing rate of non-standard employment classification that incorporates

gig work within software applications in both online (e.g., Upwork) and location-based (e.g.,

Uber) contexts. It further highlights how surveillance concepts have always been critical for

controlling society and an evolving atomised labour force even since Taylor’s notion of

scientific management. Since the introduction of platforms, algorithmic management has been

reinforced by a surveillant assemblage by tracking workers, assigning work, rewarding

performance, and sanctioning bad labour practices. The chapter reviews the pros and cons of

algorithmic management in gig work globally, identifying that platforms are further

proliferating ideals of neoliberalism through gig work, reinforced by algorithmic management

with a false notion of flexibility and workers' autonomy. This inherently leads to gig workers

devising ways that aim to subvert the power of algorithmic management. It concludes by

introducing James Scott’s concept of everyday resistance, which emphasises the hidden or

unnoticed practices of peasant farmers in Southeast Asia, enabling the development of a

conceptual model for untangling both hidden and public practices of gig workers against the

power of algorithmic management. Thus, there is an emphasis on the recognition and intention

of these practices, spatial and temporal dynamics which reflect the when and how these

practices materialise, and gradual mobilisation processes of resistances, which should be

context specific.

Chapter 3 outlines the methodological approach used in carrying out this research. This

research is a qualitative study in Lagos, Nigeria, consisting of secondary and primary data

collection methods from September to January 2018 and from July to September 2019 –

approximately five months. The secondary data collection techniques included extensive

document analyses from grey literature, policy documents, company press releases and social

media, including YouTube videos. The primary data collection techniques in Lagos were

fivefold. They included monitoring online driver forums; conducting four FGDs; participant

observation of drivers including through forty Uber and Bolt trips, attending two driver training

sessions, and listening to transport radio programmes; and SSIs comprising 25 platform drivers

(Uber and Bolt), 11 local taxi drivers and five riders. In total, 70 participants were involved in

this study across SSI and FGDs. These separate techniques facilitate the triangulation of

information and further refine mobile methods for researching the gig economy. Nvivo

software was utilised to analyse field data used in the empirical chapters four, five and six.

Chapter 4 is the first empirical chapter that attempts to answer the research question on how

ride-hailing platforms emerge in GS cities, using Lagos as an example. But first, the chapter

examines the factors differentiating the operations of the taxi industry between GN and GS

contexts to understand why informality in GS contexts is prevalent. Following this, it further

outlines four factors that define the emergence of Uber in GN contexts and how these factors

comparatively serve as motivating factors for Uber’s emergence in GS contexts. This sets the

tone for critically examining the nuances of platform emergence in Lagos and how that has

changed the taxi industry, considering the integration of algorithms in managing the labour

process. To understand how algorithmic management of ride-hailing gig work operates, it was

also critical to examine facets of analogous or non-digital management within the mobility

service in Lagos, emphasising the labour of conventional taxi drivers. Compared to GN cities,

with clear progress in innovation and strong regulatory policies in the mobility sector, the

reverse is the case in Lagos. Unions, motor park factions, corrupt union leaders and touts

(agberos) were critical in managing and controlling taxi labour. Although aspects of taximeters

were tested in Lagos, it was short-lived, and ride-hailing platforms were the first to atomise

and algorithmically manage drivers in Nigeria. This phenomenon is plausible in the Nigerian

context because it attempts to improve unemployment rates by enabling drivers to be

entrepreneurs or partners as an underlying factor for flexibility and labour autonomy. It also

gives more legitimacy to driving both as a skill and profession by assigning digital identities

for drivers with data and algorithms that facilitate a seamless digital experience with

possibilities of reducing the challenges in conventional taxi labour. Despite the pros of ride-

hailing platforms, the chapter finds that ride-hailing platforms are formalising the informal

processes of taxi labour by exploiting the existing regulatory loopholes in Lagos, which are

exacerbating unfair impacts on drivers. The chapter concludes by acknowledging that the

formalisation of gig work through platforms has led to a shift in conventional unions to

developing platform unions and groups, which become critical in drivers’ resistance examined

in Chapter six. However, it establishes a transition into Chapter five by first analysing how

algorithmic management impacts ride-hailing drivers in Lagos.

Chapter 5 is the second empirical chapter, and it analyses how algorithmic management

impacts ride-hailing gig work in Lagos. This chapter re-introduces the algorithmic management

concept by analysing its operationalisation in Lagos. Algorithms control drivers using big data

and embedded surveillance capabilities such as rider or passenger ratings that enable consistent

performance evaluation. While flexibility and autonomy are marketing precepts for platforms,

the hidden nature of algorithms and further outsourcing of power to external components create

labour exploitation. The hidden nature of algorithmic management creates what this chapter

identifies as algorithmic burdens, which are a conflation of information asymmetries, rider

biases and opacity, which transfer the risks of difficult decisions to drivers. There is a mismatch

in driver and algorithm realities, which does not incorporate the everyday experiences of

driving in the city of Lagos. Instead, drivers are entrapped in precarious arrangements, which

manipulate and incentivise them to keep driving while impacting their performance metrics

such as ratings, cancellation, and acceptance rates. The consequences of these burdensome

impacts indicate that drivers can be arbitrarily deactivated from Uber or Bolt platforms without

clear evidence and opaque or a non-existent appeal process which highlights autonomy as a

paradox and flexibility as a myth in Lagos. If algorithms are to manage ride-hailing platform

drivers, then platforms must be intentional about context specificity relating to their

experiences and how the characteristics of every city can affect gig work.

Chapter 6 is the final empirical chapter that fully operationalises the conceptual model by

introducing James Scott’s concept of everyday resistance to show the hidden and public

practices of ride-hailing drivers against algorithms in Lagos. With an emphasis on hidden

everyday resistances, this chapter first discusses public forms of resistance: online strike action

by drivers, offline protests, and public dissent through media engagements. As a follow-up

from Chapter five, the formation of platform unions and the joining of social media and

communication networks (SMCN) is central for both hidden and public resistance by enabling

drivers to make sense of opaque algorithm decisions. This further facilitates drivers to

manipulate, game and circumvent algorithmic management rules to regain power and control

the terms and conditions of labour. Compared to GN cities, it shows that public resistances

such as strikes and protests are ineffective in enabling substantial victories for drivers. The

hidden forms of resistance become critical for drivers to express how ride-hailing gig work is

unfair in the Lagos context. Meanwhile, this chapter initially delineates between what can be

classified as hidden or public resistance. It finds that because of the spatiality and temporality

of platforms, a hidden practice today may become public and vice versa because of how quickly

algorithms learn these resistance practices and how drivers replace exposed hidden practices

with new ones. It also recognises platforms as actors of reverse hidden resistances by

algorithmic managerial counteraction.

Chapter 7 concludes by discussing the core contributions of this thesis to knowledge in terms

of improving theoretical and methodological lenses. The hidden nature of these new forms of

scientific management requires new or improved methodologies that fully capture multiple

data sources. If platform gig work should become mainstream, then algorithmic management

would penetrate different work sectors with less control for workers in Nigeria and beyond.

2. Chapter Two: Theoretical Approaches for Understanding Gig

Work Management and Everyday Worker Resistances

2.1. Introduction

Gig work has always existed in freelance work and short-term contracts, even before

the advent of the internet. With technological advancements and the ubiquity of the internet,

gig work has become scalable to online environments and, more recently, platforms such as

Upwork and Uber (Healy, 2017). The rapid growth in online work and services has led to

different classifications such as platform capitalism (cf. Srnicek, 2017; Schimdt, 2015); the

sharing economy (cf. Schor, 2014; Sundarajan, 2016); the gig economy (cf. Graham et al.,

2017; Wood et al., 2018). This research adopts the notion of gig work because evidence from

the literature review indicates that digital labour progresses beyond and incorporates notions

of non-standard employment and informal work, which are prevalent globally, especially in

GS contexts. As we know it, the gig economy is based on the platformisation of labour,

consolidating old labour traditions into online and location-based services through platforms.

The development of digital networks has incorporated the power of software algorithms

which, in turn, are facilitating more efficient management of workforces, most of which are

built on Taylorist principles in scientific management but deviate because of the capacity to

manage and analyse big data in recent years (Andrejevic and Gates, 2014; Zuboff, 2015;

Shildlt, 2017). According to Healy (2017), workers must use apps developed by platform

companies – making gig work a structured and mobile capital across global borders. Despite

the difficulty in measuring the gig economy's growth, Heeks (2017) estimates about 70 million

people to be the total figure of digitally engaged gig workers globally. The growth of this type

of work has intensified academic research and public policy concern about the future of work

to critically understand its implications on the workforce, workers' agency, and other issues

around its advantages and disadvantages for workers and companies. This chapter comprises

two core arguments centred around how platform companies deploy algorithms to manage the

labour process of gig workers and how gig workers regain some control by resisting platform

power.

Lee et al. (2015), who coined the term algorithmic management, define it as software

algorithms that assume managerial functions and mediate work due to data and code. Before

discussing algorithmic management and gig work, this chapter briefly analyses how the advent

of the gig economy builds on a foundation of neoliberalism which, unlike GN contexts, has

proliferated informal labour in GS contexts. By so doing, the gig economy and algorithmic

management is another neoliberal project emerging from the GN into the GS context (Kaye-

Essien, 2020), with promises of improving workers' lives through flexible and autonomous

employment. Against the backdrop, I outline a brief history of surveillance theories drawing

from Galic (2017) to understand the hidden nature of control embedded in our everyday lives

and within technology. The thesis then draws specifically on the surveillant assemblage by

Haggerty and Ericson (2000), based on different components that reinforce algorithms in the

control and management of gig workers.

The same networks highlight control and facilitate agency in workers to make sense of

algorithms that enable them to circumvent, manipulate and game algorithmic assignments to

regain power and autonomy (Jarrahi and Sutherland, 2018; Cheng and Foley, 2019; Veen et

al., 2020). This chapter introduces James Scott’s (1985; 1989) notion of everyday resistances

as a format to categorise gig worker agencies under constructs of hidden and public transcripts

because workers intend to initiate certain practices in secret and others in public (e.g., protests).

However, the main contribution here is that delineations between the hidden and public are

blurred because of a high spatial and temporal dimension of gig work that allows the hidden to

become public and vice versa. If workers are always being watched, how are some practices

unnoticed until they become rampant? The temporality embedded in everyday resistances sets

the pace for examining different timescales in which gig workers initiate their power of

resistance. More importantly, this sets the tone for understanding how power struggles between

algorithms and gig workers in the GN may differ from workers' realities in the GS in the

subsequent chapter.

2.2. Towards the Emergence of the Gig Economy: The Flexibilisation of

Labour Based on Neoliberal Policies

The ideology behind flexibility and freedom in the workplace was shaped by neoliberal

policies perfected today in gig work. According to Harvey (2005, p. 2), “Neoliberalism is a

theory of political and economic practices that proposes that human well-being can be best

advanced by liberating individual entrepreneurial freedoms and skills within an institutional

framework characterised by strong private property rights, free markets, and free trade".

Conceived as an abstract post-welfare political-economic governance agenda between Hayek

and Friedman in the 1970s, neoliberalism became a political tool following Margaret

Thatcher’s and Ronald Reagan's elections in the 1980s, "championing the idea of free-market

in opposition to the rule of the state, big government, and trade unions" (Phelan and Dawes,

2018, p.11). In a nutshell, neoliberalism was a reaction to the welfare state and ideas that held

sway in the 1960s and 1970s that the state should be heavily involved in regulating national

economies. This phenomenon became indicative of a combination of the state's withdrawal

from areas of provision, deregulation, and privatisation (Wacquant 2012). Moving towards

fostering accumulation by dispossession, the state adopted a business-friendly capitalist model

to repress popular resistance and safeguard financial institutions overall (Wacquant 2012). In

a broader but simplified manner, Wacquant (2012, p. 68) outlines neoliberalism as “a

hegemonic economic conception by variants of market rule” and a Foucauldian notion of

governmentality.

Notions of the state, market, and citizenship are articulated through the interrelationship

between hegemonic market rule and governmentality (Wacquant 2012; Kaye-Essien 2020).

The former is a direct imposition of a flexible yet optimal market that utilises the neoclassic

economics school of thought, where supply and demand are responsible for the consumption,

production and pricing of goods and services. The latter focuses on how different regimes blend

with variable and malleable political rationality – calculated means by which the government

shapes human conduct such that it is irresistible, even by firms outside it (Li, 2007; Wacquant,

2012). This notion draws from the state's objective governance, which inspires social and

innovative policies that effectively restructure its welfare, improving its conditions and

increasing health, wealth, and longevity overall (Foucault 1995).

In drawing from the industrial relations and neoliberalism literature, a neoliberal state

must dismantle landscapes that encourage collective regimes and labour regulation, supporting

union and employment conditions known as decollectivisation while integrating

individualisation or policies that facilitate unilateral regulations (Bray and Underhill, 2009).

According to Larner (2000, p. 13), "neoliberal strategies of rule, found in diverse realms

including workplaces, educational institutions and health and welfare agencies, encourage

people themselves as individualised and active subjects responsible for enhancing their

wellbeing". However, Bray and Underhill (2009) perceive collective arrangements'

dismantling and promoting individualised arrangements as contradictory for two reasons.

Firstly, state laws, policies and support determine and protect individualised contracting,

wages, and overall conditions of an individualised arrangement. Secondly, it allows the

survival of collective institutions like unions through collective bargaining while also

possessing the power to destroy such institutions by restricting union resistance and

collectivism. These new developments also weakened labour movements based on the decrease

in real wages, firms increasing flexibility and restructuring to recover profitability, and the

collapse of the Bretton Woods fixed exchange rate system (Taus, 2013; Vidal, 2013). The

predominant logic of externalised or outsourced employment relations occurred even in core

sectors that were typically more protected from internal wage competition, such as jobs in

hospitality, sales, leisure, and retail sales (Vidal, 2013). For example, the WorkChoices Act

2005 within the Australian Industrial Relations granted the relevant minister power to intervene

in industrial disputes, including intensified dissolution and penalties against legitimate union

activities, while enabling individual employment contracts outside the scope of any collective

agreements (Cahill, 2010).

Zwick (2018) defines individual contractual agreements as more collective

arrangements that benefit consumers and financial markets, reduce middle-class employment,

reduce the burden of legal liability of employment and labour laws, and redirect employment

risks to individuals. This characteristic of inherent flexibility for firms to reduce costs typifies

a liberalist system's transition to an individualised neoliberalist system, indicating why

employers systematically push workers towards more temporary and contractual roles. In

analysing the role of local recruitment agencies in East European migration to the UK, Sporton

(2013) identifies individual arrangements characterised by informal contracts, unpaid overtime

rates and the power to fire employees as quickly as they are hired without a right of appeal.

The lack of negotiation between employers and agencies resulting from freedom in labour

choices highlights power asymmetries that have become standardised over time (Cahill, 2010;

Zwick, 2018).

Within the taxi industry discussed further in chapter four, neoliberal policies such as

deregulation and privatisation of the taxi market impacted the labour supply, tariffs, service

quality, and overall profitability for taxi companies and workers. Before the neoliberal policies

between the 1980s to early 2000s, taxi companies employed full-time workers who targeted

urban areas for their market and typically owned around 100 vehicles (Niemietz and Zuluaga,

2016). Afterwards, the employment of drivers was more varied and characterised by self-

employment; taxi firms became privatised and targeted their service according to specific

customer segments. According to Bergantino and Longobardi (2000), GN countries such as

Canada, Sweden, and the UK initiated policies to improve quality and safety standards,

effectively match demand and supply, and enhance driver flexibility. For instance, in

examining the deregulation policies, Toner (1996) examined over 256 councils in England.

From the study, it was evident that while certain local councils possess the power to allow free

entry of taxis, these did not precisely increase fares as they were fairly constant compared to

regulated councils; reduced vehicle quality standards; led to overcrowded taxi ranks; which all

contribute to a race to the bottom for drivers. However, in Sweden, the deregulation of the taxi

industry was successful compared to the UK and other GS contexts. Thelan (2018, p. 10)

identifies three characteristics of the deregulatory taxi market since the 1990s in Sweden. First,

it boosted price competition by removing restrictions across different providers. Secondly, it

led to the removal of all taxi restrictions per area, allowing freedom of movement. Lastly, it

permitted flexibility of drivers by removing the requirement of being part of a restricted taxi

consortium to offer services. By establishing a franchise model, there was a segment for taxi

operators and taxi companies, where the former controls brands without owning vehicles and

the latter contracts with operators to provide services for clients while paying monthly fees per

car. While this outcome varies across GN contexts, it is essential to note that despite

deregulation experiments, GN contexts still possessed a robust regulatory environment to cope

with the consequences.

This phenomenon has been evident in the emergence of both Private Hire Vehicles

(PHVs) in the UK and ride-hailing platforms based on pressures from collective worker groups

and government policies. In other words, indicating that the processes of taxi driving and

management in GN contexts are more formalised than informal compared to GS contexts. In

chapter four, it would be essential to differentiate the factors in the taxi sector that proliferates

informality in the GS context compared to GN contexts like the UK. The subsequent section

examines the limits of neoliberalism in GS contexts and how its shortcomings have accelerated

the proliferation of informal labour and gig work in GS contexts

2.3. From Informal Labour to Gig work: The Limits of Neoliberal Policies

in the Global South

While neoliberal policies emerged from GN contexts like the US and was somewhat

successful in generating wealth, albeit at the expense of equality, these ideologies struggled in

GS contexts, creating more wide-ranging negative economic impacts. According to Peck et al.

(2009, p.53), neoliberalism is premised upon a one-size-fits-all model of policy implementation

which assumes that identical results will follow the imposition of market-oriented reforms,

rather than recognising the extraordinary variations that arise as neoliberal reform initiatives

are imposed within contextually specific institutional landscapes and policy environments. In

a dialogue with Pierre Orelus in 2014, Noam Chomsky argues that neoliberalism has negatively

impacted the poor GN countries and several GS countries. Depicting a scenario on how

economic experts prescribe policies that potentially solve economic challenges in GS contexts,

Rodrik (2017) also argues that the major flaw of neoliberalism is that first-order economic

principles are believed to fit into a unique set of policies fostered by the Thatcher/Reagan

agenda in the 1980s. Sharing similar views as Peck and colleagues, Chomsky agrees that such

unilateral policies often emanate from GN contexts and are imposed on GS contexts leading to

dysfunctional effects, including high unemployment rates, increased debts and an uneven

healthcare system which excludes the poor (Orelus and Chomsky, 2014).

These dysfunctional effects have manifested across spatial scales, including wide-

ranging problems of regulatory coordination, uneven economic stagnation, destructive inter-

locality competition, generalised insecurity, and intensifying inequality (Peck et al., 2009).

Even in GN contexts like the US, neoliberal reforms have generated uneven results because

poor people are primarily impacted negatively (Orelus and Chomsky, 2014). However, some

contexts, such as China, have been successful due to strategic institutional tinkering and

different institutional arrangements where they provide market-based incentives rather than

state-to-private ownerships like GN contexts. According to Kaye-Essien (2020, p. 718),

because neoliberal proselytes (mostly western donors and policy entrepreneurs) worked for

certain East Asian countries (e.g., South Korea and Taiwan) and some African contexts (e.g.,

Botswana and Mauritius), governments in several other GS contexts are made to believe it can

work anywhere. Neoliberalism policies in Africa and many low-income areas in the GS,

regardless of any institutional tinkering, according to Rodrik (2017), have created more

problems that are being exacerbated in the gig economy today.

In the GS, specifically the African context, governments pursued structural adjustment

policies channelling tenets of hegemonic market rule, which highlighted privatisation,

deregulation, free market, and free trade for improvement in foreign capital as a result of the

debt crisis in the 1980s as imposed by the IMF and the World Bank (Gledhill, 2007). While

globalisation has varied in the GS, the impacts on sub-Saharan Africa have been overly

negative (Bryseson and Potts, 2006). Following the arguments that neoliberalism is not a one-

size fits all model, the lack of context specificity for its policies bolstered informal work –

indicating the need for individuals to survive outside formal employment provisions. It is,

however, critical to note that the informalisation of labour, especially in Africa, was established

long before the 1980s, especially because many countries in the region experienced rapid

urbanisation, which led to a depreciation in the formal sector, i.e., the rising population started

to outstrip formal employment (Potts, 2008). As a case example, Ghana illustrates that before

adopting neoliberal policies in the form of SAPs, most of the 2.5 million people in the labour

force were self-employed, engaging in craftsmanship, carpentry, farming, trading, and other

small-scale activities (Peil, 1972; cited in Obeng-Odoom, 2012). It was argued that the

government of Ghana, particularly the Convention People’s Party (CPP), provided over

100,000 jobs between 1957 to 1963 but failed to meet growing demands due to corruption and

mismanagement of state corporations.

Between 1957 to 1982, the economy of Ghana began to struggle, including a series of

recessions, declining GDP rates, and further erosion of formal employment, which led the

country to adopt the economic recovery programme (ERP) to improve the economy. This led

to the state withdrawing from direct employment strategies, further leading to a formal sector

decline. As Obeng-Odoom (2012) contends, adopting these neoliberal policies worsened

Ghana’s unemployment and underemployment rates; led to low salaries due to reduced public

budgets; and increased income inequality between the poor and the rich. In response to over-

regulation, especially in GS contexts, informality thrives as a choice for sustaining individual

livelihood (Rizzo, 2011). The exposure to global competition and abruptly reduced government

expenditure impacted economic activities and employment, leading to significant

retrenchments in public and private sectors (Potts, 2008).

According to Lafferty (2010), governments in several Asia-Pacific countries were

reluctant to increase the minimum wage because of the growth in flexibilisation attributes that

include contract, temporary, casual, and part-time employment as a lack of protection for

permanent workers because of a highly politicised legislated minimum wage. For example, at

the expense of jobs in the formal sector, the proliferation of informal and short-term

employment resulted from chronic unemployment and underemployment in the Philippines,

limiting unions’ bargaining power to about 5% of the entire workforce (Lafferty, 2010). As

well as neoliberalism’s direct impact on weakening collective action, over the years, workers

experienced declining bargaining power and protections because, in the public sector,

privatisation has eroded union strongholds (Lebaron and Ayers, 2013). In sub-Saharan Africa,

both the unemployment rate of 7.7 per cent or people in vulnerable employment (73%) is higher

than the global averages of 6.2 and 43.6 per cent, respectively (ILO, 2016; World Bank, 2021).

While these figures have fluctuated over the years, it reaffirms the stance of informal labour

based on how people actively engage in self-employed gigs or survival strategies to make ends

meet. According to Meagher (1995, p. 281), “informality in Africa is not a process outside the

state. The informal activity of the scale found in many African countries cannot develop

without the state’s complicity”. Moreover, informal labour patterns have coincided with

formal employment challenges and a response to the inability of economies to provide decent

jobs for workers, thereby fostering casualisation, piece-rate work, entrepreneurism, and other

unclassified statuses to make ends meet. In several cases, these informal jobs pay workers

more than formal employment (Heintz and Pollin, 2003), as traditional taxi and ride-hailing

platforms workers in this study highlight. After all, compared to GN contexts, regular payment,

safety protection and health benefits, pensions are not guaranteed under formal employment in

the public sector (Potts, 2008).

A typical reality in the mobility sector across Africa is the incessant increase of informal

transport services, even by formal employment workers, to make ends meet. This has led to

the proliferation of different classifications of work, including mainstream employment,

‘workers’, self-employment, ‘hustlers’, ‘illegal workers, and entrepreneurs, usually fluid

(Rizzo, 2011). The fluidity caused by over-regulation and deregulation based on westernised

neoliberal policies further erodes the formalisation of labour within the mobility sector,

especially the taxi industry, compared to GN contexts. The lines between formal and informal

labour are so blurred that there is little or no difference between the daladala workers (bus

drivers and conductors in Tanzania), taxi drivers in Nigeria, and underemployed white-collar

staff in Ghana. These blurred lines have made it easy for digital platforms like Uber to

permeate the existing taxi industry and formal employment, creating an entry point for them to

permeate. According to the study by Kaye-Essien (2020), the typology of Uber drivers in

Ghana is four-fold: (a) traditional taxi drivers who also use ride-hailing platforms; (b);

employees using ride-hailing platforms to support their income; (c) unemployed people before

joining Uber; (d) finally, traditional taxi drivers switching to ride-hailing platforms due to their

perceived lucrativeness.

While neoliberal policies appear to have negative impacts in regions like Africa, ride-

hailing platforms and algorithmic management remain another neoliberal project that has been

exported to GS contexts like Nigeria (Kaye-Essien, 2020) and has increasingly become another

avenue of informal work that enables economies to meet the growing labour demands. As a

growing field of research on the gig economy attests, this involves understanding its impacts,

particularly on workers. Specifically, this section ushers a basis for understanding how the

shortcomings of neoliberal policies, especially in the GS and how this has ushered in the gig

economy based on existing informal labour and how this in contexts like Lagos discussed in

chapter four of this thesis. I perceive the gig economy as repackaged informal labour from GN

contexts, but this time embedded with technological platforms that proliferate the agenda of

neoliberalism based on further deregulation and privatisation of the taxi industry, which in turn

fosters flexibilisation of labour in GS contexts. The next section unpacks the platformisation

of labour with core examples from GN and GS contexts to understand how digital platforms

are repackaging informal labour as gig work, proliferating flexibility and autonomy of labour,

and overall leading gig workers to a race to the bottom.

2.4. The Global Platformisation of Labour: Situating Gig Work in

Employment Relations

To understand the platformisation of labour, it is critical first to define platforms. The

term ‘platform’ possesses several implied meanings which are architecturally, metaphorically,

and technologically rooted. For instance, in lay terms, the train station possesses an

architectural platform for passengers; metaphorically, a TED talk event is a platform for

speakers on varying topics; and technologically, Facebook is a social media platform for

connecting with friends and family. Theoretically, the focus has been on technological

platforms which are limited to economic theory and engineering design (Gawer, 2014). The

economic perspective defines platforms as markets that are categorised as two-sided markets

or multi-sided markets and multi-sided platforms (Rochet and Tirole, 2006; Rysman, 2009;

Gawer, 2014). The commonality in these classifications is the criticality of intermediaries and

network effects that facilitate adoption processes that differ across varying business models.

For instance, according to Rysman (2009), the literature on network effects focuses on

telecommunication markets and high-end technology, while two-sided markets focus on

payment systems, matching markets, and the media. Engineering design embraces platforms

as technological architectures as a basis for innovation and a facilitator for firms to achieve

economies of scope (Gawer, 2014). Below are typologies of platforms based on a global survey

by Evans and Gawer (2016, p. 9):

Table 1: Platform typologies

Platform Typology

Definition

Example

Transaction Platforms

A transaction platform is a technology, product or

service that acts as a conduit (or intermediary)

facilitating exchange or transactions between users,

buyers or suppliers.

e.g., Uber, Airbnb,

Netflix

Innovation Platforms

An innovation platform is a technology, product or

service that serves as a foundation on top of which other

firms (loosely organised into an innovative ecosystem)

develop complementary technologies, products or

services.

e.g., Microsoft,

Oracle, Intel

Integrated Platforms

An integrated platform is a technology, product, or

service that is both a transaction platform and an

innovation platform.

e.g., Apple, Google,

Facebook etc

Investment Platforms

Investment platforms consist of companies that have

developed a platform portfolio strategy and act as a

holding company, active platform investor or both.

e.g., Softbank,

Naspers, Priceline

Source: Adapted from Evans and Gawer (2016, p. 9); David-West and Evans (2016, p. 5)

This thesis focuses on ride-hailing platforms such as Uber, which are transactional

platforms adopted due to network effects based on two-sided markets for facilitating

intermediaries. It is critical to note that the focus on ride-hailing platforms is not based on

economic or engineering design but on a social study of the internal functions such as software

algorithms that propagate gig work.

Tina Brown, a former editor of influential style magazine Vanity Fair and founder of

the news website Daily Beast, first coined the term ‘gig economy' in 2009 following the 2007

and 2008 recession (Harris, 2016). The term's formulation resulted from what Brown called

“upper-middle-class, white-collar acquaintances, who make ends meet by consultancies, free-

floating projects, and part-time bits and pieces” (Harris, 2016, p. 209). The gig economy

concept is essential for this study because it focuses on the everyday interaction between

platforms and workers. According to Zwick (2018), these include targeting and employing

economically vulnerable migrant workers, legally misclassifying workers and practising

regime shopping, i.e., manipulating local laws to follow platform principles. It is essential to

highlight that these characteristics, which broadly define the gig economy, are embedded in

the neoliberal ideology – the need for flexibility and individualisation, all for firms' benefit, in

this case, platforms. Other core characteristics of gig work include piece-rate compensation

according to specific tasks; self-provision of labour facilities or spaces such as a vehicle,

computer and physical locations for conducting gig tasks (Stewart and Stanford, 2017). These

characteristics have become the basis through which platforms in today's global cities operate,

particularly how workers' misclassification as 'independent workers', 'freelancers' or

'entrepreneurs’ diminishes their power and agency.

There are no accurate figures to determine the size of the gig economy. Heeks (2017)

estimates a global total of about 70 million registered workers in 2015. The World Bank

(2019a) estimates the total freelancer population to be less than 3% (84 million people) of the

total labour force of 3.5 billion people. In the US, Katz and Krueger (2016) estimate that

alternative work arrangements increased from 10.7% in 2005 to 15.8% in 2015, with 0.5%

working through intermediation such as Uber. According to Katz and Krueger (2016, p. 8),

these figures imply that workers increased by 8.6 million (57.2%), from 15 million in February

2005 to 23.6 million in November 2015 and less than 1 million gig workers based on digital

intermediaries such as Uber. In the UK, the Chartered Institute of Personnel and Development

(CIPD) estimates that about 1.3 million people (4%) are working in the gig economy (Taylor

et al., 2017). According to the review, an additional 58% of full-time workers engaged in gig

work, perhaps as an additional income source. A survey conducted by Mastercard and Kaiser

Associates (2019) estimated that the global gig economy has generated over $204 billion in

2018 and is projected to expand by a 17% compound annual growth rate with a gross volume

of $455 billion by 2023.

Scholars analyse gig work as an extension of non-standard employment within GN

contexts, characterised by a denial of workers’ compensations such as health insurance, paid-

leave holidays, pensions, and other benefits (De Stefano, 2014; Harris and Kruegar, 2015).

Although the definition of non-standard employment is vague, the International Labour

Organization (ILO) broadly defines it as employment arrangements different from standard

employment. On the other hand, standard employment work is done under the supervision and

control of an employer at a physical location on a full-time basis, typically with an open-ended

contract based on a five-day week (Karlberg, 2000; Storrie, 2018). The integration of standard

working arrangements as a framework within labour protection, social security systems and

collective bargaining is the norm for most industrial nations in the twentieth century, often due

to labour rights activism from union organisations (Karlberg, 2000). Karlberg (2000) classifies

significant non-standard employment as part-time work, temporary agency, contract company

employment, short-term employment, contingent work, and independent contracting. Other

similarities of gig work based on their nature in terms of flexibility and the role of

intermediaries relate to temporary employment, zero-hour contracts and temporary agency

work where third-party labour intermediaries act as labour distributors (Friedman, 2014;

O'Sullivan et al., 2015). More recently, ILO (2018) has classified it as part-time and on-call

work, dependent self-employment, disguised employment, and multiparty employment

relationships.

Workers conceived gig work as a remedy against the bureaucracy and politics of full-

time work because of the notions of so-called flexibility and freedom it produced, much of

which is a paradox in recent times. This notion of flexibility that has been prevalent since the

beginning of the neoliberal era in different forms has become a strategic selling point for firms

or platforms to transfer employment risks to their workers. According to Taylor et al. (2017),

people who sell their labour via apps or platforms such as Uber and Deliveroo, for example,

are archetypes of gig work. Schmidt (2017) broadly classifies it as online-only labour (e.g.,

freelance marketplaces, micro-tasking crowd work, contest-based creative crowd work) and

gig work, location-based digital labour or what Woodcock and Graham (2019) classify as

geographically-tethered platform work (e.g., accommodation, transportation and delivery

services, household services and personal services). The commonality across both online-based

gig work and location-based gig work is the presence of an intermediary such as a platform or

web service that facilitates interaction between users (Stewart and Stanford, 2017).

The digital gig work model attracts criticisms from scholars because of the deliberate

misclassification of workers as 'independent contractors' (Healy, 2017). One of the critical

elements in the gig economy – workers' misclassification – is not a new phenomenon, but it is

revolutionary through platform business models and working arrangements, which complicate

definitions of work. An independent contractor typically is self-employed and is responsible

for taxes, working hours and bearing the overall economic risk of work (Karlberg, 2000).

According to Zwick (2018), the court system is the only way workers can challenge

misclassification and regain labour rights. For example, in 1997, workers categorised as

independent contractors of Microsoft challenged the company over an illegal exemption from

a pension, welfare benefits and other corporate protections in the Vizcaino v. Microsoft Corp

court case (de Haas, 1998). Although there was a workers' settlement out of court, challenging

platforms based on misclassification has become the norm. The misclassification of workers

has further reduced the bargaining power compared to during the neoliberal era, such that

collective actions through unionisation have become ineffective. According to Lehdonvirta

(2016, p. 65), "… the means used to delocalise work – deskilling, codification, black boxing,

algorithmic management – also undermine organisational identities". The more work is

delocalised and dispersed; the fewer means are available for organisational identity formation.

It is important to note that such delocalisation processes apply to location-based gigs

because of their isolating nature, although they are significantly smaller than online gig work.

The absence of a human employer, wage contracts, and social protection benefits boosts the

flexibility and autonomy of work, which is becoming a paradox with algorithmic managerial

control, discussed in subsequent chapters (Lee et al., 2015; Möhlmann and Zalmanson, 2017).

Platform companies are contracting workers via apps, arguing that workers who invest their

capital or asset such as a vehicle or time and resources – in the case of remote gig work (Graham

et al., 2019) – choose when and where to work (Healy, 2017). Employment lawyers like

Bornstein (2015) argue that work law needs to replace employment law because current laws,

such as tax law, no longer accurately encompass gig workers. Of course, this has implications

for the future of work, considering the increase in non-specified contractual work like

independent contracting. Harris and Krueger (2015) propose replacing the independent

contractor classification with the independent worker classification because the latter ensures

that intermediaries contribute to at least half of social security, Medicare payroll taxes and the

right to organise. However, they are not replicable within GS contexts where most people work

in the informal sector.

According to a report by ILO (2018), GS cities and emerging economies account for

93% of informal employment. On a regional scale, according to ILO (2018), 85.8% of workers

in African countries are informal sector; 68.2% in the Arab states, and in Asia and Pacific

countries also 68.2%, to mention a few. In these regions, informal employment comprises street

vendors and hawkers, waste collectors, taxi drivers, home-based workers, plumbers, cleaners,

mechanics, hairdressers, and several other roles (Lloyd-Evans, 2008; Chen, 2012; Skinner and

Watson, 2017). Before the platformisation of labour, many of these jobs were highly temporal

and based on a pay-as-you-go service, often with payments only in cash. While informal

workers are often not legally recognised in many contexts, the term ‘employment relationships'

acknowledges certain wage workers because they are linked as employees or workers and

employers based on services rendered in return (ILO, 2003; Chen, 2012). Accordingly,

informal workers possess statuses that are not clearly defined, ambiguous or disguised such

that employment relationships do not stipulate proper legal recognition or access to social

protection benefits (Chen, 2012a). Because the income earned from informal labour is

uncertain and irregular, workers risk entering poverty without any benefits or support from the

government to assist them (Guven and Karlen, 2020). For example, the impact of the Covid-

19 pandemic led to the inability of 56.3% and 29.4% of Nigerian and Ugandan informal urban

workers to work, thus affecting their livelihoods. In GS contexts, the informal nature of work

has increasingly been integrated into platform apps today, rebranding them as utopian

innovations and with independent worker contracts that further erode the benefits and rights of

workers.

2.4.1. Spatial and Temporal Dynamics of Gig Work

In Figure 1, Woodcock and Graham (2019) compare gig work (online and offline) and

traditional waged employment, outlining the spatial (such as the need for work to be completed

at a specific place) and temporal nature of labour (the duration of labour). For example,

location-based gig work exhibits longer work durations and a high level of geographic

stickiness because it requires work in a specific place compared to cloudwork (microwork),

which exhibits low levels of spatiality and temporality. However, Woodcock and Graham

(2019) highlight similarities between traditional employment and location-based gig work

based on the need to control workers – although the strategy of control in the gig economy is

algorithmic and opaque.

Figure 1: Spatial and Temporal Conception of Gig Work

Source: Adapted from Schmidt (2017); Woodcock and Graham (2019)

Srnicek (2016) views platforms as digital infrastructures and intermediaries which

facilitate different users such as customers, advertisers, service providers, producers, and

suppliers. In this case, software developers create applications on the Apple Mac operating

system, or drivers and riders exchange cash through the Uber app based on Gawer’s

transactional platform classification above. Platforms also facilitate processes of

disintermediation and reintermediation, according to Lehdonvirta et al. (2015) and Graham et

al. (2017). Disintermediation, a term in economics, according to Rosenbloom (2007, p. 329),

“emphasises the removal or disappearance of intermediaries from distribution channels, which

if carried to the ultimate meaning of the term would result in the total elimination of middlemen

from the channel”. For example, Graham et al. (2017) identifies a 30-year Vietnamese who

started freelancing via digital platforms but soon established a trusting relationship with clients

in the US, Canada, and New Zealand outside of the freelance digital platform. This

Temporality (Job Duration)

Long

Traditional waged employment

Cloudwork (online

freelancing)

Geographic Stickiness

High

Low

Location-based gig work

Cloudwork

(Microwork)

Short

disintermediation process on the employee's side reduced unfair labour methods because there

was no need for platform ratings, commission deductions and unfair pay rates, but built an

actual working relationship based on trust and professionalism. On the other hand,

reintermediation is not a total elimination but a reformulation, realignment and downsizing or

upscaling of intermediaries in the distribution channels, favourable for the worker involved

(Rosenbloom, 2007). Graham et al. (2017) recognises this as reintermediation, where aspiring

or poorly rated freelancers can only access jobs from another freelancer in possession of

multiple jobs. What occurs is that the sole freelancer acts as an intermediary between the client

and other freelancers based on their level of expertise and ratings, often leading to further

exploitation. While processes of disintermediation and reintermediation are more associated

with online gig work or freelancing, this thesis highlights how this also applies to ride-hailing

gig work in Lagos.

Another critical factor for both segments of gig workers relates to the need for skills

and capability development. Graham et al. (2017) recognises the need for workers to upskill to

get more jobs in the presence of uncertainty and opacity from clients across the value chain. It

is critical to note that while the kinds of jobs on platforms keep expanding from coaching,

driving, legal services and many more, the level of skill, employment relations, and spatial

dispersion determines their disparities (Schor et al., 2020; Vallas and Schor, 2020). The skill

level for driving a vehicle for a transport or delivery company is comparatively low compared

to graphic design on Upwork. Although it is safe to highlight that a new entrant in a location-

based gig would be required to learn or relearn how to utilise digital maps and basic functions

embedded within the app, it is still relatively low-skilled and requires basic education compared

to a freelance programmer.

Sutherland et al. (2019) call for continuous skills and literacies to thrive in a flexible

and dynamic working environment. This observation is critical even for disintermediation and

reintermediation strategies of gig workers. For example, an aspiring web developer can take

online courses and perform low-skilled jobs like data entry to improve reputation scores while

working as a hairdresser to survive. This example also applies to freelancers who treat location-

based gigs as a side job solely for income while functionally upgrading for a targeted freelance

market. A highly skilled gig worker with an excellent reputation exhibits more power over the

condition of labour and can better navigate platform opacities. Graham et al. (2017) highlights

the lack of client transparency, such as restricted access to broad client information or the

labour process during gigs, thus inhibiting workers’ ability to upgrade their skills. For instance,

a data entry Filipino worker explained how the lack of information, such as the type of e-

commerce website a client possessed, affected job outputs (Graham et al., 2017). This

observation also indicates how the employer maintains a higher bargaining power over the

worker because of labour arbitrage (buying labour from cheap markets) (Lehdonvirta, 2016),

forcing the worker to settle for what is perceived as fair payments.

This section depicts an overview of the global gig economy built on a neoliberal

understanding based on the proliferation of flexibilisation and an autonomous working

environment. It highlights three critical factors of the gig economy: misclassification of

workers, intermediation, and skill development. More critically, it considers how the

misclassification of workers is reducing the bargaining power of workers and unions,

increasing information asymmetry, denying workers benefits, and causing workers to develop

skills that will enable them to evade rules and regulations in the gig economy. The gig economy

is evolving rapidly, and it is increasingly becoming a race to the bottom for gig workers (World

Bank, 2019a). The real challenge here is not Uber replacing traditional gigs but the myriad of

business models, which are 'uberising' or 'platformising' their labour processes to reduce labour

cost while maximising workers’ input and reducing collective actions (Nurvala, 2015;

Bornstein, 2015; Aloni, 2016). For these platform organisations to thrive, they utilise

algorithms to manage and control atomised labour processes, as discussed in the subsequent

section.

2.5. Algorithmic Management and Surveillance: Weapons of Control

The previous section has examined how the global gig economy emerged to blur the

category of non-standard employment, which is inherent in the ideals of neoliberalism based

on the flexibilisation and freedom of labour. This is increasingly digitising informal labour or

traditional gig work through platforms. Therefore, this section examines how the gig workforce

is managed and controlled. It shows that surveillance and algorithms are mutually reinforcing

instruments of control by reviewing historical accounts of surveillance and how the rise of big

data has enabled algorithms to facilitate non-human supervision and control of gig work.

2.5.1. The Value of Surveillance in Society

The growth of the surveillance culture has been rampant over the years, even before the

advent of the platforms. Its basic conceptualisation highlights the need to observe people

continuously to create a form of self-consciousness and awareness that coerces them to act

responsibly in the presence of law and order. Sometimes, these acts of control are against

peoples will leading to resistance and evasion of the law. The progression in the literature

examines surveillance in the context of different concepts, as shown in Figure 2. This section,

however, summarises the main argument based on the importance of surveillance in inducing

control of members of the society, which Galic (2017) classifies into three phases.

Phase 1 (Architectural Theories and

Discipline of Society)

- Bentham’s Panopticon

- Foucault’s Panopticism

Phase 2 (Digital Infrastructural

Theories)

- Control of Society (Deleuze

1987; 1992)

- Dataveillance (Clark, 1988)

- Surveillant Assemblage

(Haggerty and Ericsson, 2000)

- Surveillance Capitalism (Zuboff,

2015)

Phase 3 (New Conceptualisations

Surveillance from Below)

- Synopticon (Mathiesen, 1997;

Lyon, 2006)

- Mass Surveillance (Norris and

Armstrong, 1999)

- Sousveillance (Mann, 2004)

- Omnipticon (Rosen, 2004)

- Banopticon (Bigo, 2006;2008)

Figure 2: An Overview of Surveillance Concepts

Source: Adapted from Galic et al. (2017)

The first phase embraces the idea of the panopticon, indicating physical and spatial

structures such as closed institutional buildings and other territorially based settings with

capabilities of watching over subjects (Galic, 2017). As described by Bentham, the panopticon

was a spherical architectural prison where prisoners were visible to the watchtower but

invisible to each other and could not see the guards (Bentham, 1962, cited in Koskela, 2003).

Foucault (2008) depicts a plagued setting where syndics, intendants and guards assume the role

of human observers monitoring each district in the town to ensure that inhabitants obey the law

of staying indoors – failure to obey the law leads to death. Syndics, intendants, and guards are

the only groups required to move through different houses, checking for dead and infected

bodies from the plague. Foucault argues that this ideology facilitates functional power by

creating a permanent state of visibility and self-consciousness in 'the watched' (Foucault,

2008). This observation indicates the efficacy of surveillance even in Bentham's architecturally

annular prison or Foucault's plague-stricken town such that discontinued supervision or

panoptic gaze still demonstrates a dominant level of power, even though the subject can

exercise evasive power only with an integral knowledge of surveillance apparatuses. Therefore,

Bentham's design highlights that power should be visible – subjects are aware of surveillance

mechanisms – and unverifiable – subjects are insecure about when and how surveillance

mechanisms operate (Foucault 2008). As highlighted in the third phase, several scholars have

developed this metaphor of the panopticon or panopticism to conceptualise and highlight the

nature of being watched and monitored within society, including workplaces which facilitates

agency amongst the watched.

The second phase examines surveillance reliance on digital over physical technologies,

indicating an evolution from institutions to networks of opaque forms of control (Galic et al.,

2017). The similarity between the work of Deleuze (1992), Haggerty and Ericson (2000) and

Zuboff (2015) indicates the shift away from the idea of a self-disciplining panopticon –

questioning how surveillance reinforces and sometimes undermines the power structures in

networked societies (Galič et al., 2017). For instance, Deleuze (1992) views society's inherent

control as a continuous process within institutions such as the prison system, the school system,

hospitals and the corporate system, all transformed into spaces of monitoring corporations with

computerised mechanisms and codes. Compared to monitoring enclosed spaces over longer

periods in the disciplinary society, the control of society’s logic focuses on modifying how

access is perceived within such systems (Deleuze, 1992). This includes using electronic collars

to ensure a convicted prisoner remains within the prison system, handling money and profits

within the corporate system and many other examples (Deleuze, 1992). The control targets are

not directly individuals, but their access infrastructures, representations, and social behaviours,

indicating surveillance as invisible to detection. Also, Haggerty and Ericson (2000) draw on

Deleuze and Guattari's surveillant assemblages, modelled based on a rhizome plant,

demonstrating the multiplicity of surveillance apparatuses or information flows that are

interconnected with different aspects of society (Deleuze and Guattari, 1987).

Haggerty and Ericson (2000, p. 608) deconstruct assemblages as being part of

conventional apparatuses of the state, such as government rule, but more by "virtue of its own

characteristic set of operations; the tendency to create bounded physical and cognitive spaces,

and introduce processes designed to capture flows". These flows exist due to interactions with

the human body, its assemblage and within the human body, creating data-doubles that can be

analysed and traced (Haggerty and Ericson, 2000). Besides recognising assemblages such as

CCTV footage as an interconnected layer to state institutions, Haggerty and Ericson (2000)

also perceive the human body as a source of informational flows that can be captured by ICTs

such as mobile devices and credit cards linked to the Global Positioning System (GPS),

Geographic Information Systems (GIS), databases, phone conversations and in many other

ways. The sophistication of technological monitoring expands through time to capture new

bodies of informational flows such as international travellers, employees, caregivers, young

people, parolees, commuters, and many other categories (Haggerty and Ericson, 2000). The

surveillant assemblage appears to be subtle and passive and can be active in the presence of a

criminal activity or violation of law and order due to captured information flows.

In the surveillance capitalism notion, these informational flows, or what Zuboff (2015)

later classified as big data, are tools to modify and control human behaviour. At the time of

Haggerty's and Ericson's conception of the surveillant assemblage, storing information from

Fitbit activities for each employee could have been difficult because of the absence of big data

present in contemporary societies. According to Andrejevic and Gates (2014, p. 186), big data

refers to "the unprecedented size of contemporary databases and the emerging techniques for

making sense of them… with the significance of not only storing more quantity of information

but doing so in novel ways such as tracking business trends, analysing web traffic, assessing

disease distributions, predicting weather forecasts and financial markets and many other uses".

Despite managing significant information in previous years, the quality was limited to the

heavy involvement of human supervision and interpretation and the absence of predictive

analytics. For example, the ability to monitor hundreds of thousands of people was limited by

inaccuracies or lack of data capacity. However, with the novelty of big data and its algorithms,

surveillance is becoming more subtle and less independent of human interpretations and

analysis (Andrejevic and Gates, 2014). Based on this premise, the surveillance capitalism

notion uses this data not just for product and service improvements but also for a 'proprietary

behavioural surplus' integrated into machine learning mechanisms capable of predicting the

present and future of humans (Zuboff, 2015). Companies like Google, Amazon and Microsoft,

and social media platforms like Facebook are vital surveillance capitalists that can monitor,

track and predict behavioural futures with more certainty than in previous years. As Zuboff

(2015) puts it, these capitalists possess unprecedented information and power asymmetries

designed to collect information about humans for profits with no clarity on how data is used.

Despite data protection policies today, there is anxiety about always being watched because of

how technologically deterministic the world has become.

The final phase builds on the first two phases of surveillance (Galič, 2017), which has

progressed from the physical constructions of panopticism to a progressive focus on the control

of society through digital technologies and its affordances with capabilities to modify open

spaces or networked societies from human behaviour. This phase conceptualises and builds on

Foucault's panopticon, in, for example, the synopticon (Mathiesen, 1997; Lyon, 2006) and

panopticon (Bigo, 2006; 2008), and integrates the ideals of the control of society (Delueze,

1992). Developed after the September 11

, 2001, terrorist attack, for instance, Bigo (2008)

identifies how surveillance systems evolved to improve security and power by banning

criminals or individuals who openly disregard rules and regulations. While this focuses on both

the ideologies of Foucault and Deleuze, the synopticon by Mathiesen (1997) and later

developed by Lyon (2003) integrates technologies of the mass media such as the television –

empowering citizens to equally watch the observer, in this case, political representatives.

Mathiesen (1997) expressed that both ideologies could simultaneously bring discipline and

control to society. Consequently, citizens can take pictures of bad police officers and record

and observe corrupt practices by people in power. Mathiesen's perspective of the power relation

across cyberspace empowers the 'watched' to watch, leading to the appropriation of new

resistance spaces against control. For example, Mann (2004) highlights a counter-surveillance

practice termed sousveillance which means to watch from below, highlighting how 'the

watched' can similarly observe 'the watcher' through media such as audio, video and

photographs.

The conceptualisations of surveillance have shown its importance in understanding

control systems in broader society, which are transferable to workplaces today. Despite the

three classification phases by Galič et al. (2017), it is safe to agree that the first two phases are

the basis for other surveillance ideologies highlighted here.

This section highlights the historical accounts of surveillance discourses because these

are increasingly becoming critical components in algorithmic management today. More

specifically, I draw on Haggerty's surveillance assemblages and how it is evident in how

algorithms are designed to facilitate task assignments, performance evaluations, fare payments

and rewards, and sanctions or deactivations discussed in the next section. While surveillance

is critical in unpacking platform control, it is important to note that it only serves as a

reinforcing basis for examining the impacts of algorithmic management of ride-hailing

platform workers in Lagos, Nigeria. The following section examines algorithmic management

as a concept and as a set of instructions for managing gig workers globally and ride-hailing

platform drivers in Lagos.

2.6. Global perspectives on Algorithmic Management and Gig Workers’

Agencies

2.6.1. The Power of Algorithmic Management

Algorithms have become essential to our everyday decision-making through sensors

like smartphone devices, which means that they are not limited to mathematical fields but are

progressively inherent in social aspects of life. According to Shildlt (2017, p. 25), algorithmic

management is scientific management 2.0 because of how it "shifts power from a hierarchy of

managers to larger cadres of professionals who master analytics, programming, and business”.

Schildt (2017), drawing from scientific management 1.0 or Taylorism as introduced by

Frederick Winslow Taylor, is based on how traditional workplaces emphasise the need for

employees to work productively and efficiently. In simple terms, algorithms are a set of ideas,

tasks, and procedures required to solve problems. In other words, scientific management

possessed four principles as designed by Taylor, such as (a). replace the old rule-of-thumb

method by developing science for each element of a worker's labour process (e.g., using a

timer); (b). scientifically select, train, and teach each worker according to their capacities; (c).

prioritise cooperation over individualism while following scientific principles; and finally, (d).

define clear duties between management and employees to better articulate responsibilities

(Taylor, 1911, p.34).

Many of these principles have guided how organisations effectively manage and control

the affairs of their workers. The principles ensure that the labour process can be progressively

monitored and supervised for organisational benefit. Other scholars (Fernie and Metcalf, 1998;

Bain and Taylor, 2000; Bain, 2005; Steven and Lavin, 2007) have built on Taylorism principles

in call centre jobs, used in monitoring worker performances and boosting productivity. Stevens

and Lavin (2007) also examine how employers track call centre workers' working schedules

through computer systems (e.g., knowing when employees sign in) and the use of time buzzers

and occasional warnings from management staff at workstations. In the traditional working

environment, for example, monitoring and surveillance techniques range from communications

(e.g., emails, chats, phone conversations), electronic recruitment, mobility tracking, drug

testing, psychometric testing, and several other practices (Ball, 2010). According to Ball

(2010), it is typical for employees' performance to be reviewed in these ways.

These processes also draw from the surveillant assemblage because of the different

aspects that channel informational flows to employers and limit workers' power by controlling

the labour process (Haggerty and Ericson, 2000). However, it becomes controversial when

information collected is intrusive in workers' lives, demanding precise information on how

workers use time and are linked to the ethics of good data monitoring practices that could affect

trust, autonomy, and control. As technology has advanced, these processes have become more

subtle and integrated into platforms which now prioritise individualisation as opposed to the

cooperation of gig workers – creating an isolated workplace. While these possess

characteristics of neoliberal policies through the flexibilisation of labour, gig workers are

inheriting a range of risks which are otherwise difficult to challenge because of the opaque

nature of algorithms used in the labour process.

Management is no longer a human practice, but a process embedded in technology.

Gillespie (2013) highlights that algorithms play a critical role in shaping the quantity and

quality of information required to participate in our everyday lives. Accordingly, algorithms

are helpful when attached to the databases in which they function (Constantinou and

Kallinikos, 2015; Gillespie, 2014). These include the importance of algorithms in managing

social media interactions, navigating information on databases, highlighting trending news,

fashion, music and, overall, a sense of participating in social and political discourses (Gillespie,

2014). However, when thinking about the powerful ways in which ideas and notions about the

algorithm circulate through the social world, we need to think about the impact and

consequences of code (Beer, 2017). It is, therefore, no surprise that scholars attempt to

understand how algorithms shape governmental, commercial, institutional, and organisational

decision-making. As the last step in actualising complex data operations (Constantiou and

Kallinikos, 2015), algorithms have increasingly become more than technical infrastructures

(Willson, 2016). They are socio-technical infrastructures that engage in complex ways with

their surrounding ecosystems (e.g., code, human designs, platforms, and a few others) as

generative processes and artefacts.

Min Kyung Lee coined the term algorithmic management in 2015, defining it as

"software algorithms that assume managerial functions and surrounding institutional devices

that support algorithms in practice" (Lee et al., 2015, p. 1603). Through surveillance

apparatuses and big data analytics, algorithms can assign work and ensure labour process

management with little interference from human managers. Scholars such as Wood (2021) and

Kellogg et al. (2020) have identified three strains of algorithmic management that they

categorise as algorithmic directions, evaluation, and discipline (see table 2). These

conceptualisations, in a nutshell, demonstrate how algorithms control the labour process and

drivers. Algorithmic directions are based on processes that outline task assignments and

procedures for gig workers at different times and with varying degrees of accuracy. For

example, the algorithm for an Uber driver with a rider showing the estimated pick-up times,

pick-up destination, and potential arrival times following a trip acceptance. Algorithmic

evaluations are based on a set of computational structures and assemblages used to evaluate

gig workers' productivity and efficiency. These structures and assemblages are based on

reputational metrics drawn from ratings, cancellation trips, acceptances trips, hours worked,

tasks completed, keyboard strokes and others (Kellogg et al., 2020; Rosenblat, 2018; Wood et

al., 2019; Lehdonvirta et al., 2019). Finally, algorithmic discipline is an extension of

evaluations where algorithms initiate deactivations/restrictions or rewards/incentives for gig

workers based on bad or good working habits. For example, a high cancellation rate on Uber

will lead to temporary deactivation from the platform.

Table 2: Summarising Algorithmic Management

Algorithmic

Management

Definition

Highlights

Algorithmic

Directions

Algorithms are responsible for assigning

tasks to workers via an app or

smartphone device.

- Assigns tasks based on proximity

- Determines scheduling and work routes

- Determines pick-up times and arrival

times

- Determines the length of journey

Algorithmic

Evaluation

Algorithms are responsible for assessing

the quality of services based on a set of

computational structures and

assemblages that derive information

from users’ behaviours. In location-

based platforms such as Uber, ratings of

1 to 5 are outsourced to riders to score

drivers per trip.

- Records and evaluates weekly hours

worked

- Records, evaluates, and processes weekly

pay

- Evaluates frequency of trip acceptance

and cancellations

- Quantifies and assigns evaluation scores

for evaluation metrics (e.g., rating scores,

acceptance, and cancellation percentages

Algorithmic

Discipline

As an extension of evaluations,

algorithms are programmed to

discipline/punish/sanction and reward

gig workers for bad working habits and

good working habits.

- Internalises reputational thresholds as set

by the platform

- Deactivates gig workers' accounts due to

bad working habits, e.g., low ratings, rude

to riders, etc.

- Restricts gig workers on low reputation

scores (e.g., low ratings, high cancellation

rates)

- Rewards for good working habits based

on high reputational scores (e.g., high

ratings, high acceptance rates)

Source: Derived from (Rosenblat and Stark, 2016; Möhlmann and Zalmanson 2017; Kellogg

et al., 2020; Wood, 2021)

A recurring similarity across these studies is that algorithms display underlying

characteristics centred around opacity, informational asymmetries, and bias. In other words,

partial or unclear information disclosed to workers limits their knowledge of potential risks or

how to navigate potential impacts. Although other aspects have been highlighted in literature,

including algorithmic accountability, privacy, and fairness. This study only feasibly observed

these three constructs in table 2 as critical proliferators of negative impacts drivers’ experience

in the platform economy. More specifically, this research interrogates the opacities,

information asymmetries, and biases drawing from the perceptions of how algorithms operate

according to Kellogg et al., (2020) and Wood, (2021). However, recognising the opacities,

information asymmetries, and biases in this study, is critical in examining how algorithms

evaluate, control, and discipline platform workers.

2.6.1.1. Opacity

According to Burrell (2016, p. 4), algorithmic opacity is a largely intentional form of

self-protection by corporations' intent on maintaining their trade secrets and competitive

advantage". Opacity, a blanket term for remediable incomprehensibility, i.e., something that

remains a mystery or not solvable, is a strategy that proliferates secrecy and complexity in the

black box society (Pasquale, 2015). Corporations imbibe the culture of secrecy and complexity,

which according to Pasquale (2015), can be classified as real secrecy, legal secrecy, and

obfuscation. With real secrecy, there is an unseen barrier between hidden content and

unauthorised access. Legal secrecy requires employees of a corporation to keep certain

information secret, such as a bank employee not disclosing customers' account details to

friends. Finally, when secrecy has been breached, obfuscation means that corporations initiate

deliberate attempts to conceal information through ambiguity, i.e., by presenting misleading

evidence or enormous paper documents.

The opacity in algorithmic processes creates trust issues for users and sometimes leads

to discrimination and biases. Ajunwa (2019) highlights how algorithms have exacerbated age

discrimination from the platform design to platform advertisements online, automated hiring

platforms like Indeed.com, and recruitment platforms utilised by companies such as LinkedIn.

Ajunwa finds that although the Age of Discrimination (ADEA) law was passed in 1967 against

hiring older workers, over 20,857 discriminatory cases were reported at the US Equal

Employment Opportunity Commission (EEOC) in 2016 alone. While previous terminologies

like "young blood", "recent college graduate" can proliferate age discrimination, Sink and

Bales (2016) argue that terms like digital natives also signal discrimination against older

potential workers because it implies that an applicant is not conversant with technology

languages.

2.6.1.2. Information Asymmetry

Information asymmetry builds on the level of opacity that is embedded within the

platform infrastructure. In other words, the opacities serve as a basis on which digital platforms

create an asymmetrical and biased working environment for gig workers, often through

gamefic elements. For instance, Rosenblat and Stark (2016, p. 3771):

"The gamic elements of behavioural engagement tools, such as surge pricing,

the conflation of real-time and predictive demand, and blind passenger acceptance,

illustrate the multifaceted ways that Uber influences the relationship between supply

and demand. These gamic elements support the notion that Uber is not responsible for

inconsistencies in its system; rather, automated functions, such as algorithmic pricing

or blind passenger acceptance, are part of the interactive design".

This asymmetrical information indicates that gig workers can only see data displayed by the

algorithm as a competitive strategy between a customer and a worker, which can lead to biases

that can be detrimental to a worker (Möhlmann and Zalmanson, 2017). For example, drivers

often do not know a rider's destination before accepting a trip.

2.6.1.3. Bias

The bias phenomenon is not new because these are inherited based on pre-existing data

sets that reflect power structures, prejudices, and inequalities that are only reinforced through

algorithms (Katzenbach and Ulbricht 2019). In a computerised system, Friedmann and

Nissenbaum (1996, p. 332) define bias as "computer systems that systematically and unfairly

discriminate against certain individuals or groups of individuals in favour of others". Previous

studies have examined the biases within the workplace, such as racial biases (cf. Rosette et al.,

2008), gender and age biases (cf. Rupp et al., 2006; Heilman, 2012; Ajunwa, 2019), recruitment

and computerised biases (cf Friedman and Nissenbaum, 1996; Lee et al., 2019) and several

others. Hanrahan et al. (2018) observed racial, language and appearance biases based on their

perception of low ratings for black drivers, poor-English speakers and ugly people within the

ride-hailing platforms. According to Mittelstadt et al. (2016, p. 7), "Algorithms inevitably make

biased decisions. An algorithm's design and functionality reflect the values of its designer and

intended uses, if only to the extent that a particular design is preferred as the best or most

efficient option". The importance of algorithms is entwined in learning and modifying human

behaviours, either for profits or control, drawing from Zuboff's surveillance capitalism. The

efficiency of Google's search function relies on the feedback loops of human queries integral

for algorithms to improve search results (Zuboff, 2015). Despite its many advantages, it also

possesses disadvantages in that it creates too much power and control within corporations and

governments, which may be unfair to users under algorithmic management.

A commonality throughout these studies is that algorithmic management is solving

managerial and human resource (HR) challenges, creating new challenges, and compounding

old ones. This shows the proliferation of discrimination and the inherent politics embedded in

digital platforms due to their opacity and how algorithms manifest pre-existing historical

biases. In the traditional confines of worker management and evaluation, scholars like Dugan

et al. (2017) investigate the importance of algorithmic management in developing and forming

working relationships compared to conventional human resource management (HRM). Gig

work incorporates HRM-like activities such as performance management and evaluation,

management of working relationships and future assignment processes (Meijerink and Keegan,

2019). However, it differs because of the replacement of humans with algorithms to manage

these processes, which mutually reinforce surveillance as a critical component in managing

labour processes.

2.6.2. Surveillance as a Component of Algorithmic Management

Newlands (2020) identifies algorithmic surveillance, managerial surveillance and

customer surveillance as multimodal surveillance assemblages in gig work. According to

Newlands (2020, p. 5), algorithmic surveillance is "an assemblage of computational processes,

which automatically generate data, evaluate worker behaviour and assign labour activities”.

However, both definitions do not highlight the ability of algorithms to discipline workers

without any human interference. Therefore, this thesis defines algorithmic management as a

set of computational rules and assignments guiding the affairs of gig workers with the power

to assign work, outsource and evaluate performance, incentivise labour and discipline workers.

Waters and Woodcock (2017) highlight the platform company surveillance or algorithmic

panopticon from their study of Deliveroo, which includes a God's eye-view system that

monitors and is aware of workers' activities in real-time by collating quantitative data.

Although the operationality of this is less studied because of the secrecy of platform companies,

it remains comprehensive in keeping track of workers and enabling customer and managerial

surveillance as a subset. Customer surveillance as a critical aspect of algorithmic management

is more relevant than managerial surveillance because it involves customers who rate workers

for their service, which feeds back into the workers' evaluation metrics (Möhlmann and

Zalmanson, 2017; Rosenblat, 2018; Newlands, 2020).

This surveillance by customers involves a reputation system in both online and location

gig work, which is crucial for algorithmic management in limiting jobs for workers and issuing

sanctions or punishments. On the other hand, to some extent, managerial surveillance requires

a human presence, such as dispatchers for the Deliveroo or Stuart platforms that can monitor

fleets, even though algorithms are replacing most of these actions and making communication

more difficult (Woodcock, 2020; Newlands, 2020). Researched by several scholars (Goods et

al., 2019; Jarrahi et al., 2019; Wood et al., 2019; Anwar and Graham, 2020; Amorim and Moda,

2020), these surveillance capabilities enable control of workers despite the characteristics of

flexibility and autonomy they possess.

Notions of flexibility and autonomy equally characterise the ride-hailing and food

delivery perspectives; they offer different narratives about how gig work is managed, surveilled

and controlled. For example, several scholars (Griesbach, 2019; Richardson, 2019; Galiére,

2020; Cant, 2020; Woodcock, 2020) examine how algorithms manage the processes involved

in deliveries within delivery platforms. Like Uber, food delivery platforms such as Deliveroo

use algorithms to assign, monitor, evaluate and control labour processes. They also wield

flexibility and autonomy when classifying workers as independent workers. According to

Richardson (2019, p. 629), “riders with Deliveroo platforms are firstly only contingently able

to make themselves available to work, and secondly when they are available are not guaranteed

the possibility to earn anything, and thirdly when they are offered earnings these are variable

rather than fixed”. Drawing from Galiére (2020) and from the discussion above in 2.6.1, I

outline how the algorithmic management of labour within food delivery platforms can be

grouped broadly into three constructs which may also apply to other aspects of the gig

economy.

The first construct is that each worker possesses their version of the app to assign work

and evaluate workers while preventing worker collectivism. This phenomenon leads to

information asymmetries, making it difficult for workers to comprehend decision-making

processes. For instance, in studying a range of food delivery platforms across the US,

Griesbach (2019) highlights how participants complain about algorithms withholding vital

information such as the full range of orders placed by customers within a region. This is a

similar observation in ride-hailing platforms where drivers are unaware of trip destinations

(Rosenblat, 2018; Möhlmann and Zalmanson, 2017). Instead, the algorithm limits

informational access to individual orders, that workers must accept or reject. Rejection of jobs

may influence how quickly a worker can receive jobs, highlighting the temporality of gig work.

Secondly, Galiére (2020) identifies surveillance as an embedded function in the app

extrapolated from smartphone data by tracking movements through GPS signals and recording

fragmented micro tasks that workers must validate. This further expands the first point in

acknowledging the black box or hidden characteristics of algorithms in the extrapolation of

data through location signals based on speed and time, which feedback into the app via machine

learning to redefine the labour process in real-time (Cant, 2020). For instance, every assigned

job has a fixed price and a 30-minute duration to complete an order (Richardson, 2019). Waters

and Woodcock (2017) highlight a scenario where an order was delayed for over 40 minutes in

the UK, leading to the Deliveroo rider support team calling the rider to investigate the problem.

This highlights a critical difference between delivery and ride-hailing platforms because

operational managers are part of the surveillant assemblage because of their ability to ensure

continuous monitoring of workers (Newlands, 2020).

The final construct is about sanctions of non-compliant behaviours (Galiére, 2020),

where late delivery or rejection of trips can reduce labour availability for workers. Another

difference with ride-hailing platforms is that algorithmic management on delivery platforms

does not emphasise workers' ratings but more work temporality. However, this is consistently

changing with time. For instance, a Postmate worker explained how ratings from 1 to 5 were

changed to "thumbs-ups and thumbs-down" on the platform, indicating more opacity because

workers do not comprehend its ramifications (Griesbach et al., 2019). On the other hand,

performance indicators such as work attendance, late deregistration and peak periods are

critical for receiving shifts (Galiére, 2020).

In studying how platform algorithms, through surveillance, control workers, Anwar and

Graham (2019) examine how ratings, feedback and monitoring systems, integrated with socio-

economic and cultural backgrounds, suppress African workers' actions. The study's sample

found that platforms such as Upwork ensure workers work for long hours by capturing time-

stamped screenshots of their laptops every ten minutes and regular tracking of time. Workers

risk the chance of not receiving payment for a job if screenshots capture inactivity and unrelated

work activity such as browsing on social media accounts or playing games. This phenomenon

of working long hours could lead to sleep deprivation, which causes other physical and mental

impacts like back pain and eye defects (Anwar and Graham, 2019). The differences in time

zones also contribute to workers’ long hours because of how workers invest time in securing

gigs from clients abroad (Lehdonvirta, 2016; Graham et al., 2017), which, in many instances

and as a pattern of evading competitors, overpromise clients that jobs will be performed within

strict deadlines (Good et al., 2019). For example, a Nigerian gig worker was sleep deprived

because of a combination of time difference and monitoring capabilities, leading to 80 hours a

week which is detrimental to the wellbeing of workers (Anwar and Graham, 2019). Intensified

work hours could also result from workers' quest to build and sustain their ratings on platforms

like Upwork and location-based gigs like Uber because it facilitates algorithmic assignment to

workers and improves bargaining power.

For online gig work, excellent ratings enable workers to receive more work. The

bidding process for a worker with a high score is different from a worker with a low score,

such that the latter must invest more effort to attract a client (Wood et al., 2018). There are

instances where African workers are undercut by lower bids from workers in Bangladesh,

India, and the Philippines because, according to Wood and Graham (2020), African workers

stated that workers from places like India could accept jobs for as low as $1, making it difficult

to attract jobs. However, workers with a good reputation and a high number of working hours

have the bargaining power to reject client contracts or cancel jobs if they choose. Gig workers

are actively subverting algorithmic management power and evading surveillant assemblages

that ensure control.

2.6.3. Gig Workers' Agencies: Subverting the Power of Algorithmic

Management

The continuous supervision and control of gig work have facilitated worker agencies to

subvert algorithmic managerial control. Sutherland et al. (2019) also highlight the volatility of

workers achieving a high 'Job success score' and 'Top-rated badge', but algorithmic

management limits workers’ understanding of rating numbers, making the building and

sustaining of high ratings a trial-and-error process. Jarrahi et al. (2019) describe platforms like

Upwork as deliberate interfaces that project some form of autonomy for workers while

imposing boundary resources that create asymmetries of labour through the protection of

platforms from disintermediation. Some of these boundary resources limit workers from

understanding the platform's scope by restricting communications channels, not sharing

workers’ last names, and restricting the posting of links on their websites (ibid). According to

Jarrahi et al. (2019), the boundary resources also attempt to mitigate against disintermediation

or reintermediation by ensuring workers agree to non-circumvention agreements, establishing

automated monitoring channels and facilitating sanctions of accounts. Nevertheless, these

boundaries that prevent such processes create spaces of resistance for workers, which Jarrahi

and Sutherland (2018) outline as sensemaking, circumvention and algorithmic manipulations.

Workers spend hours on online forums, creating client accounts and other practices to make

sense of Upwork's ratings, procedures, timeframes, and policies. The Upwork platform model's

inaccuracies enable workers to practice circumvention acts (e.g., using a different device to

avoid Upwork's activity tracker) to regain some autonomy and power (ibid). However, there

are disparities between algorithmic management in online gig work and location gig work

because of different monitoring mechanisms such as the GPS embedded in a platform app or

an assistant in human resource work.

Literature on algorithmic management and location-based gig work arguments centre

around ride-hailing platforms discussed in Chapter five. Other platforms like delivery

platforms and Airbnb outline similar arguments about algorithmic management's operationality

in the gig economy. Cheng and Foley (2019) unpack how Airbnb hosts complain about the

lack of clarity of information and low transparency afforded to them by algorithms shaped by

penalties and rewards on the system. For example, Cheng and Foley (2019) discovered that

algorithms downgrade hosts who utilise the dispute resolution system by reducing the selection

numbers in the list without a clear explanation of how this is beneficial for hosts. This

algorithmic ambiguity created a sense of anxiety forcing agency in hosts, which Cheng and

Foley classify as sensemaking (e.g., determining what 'smart price' entails), experimentation

(e.g., decreasing the number of declined guests), manipulation (e.g., modifying words in search

results) and resistance (e.g., which helped hosts develop algorithmic competency). Similarly,

in studying 15 hosts across Brazil, Canada, the UK, and the US, Jhaver et al. (2018) examine

how algorithmic management creates uncertainty and anxiety about what hosts can control to

maximise the platform's benefits. For instance, participants assume some clarity of control over

their tenure as a host, location listings and the number of bathrooms and bedrooms in listings.

However, hosts lack clarity regarding how specific actions such as cancellations of gigs,

response time, pricing suggestions, changing profile details, and other actions impact

evaluation ratings. It becomes unfair to utilise star ratings as an avenue for evaluation when

algorithmic opacity prevents hosts from comprehending how these ratings are generated.

This phenomenon highlights how algorithms coerce hosts not to complain because it

would influence the frequency of gig tasks, giving the company more power and scope to make

more money. Aneesh (2009) recognised such processes as 'algocracy'. Unlike bureaucracy,

authority does not require legitimacy because permissible routes may be programmed or lack

alternative pathways altogether. Aneesh (2009) highlights that while the structural firms are

bureaucratic, such external processes are increasingly algocratic in the global market. Both

studies mentioned above provide some scalable context to other aspects of the gig economy

like ride-hailing, especially in terms of flexibility, autonomy, and control. However, the study

does not state the hosts' status or social class to determine if Airbnb is a part-time or full-time

gig, making the analysis limited to determining why hosts interpret algorithmic languages

differently and yield different results. Also, it is unclear to what extent hosts may be defined

as workers compared to online gig workers or those working for other location-based gigs like

Deliveroo and Uber. Regardless, it is evident that lopsided control due to algorithmic

management demonstrates facets of monitoring hosts while preventing reverse surveillance

because of the lack of information afforded to hosts.

Both Briziarelli (2019) and Cant (2020) identify workers protesting and boycotting

food delivery apps to challenge algorithmic mismanagement. Cant (2020) explains the

importance of online forums, social network groups and a unique newspaper issue called Rebel

Roo, which helps Deliveroo workers in the UK and globally make sense of algorithmic

assignments and subsequently facilitate meetings and protests. According to the Independent

Workers Union of Great Britain (IWGB), these networks have facilitated several

demonstrations and protests since 2016 across different cities in the UK, including London,

Manchester, Brighton and Sheffield, and other cities, all because of unfair pay structures,

dismissals and opacity in the labour process (Osborne, 2016; Cant, 2020). Sensemaking

strategies through online networks and Rebel Roo are circulated to about 1000 workers a month

and even across countries like Germany, France and Italy to facilitate protests (Cant, 2020). In

Italy, Briziarelli (2019) highlights how Deliveroo riders impact the delivery process by logging

in, accepting a job and refusing to deliver any orders to customers to protest against the

platform. Other actions of agency tow similar lines of circumvention and manipulation earlier

stated and what Woodcock and Johnson (2018) describe as gamification-from-below, enabling

workers to use algorithmic gamic elements such as bonuses and incentives to their advantage.

Veen et al. (2020) observed Australian food delivery pattern reworkings where workers game

locations with the highest possibility of receiving orders by creating predictive heat maps. Also,

to secure unlimited time to accept more orders and evade work shifts, workers hacked and

deactivated app auto-acceptance functions and manipulated GPS signals using location-

masking tools. Similarly, in China, Sun (2019) observes a process where riders bypass the

opacity of algorithms by virtually switching workplaces when they use multiple downloaded

delivery apps.

Many studies on delivery platforms have focused on the GN cities without integrating

GS perspectives, which offer slightly different perspectives. The studies above, specifically

delivery platforms, show similarities with ride-hailing platforms discussed in subsequent

chapters, with more scalability and contrasts to GN perspectives. However, the following

section takes a step forward through the lens of James Scott’s everyday resistance in terms of

how workers' agency could be categorised as both hidden and public transcripts and how the

spatialisation and temporalisation of gig work may affect such delineations.

2.7. Everyday Resistances of Gig Workers to Algorithmic Management:

Developing a Conceptual Model

This section enrols James Scott's (1985, 1989, 1990) concept of everyday resistance to outline

the 'hidden transcripts' and 'public transcripts' of gig workers’ resistances against algorithmic

management (see Figure 3). Below are Scott's main arguments and critical aspects from

critiques, which enhance everyday resistance to incorporate the resistances of gig workers in

the GS.

2.7.1. James Scott’s Everyday Resistances: Outlining the Hidden and

Public Transcripts

The definition of 'resistance' in different fields could signal inaccurate and contradictory

approaches that are often ambiguous (Hollander and Einwohner, 2004). Closely linked to the

concept of everyday resistance, Scott (1989, 1990) differentiates between more overt forms of

resistance as 'public transcripts' and covert forms of resistance as 'hidden transcripts'. Hidden

transcripts refer to “offstage speeches, gestures, and practices that inflect, confirm, or

contradict what appears in the public transcript” (Scott, 1990, p. 4). In his book Weapons of

the Weak (1985), Scott focuses more on peasants' everyday acts of resistance in a Malay village

against a powerful agent who seeks taxes, labour, food, rents, and interest from them.

According to Scott (1985), the practices classified as everyday forms of resistance are too

numerous to be associated with the same construct. These weak practices are often routinely

invisible and can take the form of physical and material tactics such as deliberate foot-dragging,

pilfering, feigned ignorance, sabotage, smuggling, poaching, arson, slander, anonymous

threats, dissimulations, and many others. While these hidden practices are relatively safe and

small, they require some cooperation, no formal coordination and promise vital material gains

(Scott, 1989).

On the other hand, public transcripts are on-stage practices visible through dissident

sects, social movements, unions, social movements, and other forms of an organised political

opposition that challenge powerholders openly through strikes, demonstrations, revolutions,

riots, boycotts, and many others which may not occur every day. According to Scott, these

public actions are often recorded in history through police reports, membership lists,

journalists' descriptions, and manifesto minutes (Scott, 1985). This is not to say that overt forms

of resistance are irrelevant, but covert forms are overlooked because they are often perceived

as individualistic, even though they vary and are not publicly and politically comprehensive

(Scott, 1985). According to Scott (1985, p.33):

"much of the actions of subordinate groups fall into the category of everyday

forms of resistance, that these activities should most definitely be considered political;

that they do constitute a form of collective action, and that any account which ignores

them is often ignoring the most vital means by which lower classes manifest their

political interests".

Scott's ideologies were not without criticisms, considering the technological

advancements that modify surveillance notions such that several actions of the hidden

transcripts could be uncovered. However, it is essential to highlight critical criticisms that

modify the notion of everyday resistance and how they may apply to workers' agency in the

gig economy.

2.7.1.1. Recognition, Intent and Power as Critical Components of Everyday

Resistances

Following a critical review of resistance studies, including Scott's concept, recognition

and intent are central challenges that generally underpin difficulties in classifying resistance

(Hollander and Einwohner, 2004). This perspective is particularly useful in this study to reflect

on the researcher's positionality on what counts as everyday resistances in examining the

working lives and interviews of platform drivers in Lagos. For Hollander and Einwohner,

recognition of resistance hinges on both the observers (e.g., members of the media, researchers,

onlookers at the time of resistance and the public), targets (e.g., the recipient of the resistance)

and agents (e.g., actors). This classification signals observers' awareness and unawareness by

targets in Scott's notion of everyday resistance. On this note, what a researcher may qualify as

an act of resistance would yield different interpretations for resisters and targets. For example,

a platform Deliveroo rider may disagree that manipulating the app is a way of 'resisting'

algorithmic control when asked by a researcher. However, the target, which is the platform

company, could be unaware (partially because of embedded surveillance) or completely

unaware when workers circumvent the app. The scope of 'intent' is based on three primary

points (Hollander and Einwohner, 2004). First, an actor intends to 'resist' regardless of the

scope or outcome that qualifies as resistance. For example, gig workers deliberately create

manipulation actions to subvert algorithmic control in atomised groups and collectively,

indicating an action of intent, as this thesis shows in chapter six.

Secondly, assessing resisters' intentions is difficult or impossible because resisters may

lack the motivation or terms to explain resistance processes to observers. For instance, just as

Scott (1985, p. 301) states, a thresher's need for rice may lead to stealing paddy, implying

precedence over his employer's formal property rights. Finally, Hollander and Einwohner

(2004) highlight that the resister may not be conscious of their act of resistance, which makes

their intentions vague to understand a specific action as resistance. However, everyday

resistance's dynamic nature can signal a combination of one or more intentions and hostile class

interest or a political-ideological intention (Lilja and Vinthagen, 2009). In that sense, following

a desire, solving a practical problem, and surviving, amongst other characteristics, might be an

actor's intention (Johansson and Vinthagen, 2016). These insights are critical for this research

in understanding platform workers' intentions and the observer's positionality.

Scott's notion of everyday acts or practices of resistance is useful in unpacking gig

workers' actions to resist power and control from platform companies like Uber and Bolt,

discussed later. However, Johansson and Vinthagen (2016) argue that Scott's notion passively

recognises power dynamics as a relational concept while labelling too many expressions as

resistance. A similar argument for Zemblayas (2013) questions what actions constitute

'opposition' as Scott invariably outlines several acts of resistance practices. Johansson and

Vinthagen (2016) analyse the concept of everyday resistance from Hollander and Einwohner’s

(2004) propositions, where they examine resistance as both oppositional and an activity.

Everyday resistance for Johansson and Vinthagen (2013) can be contradictory for all subaltern

groups at the same time because it is seemingly involved in power relation discourses more

than open resistance. The choice to resist is available to all subaltern subjects – those who

choose to resist, assume a position of dominance in opposition to some domination systems

(Johansson and Vinthagen, 2016). The choice to resist or not can affect power by creating new

forms of power or strengthening old ones. Johansson and Vinthagen (2016) argue that

resistance is a combination of interaction, context, and subjectivity – its emergence does not

hinge on the subject but on specific actions. In a Foucauldian tradition:

“..the exercise of power itself creates and causes to emerge new objects of

knowledge and accumulate new bodies of information…the exercise of power

perpetually creates knowledge, and conversely, knowledge constantly induces effects

of power” (Foucault, 1980, pp. 51-52).

This ideology is why Johannson and Vinthagen (2016) argue that resistance progresses

beyond recognition and intent but is a kind of act or practice (according to de Certeau, 1984)

enmeshed in power discourses that transcend different contexts. Everyday acts of resistance

and control from platforms inherently display power dynamics, as is discussed in subsequent

chapters. For example, platform algorithms that show asymmetric information to drivers give

them the power to control and manipulate drivers. This thesis will show that it becomes a power

tussle when drivers intend to manipulate or game the system in reverse.

2.7.1.2. Integrating Spatial and Temporal Dynamics

The spatialisation and temporalisation of everyday resistances are integral, even more

so because of how workplaces or gig work have become increasingly mobile and exist via

digital spaces via platforms. According to Johansson and Vinthagen (2016), sites are locations

or social spaces such as workplaces, cities, streets and increasingly digital spaces like social

media platforms (see Lu and Steele, 2018; Locke et al., 2018), where everyday resistance can

persist. For instance, drawing from the surveillant assemblage developed by Haggerty and

Ericson (2000), this thesis critically examines informational flows in platform ecosystems and

how this both facilitates control and propagates resistances through space and time. Although

Scott does not explicitly identify spatiality as a critical aspect of resistance like de Certeau

(1984), for example, both hidden and public transcripts of subordinate groups took place across

space in the form of observing powerholders before taking action.

Space is a critical concept in terms of how gig workers subvert the power of algorithmic

management, which broadly cuts across digital and physical spaces of labour. Evans and Boyte

(1986) categorise physical spaces as interconnected, fragmented, and organised places.

Physical spaces comprise the materiality of the city that shapes cities, including meeting places,

buildings and road infrastructures, mobility infrastructures and other components that make up

the physical world. Digital spaces here build on the conception of cyberspaces by Chin and

Mittelman (1997), based on the internet and the world wide web. Digital spaces are based on

big data, the internet and algorithms, which create virtual environments of power through the

web or platforms for its users.

The intersection of both digital and physical spaces is what Polletta (1999) identifies as

free spaces, namely settings outside the control of dominant groups, overcoming

individualisation of labour or smaller intimate groups for safer expressions, according to Scott,

which facilitate collective forms of resistance. Accordingly, the ideology of free spaces

facilitates the formation of movements through solidarity, networks, and skills against

dominant groups. Several authors, drawing from scholars like Lefebvre and Deleuze,

conceptualise spaces of resistance in different ways that clearly explain how resisters aim to

regain control. These free spaces can become protected or conceptual spaces beyond the control

of algorithms based on the knowledge of the resisters. As Polletta (1999) argues, free spaces

preserve and build on collective records of resistance instead of directly penetrating obvious

sources of domination which, according to Scott (1990), is a performance to deceive dominant

groups. As Chapter six shows, social media and communication platforms became part of free

spaces, facilitating resistance practices for drivers.

Scholars such as Kinsley (2014), Leszczynski (2015) and Malecki (2017) have further

sought to complicate the arguments on what constitutes digital and physical spaces. They

deviate from previous accounts that distinguish between the digital and physical spaces but

demonstrate that these notions are enmeshed due to the co-constitution of humans, technologies

and infrastructures, which are not binary. For Kinsley (2014), the epistemological viewpoint

of examining this in-between is based on what is identified as technics and transduction. To

explain technics, Kinsley (2014) first draws from code/space by Kitchin and Dodge (2011) to

demonstrate that technologies possess the power to make things happen in the real world on

their own and collaboratively with the human body. For example, the upward trend of artificial

intelligence and robotic engineering is facilitating the independence of technologies based on

scientists’ knowledge and intermittent control. Second, Kinsley thinks of technicity as “ways

in which technology and humans co-constitute each other in an ongoing associative milieu”,

where we get to understand our very being and technical supports such that there is a

dissolution of the space in-between (Kinsley, 2014).

In making arguments that converge the digital and physical spaces, Leszczynski (2015)

proposes spatial media as media based on the “information artefacts of geoweb and the

hardware/software of objects, continue to be driven by technics and pragmatics which are

embedded in Web 2.0 architectures, interactive frameworks, business correctives, and their

extension to the spatial domain”. The argument for technics is even more profound with the

introduction of Web 3.0 and its affordances that further blurs the lines between the digital

spaces and physical spaces with core examples such as blockchain technology and the

metaverse. According to Leszczynski (2015), engaging spatial media as media

epistemologically avoids the pitfalls of dividing virtual spaces and physical spaces as

previously conceived. Instead, Leszczynski (2015) proposes mediation as an ontological

framing entwined in sites of multiple conjunctions of code, technologies, space/place and

content facilitated by the practices and spaces of everyday life. Similarly, Malecki (2017)

argues that cyberspaces can be understood through infrastructures and cyberplaces (e.g., data

centres, tubes), code/space (e.g., systems that depend on energy, telecommunication and

others) and cyberscapes (e.g., location-based services which have enmeshed online activities

to the physical world). This also integrates Kinsley’s idea of technics and the co-constitution

between technology and humans, based on the ubiquity of the internet, which has demonstrated

interdependencies between real and virtual places (Kinsley, 2014; Malecki, 2017). This idea of

virtuality and immateriality also agrees with the idea of free spaces and in-between spaces

examined by Briziarelli (2019). According to Briziarelli (2019), Deliveroo riders in Italy

activated in-between space through ‘log-in’ and ‘log-out’ strategies of protests, which this

thesis builds on in chapter six.

If there were seamless communication affordances at the time of Scott's study, more

sophisticated practices could have emerged along with better access to knowledge of resistance

practices. Adas (1986) argues that the lack of technology and communication structures, low

population-to-land ratios and overall tighter bureaucratic control made it difficult to obtain the

records of everyday resistances in pre-colonial South and Southeast Asian regions. Even with

technology improvements, subordinate groups or gig workers devise practices for subverting

algorithms without clear evidence of their actions, especially for non-gig workers.

The notion of time is equally important in defining the 'everyday' in resistance.

Johansson and Vinthagen (2016, p. 427) highlight that the spatialisation and temporalisation

of social reality are not separate processes but are socially constructed, entwined, and related

to each other for the advancement of power discourses. In the sociology of work, 'time',

particularly relating to labour and workers' control, is integral. The symbolic meaning of time

is money which promotes power discourses when worker unions bargain for higher wages if

employers push for increased performance and productivity without increased pay (Adam,

2003). Similarly, temporal dimensions concerning income (work that does not commensurate

with the time spent on a gig) is a crucial argument for gig workers. The best representation for

incorporating time as a dimension of everyday resistance is evident in the surveillance research

of call centre work discussed previously. Snider (2000; 2002) argues that while time spent by

employees trying to evade work is a crime that affects capital accumulation, employers are also

guilty of impinging on workers’ time without adequate compensation.

Snider (2002) describes time theft as a Taylorist tradition of productivity and time

management discourses present in contemporary workplaces because of ideological and

technological advancements. These technological advancements facilitate improved

monitoring capabilities to prevent workers from 'stealing time' while upholding control.

However, researchers like Stevens and Lavin (2007) examine how call centre workers, for

instance, manage to steal time by inventing techniques like 'double wrap time' and 'rolling the

queue', amongst others. For example, workers hit the 'wrap up button' twice or more between

calls to prevent new callers from joining the queue. Workers are aware of monitoring

capabilities and choose to manipulate the system by stealing time, even to the detriment of their

jobs.

2.7.1.3. Cumulative Mobilisation of Everyday Resistances: Significance of Contexts

It becomes evident that Scott's concept of everyday resistance delineates between

public transcripts or overt forms of resistance and hidden transcripts or covert forms of

resistance. However, it does not categorically explain how these everyday hidden practices

may develop into public transcripts through an unnoticed challenge of power, particularly in

specific contexts like the GS. Based on the study of Latin America on armed radical

movements, religious base communities and squatters, Gutmann (1993) argues that it is

difficult to conclude that everyday resistances are always hidden. Even though everyday

resistances outlined by Scott were critical for subordinate groups in Malaysia, Gutmann’s

criticism outlines the importance of contexts which is lacking in Scott’s assertion. Further,

according to Gutmann (1993, p.87):

“Scott’s theory of resistance is ultimately and ironically a conservative one. It

does not expect or explain change. Rather, it tends to reduce social consciousness to

the acceptance of a thoroughly tragic interpretation of contemporary reality and is in its

own way as profoundly romantic as any theory it challenges.”

The context, in this case, is relevant because the interpretation of resistance to a

particular culture can be alien and thus misinterpreted by foreign observers (Hollander and

Einhowhner, 2004). On this note, Bayat (1997; 2000) criticises the notion of everyday

resistance as lacking the heuristic of analysing the urban poor's dynamics and realities in the

GS. According to Bayat (1997), the urban poor are those without direct access to resources for

their everyday survival, including the unemployed, street subsistence workers, petty thieves,

members of the underworld (e.g., beggars, prostitutes), housewives and other marginalised

groups. At least for many, the gig workers in developing countries were previously

unemployed or underemployed such that platform gig work is either a primary source of

income or an additional income source (Carmody and Fortuin, 2019).

The choice to assume the status as an independent contractor is merely a consequence

of opaque formal organisations and inefficient state bureaucracy or algocracy, forcing people

to seek informal and autonomous working conditions. Even in these informal working

conditions, independent contractors can individually or collectively navigate algorithmic

constraints in a hidden but public manner. Bayat (1997) argues that subordinate groups'

practices against powerholders are not always defensive, individualistic, hidden or quiet.

Instead, these practices are secretly offensive, i.e., disenfranchised groups place a great deal of

restraint upon the dominant groups' privileges, allocating segments of their life chances

(including the capital, social goods, opportunity, autonomy and thus power) to themselves

(Bayat, 1997, p. 56). 'Action' as a crucial intentional element for Scott ignores the intended and

unintended consequences of different types of individual practices that fail to correspond

(Bayat, 2000). For example, gig workers do not manipulate the app in defiance of algorithmic

rules. Instead, they do so because they need to regain some control and bargaining power

against algorithmic opacity.

On this note, Bayat (1997; 2000) enhances the notion of everyday resistance in the quiet

encroachment – characterised by visible and transient struggles amongst predominantly

smaller units of collective action, lacking any coherent ideology, organisation or leadership.

This phenomenon Bayat defines as “silent, protracted, but pervasive advancement of the

ordinary people on the propertied and powerful to survive and improve their lives” (Bayat,

2000, p. 545). Here, quiet encroachment indirectly embraces coping and accommodation

notions, which are not exact opposites of everyday resistance, but reproduce power through

principles of cumulative practices (Johannsson and Vinthagen, 2016). According to Weiz

(2001, p. 670), the notion of accommodation refers to “actions that accept subordination by

either adopting or simply challenging the ideologies that support subordination”. To an extent,

Scott (1985) addressed 'quiet encroachment' elements when analysing disguised trade unionism

as a form of everyday routine resistance, showing how labourers strategically improve their

bargaining powers against wealthy farmers. Accordingly, the head of a workforce, such as

harvesters, would occasionally grumble – 'your paddy is hard' 'we are losing money', to signal

the farmer's attention (Scott, 1985, p. 259). If the farmer has an excellent reputation for paying

workers, the workers may continue to work grudgingly. If the farmer (particularly a stingy

farmer) ignores the signals, the workers will strike individually and collectively until the farmer

meets their needs.

Bayat (1997) also opines that relatively short mobilisation processes as a collective in

everyday resistance affect the fellow poor compared to prolonged processes of cumulative

encroachment through the struggles and gains of agents without detrimental impacts on

themselves. Focusing on the situated techniques of the acts is critical in understanding everyday

resistance, which should be supported by recognising the intentions of the observer (Hollander

and Einwohner, 2004).

Following Adas (1986), predictions of improved communications and technology

being an integral aspect of changes in resistance practices, platforms, and algorithmic

management for gig workers present a critical case that shows gradual metamorphosis from

hidden to public forms of resistance. The discovery of resistance practice which Scott would

classify as hidden is no longer easy to keep hidden based on the ability of algorithms to learn

the behaviours of gig workers. This beckons a more flexible approach to observing workers’

resistance because subordinate groups possess a repertoire of strategies that change in response

to dominant strategies (Adnan, 2007). The flexibility in understanding workers’ strategies

further reflects Lilja and Vinthagen’s (2018) notion of dispersed resistances and Shalhoub-

Kevorkian’s (2012) E-resistances because of the irregularity of how everyday resistances

occur; they may be violent or nonviolent, hidden or public, organised or disorganised and

relatively small-scale. Lilja and Vinthagen (2018) argue that dispersed resistances are more

complex than everyday resistances; accordingly, everyday resistances are only a small aspect

of individual or small-scale resistances. As such, "small-scale resistance is not necessarily

subtle, and public resistance is not necessarily part of sustained collective action, or even

organised in connection with others at all" (Lilja and Vinthagen, 2018, p. 225).

However, despite their approach improving Scott's notion, this thesis chooses to

identify how hidden and public resistances change through time and space and how everyday

resistances are dispersed or concentrated in the GS (occurring in different places). Considering

that Scott's ideologies were not without criticism, this section has helped discuss some of the

lapses in the everyday resistance concept and how one can integrate more contemporary

understandings that have emerged with technological advancements. The following section

attempts to integrate the aforementioned developments to fashion a framework for resisting

platform control.

2.8. The Limitation of James Scott's Concept in Lagos

The implication of Scott's concept of everyday resistance to Lagos is simply framing

the gig workers' everyday struggles and strategies to resist the impacts of algorithmic

management. One might argue that the context between subordinates and powerholders was

limited to peasants and farmers in an Asian context. These concepts allowed Scott to learn

about the complex relationship between peasants and farmers, including why power exists

between them, how power is reproduced, and how power is undermined or evaded. While the

context of Lagos and the period are different, it is still relevant from an epistemological

standpoint to examine the working relations between a driver and digital platforms, especially

as algorithms invisibly intermediate the relationship. Drawing from the hidden and public

transcripts enabled this research to categorise how and what drivers in Lagos are publicly and

privately doing to resist algorithms. It is a complementary concept to algorithmic management

to unpack the nuances of driver resistances in Lagos and how these resistance strategies might

differ in different geographical contexts.

As highlighted in the previous section, Scott's theory is limited in time and scope

because it does not critically envisage a period where technology will complicate the working

relationship by creating invisible layers of control by powerholders, which expands the

repertoires of resistance for subordinate groups (Johansson and Vinthagen, 2016). As

discussed in chapter six, one disadvantage in adopting the hidden and public transcripts is that

the lines between them become blurred because of algorithms and surveillance capacities. The

risk in applying the concept is that an identified resistant practice today may become a public

practice or no longer exist tomorrow. Based on the temporality of driver resistances and

platform counter-resistances through algorithmic management, Scott's concept could not fully

grapple with the dynamics. Scholars such as Adas (1986), Adnan (2007), Shalhoub-Kevorkian

(2012) and Lilja Vinthagen (2018) have integrated technological perspective into their

understanding of everyday resistances to capture better the temporalities of the hidden and

public forms of resistance from subordinates and counter-resistances from powerholders.

This also means that there are differences in resistance strategies between GN and GS contexts

because of varying policy landscapes, technology awareness of drivers, business environment

and competitive advantages by platforms like Uber that emerge in GS contexts.

Because these platforms are ideologically constructed in the GN and brought into GS,

there is a lack of integration of Lagos's contextual specificities. However, drawing from the

algorithmic management concept also complements the shortcomings of Scott’s concept

because it facilitates worker agency, as this thesis will show in chapters five and six.

2.9. Conclusion

This chapter has set the tone for understanding how algorithmic management operates,

drawing from the surveillant assemblage for controlling workers. The review showed how

algorithms across global platforms in online gig work and location-based gigs impact workers

similarly while utilising notions of flexibility and autonomy introduced in the neoliberal era.

However, this is an autonomy paradox because algorithms monitor, supervise, and control the

labour process using gamic elements such as ratings, incentives, acceptance and cancellation

scores which are ambiguous to workers in terms of how they are constituted. The opacity

embedded in algorithmic management reinforces biases and information asymmetries that

facilitate a reduction in workers’ bargaining power which reinforces psychological coercion

and manipulation, self-discipline, and self-exploitation. The algorithms withhold information

to demonstrate functional power (Foucault 1995) – visibility of surveillance and control

apparatuses, but there is an inability to verify how these processes work. However, these hidden

channels enable workers to develop agencies by actively developing competencies from

collective sensemaking processes, circumvention, and manipulation to evade algorithmic

managerial control. These practices are not new because workers have always found a way to

regain autonomy and control over algorithmic management following the literature review.

Nevertheless, this chapter consolidates workers' actions as a framework for analysing

the everyday resistances of workers in the gig economy. The 'everyday’ in James Scott’s

resistance concept becomes essential to show how workers’ actions may be hidden or public.

More importantly, it emphasises the dimensions of space and time to ascertain when and how

hidden practices transcend into public realms of resistance and their intentions. Additionally,

it calls for reflectivity in terms of what workers intend as practices of resistance and what

practices are recognised by the researcher. This conceptualisation process progresses into the

next chapter to first capture how algorithmic management works within ride-hailing platforms

more broadly and specifically in GS contexts where studies are lacking.

3. Chapter Three: Research Design and Methodology

3.1. Introduction

In this era of digital gig platforms, algorithms as integral aspects of management and

control also create opacities for researchers. However, this led to a combination of qualitative

methods that facilitated the validity of this research. In researching the impacts of ride-hailing

platforms and algorithmic management, it was relevant to adopt a social constructivist

paradigm that interprets knowledge based on the multiple subjective realities of ride-hailing

platform drivers and pre-existing social actors within mobility gig work (Guba and Lincoln,

1994; Young and Colin, 2004; Ortom et al., 2014). This qualitative study uses a case study

approach to understand the operationalisation and emergence of platforms in Lagos, how

drivers are impacted by algorithmic management and their resistance practices against the

power of algorithmic management (Yin, 2003; Creswell et al., 2007). While Uber and Bolt are

representative ride-hailing platform cases in Lagos, they were critical as comparators to other

indigenous platforms and the realities of drivers in GN cities where they are dominant.

The fieldwork ran from September 2018 to January 2019; and from July to September

2019 following ethical approval from the University of Manchester. The data collection

methods included secondary and primary research methods, which were mutually reinforcing.

At the beginning of this doctoral programme in 2017, the secondary data sources were scarce

or non-existent, particularly regarding GS cities like Lagos. However, it was helpful to review

documents, database searches, journal articles, grey literature, online media like YouTube and

newspaper articles to establish a background for this research. Similarly, this was replicated in

relation to Lagos by reviewing archival YouTube videos and alternative media such as radio

programmes. This research utilised SSIs, FGDs and participant observations combined with

mobile and online observations for primary data collection. While using a case study approach

was dominant, participant observations, which involve mobile and online observations, were

critical in examining the everyday experiences of ride-hailing platform drivers across physical

and digital contexts, which capture critical ethnographic moments (Guba and Lincoln, 1994;

Rosenblat, 2018).

Using the six phases of thematic analysis by Braun and Clark (2006), the data collected

were effectively analysed by creating and reviewing codes reiteratively with my conceptual

model. The combination of methods enabled triangulation to improve the validity and

trustworthiness of the research (Nowell et al., 2017).

3.2. Research Paradigm: Social Constructivism

In this study, the research paradigm is attributed to social constructivism with a

relativist ontology and interpretivist epistemology stance. Young and Collin (2004, p. 375)

“propose that each individual mentally constructs the world of experience through cognitive

processes. It differs from the scientific orthodoxy of logical positivism in its contention that

the world cannot be known directly, but rather by the construction imposed on it by the

mind”. Social constructivism is enabled by social relationships, which are derived from

influences on individual constructions with varying meanings of the world (see Vygotsky

1978; Bruner 1990; Young and Collin, 2004).

3.2.1. Ontology – Relativism

As a variant of idealism, relativism posits that “reality is only knowable through

socially constructed meanings with no single shared realities, only a series of alternative social

constructions” (Snape and Spencer, 2003, p. 16). The relativist dynamic of social

constructivism defines the ontological philosophy of this research because of how this

worldview seeks subjective meanings embedded in multiple realities from the interactions with

different participants and the historical, social, and cultural norms that operate in their

individual lives (Creswell et al., 2007). This research seeks and constructs realities from

multiple participants that surround the emergence of ride-hailing platforms, the historical and

cultural norms of the management and control of pre-existing taxi labour, and an international

comparison of these phenomena and the everyday resistance to ride-hailing platform drivers’

practise.

In understanding the realities of ride-hailing platform drivers, Lagos's historical and

cultural norms are deep-rooted in transition dynamics such that the relative truths from

conventional taxi labour create a clear pathway for examining platforms and algorithmic

management of control. Guba and Lincoln (1994, p. 111) state that these “constructions are not

more or less true, in any absolute sense, but simply more or less informed and sophisticated.

Constructions are alterable, as are their associated realities”. Hence, the initial deployment of

the socio-technical transition framework and the MLP were critical in creating the awareness

that multiple realities of social actors exist across the socio-technical landscape (e.g., regulatory

policies and representatives), socio-technical regime (e.g., conventional taxis and drivers) and

niche innovations (e.g., platforms and drivers) (Geels, 2005). This was followed by adapting

the research along with the conceptual model of algorithmic management and the concept of

everyday resistance (Scott, 1985; Lee et al., 2015) to focus on the emerging realities of ride-

hailing drivers without displacing the realities from a subsequent regime.

3.2.2. Epistemology – Interpretivism

Interpreting the sparse and hidden phenomenon of ride-hailing platforms and

algorithmic management in the GN is an accurate way of constructing knowledge about the

world (Snape and Spencer, 2003). The nature of the epistemological constructivist approach is

that interpretations of reality are subjective – the construction of knowledge is only valid when

observed within a particular context from individual experiences (Mack, 2010). For instance,

a tree is not a tree without human beings assigning a name to it (Crotty, 1998). It is not

discovered; it is only constructed through human consciousness and the awareness of the world

(Scotland, 2012). According to Guba and Lincoln (1994), findings are created as the

investigation proceeds because of an active interrelationship between the researcher and the

research object, demonstrating a subjective and transactional perspective of constructing

knowledge. This further implies that the co-construction or transaction between the researcher

and participant mutually impacts each other (Snape and Spencer, 2003). For instance, interview

questions asked of ride-hailing drivers also developed consciousness and awareness of what

deserved more attention and how to understand the nature of algorithmic management. This

was evident with drivers in the FGDs realising new practices of resistance or how ride-hailing

platforms are actively manipulating them to continue driving regardless of their health and

wellbeing.

Positivist and post-positivist paradigms offer varying levels of dualist and objectivist

epistemologies such that findings are often restricted to quantitative methodologies (Guba and

Lincoln, 1994). Because the researcher perceives multiple realities from individual

perspectives in a new context, a quantitative approach may be too restrictive for new

knowledge creation. The realities that are currently known present restricted evidence and are

typical of a GN perspective. The methodology section will discuss the process through which

this would be achieved. Therefore, a qualitative research design is critical for this research.

3.3. Research Design: Qualitative Study

Scholars such as Preissle (2006) and Denzin and Lincoln (2011) have identified the

usefulness and shortcomings of a qualitative research study in critically acquiring information

from participants. The fluidity of qualitative design is why Denzin and Lincoln (2011)

acknowledge its broad nature in adapting different theories, paradigms, methods and practices

that it does not own. Its interwoven philosophies enable qualitative researchers to investigate

the 'how' and 'why' of varying people’s social constructions of the world, their experiences,

interpretations and assigned meanings of reality (Merriam, 2009). Despite the difficulty in

pinpointing a single definition, Denzin and Lincoln (2011, p. 3) recognise qualitative research:

…as a situated activity that locates the observer in the world, it is a set of interpretive,

material practices that make the world visible. These practices transform the world.

They turn the world into a series of representations, including field notes, interviews,

conversations, photographs, recordings and memos to self. At this level, qualitative

research involves an interpretative approach to the world. This means that qualitative

researchers study things in their natural settings, attempting to make sense of or

interpret phenomena in terms of the meanings people bring to them.

This perspective is known as the emic or insider's perspective compared to an etic or outsider's

perspective (Merriam, 2009).

According to Creswell (2009), the overarching reason for using qualitative research is

that an issue or a question requires exploration. It develops theories or supplements quantitative

and statistical methods that do not fit the investigation of an unclear problem. The overarching

reason for choosing a qualitative research design over a quantitative or mixed-method approach

is its ability to critically interrogate digital labour through the lens of ride-hailing platforms

from the experiences, interpretations, and realities of drivers in their natural setting. Despite

several critical reasons for conducting a qualitative study, one relevant to this study is its

importance in understanding the context (Creswell, 2009) and how drivers in Lagos interpret

the problem compared to other global settings, especially when there is little scholarship in the

Nigerian context. My perspective as a resident of Lagos reduces the etic perspective of the

realities of Lagos. However, it requires an intense process of inductive reasoning (Creswell,

2009; Merriam, 2009) from an emic perspective with drivers as the donors of information. My

role as neither a taxi driver nor a ride-hailing driver brings some biases, acknowledged through

reflectivity and interactions with insiders in the field.

Therefore, this study adopts a range of qualitative methods from an ethnographic case

study approach, interviews, FGDs, and participant observations divided into mobile and online

or virtual observation of driver experiences. As Denzin and Lincoln (2011) put it, the

commitment to using more than one interpretive practice in a study highlight that each practice

visualises the world differently.

3.3.1. Methodology: A Case Study Approach

In investigating the advent of ride-hailing platforms and algorithmic management of

drivers, this research adopts a case study approach with Lagos as the case setting. It is important

to note that this approach embodies the ideals of ethnography because it shows similarities with

what Creswell et al. (2007) consider an entire culture-sharing group. However, it is essential

to note that the case study approach is dominant according to the guidelines of exploring issues

within a bounded system such as a setting or a context, usually as a single case or multiple

cases. A case study can therefore be defined according to Creswell et al. (2007, p. 245):

As a qualitative approach in which the investigator explores a bounded system (a case)

or multiple bounded systems (cases) over time through detailed, in-depth data

collection involving multiple sources of information (e.g., observations, interviews,

audio-visual material, and documents and reports), and reports a case description and

case-based themes.

According to Yin (2003), the case study approach is a suitable methodology when

research questions start with "how?" or "why?" or "what?" the investigator cannot manipulate

unclear boundaries between the phenomenon and context; the behaviour of those involved in

the study; and finally, the researcher perceives that covering contextual conditions is relevant

to the phenomenon under study. In line with case study approaches outlined in the literature

(see Stake, 1995; Yin, 2005; Creswell et al., 2007), this research adopts a collective case study

approach because it investigates the algorithmic management of the labour process, which

causes resistance amongst ride-hailing drivers. Although a single case study approach can

similarly be adopted, investigating multiple platforms such as Uber, Bolt, and traditional taxis

indicates multiple phenomena within a single site with more depth.

Uber and Taxify (Bolt) serve as collective case studies going into the field because of

the dynamics in their operation, dominance and the competition existing between them,

explored in subsequent chapters. Oga-Taxi, which is an indigenous ride-hailing platform, was

to be included as a case study. However, platform drivers only occasionally utilised the app

compared to Uber and Bolt. This created an observation and interview question about why

Oga-taxi and other indigenous ride-hailing apps were less dominant than their international

counterparts. It facilitated the need to unravel the origins of ride-hailing platforms in Lagos

and understand why indigenous platforms were not dominating.

The city of Lagos represents the context or case setting through which a collective case

study on ride-hailing platforms was conducted. This is to create and establish a narrative for

GS contexts compared to GN contexts such as America and England, which denote the earliest

form of taxis and innovative developments. Although cities in the North provide an important

context for the emergence of taxis and ride-hailing apps, these innovations influence gig work

in the mobility sector in several GS cities. For instance, London is reviewed because it is one

of the earliest cities to claim and define personal mobility, which directly influenced Nigeria's

mobility sector. American cities are briefly reviewed to unpack the origin of ride-hailing apps

and their overall taxi regime impact. While the focus is Lagos, the knowledge of ride-hailing

platforms in GN cities is instrumental in identifying why these platforms are increasingly

globalised.

According to Yin (1984), the case study approach provides little basis for scientific

generalisations because of the limited number of subjects used during the study. It becomes

challenging to reach generalised conclusions because of the dependency on a single case

exploration (Tellis, 1997). Another criticism of the case study approach is that it takes longer

to conduct and produce a large amount of information, making it increasingly difficult to

analyse (Yin, 1984). However, this could depend on the type of case study, sample size and

structured questions. In addition, this research was conducted across six months, allowing

ample time for analysis.

3.3.2. The Study Area: Lagos as a Case Site

With a population of approximately 21 million inhabitants (Bloch et al., 2015), the city

of Lagos is a case site for studying Bolt and Uber drivers, considering it was a launching ground

for these platforms. Lagos state is one of the largest cities in the world, and a preeminent

megacity in Africa divided into five administrative divisions: Ikeja, Badagry, Ikorodu, Lagos,

and Epe (Lagos State Government (LSG), n.d). Lagos administrative division comprises five

Local Government Councils: Eti-Osa, Apapa, Surulere, Lagos Island, and Lagos Mainland

(LSG, n.d). While most of the field work took place in one of the administrative zones, Lagos,

workers' experiences cut throughout the metropolitan area of Lagos, as shown in Figure 3

below. Also, considering that ride-hailing platforms are not tied to place but embedded in apps,

riders requesting trips from drivers’ also cut across all the five administrative zones. Lagos

reflects where many GS Cities, especially African cities, would be in the next 20 to 50 years

in terms of innovation and population size. If international platforms continue to emerge in

many African cities, it could further regularise or formalise the unfair labour practices

embedded in its platformisation processes. This will further compound the existing weak

regulatory frameworks and harsh realities in places like Lagos by rendering the labour

processes and management to mere algorithms without regulatory interference from state

governments or platforms. Therefore, studying ride-hailing platforms in Lagos bridges the

existing knowledge gap and develops pathways for other cities in the GS, especially African

cities, to critically examine the impacts of the platformisation of labour and algorithmic

management as a weapon of control. Therefore, it offers a dynamic case for understanding ride-

hailing and could be seen as a point of reference for these cities because of the peculiarity in

the poor state of urban mobility and weak and corrupt regulatory frameworks.

Figure 3: Map of Metropolitan Lagos

Source: Lagos State Ministry of Physical Planning and Urban Development (2014); cited in,

Ojowolo and Wahab (2017).

Lagos State is the financial, economic and innovative powerhouse of West Africa and

home to domestic and foreign technological companies (LSG, n.d).

Ride-hailing platforms

such as Oga-taxi, Taxify and Smart cab have been rampant in the city since Uber's entrance in

2014. According to the Uber office's general manager in Lagos, Uber created 1000 jobs and

granted over 30% more rides than London in its first 16 months (Jean-Francois and

Veselinovic, 2016). Although that might have changed, Lagos serves as a case site because of

Uber and other indigenous digital technology platforms that continue to emerge in the city.

Part of what boosted the adoption of innovative technologies is the penetration of smartphones

in the state. For instance, Nigeria's smartphone penetration level has increased from 11.0

million users in 2014 to 23.3 million users in 2019 (Liu, 2015). An increase in smartphone

adoption is due to cheaper sales which dropped from $216 to $117 in 2014, increasing sales by

394% by 2016 (Smith and Tran, 2017). According to the National Bureau of Statistics, Lagos

accounts for 19.04 million out of 148.74 million voice and internet subscribers (cited in

Elebeke, 2016). Although there is no established correlation between smartphone penetration

and the adoption of ride-hailing platforms, it is evident that smartphone and internet penetration

is critical in accessing ride-hailing platforms in Lagos. In line with the city's innovative

alertness, this has led to visits investments from Silicon Valley investment groups such as – Y

Combinator and Mark Zuckerberg – to boost the tech industry in Lagos (Adegoke, 2016).

Investments have been made in tech start-ups called Andela and Flutterwave, with the latter

created to solve payment difficulties for platforms like Uber in African cities. This further

reinforces the rationale of using Lagos as a case site, with ride-hailing platforms fostering

innovative solutions for the efficient running of mobility systems and labour force

management.

Finally, urban mobility in Lagos is predominantly road-based, with over 90% of trips

dominated by high-capacity buses (Molues), taxis, private vehicles, motorcycles (Okada) and

auto-rickshaws (Keke napep) (LAMATA, 2015, cited in Oshodi et al., 2016). Besides the

private vehicles, BRT buses and Lagbus systems, many of these modes operate informally with

primarily unregulated travel fares and a lack of technological integration. As a comparator to

ride-hailing platforms, traditional taxis represent less than 1% of passenger traffic per day

because of issues with affordability, which makes them attractive only to middle- and upper-

class people (LAMATA, 2015, cited in Oshodi et al., 2016). According to the former head of

communications for Uber West Africa, Francisca Uriri, Uber's monthly riders were

approximately 267,000 and 9000 drivers (Vanguard, 2018). Although these are not accurate

representations of the daily ridership or drivers for ride-hailing platforms, it demonstrates the

penetration of technology in Lagos based on the increase in usage and why these platforms are

driving that change. More of the dynamics and rationale for studying ride-hailing platforms are

discussed in Chapter four.

3.3.3. Study Locations

Before embarking on this research, living in Lagos for over ten years influenced my

contextual knowledge of the city. Moving to Lagos in 2007/2008 from Delta State, about

420km apart, equipped me to pinpoint specific locations where conventional taxis were located

throughout the city. These included vacant parking spaces along the roadsides, airport car

parks, designated taxi parks (e.g., see Table 3), hotel and eatery environments, and other vacant

and busy corridors in the city. However, my frequent travelling out of the country for

educational purposes affected my ability to sustain my knowledge and evolving changes in the

city. More so, the ubiquity of ride-hailing platforms is such that drivers are not tied to locations

around the city but are embedded in the digital spaces via smartphone applications.

Notwithstanding, ride-hailing drivers roam the city and park in vacant car parks that are not

designated for conventional taxi drivers or restricted by traffic officials to rest or wait for trip

assignments.

In combination with my prior knowledge of the city, I conducted a google search to

pinpoint locations of ride-hailing platform offices, conventional taxi offices, popular vehicle

car parks, hotel environments and other locations listed below.

Table 3: Key Locations in Lagos

Location

Identity

8 Providence Street, Lekki Phase

Uber office in Lagos, 2018 and 2019

388 Ikorodu Road, Diroko Bus

stop, Maryland

Platform Union – National Union of Professional e-hailing driver-

partners (NUPEDP), now National Union of Professional App-

based Transport Workers (NUPABW)

Block 17B, Alausa, Secretariat

Road Ikeja

Lagos State Ministry of Transport

Plot 118, Block 25 Salem St,

Lekki Peninsula II, Lekki, Nigeria

Bolt Head office, 2018 and 2019

Plot 24, Block 113, Adebisi

Oguniyo Cres, Lekki Phase 1

Vacant car spaces along the roadside

Venia Business Hub, 8

Providence Street, Lekki Phase 1

Vacant car spaces along the roadside

350 Ikorodu Road, Maryland Mall

Uber's greenlight hub

Eko Hotel and Suites

One of the largest hotels on the Island side of Lagos. Both inside

and outside, there are both conventional yellow taxis and non-

coloured taxis, i.e., typical saloon vehicles

Guaranty Trust Bank (GTB) Car

Park, Lekki Phase 1

The vacant car park where platform drivers congregate from 9 pm

daily in 2018 and 2019

Four Points Hotel by Sheraton,

Victoria Island

One of the largest hotels on the Island of Lagos with mostly

neutral-coloured taxis parked in proximity; there is also a taxi car

park opposite the hotel

Oriental Hotel, Victoria Island

One of the largest hotels on the Island is about 550 metres from

the Four Points Hotel; neutral-coloured taxis are usually parked a

few metres from the hotel

Admiralty Way Lekki Phase 1

Daytona Car Park

Vacant car spaces along the roadside

Admiralty Way Lekki Phase 1

Tantalizers Car Park

Vacant car spaces along the roadside

Source: Author’s fieldwork (2018 and 2019)

These locations were vital to recruiting and observing the ride-hailing drivers offline

and the traditional taxis, such as the red and yellow taxis. It is also worth highlighting that these

locations also cut across metropolitan Lagos, despite ride-hailing apps not being tied to place.

The locations demonstrate residuals of platform, driver and regulatory participant activities in

the city.

3.3.4. Study Sample

This study comprised a total of 46 interviews and three FGDs, and one informal

discussion. As summarised in Table 4, the sampling of participants was in multiple stages, such

as through a snowballing technique and simple random sampling at different locations offline

or via online social media groups. The breakdown of these samples is adequately discussed in

subsequent sections.

This study involved human subjects, as stated in the Methods section below. It did not

involve drivers of personal vehicles, except when they also work with ride-hailing companies.

Taxi drivers were randomly selected from taxi ranks across the seventeen local government

areas in the city. This study does not include contract or rental taxis because it operates

differently from the typical ride-hailing or regular taxi service. Examples of contract and rental

taxis include Hertz rent a car service, Sixt rental, and other services often located within

Murtala Mohammed International Airport in Ikeja, Lagos.

3.4. Methods and Data Collection

A qualitative method was used in this study for data collection. The qualitative methods

comprise secondary and primary techniques. Before the fieldwork, desk-based research created

insights for primary research techniques such as SSIs, FGDs and participant observations

divided into mobile and online observations (see Figure 4).

Figure 4: Overview of Data Collection Methods

Source: Author’s fieldwork (2018 and 2019)

Desk Research

This creates insights and provides

access to people or settings with

knowledge about the research

Focus Group Discussion

Provides collective knowledge

about the subject of research

Semi-Structured Interviews

Provides critical individual

knowledge about the research

which corroborates or invalidates

FGDs and participant observations

Participant Observation

Mobile Observation

Observing driver actions;

experiencing ride-hailing platforms

from a rider’s perspective.

Online Observations

Observing the everyday experiences,

perceptions and struggles of drivers on

social media networks (e.g., Facebook)

to invalidate or substantiate other

methods.

3.4.1. Desk Research: Document Review and Searches

In Manchester, the initial document review was generally drawn from transport policy

documents, articles, and reputable media reports, especially in GN contexts like London and

America, because of the prevalence of existing and emerging innovative platforms in these

contexts. Online databases utilised in this study included Scopus, Google Scholar, the

University of Manchester library online database, and Google searches for clarifying details,

especially within the Lagos context where data is scarce. Keywords used in the early stages

and progressively included: ride-hailing, ridesharing, mobility, uber, socio-technical

transitions and the M.L.P. (Multi-level perspectives), scientific management, labour and

control and other concepts that emerged within the analysis.

The broader logic, in the beginning, was to understand the transitions of ride-hailing

platforms using socio-technical transitions because ride-hailing transitions, according to the

framework, were considered niche innovations capable of disrupting the taxi regime as well as

developing new rules and regulations in the landscape level (Geels, 2002; 2005). The rationale

for these reviews was to pinpoint a transition from taxis to ride-hailing platforms in cities like

London and how these impacted drivers and the overall mobility sector. While the focus was

on Lagos, the lack of comprehensive information on ride-hailing and platformisation in GS

cities redirected the review to GN cities like London and cities in America because there was

more published evidence about ride-hailing platforms in comparison to Lagos. The use of grey

literature and reputable media outlets such as The Guardian, Vanguard and technology news

outlets were used to review some of the operations of ride-hailing platforms in Lagos.

However, these outlets were not substantial enough to deduce these platforms' reality in Lagos.

Global reviews on ride-hailing platforms were beneficial in depicting how they might operate

in Lagos before going into the field. Following my fieldwork exercises that spanned five

months, I adapted the research to Lagos.

While the socio-technical transitions framework was critical in determining the depth

of research participants, it was not substantial in interrogating these transitions' hidden

dynamics. Keywords developed into: algorithmic management, algorithmic power, gig work,

algorithmic surveillance, digital platforms, platformisation of labour, ride-hailing platforms,

E-hailing taxis, the gig economy, platform economy, resistances, everyday resistances, digital

space, taxi labour. The impacts were more algorithmic and political based on drivers' stories

and experiences, leading to everyday resistances discussed throughout this thesis. The

University of Lagos (UNILAG) library provided a few books and ideas for articles on public

transport history in Lagos. For instance, an essential book, Infrastructure Development and

Urban Facilities in Lagos, 1861-2000, by Ayodeji Olokoju, a Professor of History at UNILAG,

was integral in deducing the emergence of taxis in Lagos and comparing the similarities and

differences of driver struggles compared to those of ride-hailing platform drivers. The Nigerian

Bureau of Statistics (NBS) and policy documents provided by the Lagos Metropolitan Area

Transport Authority (LAMATA) were also integral in presenting facts and information about

the previous and current existing mobility industry. Also, news outlets from Vanguard, The

Guardian, Quartz, BBC, and tech sites such as Techcrunch were integral in gaining knowledge

about taxis and ride-hailing platforms. These were also helpful in triangulating interviews from

both taxi and ride-hailing platform drivers and reviews from online driver forums.

3.4.2. Data Collection Process and Participant Recruitment

This study used FGD and SSI as critical data collection methods. This research utilised

FGDs to create ideas and themes for an initial interview guide for the study. The participants

involved were platform drivers, taxi drivers, policymakers or regulators, riders, and university

lecturers. Table 4 provides an overview of the participants of both SSIs and FGDs. Each section

would also incorporate participant recruitment and the challenges that ensued.

3.4.3. Criteria for Including Participants

The general criteria for including participants were their willingness and knowledge of

the subject area. Additional criteria for ride-hailing drivers were that they had to be drivers who

possessed smartphones, were fully registered and had passed vehicle inspections and possessed

an understanding of the job. In the beginning, the focus was solely on Uber or Bolt drivers who

were willing to be interviewed. However, in the field, it was clear that selectively picking

drivers based on the Uber or Bolt preference would deviate from a random sampling method

because most drivers register with multiple platforms. The selection facilitated choosing

knowledgeable drivers such as Charles, George, Efe, Henry, Dipo and Koffi, who understood

the framework of traditional taxis, unions and emerging ride-hailing platforms that offer gig

work. It is important to note that before any SSIs or FGDs were carried out, my participant

information sheet and consent form were shared with drivers before interviews (see appendices

B1 and B2). When drivers did not read the consent form or participant sheets before the

interview, I read them orally before beginning the session.

In that search for knowledge, other groups such as lecturers, riders, venture capitalists

and policy representatives were critical because of their in-depth understanding of Nigeria's

overall mobility sector. Table 4 below gives an overview of the criteria for recruitment for each

participant in this study.

Table 4: Overview of Semi-Structured Interviews, Focus Groups Discussions and Recruitment Strategies

Participants

Semi-

Structured

Interviews

(SSIs)

Recruitment Strategies and Selection Criteria

Focus Group

Discussions (FGD)

Observations

Platform Drivers (Uber

and Bolt) (Including

Platform Union)

Recruitment Process

- Snowballing off people who were platform

drivers such as Charles, George, Dipo, Henry, Efe

and Koffi

- Randomly selecting drivers from popular places

where they park, e.g., Lekki Phase 1 G.T.B. Bank

- Booking rides and collecting Uber driver's

contact for an interview

Selection Criteria

- Willingness to undertake an interview; when

3 FGDs

- Observed what drivers

do to get trips, including

parking in areas with

high activities

- How they make

decisions concerning

long or short distances

- I visited some of the

areas where they parked

overnight to observe

how they anticipated

trips for the next day

they refused recorded interviews, I carried on

without recording

- Being approachable and appropriately dressed

- Drivers who were expecting a trip or are in rest

mode for the day

- I took 40 trips with Uber

and Taxify (Bolt)

between October 2018

and September 2019 to

observe drivers from a

user perspective

Local taxi drivers

Recruitment Process

- I visited three major taxi parks, e.g., Falomo

under bridge taxi park, Eko hotel taxi park

V.I, and Admiralty Way taxi park

Selection Criteria

- Drivers who were assigned to a cab rank/park

were interviewed

- Willingness to undertake an interview

- Drivers who had been in the business for over

15 years

1 Informal

discussion

(not counted

as FGD)

- I visited taxi ranks to

observe the current

nature of the service and

the level of patronage

from potential

passengers

Passengers/Riders

Recruitment Process

- Snowballing from friends of friends

Selection Criteria

- People who possessed a smartphone and had

- Interviews were often

passively conducted

because I contemplated

either Uber or Taxify (Bolt) on their phones

- Riders who had used either Uber or Taxify at

least once

- Willingness to undertake an interview

adding riders to my

sample

- Interviews were

conducted via mobile,

and riders were

compensated for their

time

Transport Industry

Representatives/Ministry

Recruitment Process

- Snowballing from my network, e.g., a lecturer

from the University of Lagos, connected me to

two policymakers in December 2018

Selection Criteria

- Knowledge of the urban transport system and

awareness of the current trajectory

- Willingness to undertake an interview

Lecturers

Recruitment Process and Selection Criteria

- Snowballing from personal networks to contact

University of Lagos (UNILAG) lecturers who

were central in providing access to policy

representatives and more insight into my research

- Lecturers were key

informants to transport

ministry representatives

Ride-hailing Companies

Recruitment Process

- Trial by error; for instance, I visited three ride-

hailing companies, and I was granted only one

interview from Max.ng – a motorcycle taxi

company; Uber and Taxify refused interviews,

although Taxify granted me a drivers' training

session

Selection Criteria

- Companies that use ride-hailing technology;

three companies were selected – Uber, Taxify and

Max.ng

- Willingness to undertake an interview

- Uber and Taxify were

the prominent

companies observed but

refused access for

interviews

- I attended a Taxify

driver training for

approximately 4 hours

in November 2018 to

gain insight into their

operation; I observed

the recruitment and

onboarding process for

drivers

Other Start-

ups/Correspondents

(Venture Capitalist)

Recruitment Process

- Recruited based on an online article on the scale

of technology change in Lagos on

Nairametrics.com (contacted the journalist

directly for an interview)

- I interviewed a venture

capital start-up company

that is directly involved

in transport affairs,

especially new forms of

transport like Uber

Source: Author’s fieldwork (2018 and 2019)

Selection Criteria

- Willingness to undertake an interview

- Knowledge of the subject area

Total

Total of 24

participants

(including ten

already in SSI)

3.4.4. Focus Group Discussion (FGD)

Longhurst (2016) argues that SSIs and FGDs are similar because they allow the

interview process to adopt a conversational style instead of a 'yes' or 'no' answer compared to

a structured interview. The similarities between these methods enable the researcher to

understand the participants' values, perceptions, and experiences (Nyumba et al., 2018). The

difference with FGDs is that they involve between six and twelve people and last between one

to two hours to discuss a particular topic (Longhurst, 2016). Over the past two decades, the

number of participants for FGDs has typically ranged from six to 12 members depending on

the topic of research (Fern 1982; Wilkinson, 1998). The rationale for this research was to

generate ideas with a minimum of five participants per session that guided the development of

more research questions with emerging themes.

In this study, FGDs were integral for establishing ideas and themes that facilitated all

participants' SSI guides (Appendix C1). This is because of a lack of clarity about how ride-

hailing platforms operate, drivers’ experiences and perceptions of algorithms controlling the

labour process and overall lack of knowledge about the gig economy in Lagos. FGDs were

relevant in emphasising Lagos' culture and broad specificity compared to London and America

– where the innovation for such platforms emerged. There were three FGDs and one informal

discussion, which are outlined in the section below.

3.4.5. FGDs for Platform Drivers and Taxi Drivers

The first contact with ride-hailing platform drivers was Henry and Efe's two friends,

both Uber and Taxify (now Bolt) drivers, respectively. Henry, a customer care representative

at a commercial bank, and Efe, a full-time platform driver, were my high school classmates at

the Delta Steel Company (DSC) and Technical High School (THS) in Warri, Delta State. In

search of greener pastures, Henry and Efe moved to Lagos in 2013. They introduced me to the

everyday occurrences of ride-hailing platforms in Lagos, including popular locations where

other drivers waited for trips in September 2018. With this, I could visit vacant parking spaces

that were relatively safe with detailed security personnel at night, usually between 9 pm and

midnight. The Guaranty Trust Bank (GTB) office at Lekki Phase 1 Lagos was a parking space

by day for the bank staffers and visitors and a parking and residential space for ride-hailing

drivers at night (see Table 2). On the first occasion, I met Charles, a popular Bolt driver who

formerly drove for Uber. I introduced my motive, and we had an informal conversation about

their work in October 2018. Being the head administrator of a WhatsApp group for platform

drivers, Charles was very knowledgeable about the phenomenon of ride-hailing. He was aware

of the job's challenges, other drivers’ perceptions, and hacks to excel at being a ride-hailing

driver. Despite being a bachelor's degree holder, the country's economic difficulty had forced

him into the ride-hailing business. This was the reality for many drivers in this study.

In this regard, Charles became the gatekeeper by helping recruit drivers who became

integral for two separate FGD meetings in October and November 2018. Both FGDs comprised

a total of fifteen drivers working for Uber, Taxify and Oga-Taxi (see Table 5). The FGD in

October 2018 comprised eight participants, while November comprised just six participants.

Table 5: Focus Group Participants, October and November 2018

S/N

Gender

Pseudonyms

Participation Status (Very

Active, Active and Passive)

Abiodun

Passive

Charles

Very Active

Junior

Passive

Alex

Active

Dayo

Passive

Manix

Passive

Jude

Very Active

Henry

Passive

Efe

Active

Felix

Active

Akin

Active

Timothy

Active

Samson

Very Active

Emma

Passive

Okoro

Passive

Source: Author’s fieldwork (2018)

In the two FGD meetings, as shown in the table above, some drivers were active and

some passive. The classification of very active, active or passive in these two FGDs meant that

some drivers contributed more to the discussion, while others only contributed by affirming

their colleague's views, giving chorus answers or just observing the session.

Before proceeding to the third and final FGD, I must highlight my meeting with

traditional taxi drivers in August 2019 as an informal discussion and not an FGD meeting. An

interview with the Victoria Garden City (VGC) taxi park chairman, Isaac (August 2019), led

to an invitation to one of their general meeting days with other union members. This was their

monthly meeting which they held on the last Thursday of every month. The purpose of this

meeting was to acquire their opinions and perceptions of their struggles, considering the

existence of ride-hailing platforms in Lagos. The meeting comprised twelve members in a

small room by the car park. It was a short discussion meeting of about 15 minutes because they

needed to discuss other private issues in my absence. Taking pictures and initiating audio

recordings were prohibited. However, I was able to take down some notes in the notepad of

my mobile device. In considering the meeting length, it was challenging to identify all the

drivers and create a table with pseudonyms like the figure above. While I do not classify this

as an FGD meeting because of the length of the session and difficulty regarding full access, it

was crucial in understanding the critical role of unions for taxi labour.

Table 6: NUPEDP FGD

S/N

Gender

Pseudonyms

Participation Status (Very

Active, Active and Passive)

Dipo

Very Active

Onome

Passive

Salifu

Passive

Emma

Active

Shade

Very Active

Thomas

Passive

Ejiro

Passive

Adamu

Passive

Ibukun

Active

Source: Author’s fieldwork (2019)

100

The third and final FGD session was targeted at platform taxi union leaders was held

on 27

August 2019 (see Table 6), who were also Uber and Bolt drivers. The session took

approximately 59 minutes, comprising nine NUPEDP union members, including the president

Dipo – a key representative and founder of the platform union, with other representatives

reiterating the struggles of ride-hailing gig work in Lagos and why establishing a union was

critical.

3.4.6. Semi-Structured Interviews (SSIs)

In qualitative and quantitative research methods, interviews are the most common and

effective means for facilitating data collection between a researcher and a participant (Lechuga,

2012). Interviews are broadly divided into three types: structured, unstructured, and semi-

structured (Dunn, 2005). While structured interviews are rigid and follow specific questions

with direct responses, unstructured and SSIs are flexible. This study adopts the semi-structured

approach because of the flexibility and ability of the participants to present new and relevant

narratives. An SSI is similar to an unstructured interview, except it covers a list of general

topics, is often open-ended, and requires an interview guide (Bernard, 2006). This kind of

interview is required when the probability of interviewing a participant more than once is

limited.

As previously mentioned, initial FGDs were used in designing a semi-structured

interview guide (see Appendix C2). This was relevant in conducting the interviews, which were

face-to-face and via the telephone. Face-to-face interviews were used in the initial stages for

platform drivers following participants from FGDs. It was also critical to interview riders,

policymakers, platform company representatives and local taxi drivers. Reflections from both

the first FGD and SSI led to telephone interviews because of the difficulty in accessing platform

drivers. Several scholars (e.g., Novick, 2008; Block and Erskine, 2012; Irvine et al., 2012) have

discussed and reviewed the challenges of phone interviews. Of relevance to this research,

Novick (2008, p. 395) classifies these under three broad challenges: “the loss of nonverbal data

(e.g., gestures and actions which are lost); loss of contextual data (e.g., being unable to see

where the interviewee is located); and the loss or distortion of verbal data (e.g., reduced rapport

or probing quality)”. Irvine et al. (2012) highlights the inability to detect emotion, naturalness,

level of interest, comprehension and attention during phone interviews. However, as a

supplementary technique to face-to-face interviews in this study, the limitations were less than

101

using phone interviews as a single SSI technique. Below I discuss how the SSI was used

amongst participants.

3.4.6.1. SSIs for Platform Drivers and Taxi Drivers

At the beginning of the interview process, it was challenging to recruit platform drivers

because cars are not labelled, and drivers are always on the move. Initially, the logical way was

to request any platform taxis, seek consent and initiate an interview during the trip. This

approach was similar to Rosenblat’s (2018) study in her book Uberland – initiating interviews

after ordering an Uber meant for a journey. According to the book, some drivers refused such

interviews. This approach for this study raised ethical concerns because drivers can get

distracted during trips. Also, a driver accepting a ride with the sole purpose of delivering a

service but meeting an interrogation could pose risks for the quality of the service and quality

of responses.

In October and November 2018, at the Guarantee Trust Bank (GTB), FGD meetings

were a good source of recruiting drivers for face-to-face interviews and telephone interviews.

Some drivers, such as Dayo, Okoro, Emma and Manix, were not interviewed because they

were unwilling to continue the process. Thirty-six drivers were interviewed, 25 were platform

drivers, and 11 were taxi drivers. The first ten interviews with ride-hailing drivers were face-

to-face, while the remaining fifteen were telephone-based interviews because of the challenge

of inaccessibility.

3.4.6.2. SSIs for Transport Policy Representatives

Transport representatives here represent policymakers within the state transport

ministry at Alausa, Ikeja Lagos. I was privileged to interview two representatives within the

ministry about the incumbent taxi regime and ride-hailing platforms. Both interviews with

Lateef (December 2018) and Mustapha (December 2018; January 2019; August 2019) were

integral in explaining the mobility industry's policies, including plans and struggles in terms of

regulating existing ride-hailing platforms in Lagos. They acknowledged the struggles of both

taxi drivers and platform drivers and, as a result, agreed to a tripartite meeting between

regulators, platforms and taxi unions (including ride-hailing platforms). The interview with the

transport policy director Lateef (December 2018) lasted for 21 minutes and was audio-

recorded. Interviews with the head of taxi operations Mustapha conducted in December 2018,

January and August 2019 were not recorded because the participant refused consent. However,

102

I was able to take notes with the participant's consent. The first interview session lasted

approximately 2 hours, and the second interview lasted for 28 minutes.

In my first fieldwork phase, the election tension restricted interviews with key actors

within the transport ministry because they were not accessible. With more clarity in my second

fieldwork, more interviews from transport representatives were not relevant. It became

apparent that the focus was on ride-hailing drivers and the platform industry. Information from

these two representatives was substantial in understanding the mobility sector's historic policies

and the ride-hailing industry.

Unions such as the National Union of Professional E-hailing Driver/Partners

(NUPEDP) and the National Union of Road Transport Workers (NURTW) were represented

during interviews with Dipo, and high-ranking taxi drivers such as the VGC taxi car park

chairman were interviewed. Also, concerns about the issues of violence and extortion within

the NURTW, as advised by Dipo and other taxi drivers, informed my decision against direct

engagement with the traditional transport union. However, information from the sources above

was substantial in understanding the institutional framework and policy guiding taxi labour in

the mobility industry in Lagos.

3.4.6.3. Riders

Interviews with riders were conducted in the first phase of my fieldwork. It is important

to note that riders in this study refer to passengers on ride-hailing platforms. This is because

drivers and platforms refer to passengers mostly as riders in Lagos. However, I mention

passengers in places where I highlight traditional taxi drivers. A total of five riders were

interviewed. While the later focus was on ride-hailing platform drivers, interviews with riders

helped corroborate some drivers’ accounts or establish a need for more investigations. For

example, an interview with Teni (November 2018) and Ewoma (November 2018) validated

some manipulation techniques, such as manipulating the app to increase fares during trips,

discussed in detail in Chapter six.

3.4.6.4. Platform Companies

Interviews with this group of participants were difficult because of their global policy

of not sharing information with third parties. In this case, four platform companies, Uber,

Taxify (now Bolt), Oga-taxi and Max.ng, were approached. The primary focus was on Uber

and Taxify (Bolt), the dominant platform companies in Lagos. It is important to note that in

103

February 2019, Taxify rebranded its name to Bolt because of the aim to expand its service

beyond ‘taxis’ into other individual mobility modes, including scooters and vehicles, according

to the context (Lunden, 2019). Also, according to the co-founder Markus Villig, adopting the

name Bolt indicates speed and considers that the future of mobility is electric.

Oga-taxi was part of the initial sample. However, all efforts to contact the company

were futile. I visited their office location as seen on Google, but the company had moved from

that location. Although there were speculations of bankruptcy from drivers in the field, it was

difficult to confirm. Because interviewee Alex (November 2018) was registered on the

platform, the plan was to interview the platform owners and other drivers in the field. However,

it was difficult because only two drivers worked for Oga-taxi occasionally, and the interviewee,

Alex, highlighted the difficulty of receiving trips on the platform.

On the other hand, Uber and Bolt rejected interviews. In comparison to Uber, Bolt

responded to my participant and consent forms before refusing interviews. There were no

contact addresses, such as emails, to contact Uber, signifying a more opaque system. Although

I was able to discuss with a senior representative from Taxify informally, the information

received was in the form of rigid responses such as 'yes' or 'no' answers without further

explanation or context. It was an attempt to prevent me from asking further questions about

drivers' treatment and the platform company's overall operation. Max.ng, an indigenous bike-

hailing company, agreed to be interviewed. However, narrowing my focus to only Uber and

Bolt eliminated the need for the company. Following the definition of ride-hailing or

ridesharing, the focus was only on automobile vehicles and not motorcycles. However,

transcripts from my interview with Olamide (November 2018), who plays a critical role in the

company, gave some insight into how they use algorithms and data analytics to manage their

labour, as discussed in Chapter four.

3.5. Participant Observation

According to Gorman et al. (2005, p. 40), participant observation "involves the

systemic recording of observable phenomena or behaviour in a natural setting". In

understanding the hows and whys of human behaviours in a particular context, participant

observation, although challenging, is useful for researchers to immerse themselves in the

fundamental human experiences (Guest et al., 2017). As a researcher, asking questions from

FGD and SSI meetings, particularly in a dynamic field like ride-hailing platforms in Lagos,

requires participant observation by immersing oneself in the context. Before this research, there

104

was a basic understanding of how platforms like Uber or Taxify (Bolt) work. However, because

previous encounters with platforms were non-academic and passive, there was a lack of

curiosity and depth in unpacking specific experiences and behaviours. Assuming a researcher’s

role enables the need for active observations by asking questions about why things are done in

a certain way, note-taking, taking images, recording sounds, and uncovering behavioural

meanings of the subject (Guest et al., 2017).

Baker (2006) divides participant observation into seven typologies: nonparticipation,

complete observer, observer-as-participant, moderate or peripheral membership, participant-

as-observer, active participation, active membership, complete participation and finally,

complete membership. This study adopts the moderate or peripheral typology because my role

as a researcher maintains a balance between participation and observation, as defined by Baker

(2006). For instance, in this study, mobile participant observations were taken during 40 trips

on both Uber and Bolt; screenshots of some platform dashboards; screenshots of an online

forum (Facebook) discussion; and images of vehicles, car parks and training sessions. As a

moderate or peripheral member, my role enabled my attendance at Bolt’s driver training

session and occasional taxi driver and platform drivers’ meetings. This was relevant in

balancing my role as both an insider and outsider without any biases. The initial aim involved

total immersion as a complete participant through registering as a platform driver. However, it

was impossible because of delays in license registrations and other safety concerns, particularly

around the looming election period.

Overall, this method was relevant in triangulating participants' experiences, behaviours,

and perceptions, especially those of platform drivers in this study.

3.5.1. Mobile Observations of Platform Drivers

Mobile participant observations were used to gather complementary information from

Uber drivers.

According to Sheller and Urry (2016, p. 11 - 12), the new mobilities paradigm

led to the rethinking of fundamental ideologies in social sciences. First, it involves

examining the constitutive role of movement within the workings of most social

institutions and social practices. Secondly, work within the new paradigm examines

different modes of mobilities and their complex combinations: corporeal travel of

people; physical movement of objects; virtual travel, often in real-time transcending

distance; communicative travel through person-to-person messages; and imaginative

105

travel. The third paradigm employs mobile methods that capture many kinds of

movement and related practices and institutions, which is relevant for this research.

These involve innovative practices in motion that combines a mix of quantitative and

qualitative methods, and visual and experimental approaches, such as mobile

videography that capture events and detect inconceivable problems (Sheller and Urry,

2016).

Scholars have adapted innovative designs using mobile technologies such as GIS and

GPS-enabled devices to create spatial data for analysing mobility (e.g., Kwan and Knigge,

2006). Also, Evans and Jones (2011) examine the importance of walking mobile interviews

that develop an understanding of peoples’ knowledge about their environment and attitudes,

which indicates the importance of place and how Lagos shapes the platform experience of

drivers. If walking interviews can serve as a paradigm shift in mobility research, driving

interviews and observation are equally progress paradigms that encapsulate the phenomena of

ride-hailing platforms. Rosenblat (2018) performed participant observational methods and

interviews by taking over 400 Uber trips in both the US and Canada. However, the current

research was concerned about the unfairness of booking a trip and asking the driver for an

interview in the Lagos context, particularly with a limited compensation package for drivers.

Thus, mobile observations through driving are integral in developing new methods to study the

transient nature of the mobile workforce of platforms in Nigeria.

This study assumes the position of active participant observations with drivers

following 40 Uber and Taxify (Bolt) trips. These observations voluntarily also included

informal conversations during trips judging by the driver's willingness for a conversation. I sat

in the front of the vehicle or at the rear corner, which Nigerians often call 'the owner's corner'

– a term assigned to people perceived to be wealthy or resourceful middle-class people. After

every trip, because our conversations were informal, I took between 15 and 30 minutes to

reflect on the relevant information and write it up on my smartphone notepad. Conversations

covered various subjects from football, politics, music, gender, passenger preferences and

experiences, trip preferences (night or day) and vehicle plans (renting or hire-purchase) to

relationships with vehicle owners. Although the Uber and Taxify policies prohibit drivers from

discussing politics, my response stance was always neutral. Below are some observations made

during my trips. Mobile observations were relevant in complementing other data collection

techniques and facilitating triangulation for clarity and depth of knowledge creation (see table

7).

106

Table 7: Mobile Observations of Ride-hailing Platform Drivers

Mobile Observations

Explanation

Methodological Reflection

Drivers’ reactions when

I sat in the front or back

seat of their vehicles

Generally, sitting behind drivers (on the right side) signifies

power, with the person behind as the powerholder. It is

typically known as ‘the owners’ corner’ for many affluent or

middle-class Nigerians that can afford personal drivers.

As a rider or researcher, sitting down in front of the vehicle leads to

interactions for ride-hailing drivers while sitting behind limits

interactions or any form of interference.

The number of

smartphones drivers

possess

Thirteen drivers out of 40 trips possessed at least one

smartphone and a power bank or two smartphones for multi-

apping purposes and supplementary phones for just

communications (occasionally not a smartphone)

This further reinforces the idea that drivers work longer hours and

require their smartphones throughout the labour process. It prompts

action from the researcher to respect the timing of workers and their

ability to commit to interviews or other data collection methods.

Reflections of this further contribute to the contextual realities

involved with gig work in Lagos and drivers' awareness.

As much as ride-hailing platforms institute digital identities for drivers

through the app, these are also part of their physical identities based on

their mannerisms, devices and vehicles

The placement of

smartphones in the

vehicle

Drivers often place their devices in between their legs or just

under the handbrake by the gear. Drivers avoid placing

smartphones on the dashboards because it would expose their

identity as Uber drivers, which could attract extortionate

practices from policemen or vehicle inspection officers (VIO)

and harassment from bad riders or robbers, especially in high-

risk areas.

Power banks and

charging ports

For all the trips I entered, drivers possessed at least a power bank

for charging their phones. They also possessed charging cables

and ports for charging their phones. On six trips, drivers had

107

chargers for Android devices or Apple devices for themselves

and riders.

Refurbished vehicles

Three out of the forty trips were drivers with refurbished vehicles

such as tinted windows and shiny wheels. In most cases, these

sorts of drivers owned the vehicle. Compared to those renting or

undertaking hire-purchase plans, they took care of the vehicle

more. I would argue that hire-purchase drivers were also careful

when the vehicle owner was lenient and respectable. However,

most drivers are not loyal because they believe it is complicated

to acquire the vehicle at the end of their contracts because of how

vehicle owners threaten to take the vehicles from them. Also, the

vehicles may be worn-out by the end of the contract, which may

not be adequate for the job.

The cleanliness of the

vehicle and services

rendered

Some drivers' vehicles were neat and comprised items like

sweets, packet chips (or what is called 'kpekere'), bottled water,

and bottled soda (e.g., Fanta, Coke, Pulpy drink, and others).

My perception of this was to ensure drivers were comfortable and to

attract 5-star ratings from riders.

Dress code

Drivers are always dressed casually in just a T-shirt and jeans.

Few drivers dressed formally in tucked-in shirts and plain

trousers.

108

Ride-hailing platform

experience

There was a difference between the Uber and Taxify (Bolt)

experiences. Just like drivers perceive Uber riders to be more

respectable and civil, drivers on Uber were more professional

than Taxify, even though most drivers work on both platforms.

My perception of this is that Uber is a more stabilised brand

globally. Hence both passengers and drivers felt more secure

than Taxify (Bolt). The registration process for drivers creates a

better sense of structure because they are required to pass a

100-word assessment, unlike Taxify. My trip with Abiodun

(August 2019) asserted that Uber was more structured and

secure because he could not infiltrate the platform.

When drivers speak about their preferences across platforms, it is

worth observing how drivers treat their riders to corroborate or negate

drivers’ stories and extract clearer meanings. Considering Bolt and

Uber were two international platforms, there were still nuances across

these platforms and variations in drivers' and even riders’

perceptions.

Route selection

A driver may choose to follow a longer route if a rider is not

conversant with the route to their destination. I observed that

drivers were more likely to manipulate the app when the

passenger was not paying attention or unaware of the

destination.

This prompts the researcher to observe when drivers steer away from

the platform digital map directions. It would help understand why

these occur and how to respond in such situations.

Phone usage on trips

When drivers fiddle with their phones during trips, it is because

of these main reasons: receiving or rejecting a call, checking the

trip map, rejecting a trip request, connecting the phone to

charge, and manipulating the app. It is often challenging to

capture a driver when they are manipulating the app during the

This prompts the researcher to be more observant and careful not to

label every fiddling attempt as an avenue to cheat their rider.

109

trip, especially when the passenger is sitting in the back seat.

This makes riders often suspicious when drivers use their

phones during trips. In my experience, based on my knowledge,

I often ask to see the before and after trip distance on the map

on drivers’ phones. I also instruct drivers to start and end the

trip in my presence.

Vacant car parks/spaces

Other observations were at vacant parking spaces during the

night. I observed drivers sleeping in vehicles in central areas like

Lekki Phase One to access profitable trips the following

morning.

These vacant spaces were temporary resting places for drivers who live

in distant locations such as on the Mainland, Ajah, Sangotedo and other

places. These spaces facilitated networking between drivers,

friendships, and recruitment for their WhatsApp groups. The G.T.

Bank's parking lot in Lekki Phase One was at the centre of party clubs,

shopping centres, offices, and eateries. It was a strategic location for

gaming trips; it is a central area connecting to critical leisure places for

potential riders.

Source: Author’s fieldwork (2018 and 2019)

110

3.5.2. Online Observations of Platform Drivers

With the ubiquity of the internet and advancement in information and communication

technologies (ICT), online environments such as social media networks are increasingly

entwined in the everyday offline experiences (Garcia et al., 2009). Since the dot-com bubble

period, scholars such as Rheingold (1993) and Escobar (1994) have begun reconceptualising

ethnography as part of cyberspace and virtual communities. According to Escobar (1994),

cyber-ethnography acknowledges the potentiality of cyberspaces to transform culture and the

structure and meaning of society based on biological technologies, computers and information.

While cyber-ethnography does not integrate social network environments today, it provides a

progressive platform that introduces creative methodologies which examine online

environments. Other scholars have conceptualised alternative ethnographies such as internet

ethnography (Sade-beck, 2004); digital ethnography (Murthy, 2008); virtual ethnography

(Hine, 2008); expanded ethnography (Beneito-Montagut, 2011); and other classifications that

merge online and offline environments methodologically.

According to Hine (2008, p. 257), “virtual ethnography transfers the ethnographic

tradition of the researcher as an embodied research instrument to the social spaces of the

internet”. This indicates that with technology, behaviours, experiences and everyday realities

can be examined without a physical presence. Murthy (2008) highlights digital videos, blogs,

online questionnaires and email interviews, and social networking sites as critical platforms for

examining online environments. While there are valid arguments about virtual sites differing

from offline conventional real-life settings (Hine, 2000; 2008), Garcia et al. (2009) argue that

virtual reality is part of conventional reality, which incorporates human actions, experiences

and perceptions. Therefore, virtual reality exists in offline and online environments and should

be examined based on field study definitions of the research topic (Garcia et al., 2009).

In this study, social networking sites such as Facebook and Twitter were critical in

observing contextual information, which served as a medium to corroborate or negate stories

from other qualitative data collection techniques. Accordingly, three characteristics of social

networking sites, as categorised by Murthy (2008, p. 845), are relevant here:

a. Ethnographers can invisibly observe the social interactions of page members.

b. Social network sites contain vast stores of multimedia material regarding even the

most marginal social movements or groups. This was relevant in observing images

111

and videos of conflicts, harassment, and advertisements in ride-hailing platform

groups.

c. These networks contain virtual gatekeepers with chains of friends who are potential

research respondents. For instance, this is relevant to meeting with the interviewee

Koffi (September 2019), who became a gatekeeper to the PEDPA platform union

and an active participant in this study.

In understanding the phenomenon of gig work through ride-hailing, Rosenblat (2018)

spent hours daily reading and observing the virtual realities of ride-hailing drivers and further

comparing observations to interviews. I applied to join a platform drivers’ private group on

Facebook for my observations. After my interview with Charles in October 2018, the intention

to join arose when he explained the importance of online social media and communication

networks (SMCN) such as Facebook, WhatsApp and Telegram for their work. I could not join

a private WhatsApp group through my gatekeeper because of a lack of platform driver

requirements such as an Uber or Taxify (Bolt) dashboard screenshot, licence and registration

forms. WhatsApp platforms' administrations appear to be stricter than Facebook open groups

because drivers within those communities are suspicious of intruders who may expose their

strategies. Between October 2018 and January 2019, I passively observed only two public

groups without explicitly joining the community.

Despite being a public group, these groups' administrators screen members before

accepting their request to join the platform. These groups' names are UBER/BOLT PARTNERS

with over 12 000 members and Uber and Taxify Drivers in Lagos with over 18 000 members.

I was able to join one private Facebook platform group with over 17 000 members on 27

April

2019. I joined the private Facebook Uber and Taxify Lagos Driver and Partner forum group

by submitting a consent form and participant information sheet (see table 8 below and

appendices B1 and B2). After three days, I was accepted onto the platform. It was critical to

join a community, particularly a private community of platform drivers, to closely observe their

everyday decisions, experiences, and perceptions of their job. As a member of these online

communities, I occasionally replied to posts from other drivers. This action enabled me to gain

trust within the community, which subsequently facilitated recruitment for telephone

interviews. As Hine (2008) opines, authenticity and trust are critical aspects of doing virtual

research because of the ease of possessing fake or multiple identities and the inability of these

community members to validate the researcher's identity. In my call for participants, 28 people

112

publicly replied to the posts out of a maximum of 15 drivers that were requested. An additional

12 drivers sent direct messages showing an interest in the study.

The driver community comprises different categories of people such as platform

drivers, vehicle owners, car sales and rental companies, ride-hailing platform workers and other

unknown members. With over 150 posts a day, I visited the group at least three times a day. It

is important to note that this was a continuous process, i.e., when I needed to crosscheck stories

from drivers, I browsed through these online communities for contextual information. A

summary of posts on platforms includes the categories shown in Table 8.

113

Table 8: Overview of Online Observations

Facebook

Groups

Observed

Over Time

Members

Created

Accessibility/Status

Examples of the Type of Observations

from Facebook Groups.

Methodological Reflections

UBER/BOLT

PARTNERS

14,900

October 2016

Public/Active

- Drivers that complain about a

passenger; a route; a platform

company's unfairness; a vehicle

breakdown, and many other

complaints

- Drivers that advertise about a

platform company promo; a new

platform company; vehicle sales;

an Uber or Bolt account for sale.

- General advice on how to improve

earnings on the job and other

advertisements

- Platform union recruitment posts

- An invitation to a private WhatsApp

group by union leaders or vocal

These online observations enabled the

researcher to pay attention to the

behavioural patterns, activities,

reactions, perceptions, and overall

experiences.

Behavioural patterns: By taking note of

their daily routines and how decisions

are created to achieve their aims.

Activities: Acknowledging that these

SMCNs are mobile workplaces where

different activities occur, which do not

always include drivers. E.g.,

advertisements from random users,

researchers like me, and new platform

companies in search of drivers.

Uber and

Taxify

Drivers in

Lagos (e-

drivers)

48,400

August 2018

Public/Active

Uber &

Taxify Lagos

Driver &

Partner

forum

May 2017

Private/Defunct

Uber and

Bolt Drivers

in Nigeria

11,800

July 2020

Private/Active

114

drivers intending to build a network

with like-minded drivers.

- Advertisements for offline or courier

trips.

- Posts (pictures and videos) on riders

that have defaulted and how and why

colleagues should avoid such riders.

Reactions: Taking note of reactions

such as the varying aspects of the ‘like’

buttons and comments on posts. These

reactions possess implied meanings that

bolster the quality of data collection.

Perceptions and Experiences:

Understanding the varying perspectives

of different categories of drivers such

as new, intermediate, old and also,

being able to compare these

perspectives to workers’ ride-hailing

platform drivers in GN cities.

Source: Author’s fieldwork (2019)

115

Observing drivers included taking screenshots of posts relevant to my study and saving

posts in different folders on the platform for future review. It was difficult to obtain consent

for every screenshot used in this study. However, the names and pictures of the drivers from

screenshots were censored to protect drivers’ anonymity. As much as these online driver

communities were virtual realities of everyday experiences, they were also avenues for

continuous learning about the evolving realities for ride-hailing drivers in Lagos.

3.6. Method of Data Analysis: Thematic Analysis

The data collected in this research were analysed using the Nvivo software by coding

qualitative data into codes to facilitate thematic analysis. Thematic analysis, according to Braun

and Clark (2006, p. 79), “is a method for identifying, analysing, and reporting patterns (themes)

within data”. Accordingly, two ways of identifying themes are inductive and deductive

approaches, where the former is based on data, and the latter is based on theoretical lenses

(Braun and Clark, 2006). Although initial research questions were based on a deductive

approach using the sociotechnical framework and MLP (Geels, 2005), as explained in box 1

below, the thematic analysis is overtly based on an inductive approach, allowing the qualitative

data to determine the themes of this research. This back and forth characterises the flexibility

of thematic analysis, which, according to King (2004), is suitable for an experiential or

exploratory approach that can be adaptable to the changing needs of the study. Thematic

analysis and its principles are easy to comprehend for a lay audience because of the ability to

codify complex or broad data in simple language (King, 2004; Braun and Clark, 2006).

Table 9: Box 1 showing Socio-technical Transition and Multi-Level Perspective (MLP)

Box 1: Socio-technical Transition and Multi-Level Perspective (MLP).

Socio-technical or technological transitions, according to Geels (2002, p. 1257), "are major technological

transformations in the way societal functions such as transportation, communication, housing, feeding, are

fulfilled." According to the theoretical framework, technologies operate across three separate layers known

as the socio-technical regime, landscape, and niche, which are often consolidated into three levels known as

the multi-level perspective (MLP), as seen below. The MLP is a quasi-evolutionary theory which is rooted

in the historical analysis of technological change and how interactions between the agency and across the

three levels that explain it (Geels and Schot, 2007; Raven et al., 2012). These three levels interact through

learning processes, performance improvements, and robust support groups. (Geels and Schot, 2007). Actors

116

can build internal momentum at the niche level, which may variably create changes at the landscape level,

which affects an existing regime through a ‘window of opportunity' such that it is either displaced or

substituted (ibid).

Source: Adapted from (Geels, 2002, p. 1263)

Landscapes

The socio-technical landscape, which is similar to Braudel's concept – longue dureée, is used in a

metaphorical sense because it is embedded in deep structural trends where technological trajectories are

situated (Geels, 2002; Geels, 2011). These structural trends comprise heterogeneous or external factors,

which may be material (highways, factories, electricity infrastructures, spatial arrangement of cities) or

immaterial (social and cultural values, political culture and coalitions, oil prices, wars, environmental

problems, emigration and so forth) (Rotmans et al., 2001; Geels, 2002; Geels, 2011). For example, the car

117

is a combination of networks across space with characteristics such as steel and plastic, concrete (roads),

law (traffic rules), and culture (the value and meaning of personal mobility).

Regimes

Rip and Kemp (1998, pg. 338) define regimes as "the set-rule or grammar embedded in complex engineering

practices, production process technologies, product characteristics, skills and procedures, ways of handling

relevant artefacts and persons, ways of defining problems – all of them embedded in institutions and

infrastructures”. The stability of an existing socio-technical system is supported by a well-established socio-

technical regime (Geels, 2004). Hence, the production of various elements in socio-technical systems is a

result of a semi-coherent set of rules that orient and coordinate activities of diverse social groups of existing

regimes (Geels, 2011). Regimes thrive through lock-in mechanisms (see Urry, 2004; Klitkou et al., 2015),

such as rules which enable legally binding contracts, capabilities and competencies, lifestyles and user

practices, favourable institutional arrangements and regulations, cognitive routines and shared beliefs.

Niches

Niches, on the other hand, are guided spaces where new innovative ideas are harnessed and eventually

emerge through radical changes due to a loophole or window of opportunity in an existing regime (Geels,

2002). In the world today, every regime began as a niche, and every regime faces some form of disturbance

from novelties vying to become a regime. Some examples of niches include small markets with loyal

customers who support technologies regardless of their level of diffusion (electric vehicles, driverless cars,

3D printing); research laboratories and subsidised demonstration projects (military funding for GPS) (Geels,

2011). Research in strategic niche management and technological innovation has identified three

fundamental processes for niche development (Schot and Geels, 2008; Rip and Kemp, 1998):

a. The articulation of expectations and visions. Definite plans for technological innovations create

awareness and thus instigate directions for learning processes and progressive planning and

protection.

b. The building of social networks. This process creates a legitimate network for further bolstering

awareness of new technology by facilitating interactions between actors which may lead to the

provision of necessary resources such as money and skilled individuals.

c. Learning processes at multiple dimensions. Finally, with resources at their disposal, equipping

individuals in different fields becomes a priority. These dimensions include technical aspects and

design specifications; market and user preferences; cultural and symbolic meaning; infrastructure

118

and maintenance networks; regulations and government policy; and societal, and environmental

effects.

It is important to note that the MLP framework was critical in the first phase of my fieldwork because of the

intention to unpack the transition from traditional taxis to ride-hailing platforms. It was relevant in the

recruitment of interviewees in the field especially leading to an initial holistic understanding of the taxi

sector in Lagos. However, it was too rigid in interrogating the everyday experiences and challenges of

drivers, which were often algorithmically driven and their everyday resistances.

This research adopted the six-phase pathway for doing thematic analysis by Braun and

Clark (2006), as shown in Figure 5, to establish credibility, dependability, confirmability,

transparency and transferability of analysis (Nowell et al., 2017). The cycle in Figure 5

contends that the six phases of thematic analysis are a continuous and reiterative process.

Because of the novelty and sparse knowledge of ride-hailing gig work in Lagos and African

cities, generating initial codes involved a reiterative process that compared data to other

qualitative methods used in this study. In familiarising myself with the data, I created

temporary codes generated manually from FGDs, as seen in Tables 10 through 14 below. This

familiarisation process also involved manual temporary codes generated from initial SSIs with

key informants such as Charles, Henry, Efe and others mentioned above. The FGD and SSI

interview guides in Appendices C1 and C2 show temporary codes that were later integral nodes

for the NVivo analysis process. At the time of this coding exercise, the goal was to categorise

themes along with the framework of the MLP, which covered niche innovations, socio-

technical regimes, and landscapes (Geels, 2005). However, as mentioned, this limited the

intricacies and granularities of driver experiences and the impact of platformisation and opaque

management that were devoid of broad socio-technical aspects of mobility. Despite removing

the MLP framework, keeping its ideals in the background was critical in targeting proponents

in the field and establishing thematic narratives that capture the history of taxi labour and the

background of innovative platforms, unions and gig work, workers’ struggles, and resistance

practices. This explains the reasoning behind categorising qualitative data according to

participant groups to understand the commonality and differences across these nodes.

119

Figure 5: Six Phases of Thematic Analysis

Source: Braun and Clark (2006)

After the transcriptions from SSIs and FGDs, the emerging themes focused more on

ideals of algorithmic management (Lee et al., 2015; Möhlmann and Zalmanson, 2017) and the

impact it has on their decisions, income, and relationships with colleagues as well as platform

companies. The underlying idea was that drivers did not feel in control of the labour processes,

and algorithms limited this experience. Initial codes on taxi labour, state officials and others in

the following tables were critical for understanding the history and transition process from taxis

to platforms by highlighting the role of the state and unionisation in analogously managing and

controlling labour. This was a rigorous process that introduced the fourth and fifth phases based

on the guidance of concepts outlined in Chapter two. Because the studies on algorithmic

management and the emergence of platforms were sparse in GS cities, it was critical to adapt

a core understanding from GN scholars and theoretical concepts based on other authors (Scott

1985; 1989; Rayle et al., 2014; Lee et al., 2015; Rosenblat and Stark, 2016; Möhlmann and

Phase 1:

Familiarising

yourself with

your data

Phase 2:

Generating

initial codes

Phase 3:

Searching

for themes

Phase 4:

Reviewing

themes

Phase 5:

Defining and

naming

themes

Phase 6:

Producing

the report

120

Zalmanson, 2017; Jarrahi and Sutherland, 2018) to establish framework of concepts for further

streamlining my codes. According to Nowell et al. (2017), triangulating information cuts across

all the phases of thematic analysis to ensure trustworthiness in the process. In this study, all the

phases underwent reiterative processes and triangulations from participant observations and

other secondary data sources in line with the utilised concepts. This also indicated that the

historical perspectives of theoretical concepts were critical in examining changes in control

and management that were not digital. For example, surveillance concepts provided a basis for

critically understanding how platforms enable workers' control through algorithmic

management today.

Considering the flexible nature of the thematic analysis, one of the disadvantages is that

it becomes challenging to determine comprehensive data and a theoretical lens for emphasis

(Braun and Clark, 2006; Nowell et al., 2017; Kiger and Vapio, 2020). For example, the

qualitative data could be adaptable to other theoretical aspects like labour process theory,

disruption theory and resilience theory. However, the intriguing nature of gig work, gradual

determinism on platforms and management of workers by algorithms informed the direction

of analysis and research.

121

Table 10 were preliminary themes developed at the beginning of the data collection stage to understand the overall picture of ride-hailing

platforms in Lagos. It is important to note that this perception was based on the lack of knowledge about the field and an exploratory method of

learning about ride-hailing platforms and their components in Lagos.

Table 10: Focus Group Discussion Themes and Justification

FGD Preliminary Themes

Justification

Platform choice/preference of

registrations

To understand the choices behind driver’s platform choices and the corresponding registration process

Comparing ride-hailing gig work and

previous work experience

To understand why drivers leave their previous jobs or use ride-hailing work as a supplement

Vehicle specification and gadgets

(including network providers)

To get an idea of the identity of ride-hailing drivers through the type of vehicle or devices they use for work

Vehicle arrangements

Examining the types of vehicle arrangements (e.g., hire-purchase, weekly rentals, personal vehicle ownership)

Relationships with vehicle owners

(e.g., in rental vehicle arrangements)

Especially for drivers on hire-purchase, this helps understand how the relationship with vehicle owners affects their

work in terms of maintenance agreements, weekly fee payments, and other characteristics

Driver versus passenger relationships

To understand how relationships between drivers are formed and, in some instances, sustained. Also, to what extent this

improves or affects the labour process of drivers. For example, how and why a passenger would rate a driver 1 star or

report a driver falsely via the platform.

Weekly vehicle earnings and expenses

To understand how much drivers make after expenses and what this means for drivers with different vehicle

arrangements

The safety and security of ride-hailing

work

To grasp how safe and secure ride-hailing works compared to the taxi regime

Health and well-being of workers

This finds how the decisions of drivers and the nature of the job affect their health and wellbeing (e.g., tiredness,

skipping meals, accidents, death)

Communication channels of ride-

hailing platforms

To understand the ease or difficulty of communication channels when drivers require to resolve an app discrepancy or

conflict with a rider

122

Perceptions of gamic elements (e.g.,

ratings, acceptance, cancellation

scores)

To understand how drivers perceive that these elements work, especially in motivating or demotivating them

Gaming strategies for drivers

What drivers consider gaming and the different strategies they utilise in evading algorithmic rules

Manipulative driver techniques

(reasons behind them)

To understand practices drivers utilise in manipulating the app and their reasons behind it

Trip frequency and peak periods

To understand how algorithms assign trips and drivers' awareness of the key periods and seasons that they are more

likely to get jobs

Impact of physical attributes on ride-

hailing gig work (traffic congestion,

bad roads, driver behaviour)

This seeks to find how the realities of the city can affect the rigidity of algorithm rules (e.g., trip assignments)

Accuracy of digital maps versus

knowledge of the city

To determine the accuracy of digital maps on platforms compared to drivers’ knowledge of the city

Dangerous areas in the city

This creates awareness of the dangerous areas in the city and why drivers lack the freedom to decline trips from these

areas

Social Media and Communication

Network Groups (SMCN)

This is critical in understanding the importance of SMCN groups to every aspect of ride-hailing gig work, especially

resistance strategies

Traditional taxi unions versus platform

unions

To understand how these unions differ and drivers’ perceptions

Airport experiences

This helped me understand why platform drivers were harassed by state officials such as FAAN and why airport taxi

unions were seemingly more organised compared to mainstream taxi unions.

Table 11 builds on the data established through FGDs by enabling me to model questions according to specific actors in the field. It is also

critical to note that, during this exploratory pathway, the data emerging was linked to worker resistance and difficulties in the reconciliation of

opaque management. While the intentions were to ascertain the discrepancies between control from the app and the realities from drivers, the

algorithmic management concept was not immediately integrated until the second phase of fieldwork between July to September 2019.

123

Table 11: NVivo Themes and Justification for Platform Drivers

Nvivo Themes for Ride-

hailing platforms

Justifications

Ride-hailing platforms

(Drivers and Experiences)

This parent node has an overview of driver experiences and general occurrences in Lagos.

Descriptive and Demographic

Questions

Background: This helps understand the operationalisation of ride-hailing platforms in Lagos compared to other jobs and

other contexts.

Platform registration,

preferences, job satisfaction

Nature of the job (Terms of

Labour conditions)

Before the advent of Ride-

hailing platforms (perceptions)

History: This helps to understand drivers' perceptions before ride-hailing platforms, including conventional taxi gigs. It also

sheds insight on why and how drivers favour international platforms such as Uber and Bolt over domestic platforms. More

about the history was coded under the category of the ‘state and unions’ and Taxis and other forms of transport in Lagos’

Domestic versus International

platforms

Platform challenges for drivers

Platform and Algorithmic Impacts: This distils the unique challenges platform drivers face in Lagos as specifically as

possible. This section typical examines how platforms impact drivers. While one node mentions algorithms, it cuts across

drivers’ stories, highlighting algorithmic management's opacity in Lagos.

Algorithmic assignments,

decisions and information

asymmetries

High-risk Locations for drivers

Opaque platform governance

(Platform manipulations,

incentives, deactivations).

Platform marketing strategy

(Incentives, competition,

commissions)

General impacts on mobility

and recommendations

Resistance Strategies 1

(Coordination, Networking,

Communication)

Resistance/Coping strategies: These categorise the everyday resistance practices observed. I further streamline these into

both hidden and public resistances

124

Resistance Strategies 2 (App

manipulation, gaming, City

Knowledge)

Based on further exploration, table 12 focused on the analysis of responses from state representatives such as policymakers and union

leaders. This is because compared to other mobility modes such as the BRT, danfo, Okada and others (described in chapter four), the taxi

industry lacks research, including the drivers of the vehicles. Therefore, it was critical to interview union representatives to understand the

existing taxi industry regulations and how these apply to platforms. Also, considering these representatives possess an overarching knowledge of

existing and potential regulations that impact the taxi industry and ride-hailing platforms.

Table 12: Overview of NVivo Themes for Policy and Union Representatives

Nvivo Themes/Nodes for

States and Unions

Justifications

States and Unions

These nodes give an overview of the state and union perceptions of gig work and corresponding taxi labour in Lagos. The

overarching aim contributes to the history and culture of gig work in taxi labour.

Evolution of Taxis in Lagos

History and Constitutional Dynamics: This helps understand the evolution of taxi labour, regulations, and evolving labour

laws before the emergence of Uber and more recently. It also outlines the weaknesses of local institutions and regulations

and how international platform companies are exploiting these weaknesses. International companies operating in Nigeria

with the same constitution as their country.

Shortcomings of the traditional

taxi industry

Labour laws and regulations in

the taxi industry

Local vs foreign constitution

differences

125

Synergizing Conventional taxi

unions and Platform unions

Unionisation: These include discussions about the existence of traditional taxi unions and plans to consolidate these

platforms with emerging platform unions under a more formalised umbrella.

Growth of Platform unions and

platform governance

Driver challenges (union and

state leader perceptions)

Platform driver challenges and Impacts: These give an overview of drivers' impacts from the perceptions of union and

transport officials.

Safety and Security of drivers

(State and union perception)

Resistance and coping

strategies

Resistance and Coping strategies: These are also perceptions from union leaders about why platform drivers engage in

resistance practices

The Future of Labour in the

taxi industry

Future of Labour: Union leaders and state officials embrace innovative platforms but advocate for regulations that protect

the future of gig workers.

Table 13 focuses on interviews with traditional taxi drivers in Lagos. The idea to interview taxi drivers was due to acknowledging the

history of pre-existing services similar to ride-hailing platforms based on the categories of vehicles or four-wheeled taxis. It was also relevant in

comparing challenges and resistances between ride-hailing platform drivers and taxi drivers. This is because certain experiences were not exactly

new with ride-hailing platforms. The app as an intermediary and algorithms as invisible managerial control did not exist with traditional taxis.

However, manual processes of burdensome labour from the state and unions discussed in Chapter four existed before Uber’s emergence.

Table 13: Overview of NVivo Themes for Taxi Labour

NVivo Themes/Nodes

for Taxi Labour

Justifications

Taxi Labour

These nodes give an overview of conventional taxi labour before the emergence of ride-hailing platforms and how taxi drivers are

surviving since platforms emerged. It outlines the history and culture of taxi drivers in Lagos.

Descriptive and

demographic questions

126

History of transport

and taxis in Lagos

Background and History: This helps understand the evolution of taxi labour, regulations, and evolving labour laws before the

emergence of Uber and more recently. It also helps triangulate platform driver stories on how platforms are changing the taxi labour

and broader mobility in Lagos.

Nature of the work

(before and after Uber)

High-risk zones

Taxi driver

requirements

Taxi labour and

Unionisation

Challenges for drivers

Challenges and Impacts: These themes outline how the emergence of ride-hailing platforms has impacted taxi labour. It also

demonstrates why technology is essential and the need for stricter regulations that enable the progression of their work.

Perceptions of

technology in taxi

labour

Regulating taxis and

platforms

(perceptions)

Resistance and coping

strategies

Resistance and Coping Strategies: While there is not much being done by the state, drivers are learning to adapt, resist and, in a

few cases, transition to ride-hailing platforms.

Table 14 was targeted at platform companies, venture capitalists, and technology journalists that understand the processes of platform

emergence and the reasons for successes and failures. As mentioned, it was impossible to get interviews with the Bolt or Uber platforms. However,

I interviewed Max.ng, a bike-hailing platform and a venture capitalist representative who gave me insights on hidden details of the algorithmic

management of platforms. While this was substantial and somewhat transferrable to Uber and Bolt, it was limited in critically understanding the

difference between the algorithmic management procedure at ride-hailing platforms and bike-hailing platforms.

Table 14: NVivo Themes and Justification for Platform and Venture Capitalists

NVivo Themes for

Platforms, Venture

Capitalists

Justification

127

Platforms and Venture

Capitalists

The idea of these sections was to interview platform companies such as Uber, Bolt, Oga-taxi, Max.ng to understand the hidden

nature of platforms. While Max.ng was later dropped from the sample, an interview with the company gave some insight into how

algorithms, data and surveillance are critical components of platform business models. A venture capitalist representative was

critical in narrating the historical and current landscape of technology innovations in Lagos.

Platform Emergence

(Domestic and

International)

Refer to the ‘history and constitutional dynamics in Table 12. Some specific insights outline the onboarding, requirement and

recruitment process for max.ng drivers.

Data Technicalities

and Company

Statistics

Algorithms in Action: This shows insights into how algorithms manage drivers through ratings and other surveillance apparatuses

spy on even how riders accelerate during rides. The platform also uses the god view like Uber and Bolt platforms to achieve

monitoring and control of its drivers

Challenges for

Platforms

Here, it shows an overview of why past innovative platforms failed from a tech-oriented person and highlights some of the struggles

Max.ng faces, which is a representation of tech platforms in Lagos such as financing, regulations and law enforcement harassment

Impacts of Platforms

on the Taxi industry

Refer to ‘challenges and impacts’ under taxi labour. The difference is the insight and perception of venture capitalist or tech

personnel.

Perceptions of drivers’

resistance and Coping

strategies (perceptions)

Refer to ‘challenges’ and ‘resistance and coping strategies’ in the previous tables. The difference is the insight and perception of

venture capitalist or tech personnel.

Changes in urban

mobility (perceptions)

Following the overview of different participant categories, I later developed sub-nodes by drawing from these categories. The two sub-

nodes I created were called hidden resistances and public resistances. Sub-nodes under broad hidden practices were created under the hidden

resistance node, including online SMCN & sensemaking, circumvention, manipulation of algorithms and gamification-from-below. Subsequently,

platform unionisation and mobilisation, protests, and media engagements were created as sub-nodes under the public resistance node. Extracting

data to these codes was based on the revelations of drivers in the field, perceptions of union representatives, and my understanding of James Scott's

concept of everyday resistance. The overview above was relevant in comparing some of these resistant practices to traditional taxi drivers identified

in interview data from parent nodes.

128

3.7. Limitations of the Study

At the time of initiating this study in 2017, scholarship on the emergence of ride-hailing

platforms was scarce globally. Comparatively to GS cities, there was a growing scholarship in

GN cities on the platformisation of labour with ride-hailing platforms such as Uber as integral

case studies for understanding the gig economy. This presented an initial limitation where

preconceived notions of ride-hailing platforms and the gig economy were deduced from how

they operated in GN cities instead of GS cities. Although this became a gap in the literature,

the initial phases were difficult to understand.

The difficulty in the initial phases of desk research persists in the paucity of information

and difficulty accessing credible information within the Nigerian context and Africa in general.

According to Akanbi (2016), data production remains poor because of Nigeria's weak

administrative and institutional system. This phenomenon creates a human resource gap within

Nigeria's data ecosystem, which often reflects the quality of information (Akanbi, 2016). The

paucity of information has affected the quality of research and policymaking in Lagos. In

reverse, this made it challenging to review credible archival resources about the innovation of

platforms in Lagos even before the emergence of Uber in 2014. With early innovative practices,

even the pre-Uber era was examined from online archival sources such as digital footprints

from press interviews and online news. A few studies, such as Olukojo (2003), Agbiboa (2017;

2018; 2020) and Albert (2007), were critical for this study in deducing how traditional mobility

platforms, specifically taxi labour, develop a pathway for innovative platforms like Uber.

Being a resident of the city was critical in understanding the management of taxi drivers and

other mobility services before Uber. However, sparse scholarly accounts analysed the

importance of taxi driving as a profession and the analogous form of management of taxi

labour, which is comparable to gig work in Lagos.

Due to time constraints and the availability of information on platforms in Lagos, this

thesis could only focus on two major platforms in Lagos: Uber and Bolt. While these were

critical in showing a disparity in their operations and modalities globally, more detail from

indigenous platforms could have demonstrated a disparity in management and resistance of

drivers. Although some indigenous platforms were mapped in Chapter Four, only two

interviewees mentioned an indigenous platform as an alternative to Uber and Bolt, even though

they occasionally used it because there are little or no jobs on such indigenous platforms. The

usability of driver participants was reflective of their experiences on Uber and Bolt, compared

to Oga-taxi, which were critical in understanding gig work and algorithmic management in

129

Lagos. At least, this brought about calling for international platforms to be context-specific,

i.e., for Uber and Bolt to adapt to the everyday experiences of ride-hailing drivers in Lagos and

the realities of the city.

While SSI and FGDs were critical in developing a story for ride-hailing gig work in

Lagos, it was not entirely representative of all ride-hailing drivers in Lagos because of the small

sample size. A mixed-method approach could have improved the reach of drivers by sharing

quantitative surveys that can determine different variables of pay ranges, vehicle ownerships,

collective organisational participation, and other such data. Such quantitative information

could have also been applied to Abuja, Nigeria – the other city where Uber and Bolt are

dominant, just for more contextual data on how the questions in this thesis about Lagos differ

or align even within the country. Overall, it presents an overarching case of Nigeria. However,

by triangulating through online sources, mobile observations and other means mentioned in

detail in the methodology section, this thesis considerably presents the city of Lagos as an

exemplary case site for Nigeria with substantial detail comparable and scalable to both GS and

North cities.

Finally, the challenge of information scarcity and paucity also created conceptual

difficulties initially. In the beginning, the socio-technical framework and the multi-level

perspective introduced by Frank Geels (2002, 2005) was to outline how ride-hailing platforms

as a niche development have changed Lagos's taxi regime. In a nutshell, socio-technical or

technological transitions, according to Geels (2002, p. 1257), "are major technological

transformations in the way society functions such as transportation, communication, housing,

feeding, are fulfilled." According to the theoretical framework and as noted in Box 1,

technologies operate across three separate layers known as the socio-technical regime,

landscape, and niche, which are often consolidated into three levels known as the multi-level

perspective (MLP).

With more insight from the fieldwork and analysis, the MLP was limited in analysing

the inherent nature of driving as a profession and a skill that displays drivers' management

practices, experiences, and perceptions. Other concepts like disruption and the labour process

theory were in contention because of the novelty and multiplicity of emerging ride-hailing

platforms in cities. However, like the MLP, such concepts were too structural to explicate the

dynamism of ride-hailing gig work in Lagos. This led to adopting and developing the

algorithmic management concept. Although the concept is novel, it captures such dynamisms

of gig work in general and introduces a foundation for why resistances and coping mechanisms

are on the rise. More so, it acknowledges the socio-technical changes and atomised labour

130

process while focusing on the novelty of being managed by algorithms. Therefore, at the

beginning of this thesis, the conceptual limitation serves as contributions to structural or rigid

theories based on the examples above and from the empirical chapters.

3.8. Methodological Challenges in the Field

A significant limitation of this study was initially adopting a deductive approach from

the literature in the context of Lagos. I was unclear initially whether the problem was simply

about residents adopting the technology, platform companies’ penetration difficulties, or the

overall impact on the mobility system in Lagos. This reasoning was why my initial theoretical

lens, the socio-technical transition framework, was intended to inform my fieldwork plans.

However, after visiting the field, my problem and focus for the study became clearer.

Initially, there were difficulties accessing platform drivers because they were on a trip

or strategically searching for trips. This phenomenon often impeded face-to-face interviews in

the initial stages, creating a need for more mobile phone interviews within Lagos, usually at

the end of their self-assigned working hours or a work-free day. It is important to note that

drivers renting a vehicle or part of a hire-purchase plan were more rigid than drivers who owned

their vehicles. Perhaps, face-to-face interviews could have worked during their break times in

the day, but the monetary resources at my disposal were insufficient to compensate for their

time, considering I had to calculate any potential trips they may have during our interviews.

The strategy for compensating mobile phone interviewees was cumbersome in the beginning.

This is because the compensation was based on a certain number of transferable credits or data

per minute, which several drivers thought was too small for compensation. However, it was

difficult to stop some interviewees who spoke longer than required. I decided to reduce the

compensation per minute after the 45-minute mark to address the challenge. It was unsafe to

navigate the city from late December 2018 to January 2019 because of political election

tension. Overall, it affected the opportunities to conduct interviews with transport ministry

workers earlier and subsequently. However, with more clarity in the second phase of my study,

the need to recruit policymakers was minimally required for the study.

The first phase of FGD meetings at the GTB car park was challenging because of the

ease of access by other drivers and pedestrians. Because it was also a residential location for

drivers, they would walk into the discussion session to observe or join the session. This meant

I had to read the consent forms multiple times to drivers who stayed longer than 10 minutes,

mainly if they wanted to contribute to the discussion (see appendix B1). Despite the challenges

131

and limitations throughout this study, I reflected and created relevant solutions for validating

my findings through triangulation with other data collection strategies.

3.9. Reflexivity and Positionality

Researchers' reflexivity involves how research processes shape research outcomes

(Holland, 1999; Hardy et al., 2001). Extensively, this includes data collection methods,

analysis, developing conclusions and disseminating information inclusive of a lay audience

(Corlett and Mavin, 2018). Therefore, it is essential to question how my research is situated in

reality with a subjective understanding of how knowledge is formed through a reflexive

process. Following Corlett and Mavin (2018, p. 378), I base the reflexivity in this study on

three classifications: firstly, reflecting on new ways of thinking about how phenomena question

our understanding of reality and the nature of knowledge using alternative paradigms and

perspectives; secondly, reflecting on my relationships with the research subjects, participants,

data and how this influences or negates the research context; and finally, questioning what can

be defined as valuable or valid for my research.

Growing up as a Nigerian male in a middle-income family with both parents as

professors, my drive to pursue this research was initiated based on my educated background.

As an undergraduate student, I studied Geography and Resource Development, and through

this, I was exposed to different kinds of fields under Human Geography. This facilitated a

master’s degree at University College London (UCL), where I examined the mobility-related

social exclusion of disabled people in Lagos. As a Geographer by discipline, building on my

master’s dissertation was critical because the intention was to examine how infrastructure could

reflect the inclusivity of disabled people in Lagos. However, probing further in the initial

phases prompted an alternative pathway and positioned my research within the praxis of

mobility and labour. Access to data was also a critical defining factor for metamorphosing my

research pathway. However, it was a similar challenge with ride-hailing platforms, but the

choice to pursue it was due to the novelty in GS cities and the potential contribution it would

have. Although this deviates from mobility-related exclusion, the idea that platforms use

algorithms to control labour without the inclusivity of drivers’ experiences and active

participation remains an aspect of exclusion that is inherently digital in this case.

Using the MLP in the initial research stage developed my understanding of the varying

realities across GN cities and their application to GS cities. I understood the context of certain

GN cities, such as a developed policy and regulatory framework for the overall mobility

132

system, better infrastructure and other characteristics compared to cities like Lagos. This

created an awareness of not adapting methodologies for researching those cities because of the

varying context specificity, which captures different realities and nuances for Lagos drivers.

Creating a story of drivers' stark realities informed the need for qualitative methodologies to

unpack those realities using a combination of qualitative methods. This phenomenon was why

my exploratory interviews with other samples, such as traditional taxi drivers and passengers,

remained relevant in understanding different views of platform drivers' realities.

In considering the research area's novelty, it was essential to adopt compelling data

capturing approaches, for example, adopting a mobile participant approach from a platform

perspective. There were three attempts at mobile participant observations, which all failed.

Firstly, I issued notepads to five drivers to record their decisions before and after every trip and

the messages they receive from the app just to capture their experiences and perceptions. I

later created a WhatsApp group on the 22

of November 2018, where drivers would post

dashboards, in-app messages, trip requests and other granular details. However, only one driver

was inconsistently active. This led to cancelling the observation exercise after three days.

Secondly, I tried to register as an Uber driver but failed to register because I did not

possess a valid licence. Getting a valid license would have taken between six and eight months

which was not conducive for my research. The final attempt was to collect tracked trips with

trackers from one of the key platform drivers that possessed a tracking app. However, I failed

to compile this information because the driver was sceptical, not in terms of withholding

consent, but by always postponing our meetings. This led to adapting to a rider-led observation

method. The need to adjust to a user perspective was also because of my awareness of the

politics and difficulties of getting a licence and the lack of safety of ride-hailing platform

drivers, especially during the period leading to the presidential elections.

Thirdly, I attempted a survey experiment where I posted a Google form link to

Facebook groups with 18 questions, including a question on consent. The goal was to

triangulate data on working hours, income range (before and after expenses), and vehicle

ownership models highlighted in SSI interviews. Other questions were around flexibility,

previous employment, the reason for adopting platform work, challenges, and years of driving

for platforms. This experiment was targeted at 50 to 100 drivers. However, this also failed

because I received only 11 responses, of which the relevant responses are in Appendix C4. The

reason for this failure was the lack or inability to compensate drivers beyond a N500 mobile

top-up unit, which drivers complained was too small.

133

I differentiated my role as a researcher, a city resident, and an active ride-hailing

platform user. The relationship with most of the participants was both cordial and professional.

This facilitated my awareness of the job's difficulty, enabling friendliness and cordiality in

engaging participants. An essential part of this was assuring participants of the kinds of

compensation they would receive, knowing that their motivation for working on a ride-hailing

platform was to improve their livelihoods. I am certain that a few drivers would have refused

to partake in this study if compensations were not involved. I had a few experiences when

about four drivers declined to participate in interviews after knowing about the compensation

– exclaiming that they would prefer 'cash rewards', which would have deviated from my ethical

clearance from the University of Manchester.

3.10. Ethical Considerations

According to the UK Foreign Aid and Commonwealth office, Lagos is classified as low

risk (see figure 6). Also, the researcher is familiar with the terrain. In line with the ethics

committee's ethical document, this research overall falls into the medium-risk category.

Accordingly, before embarking on both phases of fieldwork, ethical approval was sought from

the University of Manchester Data Management Planning (DMP) to ensure I followed

protocols on research practices and critically adhered to General Data Protection Regulation

(GDPR) laws. Ethical approvals were granted in July 2018 and July 2019, respectively (see

appendices A1 and A2).

134

Figure 6: Risk Zone Classifications in Nigeria

Source: UK Foreign Aid and Commonwealth Office (August 2018)

In my study, I consider the novelty of platform research, especially in a GS context

such as Lagos, and the ethical considerations cut across the different techniques used in this

study (see Appendix B3). A summary of my ethical considerations indicates that they cut

across participants’ informed consent, confidentiality and anonymity, risk of harm, public

versus private data, and payments and incentives (Madge, 2007; Allmark et al., 2009;

Longhurst, 2016; Henell et al., 2020).

• Informed Consent: In both SSI and FGD methods, participants requested were

asked for consent using printed papers and verbally. The consent and

information sheet were sent to their inboxes for participants recruited from

Facebook groups at least a day before. FGDs and some other SSIs were

requested before interviews in an audio format.

• Confidentiality and Anonymity: Within the consent forms, I instructed

participants about the confidentiality of SSIs and FGDs with necessary email

addresses or phone numbers from the University of Manchester to contact if

confidentiality was broken. The names of participants in this study were also

135

anonymised using pseudonyms (fake names) to protect their identity and ensure

the confidentiality of information provided regardless of its sensitivity. This

research chose not to use participant codes such as Participant 1 (P1), or Ride-

hailing driver 1 (RH1) because it reduces the identity of these participants to

mere codes, which is already prevalent with algorithmic management.

• Risk of Harm: The information collected in this research was not primarily

sensitive compared to research about vulnerable people, medical research,

bereavement research and others. Discrete information about platform

companies or platform driver-led start-ups remained confidential. In places

where participants' names were mentioned, it was pseudonymised, and the data

generalised to prevent any harm.

• Public versus Private Data: In joining Facebook public and private groups,

the data presented on the platform cut across public information and private

because of different accessibility restrictions. However, some groups (both

public and private) required me to officially join by clicking the button and then

being reviewed by administrators. Therefore, any information collected or

observed, such as screenshots of messages or different occurrences, was treated

as private by censoring participant names, paraphrasing quotes or summarising

in my words.

• Payment and Incentives: Participants were duly compensated for participating

in SSIs and FGDs by paying for their meals, mobile top-up units, and transport

costs. No cash payment was directly made to participants. However,

considering most of the interviews were via the phone, cash transfers were made

using a secured banking GTB app, with narratives such as “meal compensation

for interview or mobile or data units’ compensation for your interview time”

after each interview. Drivers that opted directly for mobile unit top-ups also

received compensation via the banking app. In the final FGD with platform

union representatives, I contributed snacks, drinks and smartphone power banks

to union drivers as a form of compensation.

136

3.11. Health and Safety Precautions

Despite the ethical approval (see Appendix A) for this research, the exploratory nature

indicates that entering the field, the researcher did not possess complete control over the

evolution of topics from the responses and actions of participants (Dickson-Swift et al., 2008).

As is widely considered, researchers face emotional and physical risks in the field, which can

hamper the research process (Bloor et al., 2008; Williamson and Burns, 2014). Considering I

am a resident of Lagos, my awareness of risk was high in both recruiting and interviewing

drivers. The face-to-face interviews and observations of the conventional taxi drivers were

undertaken during the day and mostly with prior notice to the targeted interviewees. In

situations with no prior notice, I approached participants in a busy environment and assessed

their initial responses before further discussions. FGDs that took place at night were between

9 pm and midnight and were in a controlled bank (GTB) car park with well-lighted streets, a

busy environment and detailed security personnel in sight. In addition, I was accompanied by

two friends as a further safety precaution.

Despite the difficulty of conducting face-to-face interviews with ride-hailing drivers,

my awareness of violence and insecurity in the election period (Bloor et al., 2008) informed

my decision to adopt a hybrid interview process with mobile telephone interviews as the

dominant form. Overall, in this research, the necessary health and safety precautions were

observed to avert potential risks and harms to the researcher.

3.12. Conclusion

At the time of conducting this study, scholarship on the emergence of ride-hailing was

lacking, especially within GS contexts. Based on the scarcity of information, early studies from

GN cities, in some cases, focused on Uber as a solution for reducing car ownership in cities

and thereby reducing carbon emissions. These studies developed into Uber disrupting mobility

systems, particularly the taxi industry globally. Likewise, this led to examining the business

model, labour dynamics, and how it impacted taxi industries, legal transport frameworks, and

drivers' livelihoods. Even in GN contexts, these impacts were indicative of the severity of the

problem and how several GS cities would struggle due to weak legal frameworks, corruption,

and poor transport planning. With time, it became evident that digital gig platforms such as

137

Uber were critical for rendering mobility and employing workers based on precarious

independent contracts.

Scarcity of information was a critical challenge in the Lagos context, reflecting the

hidden nature of platforms. This was compounded by the sparse data on taxi labour and the

history of innovativeness in Lagos. Also, this indicated the lack of data on the everyday lives

of drivers within the taxi industry. Ride-hailing platforms displayed their global strategy of

hoarding data from the public in cities. Similarly, this was a subsequent difficulty with ride-

hailing platforms in Lagos because these platforms would not communicate with researchers

or state government officials. This phenomenon made it difficult to access information critical

for this study. In navigating these difficulties, this study adopted a combination of qualitative

methods to improve the knowledge gap in emerging research. These methodologies included

SSIs, FGDs and participant observation which embodied mobile and online or virtual

observations. In other words, I joined online communities where drivers shared their everyday

lived experiences, which were critical in deducing the operationalisation of platforms in Lagos.

Other methodologies adopted in this research were critical for triangulation and creating

information necessary for bridging the knowledge gap between platforms and researchers,

drivers and regulatory bodies in cities.

Consequently, this also led to practical methodologies that were adaptable within the

African context and GS cities, which are vastly different to GN cities because of the similarities

between informal mobilities and gig work that existed before ride-hailing platforms. In

addition, thematically analysing the data extracted from the data collection was vital in

understanding the emergence of ride-hailing platforms in Lagos, examining the impact of

algorithmic management and the resistances of ride-hailing drivers, all of which are the basis

for the research in Lagos.

138

4. Chapter Four: The Emergence of Ride-hailing Platforms:

From Global North to Global South Cities

4.1. Introduction

In many cities today, ride-hailing platforms such as Uber have emerged as a

technological solution for seamless mobility and an effective medium for assigning and

managing labour compared to the taxi sector. In the previous chapters, this thesis has defined

and examined gig work, algorithmic management, and resistances across global gig platforms,

including online-based gig platforms and location-based gig platforms. This chapter first

examines the differences between the taxi industry in GN and GS contexts using three factors:

the regulatory environment, dynamics of trade unions, and the existing digitised environment.

Considering the influence of the UK and the US over Nigeria in terms of colonialism, labour

laws, fashion, language, and entertainment, these served as core comparators to GS contexts.

The goal was to establish why the taxi industry in GS contexts, specifically Nigeria, has

operated informally and further proliferated different impacts from platforms.

Against the backdrop of Uber's operational model, the chapter introduces four factors

that explain the political economy of Uber's emergence in GN contexts, emphasising the US

as a reinforcing context for enabling technology emergence. Based on a progressive, innovative

environment, strong policies despite early regulatory battles, the hype of platforms, and ease

of access to capital funding, Uber in San Francisco facilitated the emergence of Uber in the US

and globally. In reverse, these factors also serve as motivating factors for Uber to emerge in

other contexts like Lagos and appealing aspects for Lagosians. For instance, the low access to

funding in GS contexts leads to reliance on venture capitalists in GN contexts. Also, the hype

that ride-hailing platforms create flexible employment for the unemployed and underemployed.

Uber’s introduction in GS cities exploited the weak regulatory policy loopholes in African

cities and highlighted the poor state of the taxi industry as well as formalising the informalities

within these contexts (Carmody and Fortuin, 2019; Kaye-Essien, 2020). South Africa, Ghana

and Egypt provide core examples of Uber’s impact on the taxi industry, the differences with

GN cities and resistance from ride-hailing drivers. Considering that taxis have always been

gigs, this chapter identifies how ride-hailing platform gigs are different in terms of controlling

and managing ride-hailing gig workers and the processes of control it introduces in the GS

context, especially in Lagos, Nigeria.

139

This chapter, therefore, constructs a historical overview distinctly for four-wheeled

taxis drawing from interviews with the director of transport policy and head of taxi operations

(October 2018; August 2019) in what were classified as first-generation taxis, second-

generation taxis and third-generation taxis or ride-hailing platforms. Uber, at the forefront,

provided utopian ideologies from driving as a profession to driving as a skill by enabling

professional drivers, the unemployed and underemployed, to delve into earning money as a

side gig or full-time gig. Although it replicated informal taxi labour taxis, it became easy to

adopt ride-hailing platforms subsequently based on the hype of flexibility and autonomy of

labour – being your own boss, which transcended control from the state as well as traditional

taxi unions (Rosenblat and Stark, 2016; Mäntymäki et al., 2019). This sets up the subsequent

chapter to examine how algorithms atomise the labour process for drivers based on

surveillance, evaluation metrics and gamic elements.

4.2. Comparative Factors of Taxis: Global North vs Global South

This section discusses three factors that exacerbate the informalisation of taxi drivers

in GS contexts with GN contexts. These three factors form the background for examining the

differences in the taxi industry between GN and GS contexts and how this has contributed to

the informal and semi-informal mode of operations for taxis in GS contexts. Table 15

summarises the differences exposed in this section, where a formal regulatory environment,

formalised processes of trade unions, and an existing digitised environment in GN contexts

have been lacking in GS contexts, thereby consolidating the informalisation of taxis in the

latter.

Table 15: Comparative Factors of Taxis

Factors

Regulatory Environment

GN contexts tend to possess more

robust regulatory frameworks that

include workers and unions such that

progressive laws attempt to protect

workers from exploitation.

On the other hand, the existing laws

are porous and enable informal

processes in the taxi sector, such as

the irregularities in the taxi industry

and low entry barriers for drivers.

140

For example, fares are also adequately

regulated, which prevents drivers from

being exploited as well as exploiting

passengers.

For example, fares are not adequately

regulated, and drivers often exploit

passengers and can be exploited when

they are unaware of the driving routes.

Dynamics of Trade

Unionism

There are formalised trade unions

which are subject to the rules and

regulations guiding the affairs of

drivers. This means that there are fewer

corruption practices.

There are also formalised unions, but

they operate informally and often

outside the rules and regulations

guiding driver affairs. The trade

unions are associated with violent

practices, corruption, and opaque

management of drivers, such as

unclear dues collection.

Existing digitised

Infrastructure

GN contexts possess existing digitised

systems that facilitate the ease of fare

calculations through taximeters,

telephone booking via a dispatcher,

surveillance through CCTV and option

for installed cameras, and integrated

data storage for information of drivers

and others.

GS contexts often lack the presence of

a digitised environment that can

improve the efficiency of the service

and management of drivers. Until

today, many traditional taxis are

failing to innovate and thus getting

left behind.

4.2.1. Regulatory Environment

Since the Victorian era, taxi vehicle regulations in the Global North have intensified

over the years. For example, in the UK, regulation was aimed at entry, price, and quality of

vehicles. According to Toner (1992), entry regulations were summarised in three arguments.

The first reason was to reduce the unrestricted entry of taxi vehicles to increase the occupancy

rate and improve fares for drivers. The second reason is that free-roaming of taxis will lead to

an increased demand for rank spaces and traffic congestion, particularly around city centres.

Finally, regulating entry restrictions helps maintain the continuity of ownership and supply,

which impacts vehicle quality by reducing the administrative burden of more regulations.

According to Gwilliam (2005), economic regulations are fare-related because of

passenger exploitation due to administrative difficulties of providing differential access to

141

crowded taxi stands, asymmetry of information in an irregular market, and the vague nature of

transactions. There is also the regulation of fares to prevent weak bargaining positions by

ignorant passengers and exploitation by drivers (Toners, 1992). Different rules guided the

operations of PHVs in the early stages. These included prohibition from plying for hire –

booking was only through the telephone; the knowledge test was not compulsory for PHVs;

fares had to be agreed at the time of hiring the vehicle; it must not look like a hackney carriage

vehicle and must be suitable and safe, and comfortable for plying on roads (Butcher, 2016).

The regulation of fares was critical in reducing extortion or overcharging fares and conflicts

between drivers and passengers. For example, the Hackney Carriage Act in 1853 by the

Metropolitan Police was enacted to prevent extortion of passengers. While fare books, paper

maps, and guidebooks were critical tools for introducing transparency in the taxi sector, the

taximeter was a turning point in history for solving this problem (Dobraszczyk, 2008). The

taximeter was a real-time radio device for accurately calculating fares along distances. It was

also critical to control the behaviours of taxi drivers by relinquishing extortionate power over

passengers.

Quality regulations were enacted to address drivers' knowledge, insurance, vehicle

standards and the operator's fitness to work (Aarhaug, 2014). On the other hand, quantity

regulations address congestion problems – by limiting the number of licences available at a

particular time (Gwilliam, 2005). It is important to note that in London, for instance, the

regulatory framework prevents PHVs from being accessed by hailing or from a rank/stand

because, unlike the hackney carriage, a taximeter is not compulsory for a PHV and does not

embody the characteristics of a hackney carriage (Butcher, 2016). The Public Carriage Office

(PCO) implemented policies to ensure that taxi vehicles were in good condition and made it

mandatory to install taximeters (Georgano and Munro, 2008). The knowledge test introduced

in 1865 was critical to ensure that drivers admitted into the sector were professional,

knowledgeable about streets and landmarks, and possessed a general knowledge about London,

which is still relevant today (Garner and Stokoe, 2000).

In recent times, taxis and PHVs are regulated by the Department for Transport (DFT)

with statutory policies differentiating the former from the latter, some of which include vehicle

driver licensing (e.g., criminality checks for drivers), vehicle licensing (e.g., criminality checks

for vehicle proprietors), PHV operator licensing (e.g., booking and dispatch staff), and several

other policies (DFT, 2020).

Comparatively, many GS contexts possess a weak regulatory environment with poor

policies and a porous and weak regulatory environment that exacerbates informality in the

142

mobility sector, including the taxi industry. Like GN contexts, some regulations and policies

guide the affairs of the taxi industry in the GS to regulate entry, quality of service, level of

professionalism, and earning capabilities. However, the lack of enforcement, corruption, and

political bottlenecks have exacerbated the informality in the sector. For instance, in Ghana,

before plying for hire, the 1993 Local Government Act required taxi operators to obtain permits

from the district assembly, which according to Kufour (2019), is a low entry barrier for taxis.

This phenomenon does not limit taxis to their jurisdiction because there is also no enforcement

from the assemblies, further proliferating free entry to other jurisdictions without localised

knowledge and unrestricted fare regulation for drivers (Kufour, 2019). In Africa, such a low

barrier of entry often creates an informal mode of service with deregulated quality of service,

fares, and overall safety and security of drivers.

South Africa, which appears to be more progressive in its regulatory efforts for taxis,

equally experiences the informalisation of taxi labour. As an indication of informality, in 2005,

one in 10 drivers possessed mainly verbal contracts, partly due to low literacy levels in the

sector (Ingle, 2009). It also indicated that it was not as stringent in determining the similar

implement tests like the knowledge. To formalise the taxi industry, the Department of Labour

instituted that taxi drivers should be entitled to minimum wage, paid leave, and unemployment

insurance in July 2005 (Ingle, 2009). However, the owners of taxis and operators were

reluctant to implement the policy based on a sectoral determination by the Department of

Labour. Also, it was found in South Africa that while 45,000 drivers were registered legally,

about 10,000 drivers operated despite the extensive licensing efforts by the Gauteng

Department of Roads and Transport (Fobosi, 2019). These examples clearly show why and

how the African regulatory environment contributes to the penetration of ride-hailing platforms

and varying impacts compared to the GN contexts.

4.2.2. Dynamics of Trade Unionism

Over the years, there has been a decline in trade union memberships in GN contexts

such as Germany, the US, New Zealand, Australia, and other countries, which have exposed

workers to harsh market forces such as intense work disciplines, derecognition of trade unions,

the decline in relative wages, and contingent contracts (Heery and Abbott, 2000). In the UK,

trade unions have declined from the peak of 13.2 million members to 6.6 million members

(Elliot, 2021). While the increased level of insecure employment has been rampant over the

143

years, there have been gradual increments since the pandemic and economic meltdown in 2020

(Ibid).

In the UK or US, the registration of unions in the mobility sector is organised,

coordinated, and legally recognised by the state or context where it is situated. It also means

that worker rights requests must align with the existing legal and regulatory framework. As

part of an formal regulatory environment in GN contexts, the state has legalised and recognised

trade unions that can influence the law through peaceful protests, strike actions, court rulings,

and negotiations. In the US, the National Taxi Workers Alliance (NTWA) has advocated for

taxi worker struggles since 2011, with branches in New York, San Francisco, Philadelphia,

Austin, and a few others (NYTWA.org, n.d) which are legally recognised by the government.

Recently, the union has run nationwide campaigns to protect taxi drivers’ rights, their full-time

pay and full-time work that has been under threat since the advent of ride-hailing platforms.

Despite the decline and uncertainties of trade unionism in GN contexts, some unions

still advocate for taxi workers with legitimate requests and without any interference from

employers or the state government. For example, the Rail, Maritime, and Transport (RMT)

workers union in the UK were responsible for taking Uber to court because of the perception

that the Uber app is a taximeter, as highlighted in the subsequent section (Topham et al., 2015).

In Africa, unions can collectively bargain for their workers’ rights through protests,

dialogue and negotiations, strike actions, and demonstrations. While trade unions aimed to

operate more formally in advocating for their workers, they were often undermined by a large

informal labour workforce (Sutcliffe, 2012), non-compliance with labour regulations (Fobosi,

2019), corrupt practices, political bottlenecks, and lack of freedom of association due to an

alignment with government representatives. The Congress of South African Trade Unions

(COSATU) comprises over 33 unions with a membership of about 2 million workers (Aliu

2015). The union has been an integral counterforce against neoliberal policies that have led to

insecure employment and safety protections for workers. However, its alignment with the

African National Congress (ANC) governing political party has undermined the union’s

credibility, workers’ freedom of association, trust between union leadership, internal

coherence, and discipline (Aliu, 2015). Because of the blurred lines between informal and

formal labour, unionising across traditional employees and informal workers are cumbersome.

According to Schillinger (2005), informal workers are typically self-employed, and unlike

traditional workers, it is challenging to recruit them because of the difficulty in locating them

in large numbers. In addition, it is a challenge for unions to pay membership dues sometimes

due to the frequency of their work.

144

In many cases, there is a lack of synergy in advocacy requests because traditional

workers often negotiate based on wages, working conditions, safety protections, and other

realities that are entwined in an employer-employee relationship, often leading to a collective

identity crisis (Schillinger, 2005; Fobosi, 2019). For example, taxi drivers, operators and

marshals complain about misrepresentation under the South African Transport and Allied

Workers Union (SATAWU), mostly about taxi impoundment, traffic fines, and less about

precarious working conditions (Fobosi, 2019). Fobosi (2019) argues that taxi drivers often

focus strike actions on the state and not their employers due to intimidation by the latter.

Evidence from chapter six of this research will show in more detail why collective

bargaining by taxi unions is not often representative of all workers but aligned according to the

public interest, political ambition, corruption, and complicity from the state. These three

factors in this section contribute to the proliferation of informalities in the taxi sector and the

emergence of ride-hailing platforms, equally affecting the formalisation of platform unions.

The regulatory environment is critical across these factors, and it has shown that the taxi

drivers, ride-hailing drivers and informal workers in general in GS contexts are experiencing

the brunt of failed neoliberal agenda while still vulnerable to the impacts of new neoliberal

ideologies that further dissipate the power of collective bargaining.

4.2.3. Existing Digitised Infrastructure

The taxi system in the places like the UK/US already comprises a digitised environment

where taxi fares are calculated using taximeters on real-time map distances. This is a

mandatory requirement, especially for taxis and some PHVs in the UK, which is also indicative

of an upright regulatory environment, as discussed. In recent debates following Uber’s

emergence, a key argument by drivers was that the app is equivalent to a taximeter device

because it also calculates fares (Lomas, 2015). However, it was ruled in favour of Uber by the

UK High Court of Justice, citing that “the driver’s smartphone with the driver’s app is not a

device for calculating fares by itself or in conjunction with Server 2 (i.e., the algorithm

responsible for calculating fares) and even if it were, the vehicle is not equipped with it”.

[

]

Another aspect is the use of tele-booking since the 19

century in England. Following

the introduction of the PHVs in 1961 as a solution to insufficient taxi drivers due to the

difficulty in passing the knowledge test, by Michael Gotla, a law graduate and proprietor of

Welbeck Motors Limited (Motor Sport, 1961). Since then, it became compulsory for their

145

vehicles to be pre-booked before any trips. This was another component separating taxis from

PHVs because taxis were allowed to ply for hire, unlike the latter. It is illegal for drivers to

pick a trip that was not pre-booked and can be reprimanded by the taxi company or the operator.

An integrated data storage infrastructure can store information about defaulters as well as other

aspects of taxi labour. With this capability, the DFT can also develop and project policies

critical to improving the taxi industry.

The third component is the use of installed CCTV cameras in vehicles. For instance,

to reduce crime rates and enhance drivers’ and passengers’ health, safety, and security, TFL

provided guidelines for taxis and PHVs to install CCTV cameras (TFL, 2021). In 2020, only

5% of authorities required taxis to have CCTV-fitted cameras and 4% for PHVs, indicating

low uptake rates (DFT, 2021). Despite the low uptake rates, the UK possesses security CCTV

cameras that are integral for monitoring the behaviours of drivers and passengers. While this

may not be as detailed as personalised CCTV cameras, in cases of an investigation, relevant

footages from integrated databases are critical in solving the case.

By comparison, these existing digitised environments are not often a reality in many

GS contexts. This indicates that drivers are exposed to safety and security risks compared to

their GN counterparts. While it is compulsory in places like South Africa, Rwanda and Kenya,

drivers in South Africa and Rwanda revolted against using taximeters because they argue that

it reduces their fares (Georg and Rose, 2016). In South Africa, there were also reports of

drivers prolonging trips to increase fares and issues of faulty taximeters. The challenge is

exacerbated by unlicensed and unregistered metered taxis, to prevent licenses which they argue

only reduce the quantity and not the quality of taxi services. The Rwanda Utility and

Regulation Authority (RURA) started impounding vehicles without these meters in 2019 and

issued a fine of RWF200,000 (Igihe, 2019). With only a few drivers adopting the technology,

other drivers pleaded for more time to adapt to using taximeters (Igihe, 2019).

While there is not enough evidence in African countries concerning the utilisation of

technological enhancements like CCTV and adequate databases in urban cities and in the

mobility sector, this research on Nigeria suggests the absence of these apparatuses. These three

factors have historically shown the differences between the taxi industry in GN and GS

contexts, especially how the weak regulatory frameworks, poorly formalised unions, and a lack

of a digitised environment have exacerbated the informality of the taxi industry in GS contexts.

While more evidence will be discussed in subsequent sections, the next section briefly

introduces the operationalisation of ride-hailing platforms like Uber before outlining the factors

that led to its emergence in GN contexts.

146

In sum, it is essential to note that experiments of neoliberalism through deregulation

and privatisation of taxi firms and authorities remain negatively impactful in GS contexts

compared to GN contexts. GS contexts such as Nigeria have struggled to bounce back, to fully

manage, regulate, and take control of the taxi industry and further exacerbated informal labour

based on the poor quality of service, lack of safety protection nets, and a weakened trade union

base characterised by decentralised power, corruption, and political bottlenecks. This section

has evidenced why informality in the taxi sector is still occurring and how these have

contributed to the emergence of ride-hailing platforms discussed in subsequent sections.

4.3. The Operation of Ride-hailing Platforms: A Brief Overview of Uber

Ride-hailing platforms seem to be impinging on the taxi regime by redefining the mode

of access, driver management and passenger coordination, and data management for managing

information and predicting demand. The app renders developments in traditional taxi services

redundant because it can provide fares and payment through credit cards and provides

geolocation of trips and details about riders and drivers (Dudley et al., 2017). Accordingly, this

has sparked political debates in London, contending that the Uber app is a taximeter and should

operate as a taxi. These occurrences are central to the dynamic pricing algorithm, surge pricing

and rating system, which are critical components for algorithms to manage gig workers and

thereby facilitate control.

Data is essential for algorithmic functionalities. It is the blueprint through which Uber’s

routing algorithm plots the most efficient path for a trip, informs algorithms to predict potential

demand locations, and efficiently matches riders with driver-partners (Srnicek, 2017). In terms

of its base pricing, Cohen et al. (2016) claim that it exhibits similarities with conventional taxis,

such as a minimum total fare, defined price per mile, and a fixed fee and price per minute. It

is, however, different when demand exceeds supply because of its dynamic pricing system,

called the surge pricing algorithm, which institutes a multiplier effect on the base price x1 from

monitoring rider demand and available driver supply. It does this by extracting information

from traffic data and the activities of drivers and riders when they interact with the app

(Srnicek, 2017). For instance, a study carried out by Hall et al. (2015) shows how the algorithm

balances demand and supply and issues trips to riders with the highest interest (see Figure 7).

In summarising Figure 7, surge pricing activates by increasing fares when demand exceeds

supply (blue) in order to instigate drivers to meet demand. This was a study carried out on New

147

Year’s Eve to see how the surge multiplier increases to balance supply. What can be deduced

is that opening the app (red lines) in itself triggers a potential surge zone (Hall et al., 2015).

Figure 7: A Practical Example of Surge Pricing

Source: Adapted from Hall et al. (2015)

The Uber labour administration can be classified as two-sided algorithm management

for drivers (Ofstad, 2017). On the one hand, through various algorithms, Uber exerts substantial

control over its drivers; on the other hand, it portrays them as freelancers with high work-time

flexibility. The implication of the surge pricing algorithm, in most cases, psychologically

manipulates drivers, like a videogame, to continue driving. According to Schieber (2017), Uber

utilises a similar algorithm to the Netflix feature, which encourages binge-watching after the

next programme automatically loads. The algorithm in Uber sends drivers their next fare before

completing the trip, encouraging them not to log off because they are close to reaching their

goal of the day (Figure 8) (Schieber, 2017). Rosenblat and Stark (2016, p. 3771) argue that

Uber influences the relationships between supply and demand through gamic elements and

behavioural tools such as blind passenger acceptance, surge pricing, and a conflation of real-

time and predictive demand for drivers.

148

Figure 8: Nudging Drivers to Keep Driving

Source: Scheiber (2017)

Part of these gamic elements includes the ratings, cancellation, and acceptance rates

which are facilitated by the platform algorithm. Uber drivers’ employment eligibility is

determined by how riders rate them, such that drivers need to score at least 4.6/5 and above to

continue driving (Rosenblat and Stark, 2016; Rosenblat, 2018). Drivers with consistently low

ratings receive deactivation notices without proper clarity on the decision (see Figure 9).

Riders, on the other hand, experience less algorithmic control from the app; ratings may not

matter to drivers who have not made enough money in a day. Also, high trip cancellation rates

or low acceptance rates on the platform signal a lack of productivity from a driver and could

lead to a temporary deactivation or limitations of trips. Compared to traditional taxis, drivers

seem to experience stricter controls, but this time from platform companies that utilise

algorithmic management as a weapon of control.

149

Figure 9: Snapshot Example of Driver Ratings and Trip Metrics

Source: Adapted from Rosenblat and Stark (2016)

Traditional taxi services are responding by developing apps to facilitate mobility with

similar apps like ride-hailing platforms, indicating a transformation at the regime level because

of broader landscape developments in consumer preferences. These services are classified as

E-hail services (Shaheen and Chan, 2016). Examples include Mytaxi, Flywheel, Hailo, and

other companies (Shaheen and Chan, 2016). For instance, with the Mytaxi app, taxis can be

requested, showing arrival times and progress in traffic (Gösseling, 2018). The difference,

however, is that they indirectly still operate as a taxi, in that professional-rated drivers can be

150

selected via the app, unlike Uber, where it is anonymous (Gösseling, 2018). They are still

subject to traditional taxi regulations and cannot utilise surge pricing to meet the demand for

ride-hailing apps (Chan and Shaheen, 2016).

This section has briefly highlighted what constitutes Uber and how it works, especially

in a GN context. Subsequent sections in the case site Lagos will unpack more of the intricacies

and display of algorithmic management. However, the next section critically discusses factors

that led to Uber’s emergence in the GN contexts with core examples from the US and how this

paved a pathway for its arrival in GS contexts.

4.4. The Factors of Platform Emergence in Global North Contexts: From

the United States to Africa

The emergence of ride-hailing platforms in the GN contexts, especially in the US, was due to

a series of factors as analysed in this section and summarised in Table 16. These factors

consolidated their power to emerge in other contexts, especially the African context in the GS

where, as Table 16 also summarises, the context is quite different and leads to transfer of

platforms from the GN rather than indigenous innovation. In other words, examining the

political economy of platform emergence based on how platforms like Uber navigated the

challenges faced in the US will help understand how this became a motivating factor in shaping

the platformisation of labour globally.

Table 16: Comparing GN Factors of Platform Emergence to GS contexts

Factors

GN (e.g., US, UK)

GS (e.g., Nigeria; Ghana)

Innovative

Environment

Due to an innovative enabling environment

based on strong regulatory frameworks and

better access to venture capital funding, GN

contexts are more experimental.

Due to a low-innovative enabling environment

based on weak regulatory frameworks, political

bottlenecks, and poor access to venture capital

funds, GS contexts are less experimental; for

example, in relation to indigenous platform

emergence.

Venture

Capital

Funding

GN contexts have better access to venture

capital funding and more opportunities for

innovative ideas.

GS contexts have little or no access to venture

capital funding to develop ideas or compete

against international platforms. Hence, it is

overly reliant on funding from GN venture

capitalists.

151

Regulatory

Battles

GN contexts experience many regulatory

battles due to the pre-existing strong

regulatory environment, which also played

a part in Uber’s rollout in 2009.

Less regulatory battles based on the pre-existing

poor and weak regulatory environment with

inadequate policies that stifle the growth of local

ride-hailing platforms.

Hype and

Rhetoric

Platforms from GN contexts possess higher

levels of hype, validity, and reliability with

tales of flexibilisation of labour, solutions to

unemployment, and improved safety than

traditional taxis and others.

Platforms from GS contexts possess lower levels

of hype, validity, and reliability because there is

no substantial track record of developing

successful platforms. Hence, the drivers and

policymakers are more reliant on platforms from

the GN context to fulfil the hype of creating jobs

for the unemployed, reducing congestion,

improving safety and others.

4.4.1. Innovative Environment: Early Experiments in the United States

In the US, the invention of the internet and mobile phones brought about innovative

processes in accessing taxis and has transformed mobility since the 1990s. For instance,

carpooling and vanpooling were easily facilitated with the help of telephone-based systems by

an operator or by registering online via the website (Kodranksy and Lewenstein, 2014). This

further facilitated policies to improve modal share for these novelties by developing high-

occupancy vehicle lane facilities (Shaheen and Chan, 2012). Levosky and Greenberg (2001)

highlight ridesharing experiments in five cities using telephone-based systems and mobile

phones to facilitate the real-time matching of drivers and passengers. In these experiments,

between 1993 and 1999, it was found that SMS interfaces embedded in mobile phones and

email because of web browsing facilitated more rides (150 rides out of 500 requested rides).

This facilitated the embeddedness of platforms such as push-email features and GPS

technologies in Blackberry (2002) and Apple (2007) – the first and second smartphones of the

information era. The introduction of smartphone devices further enabled faster and easier social

interactions through social networks and email responses, stimulating travel. For instance,

Logan Green and John Zimmer introduced Zimride – a carpooling ideology based on the

Facebook social network launched in 2007 due to poor vehicle occupancy in personal mobility

in US cities (Gustin, 2012). With the introduction of the 3G iPhone in 2008, the Facebook

152

platform became one of the earliest mobile apps embedded in the smartphone app store with

tracking capabilities (Chin, 2008). Following safety concerns at that time, the smartphone with

embedded features could enable people to share their locations on Facebook, for instance,

while matching riders in real-time (Chan and Shaheen, 2012).

With the invention of smartphones and other developments, such as the app platform,

a small network of actors and markets began to expand. For instance, the first ride-hailing

service app, called 'ride charge', was invented in 2008 to facilitate on-demand real-time taxi

hire by pushing a button (Kincaid, 2008). Accounts as to why this technology did not emerge

at the time are unclear. Speculations show that market investments and access practices were

not as forthcoming. For instance, passengers did not receive an in-app notification when a taxi

was dispatched, and it did not integrate toll estimates based on distance travelled (Kincaid,

2008). This phenomenon, in turn, signalled a potential limitation with the need for telephone

operators in taxi-type services and new practices of consumers, which became evident in ride-

hailing apps from 2009 onwards.

In 2009, ride-hailing platforms were introduced as facilitators of urban mobility and

employers or intermediaries of labour due to ICT developments based on the internet and GPS

location technology. Uber (launched as UberCab), founded by business and engineering

graduates Travis Kalanick and Garret Camp in 2009, became the first official ride-hailing app

(Flores and Rayle, 2017). It is important to note that UberCab launched as an exclusive on-

demand limo service, which was licensed by the California Public Utilities Commission

(CPUC), while local taxis were licensed by the San Francisco Municipal Transport Agency

(SFMTA) (Flores and Rayle, 2017). The platform was invented due to difficulties in the taxi

regime in the US at the time. Before then, pre-booking taxis took over 30 minutes to arrive in

2006 and even until 2012. On taking the first UberCab trip in San Francisco, the fare cost

significantly more than a traditional taxi but gained traction amongst tech-savvy people in the

Bay Area (Hartmans and Leskin, 2019). Following the UberCab experience, Zimride founders

launched Lyft in May 2012 as a competitor for UberCab and for traditional taxi fares, based on

its lower-priced model and an established network (Gallagher, 2013). Also, Sunil Paul, a serial

entrepreneur, launched Sidecar in beta-mode in June 2012, granting more power to users by

displaying cheaper fares based on the average amount from previous passenger fare payments

(Constine, 2012). This competition forced Uber to launch a lower-priced product known as

UberX (see Figure 10). For instance, Rayle et al. (2014) found, in their study of 380 users, that

53% of rides were with UberX. Cohen et al. (2016) identify 50 million UberX sessions,

affirming its popularity within mobility systems.

153

Figure 10: Uber Product Classification

Source: Derived from Cohen et al. (2016); Uber.com

Because of strong laws, ease of access to funding opportunities, and an all-around

enabling business environment, the innovativeness in GN is much higher compared to its GS

counterparts. According to a report published by the World Intellectual Property Organisation

(WIPO) (2021), the Global Innovation Index (GII), which uses seven pillars to assess the level

of innovations in countries such as: institutions, human capital and research, infrastructure,

market sophistication, business sophistication, knowledge and technology outputs and creative

outputs, is critical in establishing knowledge about the innovative capacities of countries. In

the 2021 ranking, countries like the US and UK ranked third and fourth out of 122 countries

surveyed. The least-ranking pillar in all seven categories was 23 out of 132 for infrastructure

in the US. The least ranking category for the UK was Business Sophistication at 21 out of 132

countries. This survey suggests innovative strengths and capacity of contexts, with GN

contexts more dominant and powerful in this regard. For instance, this facilitates the ease of

154

innovating to solve critical challenges such as the post-recession unemployment crisis in the

US, which has since led to several gig work and online work platforms.

Based on early experiments in an innovative environment like the US, ride-hailing

platforms like Uber have continued to expand into global markets. On the other hand, contexts

like Nigeria ranked number 118 out of 132 countries globally (WIPO, 2021). Across all seven

categories, business sophistication scored the best, ranking 76 out of 132 countries. While

innovativeness has increased in Nigeria in the last few years, the market shareholders in the

ride-hailing platform sector remain Uber and Bolt. Outside of the ride-hailing sector, most

start-up firms and innovations rely on external funding from VCs/organisations in the US and

other GN countries discussed in the next section.

4.4.2. Venture Capitalism: Ease of Access to Funding

Access to funding has been at the core of Uber's growth and global expansion. As a

global world power, innovative powerhouse, and enabling business environment, the US paves

the pathway for developing new technologies. It is essential to understand the influence of the

US and, specifically, San Francisco in the emergence and expansion of Uber.

The US possesses a significant share of venture capitalist funding and investors with

locations such as San Francisco, New York, Boston, Seattle, and several others (Adler and

Florida, 2021). Extrapolating from archived data by the PWC, Florida (2017) found that the

Bay Area increased venture capitalist investment from 22 per cent in 1995 to 46 per cent in

2015, bringing in over $30 billion in that year. Accordingly, Florida (2017) argues that San

Francisco, San Jose, New York, Boston, and LA contribute over 70 per cent of venture

capitalist funding in the US. Other locations outside of the US include London, Beijing,

Singapore, Barcelona, and others. In a 2021 mapping of access to funding by Crunchbase, the

US alone amounted to about $269 billion in venture funding, with China a distant second with

$60.6 billion (Glasner 2021). There is a greater level of hype, trust, and validity surrounding

tech hubs like Silicon Valley in San Francisco, driven by successful companies like Apple and

Alphabet that emerged from that context. The success and global domination of initial start-

ups like Apple create a ripple effect by attracting even more investors to commit funds that

propagate technological ideals that align with their profile. However, why is San Francisco

critical in the development of several start-ups that have eclipsed the world?

According to Adler and Florida (2021), popular tech companies are often concentrated

and emerge from high tech centres such as Silicon Valley and other global cities like London

based on population, talent pools, access to universities and economic size, which possess high

155

attractiveness for venture capitalists. It is also important to note that the geographic

concentration of tech companies in places like Silicon Valley facilitates job mobility between

firms and better matching of employees and increases worker incentives which reduce selection

issues for employers based on specific human capital (Adler et al. 2019). In this case, the

proximity to critical resources has bolstered the comparative advantage of Silicon Valley such

that new venture capitalists are convinced to raise large sums of capital even when the company

is not publicly listed, just like with Uber before now (Gupta and Wang 2016). This

phenomenon improves the chances of expansion and scalability for a business and different

series of funding, which has been critical in Uber's development and global expansion.

At the early stage of Uber's journey, Garrett Camp and Travis Kalanick raised $200

thousand in funds to kickstart the platform's operations, followed by seed funding of $1.2

million in 2010. Notably, in 2011, the platform acquired series A and B financing of $11

million and $32 million, respectively, i.e., funding from external investors that subscribe to the

ideals of the business through stocks or ownership percentage to facilitate the expansion of the

business (Bonini and Capizzi 2016; Crunchbase.com). This funding was critical for its

expansion into other states in the US, such as New York, and internationally in Paris, France,

in December 2011 (Tsotsis, 2011). Since then, the company has received series C to series G

funding from notable investors such as Menlo Ventures, Goldman Sachs, Microsoft, Google

ventures, and many others. Uber as a beneficiary of the influence of the US and Silicon Valley

on its business model, has raised over $25.2 billion after 33 funding rounds between 2010 to

2020 (Bonini and Capizzi, 2019). A key facilitator in the Uber expansion journey was

achieving unicorn status, i.e., its market valuation reached $1 billion in 2013 (Brail 2020).

Other ride-hailing platforms such as Didi-Chuxing, Lyft, and Grab achieved unicorn status in

2014. Bolt achieved this status in 2018 (Brail, 2020), although it had already emerged in

Nigeria in 2016. This phenomenon also highlights the significance of the US and Silicon

Valley in Uber's radical expansion across GN contexts, emerging in over 10,000 cities and 74

countries compared to Bolt, which is present in 45 countries and over 400 cities (Uber.com;

Bolt.com). Tallinn in Estonia is growing to be part of the cities with concentrated efforts and

attractiveness for innovation and funding, especially with Bolt at the forefront of development

with a recent market valuation of $8.4 billion (Lunden, 2022). This phenomenon further

illustrates the concentrated nature of capital based on the headquarters of ride-hailing unicorns

across only ten cities, extending to other significant networks across 29 cities globally (Brail,

2020).

156

The ease of access to venture capital in GN contexts like the US, alongside other factors

discussed, has been a critical and motivating factor for the emergence of Uber in GS contexts.

It also indicates that GS contexts rely on both the innovation of these unicorns and external

venture capital funding from GN contexts and other upper-income GS contexts such as China.

In Africa, a few platforms have achieved unicorn statuses, such as flutterwave, Andela, Jumia,

Opay and others. However, the backend of some of these companies largely revolves around

foreign VCs or foreign founders and co-founders. For example, Opay, which reached over a

$2 billion valuation, was founded by Chinese billionaire Yahul Zhou in 2018 (Iwayemi, 2022).

Flutterwave, a fintech company that facilitates payment in Africa, has headquarters in San

Francisco and Lagos, Nigeria, with crucial investors such as Avenir Growth Capital and Tiger

Global, a US hedge fund and investment firm (Kene-Okafor, 2021). While there is no African-

based ride-hailing unicorn yet; this has consolidated the argument that GS contexts are

intensely reliant on tech hubs in GN contexts. With more foreign VCs now funding tech

platforms in Africa (Kazeem 2017a), it remains to be seen in a few years if Silicon Valley and

other clustered tech hubs will remain powerful because of a post-pandemic global shift in

working patterns.

4.4.3. Regulatory Strength and Permissibility: The Battles Within

GN contexts like the US with strong regulatory environments also indicate adherence

to pre-existing policies. In other words, the policies do not provide preferential treatment to

new entrants like Uber. According to Thelan (2018), two core political-economic arguments

explain why and how Uber emerged in the US. The first suggests that the advent of new

business models such as Uber pushes deregulation, such that governments without awareness

or weak regulatory framework are overshadowed and which leads to the exploitation of the

gaps while establishing a solid customer base, making it difficult for governments to recover.

While the US has always been heavily regulated, the government initially lacked awareness or

knowledge to curb its surge. This thesis will show that GS contexts like Nigeria have

experienced a worse fate because of the already weak regulatory framework.

The second argument is that liberal economies like the US, despite strong regulatory

background, are more receptive to technologies or start-ups compared to more coordinated and

less permissible markets across Europe (Thelan 2018). In addition, relaxed rules and

regulations exist for tech start-ups, even more so post-2007-2009 Great Recession. In

California, for example, a non-compete clause enables start-ups and more established firms to

157

hire the best talents within the jurisdiction (Adler and Florida, 2021; Brail, 2020). There is also

a greater degree of permissibility in hiring staff and firing them without recourse, thus allowing

lean platforms like Uber to continue to innovate competitively but also at cheaper costs.

However, this does not imply that companies are entitled to steal a competitor's intellectual

property, as this became a significant case between Waymo – the self-driving Google company

and one of their staff hired by Uber (Lobel 2017). This further indicates that, as it was relatively

straightforward for Uber to emerge in the US, it also experienced regulatory battles, shaping

the platform today.

In the US, California, where Uber was founded, there is evidence of regulatory battles

between ride-hailing apps and regulatory bodies. Because platforms were impinging on the taxi

industry and pre-existing laws that defined the taxis and rights of workers, ride-hailing

platforms faced regulatory battles in the US and globally. For example, incumbent taxi actors

SFMTA and CPUC were protesting the nature of the service (Flores and Rayle., 2017;

Conneran et al., 2017), which was one of the first regulatory challenges experienced by Uber.

This was because the platform was not adequately defined. It operated as a taxi but was

classified as a technology service, leading to massive protests from drivers. The protest from

taxi drivers was that UberCab accepted street hailing, making it a taxi service by law and not a

limo service, leading to both the CPUC and SFMTA issuing a notice to 'cease and desist'

(Flores and Rayle, 2017).

The phenomenon of Uber's emergence was also due to the scarcity of taxis based on

exorbitant licence fees and high entry barriers for drivers. Compounding this was Uber evading

the rules and routines of the medallion system, such as insurance requirements and coverage,

driver certification to ensure safety, alcohol testing certification and other requirements

(CPUC, 2012; Flores et al., 2017; Conneran et al., 2017; Dudley et al., 2017). The Medallion

taxi system, which are yellow taxicabs, was typically situated in New York since 1907 (Teo,

Kimes and Yong, 2019). In 1971, the New York City Taxi and Limousine Commission (NYC

TLC) was established as a solution for monitoring and regulating licenses yearly, ranging from

11,787 to 12,187 medallion licenses (Skok, 2003). Like the knowledge test in England, drivers

were required to attend training schools, write English proficiency tests and write a final

examination which affirms their knowledge of the city and ability to convey passengers in the

city (Skok, 2003). Due to the scarcity of medallion taxis and artificial inflation, the value

increased from $2,500 in 1947 to $1 million in 2014 (Khafagy, 2021; Van Zuylen-Wood,

2015). The limited yearly licenses and increase in the value of medallions indicated that there

were often more qualified drivers than licences, which affected the services. For example, in

158

1994, there were 40,000 drivers with only 11,787 licenses available (Skok, 2003). The

medallion taxi system in the US was unstainable for drivers and their livelihoods because of

the cost of securing a medallion, coupled with exorbitant taxi fares for taxis in San Francisco

(see Lam et al., 2006, pp. 30-39). While the value for medallions has depreciated to $80,000

and is now regulated to 13,500 licenses per year, there are more drivers in debt on average of

$600,000 (Khafagy, 2021).

Lyft and SideCar also received a similar notice, and eventually, they were all fined

$20,000 for not adhering to the regulation (Chen, 2012). According to the CPUC, Uber

violated two sections of the Public Utility Code: (a) operated as a charter-party carrier of

passengers without an operating authority. (b) advertised as a charter-party carrier without

including the number of a permit or certificate issued by the commission (CPUC PSG-3018,

2012). In response, UberCab changed its name to Uber both as a response to the regulatory

bodies (indicating that it is not a taxicab service) and in response to competitors based on a

lower-priced service and other subsequent products (Flores and Rayle, 2017). Alongside

California, regulators in about 13 other cities, including New York, Los Angeles, Washington,

and Chicago, further drafted guidelines that address the concerns of taxi drivers based on (a).

a luxury car must not use a GPS device as a meter in calculating fares; (b). while driving, an

electronic hailed trip must not be accepted via a smartphone; (c). a request made less than 30

minutes in advance must not be accepted by a driver (Chen, 2012). These guidelines were

targeted at constraining the power of Uber, but instead, it increased its visibility to the world.

In addition, despite the regulations and enforcements, Uber was successful in lobbying in at

least 50 states and cities to reinstate their service in places where it was banned (e.g., New

York, Austin, Philadelphia) or continued to operate without clear evidence of compliance (e.g.,

Seattle, Chicago) (Helderman, 2014; Dubal et al., 2018).

So far, Uber has experienced regulatory challenges where the platform has been banned

or experienced legal challenges globally, especially between 2010 to 2018, even in contexts

such as Sweden, which has possessed a deregulated market since the 1990s, which created a

fluid system for taxi operators and drivers (Thelan, 2018). Although Uber had a relatively

straightforward emergence in Sweden, the platform had to succumb to specific rules and

regulations, such as the requirement of commercial licensing. Other examples of more intense

regulatory battles were in parts of Europe (e.g., Bulgaria, Denmark, Germany); North America

(e.g., Oregon in the US, Vancouver in Canada); and Asia-Pacific (China, Northern Territory in

Australia, most of Japan) (Hao 2017). These regulatory issues border around worker status,

insurance, employee benefits, taxi licensing, safety, public goods (e.g., congestion), and

159

consumer protection (e.g., accessibility) (Martini, 2017; Dubal et al., 2018). While there were

protests, demonstrations, and attacks from taxi drivers (Vilakazi and Chisoro, 2017), it is

notable that between 2012 and 2018, there was no ban from any African country. This is partly

because of the already weakened regulatory environment, perceived employment alleviation,

and buying into aspects of the hype and promises of the platform discussed in a separate

section.

However, these regulatory battles have today shaped the emergence of Uber in GN

contexts and more globally. For drivers, there is a gradual process of understanding their rights

and how best to channel their grievances. As this study demonstrates in subsequent chapters,

stories of Uber's regulatory challenges in GN contexts are awakening African regulators either

by defining policies that should shape the taxi industry or by redefining existing policies that

formalise the existing informal practices.

4.4.4. Hype and Rhetoric of Ride-hailing Platforms

The emergence of Uber increasingly generated hype through good and bad press

throughout the world. Its regulatory battles, as mentioned above, are some examples of bad

press the platform experienced. In terms of good press, academic research and news outlets

have continuously highlighted how ride-hailing platforms contribute to the public good, solve

unemployment and improve workers' safety and security. Below, I briefly highlight what this

means and how this rhetoric contributed to propagating its business model globally.

As part of the hype, Uber was said to be beneficially for the public good. It was

integrated into the smart city agenda, notably on reducing traffic congestion (e.g., more people

leaving their vehicles at home), reducing carbon emissions based on reduced driving, improved

wait time compared to taxis and a cheaper first and last mile mobility conduit (Uber 2016;

Conner-Simons, 2017; Lomas 2017b). It is important to note that some of the services alluded

to in Uber's argument for the public good were based on their carpooling service. However,

only the brand image was recognised, with not enough knowledge about the nuances of its

different products.

Studies showed that ride-hailing platforms could reduce congestion and improve

efficiency (Rayle et al., 2014; Cramer and Krueger,2016). For instance, in conducting a survey

in San Francisco, Rayle et al. (2014) found that vehicle occupancy was higher for passengers

(1.8) in the UberX product than taxis (1.1). This indicates that UberX shows more significant

potential for reducing traffic congestion by reducing the number of cars on the road (see Li et

160

al., 2017). Similarly, Cramer and Krueger (2016) highlight that Uber exhibits high vehicle

utilisation and increases efficiency because drivers spend less time and resources roaming the

streets for fare-paying riders, potentially reducing traffic congestion.

Compared to taxis, Rayle et al. (2014) also found that 35% (of 380 users) waited ten

minutes or less, compared to 90% for ride-hailing workers, and 67% waited for 5 minutes or

less. In line with ride-hailing platforms reducing congestion and a mobility conduit and

substitute, Li et al. (2017) found that peak surge pricing could potentially delay or divert peak-

hour demands causing riders to use public transit because of high prices in that period.

Similarly, Shaheen and Chan (2016), from their study in San Francisco, show that 40% of

people reduced their driving due to ride-hailing apps, and 33% chose to use public transit (bus

or rail) in the absence of it. Interestingly, the absence of ride-hailing apps also induced the use

of taxis (33%) in their study.

Another aspect of the hype of platforms was their perceived safety and security based

on the platform technology. The perception of safety and security developed from the

availability of data and surveillance – the possibility of tracking and solving complex situations

or preventing them from happening because of data. From a technological point of view, based

on the possibility of real-time identification and recognition, platforms appeared to be safer

and more secure than local taxis, despite the lower barrier of entry for drivers. Riders can

recognise a driver's face, phone number, plate number, and type of car before a trip. The

perception by users was that the platform would improve safety and security and limit conflict

amongst users. Starting as a luxury service, drivers had to offer professionalism by first passing

the assessment test, vehicle inspection before accessing the platform, learning how to use

digital maps, and treating passengers with respect. Also, the integration of an evaluation

system based on 5-star ratings bolstered the safety and security of users, especially passengers,

as it was critical to prevent deviant behaviours (Rosenblat and Stark, 2016; Young and Farber,

2021).

Paradoxically, as this research shows, there have been increasing concerns about the

safety, security, and even privacy data issues that often expose drivers and passengers to risks

(Rogers, 2015). It is worth noting that aspects that propagate safety and security on platforms

like Uber do so in favour of passengers and other actors because of its business as usual – the

customer is always right. While passengers have experienced attacks, harassment, and assaults

on the platform (Rogers, 2015), it is drivers bearing the brunt of these risks as they have a lower

form of bargaining power. Nevertheless, these also feed into the hype of Uber because pre-

161

existing taxi services, especially in GS contexts, experienced similar challenges without the

infrastructure that facilitates investigations that prevents deviant drivers or passengers.

The final aspect this section will briefly discuss is the perception of unemployment

alleviation or employment provision. According to the US Bureau of Labour Statistics (BLS)

(2012), the unemployment rate was about 10% following the great recession, almost as high as

the 10.8% record in 1982, with over 30 million people losing their jobs (Kalleberg and

Wachter, 2017). There became an increasing lack of job security, and Uber and other rising

gig platforms provided a short-term solution to gainful employment with increased

flexibilisation of labour but without adequate safety nets. Following the great recession

between 2007 and 2009, ride-hailing platforms like Uber were critical in implementing some

of the ideals of neoliberalism which were the flexibility and autonomy of labour, as highlighted

in chapter two. Uber claimed that the average earnings for drivers, specifically on UberX, were

$74,191 a year in San Francisco and $90,766 in New York, respectively (McFarland, 2014).

This was more than taxi drivers' earnings of $30,000 a year and, at the time, more than tech

workers' average salaries of $87,811 a year (D'onfro, 2014). While verifying these earnings is

difficult, potential drivers perceived that with a low barrier of entry and more wages, becoming

an Uber driver is a lucrative business. In essence, this was a contributing factor that expanded

the for-hire taxi sector in New York from 47,000 in 2013 to over 100,000 in 2018, with Uber

constituting two-thirds of the latter (Bellafante, 2018).

Despite its negative externalities, ride-hailing platform work that emerged from GN has

facilitated employment creation which is seen as an alternative solution to high unemployment

and underemployment as well as poverty alleviation in GS contexts (Meagher, 2018; Pollio,

2019; Keskinen and Winschiers-Theophilus, 2020). According to Keskinen and Winschiers-

Theophilus (2020), platforms like Uber emerge in GS contexts where there are no safety nets

for informal workers and even some formal workers, thus are seen as a viable means of earning

a livelihood. Added to the high profile of claims around reducing congestion and improving

safety – which are key issues in GS cities – this has created a favourable environment for entry

of ride-hailing platforms from the GN.

4.5. Ride-hailing Platforms in the Global South: Opportunities and

Challenges in Africa

162

The arrival of Uber in Africa is also linked to the factors discussed in section 4.4 above.

More specifically, the lack of innovation, aspects of hype and rhetoric, poor access to venture

capital and weak regulations were critical motivating factors for Uber and other ride-hailing

platforms to exploit. Before the advent of digital and information and communication

technologies in most African cities, people always shared cars in different informal ways, with

vehicle drivers and conductors acting as intermediaries for accepting payments and passengers.

In Nigeria, the vehicles used were Kabu-kabu (shared yellow taxi) and Danfos (local

minibuses) which are used even today; in South Africa, they are generally called Kombi or

minibus taxis (Barret, 2003); and in Kenya and Uganda, they are called Matatus or minibus

taxis (Mutongi, 2006). It is, however, essential to note that most of these utilised minibuses

would not usually fall under the definition of a ‘taxi’ as defined in places like England.

[

]

More generally, they would fall under the classification of ‘buses’ or ‘minibuses’,

which can carry nine passengers or more (Cooper, 2010, cited in Aarhaug, 2014; Butcher,

2016). Therefore, the term ‘taxi’ is socially constructed and can broadly be used in generally

defining the African perception of taxis. While it is critical to observe taxi-type or innovative

services in this study, it is important to note that the core focus is on traditional four-wheeled

taxis or sedans in African cities because of the similarities with ride-hailing or ride-sourcing

platforms as used widely globally (Rayle et al., 2014; Rosenblat, 2018). However, because of

how loosely ‘taxis’ are used in an African context, I would briefly highlight the different

informal modes, especially in the section below discussing taxis in Lagos.

With the impact of globalisation, ride-hailing platforms have permeated GS cities,

especially African cities. The world is now more connected than in the 20

century with ICT

development, the ubiquity of the internet, cloud services, and data analytics that connects

industries, companies, and individuals through data flows, digitally enabled capital, and other

facets beyond borders (Luo 2021). For platforms to realise the disparities in the quality and

price of production, platforms tend to find other suppliers of goods and services across borders

to reduce overall costs while improving the functionality of their product as well as its

competitive advantage over competitors (Hill 2008; Katerina et al., 2014). This phenomenon

is part of the strategies of multinationals like Uber to benefit from low-wage, low-skilled, and

informal labour instead of seeking access to high talent pools across varying wage and skill

levels (Branstetter et al. 2018; Brail, 2020). As highlighted, Uber possesses the capital and

technology to penetrate the African context, with less challenging hurdles to evade due to weak

regulatory frameworks, poor innovation tendencies, and issues of unemployment and the

163

nature of informal labour. In other words, Uber’s motivation was to exploit these loopholes by

providing improved technology, solution for the unemployed and underemployed, and data

infrastructures for effectively managing the workforce, including solutions for taxing gig

workers.

In 2019, Pierre-Dimitri Gore-Coty, the previous Europe, Middle East, and Africa

(EMEA) regional manager, stated in a BBC interview (2019) that Uber was present in over

seven African countries, with five million riders and over 150,000 drivers. In considering the

hidden display of information from platform companies, it was difficult to validate the claim

of over 150,000 drivers and five million rides. The lack of validity is because, in Egypt, it was

reported that there had been over 4 million rides and 157,000 drivers in that country alone

(Reuters, 2018). Recent research from the Uber website shows the platform's presence in 60

cities across eight African countries, including South Africa, Nigeria, Kenya, Egypt, Morocco,

Tanzania, Ivory Coast and Uganda (see table 17). Below, I discuss the emergence of Uber in

African cities in countries such as South Africa, Egypt and Ghana, including the impacts on

the taxi industry and resistance from drivers and existing unions. While the section below does

not delve into the nuances of these cities, it sets the tone for examining the emergence of ride-

hailing platforms in Lagos in subsequent sections.

Table 17: Regions Using Uber in Africa

Countries in Africa

Number of cities

City Examples

Egypt

Cairo, Alexandria, Delta etc.

Ghana

Accra, Kumasi, Tamale etc.

Ivory Coast

Abidjan

Kenya

Nairobi, Mombasa, Eldoret

Morocco

Casablanca

Nigeria

Lagos, Abuja. Kaduna, etc.

South Africa

Cape Town, Johannesburg and Pretoria etc.

Tanzania

Dar Es Salaam, Dodoma etc.

Uganda

Kampala

In South Africa, Uber first launched in Johannesburg, Pretoria and Cape Town in 2013

(Pillay, 2016) after closing a series C funding round of $258 million from Google Ventures

164

(GV). Uber was advertised as a safer and more reliable taxi than local taxis while leveraging

off the entrepreneurial landscape and global tourism of Cape Town, offering tales of

empowerment for women and opportunities for the self-employed or underemployed people

(Pollio, 2019). The ideals of its operation were entirely different from other cities such as

Lagos and Nairobi, except that there have been intense taxi wars between local taxi drivers in

both South Africa and Kenya. South Africa, which is the regional Uber headquarters of Africa

with Alon Lits as the general manager, acknowledges that the platform’s expansion strategy is

adapted to the mobility needs of a city – for instance, the launch of the CHAP CHAP product

in Nairobi in 2018 (Dahir, 2018). Uber Chap Chap, which means hurry hurry in Swahili, are

small, fuel-efficient low-cost vehicles with a minimum cost of 100 Kenyan Shillings (about 99

cents), cheaper than Uber X at 150 Kenya Shillings (about £1.13) (Shu, 2018). The manager

also highlights that with time, it plans to adapt its technology to other sectors according to the

cities’ needs, such as Keke Napeps (three-wheeled auto rickshaw) in Lagos and Boda-Bodas

(motorcycles) in Uganda (Dahir, 2018). However, Uber has a similar capitalist strategy where

drivers are known as entrepreneurs or independent contractors in terms of its marketing

strategy and governance. Pollio (2019) examines how Uber exerts its ride-hailing market

locally in Cape Town by encouraging drivers to perceive the job as entrepreneurial

development, while drivers with or without the affordances of the software application devise

multiple technical, ethical, and economical means of working on the platform.

This phenomenon has been a selling point for platforms like Uber because, according

to a driver from Pollio’s study (2019), call operators for local taxis are often corrupt; bribing

an operator with cigarettes or small tips is the norm to get trips. Therefore, it is plausible for

drivers to sign up for platforms with clear data displays and less corrupt practices which are

not limited to human errors. Emerging in contexts with highly informal transport systems, such

as is the case with many African cities, appears less difficult for platforms like Uber and Bolt.

Compared to local platforms, international platforms like Uber have access to more capital and

funding, which enables efficient running of its services, including marketing strategies and

management that can sustain platforms, provide more efficient service, and enable continuity.

For example, during FGD (November 2018), drivers in Lagos preferred international platforms

because they are more stable and good alternative sources of income, possess labour continuity

and are less corrupt than existing local platforms. Considering that such contexts possess a

rather analogous taxi system and increasing unemployment rates, platforms like Uber emerge

with promises of alleviating these problems with a track record from GN cities compared to

local platforms. Also, in Cape Town, Carmody and Fortuin (2019) observed a similar trend of

165

casualisation and informalisation studying Uber and Bolt (Taxify), such that the tenets of the

gig economy, such as virtual accumulation and informationalisation through datafication, may

be giving rise to a new definition known as ‘uberisation’. Drivers utilise their assets, while

riders reap the benefits and platform companies benefit from virtual capital by leveraging these

assets for capital gains in the form of ‘commissions’.

In 2016, Uber launched in Ghana and signed a Statement of Understanding (SOU) with

the Ministry of Transport, which was the first of its kind in Africa (Mourdoukotas, 2017). In

the presence of the Minister of Transport Fifi Kwetey, it was acknowledged that the benefits

of Uber would include updating regulations or self-regulation; improved mobility efficiency

by providing safe and reliable transport options; reduced congestion by enabling more people

to use the platform, and finally the creation of thousands of entrepreneurship opportunities for

drivers (Uber Africa Newsroom, 2016). Unlike most cities, Uber did not battle with regulatory

bodies in Ghana because they acknowledged the challenges of the local taxi industry and the

extent to which they could learn from technology (Kaye-Essien, 2020). However, it is unclear

to what extent Uber shares its data with the ministry based on its policy of not sharing

information with third parties. Uber currently wields power to improve or destroy the taxi

industry in Ghana by implementing its ideals from GN cities with laws that do not adequately

reflect drivers' experiences in Ghana. Before Uber, the transport industry in Ghana benefitted

from some level of formality, such that the main transport union, the Ghana Private Road

Transport Union (GPRTU), has approximately 90% of the commercial drivers in Ghana as

members, while the remaining drivers are in the opposition unions, the Ghana Cooperative

Transport Union (GCTU) and the Progressive Transport Owners Association (PROTOA) (Finn

et al., 2009).

The GPRTU is responsible for regulating the fares of its affiliated fleets based on

negotiations with the government on most occasions (Kufour, 2019). However, it is unclear

how this happens because taxis in Ghana, based on observation, do not possess taximeters;

neither do buses possess an e-ticketing system. The loophole Uber exploited is classified as

‘floating taxi drivers’ because these drivers are unaffiliated with any union, thus negotiating

fares before trips (Kufour, 2019). What is important to note here is that Uber, via the price

algorithms and dynamic pricing, defines its prices per trip which is already an informal method

for offering ‘floating taxi drivers’ only more cheaply, while formalised taxi drivers negotiate

fares with the government regularly. The ILO encourages a formalisation agenda to ensure that

workers in the informal economy are protected, but Uber’s business model and hidden politics

have made this problematic globally. There is also a growing perception from the Secretary-

166

General of the Ghana Trade Union Congress, Yaw Baah, that Uber may be secretly formalising

the informality in the taxi industry (Houeland, 2018). In other words, Uber is no different from

traditional taxis. However, through its technology and business model, they have stabilised

traditional informal practices on taxis while creating a race to the bottom for drivers, but in the

absence of worker collective voices or unions and policies from the state that can regulate their

affairs.

In the case studies above, including in other countries like Egypt, unions and drivers

have started to push against the ideals of Uber and its impacts on drivers. On the other hand,

platform drivers protest against the precarity of labour but less violently because contracts are

not necessarily binding; drivers cannot fight software applications. In Ghana, there has been

resistance from drivers concerning the agreement signed with Uber. In fact, in 2016, most of

the registered drivers conducted a nationwide strike following the introduction of Uber and the

BRT initiative in Accra (Ghanaweb.com, 2016).

Recent strike actions for Ghanaian taxi drivers

indicate their lack of involvement in decisions that affect their labour, including the issue of

low fares, which are not commensurate with the driving effort (Ayamga, 2021). In South

Africa, about 4,000 Uber drivers collaborated with the South African Transport and Allied

Workers Union (SATAWU) to advocate for their rights as workers and fight against unfair

termination (Houeland, 2018). Although the case was lost eventually after winning the first

round, it is commendable to see drivers collaborate across boundaries to stage a regulatory

battle against Uber in African cities.

Similarly, in 2018, 42 taxi drivers filed a lawsuit against Uber and Careem in Egypt,

stating that the companies operate illegally as registered call centres and an internet company

while using private cars as taxis (Reuters 2018). This led to a temporary ban on these

companies, which was later lifted by a top administrative court in February 2019, just before

Uber acquired Careem at 3.1 billion dollars the following month (Lomas 2018). The dynamic

of ride-hailing in Egypt is entirely political because of the intrusion of the government.

Ministers under President Abdel Fattah el-Sisi in Egypt have been requesting access to Uber’s

internal software ‘Heaven’ (Walsh, 2017). The Heaven software provides data about drivers

and riders and tracks journeys across the digital map (Walsh, 2017). This is one of Uber’s

ultimate strategies to keep this information from third parties or the public.

These case studies also show how protests in GS cities were not impactful until

recently. GS cities such as those in South Africa and Ghana and other GS countries rely on the

successes of resistances in regions such as the United Kingdom (UK). For example, the recent

high court ruling in London, which orders drivers to be classified as workers, has boosted

167

unions and capacities in countries like South Africa to file class action suits against Uber

(Butler 2021b; Cheng 2021a). However, except for the Egyptian case, it is worth noting the

slow pace of suing ride-hailing platforms in African cities, which often do not yield tangible

results. This can be linked to an enabling environment and robust legal and regulatory

frameworks in cities like London, which have been developed over time, compared to South

Africa or Ghana, with weak legal and regulatory frameworks. It is also noteworthy that the

recent class action against Uber in South Africa was led by a British law firm, Leigh Day,

though in collaboration with South African-based peer Mbuyisa Moleele Attorneys (Cheng,

2021b). This further indicates the reliance on GN ideologies in battling ride-hailing platforms.

The Ghana case presents indications of how platforms exploit regulatory weaknesses in the

taxi industry, which are prevalent in several African countries

On a positive note, Uber has paved the way for other global players like Bolt (Taxify)

to emerge in the market and boosted indigenous innovative ride-hailing platforms to address

contextual challenges in African cities. Since Uber launched in South Africa, it is arguable that

other African cities like Cairo have boosted their innovative capacities to produce domestic

platform companies that localise the technology of ride-hailing. For instance, Zebra Cab, a

traditional metered taxi company, launched in 2011 in Johannesburg but later scaled up its

services to a platform to compete with the likes of Uber and Bolt (Atolagbe, 2016). ‘Taxi Live’,

which raised about R2.3m, was also introduced in Durban to advocate for the needs of metered

taxi drivers (Timm, 2019). Others are adopting the technology based on vehicle mobility needs,

such as boda-bodas motorcycle taxi in Kenya or Safemotos in Kigali, which is also a

motorcycle company that admits the technology is like the Uber experience (Chonghaile, 2015;

Houeland, 2018). This shows the influence of Uber on the African continent based on the

emergence of other ride-hailing platforms in different cities (see Figure 11).

168

Figure 11: A Map Showing the Growth of Ride-hailing Platforms in Africa

Source: Briterbridges.com (2018)

While drivers are aware of its power of data and surveillance, it is also a reminder that

rules are transient, such that drivers find ways to circumvent the app, in this case, by working

out of the confinement of the app just to maximise income. This phenomenon is because

workers are starting to realise that the informalisation of work in the GS cities is putting them

in a precarious situation because of the dynamic nature of gig work. This also obscures

traditional labour and employment practices while being embedded in the very materiality of

the city and its transport system. However, this calls for a careful analysis that would create

policies and regulations for the nature of work in the platform economy such that informality

is clearly defined and benefits its workers. Additionally, there is a need for humans to interfere

with algorithms that operate linearly and could create unfair environments for workers through

metrics like ratings. It is important to note that discussions in this section are not exhaustive

but indicative of the plurality of the gig economy in African cities in the GS.

These African cases highlight why GS cities, particularly African cities, are lagging

behind GN cities in a push against the realignment of policies and practices which reflect the

contextual realities of drivers in these cities. This section has highlighted how Uber emerged

169

in African cities, paved the way for other ride-hailing platforms, and exploited the weak

regulatory frameworks in these cities leading to direct impacts on drivers. The next sections

will outline how the shortcomings in the traditional taxi industry have facilitated the entrance

of ride-hailing platforms in Lagos, Nigeria. While the subsequent section identifies the nuances

of how Uber and Bolt emerged, including dismantling traditional processes of taxi driving, it

creates a background for examining the impacts of algorithmic management in the following

chapter.

4.6. Taxi Driving as Gig Work in Lagos: From the Kabu-kabu to Modern

Taxis

This section highlights how taxi driving has always been gig work, but that some of its

features change with the emergence of digital platforms. Comparing Kabu-kabu services

(shared taxis) to today's platforms shows how taxi gig work is different from ride-hailing gig

work, notably how workers were controlled and managed. Understanding this shift highlights

the changing role of unions in managing taxi drivers and how the intermediation of Uber has

replaced physical layers of control with digital ones, using algorithmic management as a

weapon in controlling and managing the labour process of drivers in Lagos.

Table 18 summarises the different mobility modes in Lagos as a background for

situating ride-hailing platforms in Lagos. While the core focus in this study is ride-hailing

platforms Uber and Bolt, it was important to review traditional taxis as part of history as well

as to highlight other mobility modes that are informal or shared services such as the Kabu-

kabu.

Table 18: Informal Transport Modes in Lagos

Mode of transport

Definition

Danfo

The danfo, which means hurry in Yoruba, are informal mini-buses or shared taxis

that are cheaper and faster in conveying passengers through Lagos. It possesses a

driver and a conductor, i.e., an individual that helps in facilitating passengers for

trips.

Vehicle capacity: 14 to 18 passengers

170

Bus Rapid Transit

(BRT) buses

These are large buses launched in 2008 with two corridors of over 35.5km, with 20

per cent separated by road markings and over 65 per cent physically designated

pathways for the buses (Mobereola, 2009; Otunola, 2019).

Vehicle capacity: 21 to 42 passengers

Yellow taxis

This is one of the oldest taxi services in Lagos, recognised by its yellow colour, which

is symbolic of the state. They are used for transporting passengers throughout the

city.

Vehicle capacity: 4 passengers

Red Cabs

These are small Suzuki taxi vehicles, typically with at least 1,300 cc engine capacity.

Although these are no longer actively operational in Lagos, they were part of the

second generation of taxis between 2009 and 2015 facilitated by Governor Tunde

Fashola’s regime.

Vehicle Capacity: 4 passengers

Kabu-kabu

Kabu-kabu services are shared taxis in the form of private vehicles like Peugeot 504

or Toyota vehicles and mostly old yellow taxis. While in the 1980s majority of Kabu-

kabu taxis were typically unregistered yellow taxis, in recent times, private vehicle

owners, including blue-collar workers, sometimes illegally used their vehicles to

convey passengers for extra income.

Vehicle capacity: 4 passengers, but drivers tend to carry above the capacity.

Car-hire taxis

Car-hire taxis are non-coloured taxis, i.e., like a typical private vehicle without the

yellow Lagos colour. They offer the same service as the other taxis; only they are

more exclusive and not restricted from entering residential estates or locations like

the airport. Also, passengers can rent their services for more extended periods or

simply pay per trip. This also applies to the other taxis, but passengers prefer the

exclusive nature of car-hire taxis.

Vehicle capacity: 4 passengers

Okada

These are commercial motorcycles (two-wheeled) that operate as an informal mode

of conveying passengers across relatively short distances. While Lagosians recognise

them, Okadas are controversial and have experienced intense regulations, including

banning their operations in 2018 and 2021. With the aim of addressing robbery,

chaos, and disorderliness in line with the transport reform law of 2018, Okadas were

171

restricted to certain corridors in 2020 and ultimately banned in 2021 (Guardian,

2021). However, many of these services still operate illegally in the city.

Capacity: 1 to 2 passengers.

Auto-rickshaws

(Keke-napep)

These are three-wheeled tricycles typically called Keke-napep or Keke Marwa. The

majority of these vehicles in Lagos are also yellow colours, symbolic of Lagos State.

Capacity: It can convey 4-passengers at a time.

Source: Derived from Moboreola 2009; Otunola 2019; and Guardian 2021

4.7. First Generation Taxis: Challenges in Managing Taxi Work

Nigeria experienced a plethora of bad economic impacts from the oil boom in the late

1970s following the civil war, then a decline in exportation and a reduction in the Gross

Domestic Product (GDP) by an average of 1% per year between 1980 and 1987 (Lewis, 2006).

During General Ibrahim Babangida’s era, there were attempts to stabilise the Nigerian

economy by initiating an agreement with the International Monetary Fund (IMF) and the World

Bank, which brought about the Structural Adjustment Programmes (SAPs) (Lewis, 2006).

However, Ekanade (2014) argues that introducing SAPs and other neoliberal policies was

detrimental to Nigeria's recovery and national development. Neoliberal reforms through SAPs

were concerned with market efficiencies and not solving social problems, leading to poverty,

hardship, high unemployment rates, declining purchasing power and overall discontent in

Nigeria (Ekanade, 2014). Informality in terms of the nature of jobs picked up by citizens, such

as small-scale manufacturing, informal health services, vehicle repairs and maintenance, taxi

driving, tailoring, furniture making and several other practices, increased between the 1980s

and 1990s (Yelwa, 2016). With regards to taxi driving, this led to more unlicensed vehicles

and difficulties in tracking private vehicles that unemployed Nigerians used as a means to make

more income. There were high levels of second-hand vehicles, initially referred to as Belgium

cars, because of the rate at which West African migrants exported cars from Europe, which is

still a trend today (Ezeoha et al., 2019). They were later renamed and are generally known as

'Tokunbo' vehicles because Nigerians in the diaspora became involved in the exportation of

second-hand vehicles as a wealth creation business.

[

]

172

Figure 12: Picture of a Typical Kabu-kabu Taxi

Source: Ebenezer (2018)

The Kabu-kabu services are operated using private vehicles and yellow taxis such as

saloon vehicles like the Peugeot 504 and Toyota vehicles (Oyarinu, n.d).

Typically, these were

meant for taxi services but were often used as shared taxis because of their expensive fare

compared to other informal modes (see figure 12). As a coping mechanism, some of the drivers

involved with Kabu-kabu jobs were characterised as unemployed or underemployed workers,

such as civil servants who needed to make an additional income to survive in difficult times

(Olukujo, 2003; Ekanade, 2014; Onatere-Obhruhe, 2016). The impacts of SAPs from the 1980s

also brought about smaller motorcycle taxis known as Okada, which enabled workers to utilise

their vehicles similar to the Kabu-kabu for additional income, but with more efficiency because

these smaller vehicles could evade congestion compared to other modes (Ogunbodede, 2008;

Onatere-Obhruhe, 2016).

According to the interview with Mustapha (August 2019), the first-generation taxis

were the traditional yellow taxis (see Figure 13 above) and black taxis, which came into

173

existence during the regime of the former governor of Lagos State, Lateef Jakande (1979–

1983). The taxi industry was not adequately regulated, and there was free-market entry and

exit into the taxi business, which continued until 2015, when the new taxi scheme was reviewed

(Interviewee Mustapha, January 2019). Most of these taxis were said to operate illegally,

according to Mustapha, because they were not licensed. Unlike the informal buses, taxis used

for Kabu-kabu service were better organised and regularly plied the roads. In preventing illegal

taxis and buses from being impounded by law enforcement agencies, the National Union of

Road Transport Workers (NURTW) and the Road Transport Employers Association (RTEAN)

were formed (Oyaniru, n.d).

The yellow taxis that were sometimes used as Kabu-kabu represented the dominant

form of mobility, especially for affluent people in the 1990s, until they were banned from

entering estates under the regime of President Obasanjo and Lagos governor Bola Tinubu in

1999 (interview with Olawumi and Femi, October 2018). This was because of the frequent

kidnappings and robberies that impacted the transport industry from the 1980s, especially

among taxi and danfo drivers. Danfo, which means hurry in Yoruba, were mostly yellow-

coloured buses and were informal means of moving around Lagos, which were faster and

cheaper, especially during peak traffic congestion. These buses created self-employment for

both drivers and conductors, who were responsible for admitting passengers, quoting fares and

instructing drivers in navigating the roads. Conductors were mostly unemployed and unskilled

workers ranging between 20 and 50 years old, often referred to as agberos (Lawuyi, 1988;

Agbiboa, 2017). These conductors risked their lives by hanging onto bus tailboards during trips

to facilitate passengers boarding buses. For example, some phrases from conductors include:

"Oshodi! – A shout for the destination of the bus and an invitation to the

commuters going that to a similar route."

"Ketu-Mile 12 N20 – A call inviting the commuters going to the route of Ketu-

Mile 12 with the information of the fare they will pay."

"Wole Sowo e (enter to your side) – Merge to the main road but put keep to the

lane on your side." (Adedun, n.d, pp. 212 – 214)

Conductors were as important as drivers within informal transport services such as the

Danfo. The Kabu-kabu adopted a similar approach for recruiting passengers, except that there

was often no requirement for conductors during trips because of their small and enclosed

nature. Also, the fact that everyday citizens began to utilise their vehicles as shared taxis

174

demonstrated the influence of Kabu-kabu and its impact on mobility. This is because yellow

taxis were not allowed to enter residential estates and exclusive locations, and passengers had

to meet with drivers at designated areas and ranks (interviewee Kehinde, October 2018).

Regular commuters who offered Kabu-kabu style services facilitated access to enclosed spaces

like residential estates. Consequently, this made the car-hire business blossom, i.e., taxis that

did not bear the Lagos state yellow colour. Taxi drivers started to use private vehicles without

changing their colour to avoid law enforcement agents' sanctions, and to improve passengers'

comfort and to improve access to prohibited spaces.

Interviewee Kehinde (October 2018), an elderly taxi driver in his 60s, narrates the story

of the ban on Danfos and the introduction of car-hire vehicles back in the 1990s.

Way back, danfo was not allowed on the Island/ Ikoyi, but now it has been

allowed. Before if you enter, they will impound you immediately. It was only taxis that

could enter these areas. The time they produced even and odd numbers, that is when

car-hire came out. The big men do not enter taxi again. All the places taxi used to enter

are not allowed anymore. I can be arrested if I go to the airport. From the local to the

international airport, they allow commercial vehicles to cross, but they do not allow

yellow taxis.

It is important to note that the car-hire business was relevant across other informal

modes based on the vehicle ownership patterns. According to Lawuyi (1988), a taxi driver may

begin the business as an owner-driver but later become a taxi owner after acquiring new fleets

outsourced to potential drivers on a hire-purchase or a rental basis. These precarious

ownerships (Agbiboa, 2017) further empowered both taxi owners and drivers because the

former would make more income while the latter would have a job. Drivers on hire-purchase

plans pay taxi owners daily or weekly by working tirelessly and strategically to meet their

targets. As Agbiboa (2020) argues, the informal transport sector is a site of indigenous

entrepreneurship and creative adaptations – workers always find a way to survive using the

resources available to create separate realities from the consequences of failed neoliberal

policies, which exacerbated taxi struggles and poverty in Lagos.

175

Figure 13: Yellow Taxi and Car-hire Parking Arrangement in Eko Hotel Taxi Stand Victoria

Island, Lagos.

Source: Author’s fieldwork (October 2018)

To book yellow taxis or car-hire taxis, passengers must either hail the vehicle from

ranks or on the road and haggle for trip fares. The level of organisation with yellow taxis is

evident in most of the parks and car-hire taxis even today (see Figure 13). Rides are booked on

a turn-by-turn basis such that drivers write their names on rosters when they return to the park.

If a driver's number is up and does not accept the passenger's fare, the next driver can take the

opportunity.

[

]

Before moving to second-generation taxis, it is important to unpack the role of

unions in the management and control of mobility, particularly taxi labour, in the next section.

4.7.1. The Role of Unions in Controlling Taxi Labour

176

The efficiency in managing the overall mobility workforce in Lagos was galvanised by

transport unions, facilitated by a semi-informal arrangement with the state, which cuts across

the 37 local government areas (LGAs).

Between 1976 and 1986, a nationwide reform of LGAs in Nigeria instituted by the

federal government utilised two critical objectives for managing mobility services in Lagos

(Olowu, 1986, p. 290). The first objective was an indication of the return to civil rule such that

the establishment of local governments came before decolonising and political independence

practices. Considering how windfall gains from oil revenues ended up in major cities in

Nigeria, the second objective was created to ensure the decentralisation of social and economic

developments reached the grassroots, such as LGAs (Olowu, 1986).

Specific to the taxi industry, this phenomenon, according to Albert (2007, p. 127), led

to exclusive power for local governments towards the "collection of rates, fees and taxes, to

boost internal revenue and control of socio-economic activities and the establishment and

maintenance of markets and motor parks”. The idea was to ensure that the 37 LGAs across five

administrations in Lagos State could manage an atomised mobility workforce, including

minibuses, Okada and traditional taxis in over 4,000 motor parks (Albert, 2007).

The establishment of the National Union of Road Transport Workers (NURTW), 1978,

meant they were now gatekeepers and managers of these motor parks, whose tasks included

collecting dues, managing operations and being responsible for lobbying the government for

the affairs of commercial drivers in Lagos (Oyarinu, n.d.; Agbiboa 2017). Other unions

responsible for managing and interceding for taxi services were the Road Transport Employers

Association (RTEAN), Lagos State Taxi Drivers and Cab Operators Association (LSTDCOA)

and Self-Employed Commercial Drivers Association of Nigeria (SECDAN) (Albert, 2007).

According to Albert's critical study on unions in Lagos and Ibadan, motor parks were mainly

divided into three management categories under unions. The first strand comprised parks of

rectangular enclosures or gated environments where taxi drivers paid dues to unions whenever

they were moving out – with an embargo such as a rope or spiked plant preventing movement

if drivers did not pay.

[

]

The second category was motor parks established by former union

representatives that may have lost elections but clamoured for official recognition by going

against the state's rules. This phenomenon explains why the VGC drivers expressed a partial

disconnection from the state in my study. They operate under the LSTDCOA, which has been

revamped since the second tenure of Governor Fashola's regime. The final category was formed

out of potential spaces along the highways with high patronage from passengers, usually by

independent NURTW unions and taxi unions. These atomised groups facilitate the

177

management of taxi labour across Lagos such that each formal or informal park self-organises

by creating positions of power such as a chairman, vice chairman, general secretary and many

other roles. For example, in the field, I interviewed Kehinde (October 2018), Chairman of Eko-

hotel taxi car park, and Chairman and Vice Chairman Tom and Isaac (August 2019) of the

VGC car park. Unions as critical managers of these contested spaces (Agbiboa, 2017)

facilitated political antagonisms between them and led to the dissipation of formal ownership

by the state into informal and private ownership (Fourchard, 2011).

According to Albert (2007), the relationship between local governments and the

NURTW was informal because they handed tickets and power to the union to collect fees and

union dues and manage the overall affairs of workers. This was due to conflicts and friction

between local government staff and these unions. The NURTW and RTEAN collect dues at

higher rates and remit less than the required percentage to the government. Agbiboa (2020)

highlights a 'Janus-faced' leadership of unions, such as the NURTW, which act as a supportive

network for solidarity, accumulation, and survival when business is slow for many commercial

drivers on the one hand, and a predatory organisation on the other hand. Agbiboa (2020)

analyses the politics and violence involved within the NURTW as currency for imposing

extortionate power by collecting membership dues from drivers, using these levies as

weaponised oppression and subversion of power through bribery of state officials such as the

police. The power of the NURTW and RTEAN is embedded in their involvement in political

campaigns and elections throughout Nigeria, based on their capability to dissuade the masses

from voting for the opposition by mobilising voters, engaging in violent misconduct,

mobilising political thuggery and liaising with law enforcement agents to achieve their goals

(Fourchard, 2011; Omobowale and Fayiga, 2017; Agbiboa, 2018). Union chairmen are often

aligned with state governors, influential or ‘big men’ who possess the clientele and resources

to enable wealth redistribution, food distribution and contracts as rewards for political gains

(Fourchard, 2011). There are also aspects of Godfatherism (Albert, 2005). This is also reflected

in intra-union and inter-union elections in the quest for leadership or political positions such

that, according to Agbiboa (2017; 2020), political frictions and warfare develop between the

NURTW and RTEAN, resulting in violent clashes amongst area boys and hidden law

enforcement agents who are attached to these unions.

In most cases, these unions use area boys or agberos (also known as touts) to forcefully

retrieve levies from commercial riders. According to Agbiboa (2017), area boys are divided

into agberos, who are indigenes of Lagos state, and the erukus, who are inter-state immigrants

in search of a better life in Lagos. The latter aspires to be like the agberos because union leaders

178

are granted business investments, travelling opportunities, and overall benefits. This

phenomenon is why platform driver and union leader Dipo posits that traditional unions are

powerful but corrupt because of the misuse of monthly dues from union members and extortion

from vehicle operators without remittance to the state. Agbiboa's research corroborates Dipo’s

assertion that union leaders make between 5 to 10 million Naira ($13,888 - $27,776) daily

through abgeros who facilitate the collection of levies from commercial drivers. Therefore,

according to Agbiboa's research, most unions aspire to become leaders at the national, zonal,

city or local levels (Albert, 2007; Agbiboa, 2017; 2019; 2020).

While in Lagos State, the NURTW does not express total control over traditional taxis,

they still control the majority of mobility services throughout Nigeria and, if left unchecked,

could permeate aspects of emerging ride-hailing platforms (Interviewee Dipo, 2019). Because,

throughout the city, there are politicised spaces of control (Albert, 2007), there are possibilities

for taxi services or the Kabu-kabu to encroach in parking spaces where agberos extort other

informal modes like Okada (motorcycle taxis), thereby putting them at risk. The transport

policymaker Mustapha (August 2019) highlighted the purpose of these unions and the

difficulty of controlling them.

NURTW and RTEAN are transport unions with their duties basically to collect

tolls from all vehicles; the tolls they collect is used for themselves, they do not remit the

tax to the government and the government have not been able to work in that angle…

but the government is looking away because of political reasons.

Due to the power struggles between unions, it remains unclear if taxis are a formal or

informal form of transport. While the interview with Dipo (2019) highlighted that the taxis

operate only under the LSTDCOA, interviews with Deji (October 2018) indicated being under

the RTEAN. The interview with Mustapha (August 2019) depicts most taxis in Lagos as

operating illegally. He also highlights that while yellow taxis operate formally by paying taxes

and obeying regulations, several yellow taxi and car-hire services hide under the NURTW and

RTEAN union, causing difficulties in legitimately recognising drivers. In corroboration,

observation from the field showed that unions such as the LSTDCOA and RTEAN are central

to managing the affairs of taxi drivers but are operating independently by self-organising

through monthly meetings and donations of dues which are integral for the management of

designated parks (Interview with Tom and Isaac, August 2019).

179

4.8. Second-generation Taxis: Attempting to Control and Measure Taxi

Work

In 1999, Nigeria experienced a transition from four successive military regimes from

1966 to 1999 into democratic rule (Gandi, 2006), leading to a civilian administration in most

cities, particularly Lagos city. Bola Tinubu became the first democratic governor of Lagos

State in 1999 with radical plans to make Lagos a world-class megacity, including a 10-point

agenda known as the Lagos State Economic Empowerment and Development Strategy

(LASEEDS), where the development of transportation and improvement employment were

critical elements (Filani, 2012; Kuris, 2014). Succeeding governor Tinubu, governor Fashola's

administration, between 2007 and 2015, continued implementing the LASEEDS agenda to

develop Lagos into a world-class megacity. This transformation brought about initiatives such

as the Bus Rapid Transport (BRT), upgrading the transport infrastructure and an overall plan

to develop integrated transport systems, formalise the process of registering mobility modes

such as licensing taxis and other private modes, and the possibility of having more control over

workers through technology (Mobereola, 2006; 2009; Kuris, 2014). Between 2007 and 2015,

the second generation of taxis included eight taxi companies such as Red cabs, Corporate cabs,

Metro cabs, Orange cabs, Easy-cab, E-taxi (Eko taxi), and K-kabukabu and so forth, with a

combined fleet of over 1,255 (Interviewee Lateef, October 2018). Baba Tunde Fashola's

administration facilitated much of these developments between 2009 and 2015, where taxi

ownerships were taken from private individuals to corporate companies because of a lack of

funding (Nairaland, 2009a).

Companies such as C&I Leasing Motors, I-Trans Logic Limited,

Basscomm Limited, Cash-Link Investment Limited, Corporate cabs and many others partnered

with the Lagos Taxi and Cab Operators Association. Funds were being provided through the

Union Bank of Nigeria and the Lagos-Microfinance Scheme to create jobs for over 1,000

drivers with the possibility of providing comfortably for their families, like drivers in England

and America (Nairaland, 2009b)

According to an interview with Lateef (October 2018), a regulation was passed for

corporate companies willing to venture into the taxi business. This included, in 2009 and 2015,

the new taxi scheme:

• Financial stability with proper cash flow analysis of the company

• Capabilities for metered taxis and well-spaced taxis

180

• A registration fee of about N5 million (£10,000) and a yearly renewal of N1.5 million

(£3,239). This regulation passed between 2011 and 2012, where franchisees could start

a business with that amount by funding at least 100 vehicles.

[

]

• Vehicles must have an engine capacity of at least 1,300 ccs. This regulation was passed

to kick against Red cabs using Suzuki car brands known for their small engines (see

Figure 14).

• Drivers must be trained at the Lagos drivers institute (LASDRI) with certificates and

must attend once a year.

• They must be able to afford comprehensive insurance to safeguard their investments.

• Taxi vehicles admitted into the scheme must not be more than 12 years old.

Figure 14: Red Cabs in Lagos

Source: Bassey (2014)

The idea of the modern taxis came from US taxis, where drivers would purchase a

Medallion license which was tradable via the stock exchange after acquiring some value

(Interviewee Lateef, October 2018). Furthermore, the licences were tradable and could be used

to access bank loans. This phenomenon, however, was not properly implemented because of

181

the lack of enforcement and the high rate of informality from struggling corporations. Besides

funding issues, the looming security issue that started in the 1980s and its rickety nature

contributed to its low patronage and the struggles for taxi drivers. Most of the companies that

took part in the scheme had poor knowledge of the transportation business, leading to poor

management of drivers and high overhead costs (Interviewee Lateef, October 2018).

Taximeters were introduced to ensure that both drivers and passengers were aware of the cost

of trips without causing conflicts. However, the metering system was not profitable for

corporations and drivers because of the inability to calculate fares accurately when stuck in

congestion, which had always been a 'wicked problem' in the transport industry in Lagos

(Interviewees Lateef, October 2018; Mustapha August 2019). Despite this being a failed

experiment in yellow taxis, the Metro taxi adopted this approach for their fleets.

Metro taxi came into existence in 2010 when the CEO and entrepreneur Imorou

Moumoumi claimed that the company was the first to have meters and fitted cameras to ensure

safety and security for both riders and passengers (see Figure 15). Like the typical black taxis

in England, Metro taxis also utilise an online dispatch system connected to GPS, which

facilitates passengers booking a ride in approximately one minute by just calling an operator.

This technology was expected to efficiently balance the demand and supply of taxis by calling

an operator who then dispatched a ride to passengers and effectively supervised taxi labour.

Taiwo, who had been driving for over 30 years with yellow taxis, now in his 60s,

explained that:

E get one time metro come. Dem dey put taxi meter for car. We no get meter.

We no dey use all this thing. We use am long time ago, but we don cancel. About 10 to

20 years ago. Some passengers no like that package. If go slow dey the passenger don

je gbese be that. If traffic dey and the thing dey read, the passenger go shake self

(Taiwo, October 2018).

[

]

In a nutshell, Taiwo highlights the infeasibility of taximeters in a city like Lagos. Other

taxi drivers such as Jeff (November 2018), Kehinde (October 2018) and Ahmed (November

2018) attested to the infeasibility of the Metro taxi model in assigning and monitoring the

labour of drivers. Within two years of its launch, despite the innovation of Metro taxis and the

adoption of a similar model by companies for other taxi firms such as Red cabs, many of these

taxi companies failed (Interviewee Lateef, October 2018). A core reason for this was that most

taxi companies were decentralised – they refused to unite under one body, making decisions

182

concerning regulations and driver warfare challenging to implement (Interviewee Mustapha,

August 2019). The only surviving company to date is the yellow taxi and car-hire service,

which survived because of a well-organised structure due to self-organisation and support from

the LSTDCOA, which ensures car loans to sustain the business (Interviewee Mustapha, August

2019).

Figure 15: Metro Taxi and the Introduction of Taximeters in Lagos

Source: Africareport.com (2012)

It is important to note that Corporate Cabs emerged in Lagos two years before Metro

taxi. While there is little or no information about their entrance, residual material on the

Nairaland website indicated that all their vehicles integrated the use of taximeters and

dispatchers in facilitating trips.

[

]

The takeaway is that most drivers did not also subscribe to

its managerial style, deducing from the responses from drivers about Metro taxis and the

challenges surrounding the use of taximeters.

Following Metro taxi's entrance, two critical platforms emerged before Uber, Easy Taxi

and Afrocabs. The Easy taxi platform was founded by Tallis Gomes, a 23-year-old

entrepreneur from Brazil (Watson, 2014; de Souza Silva et al., 2018). The entrepreneur created

the platform to optimise taxi access considering the long waiting times, violent reputation in

Rio de Janeiro and large numbers of unregistered taxis. The recruitment of taxi drivers began

183

manually, visiting taxi ranks and convincing drivers to become a part of the initiative in Brazil

(Watson, 2014). This was a similar experience in the Lagos market when the company tried to

convince traditional taxi drivers to join the platform. More so, the insecurities of kidnapping

affluent people and expatriates were high in 2013. Kidnapping crimes were frequent in

traditional taxis, the Kabu-kabu service and sometimes the minibuses like the Danfo, locally

known as one chance.

[

]

According to Control Risk Mapping (CRC), Nigeria was third behind

Mexico and India for kidnapping-related crimes (Perlberg, 2013). Easy taxi improved the

technology of Metro taxi cabs by including a taxi app, giving drivers perceived control over

the labour process and reducing exposure to such risks. In an interview with ENCA news in

2014, the former CEO of Bankole Cardoso stated:

We are effectively skipping the calling a cab-metro minicab kind of format that

is used in other countries. We are leap-frogging that idea and bringing a new

technology to Nigeria.

Similarly, Afrocabs, which became the first indigenous ride-hailing platform, presented

the same technology but perceived a local advantage because of a better understanding of the

Nigerian taxi industry in early 2014 (Mátùlúko, 2016). The Afrocabs platform integrated a

price haggling feature in the app because of the culture of negotiating fares before trips and

accepting cash payments, creating a balance of power between a driver and a passenger. Before

the dominance of platforms, this observation of haggling, locally known as pricing, was

frequent in Lagos. For example, in an interview with Tom (August 2019), a father of three

children and the deputy head of a taxi car park:

When one customer come and say I am going to say Abraham Adesanya, maybe

I do not know there, I would say, give me N4,000. The passenger might say, 'Abraham

Adesanya here? That means you do not know there', I will carry it, for me to know there

in case of another day, I will be able to charge lower or higher. (Tom, August 2019)

Besides vehicle owners managing taxi services, drivers were responsible for the

decisions during the labour process, such as taking shorter routes, evading traffic zones,

targeting areas such as Eko hotels in Victoria Island, offices for potential high-paying

customers, and actualising inter-state trips for more income and many other practices. The

technological changes in the taxi industry did not thrive for long because, according to

184

interviewee Abiola (September 2018), most of these drivers were elderly, not well educated

and thus less enthused about utilising smartphone apps. Further interviews with transport

policymakers Lateef and Mustapha reiterated that the failures of several taxi firms were due to

the difficulty in managing drivers and the lack of understanding in doing taxi business in Lagos.

With issues of mismanagement of drivers and the service, a lack of transparency and

accountability from unions and the state, the platformisation or uberisation of the taxi industry

became imminent.

4.9. The Uberisation of the Taxi Industry in Lagos: Third-generation Taxis

The early experimentation of ride-hailing platforms in Lagos created a pathway for

international and indigenous ride-hailing platforms to emerge into the mobility space.

Uberisation, often used interchangeably with platformisation, is derived from the ride-hailing

platform Uber as a transition to a new economic model capable of disrupting incumbent

arrangements because digital technologies are intermediaries between service providers and

potential consumers, normally at lower transactional costs (Dudley et al., 2017; Radu and

Psaila, 2017; Lakemann and Lay, 2019; Stehlin et al., 2020). On the other hand,

platformisation, according to Nieborg and Poel (2018, p. 4276), "is defined as the penetration

of economic, governmental, and infrastructural extensions of digital platforms into the web and

app ecosystems, fundamentally affecting the operations of the cultural industries". However,

this chapter adopts the uberisation term because of the impact of Uber in Lagos as the first

platform with an accurate deployment of digital technology and algorithms in managing

drivers. This section identifies how platforms such as Uber, Bolt and other indigenous

platforms emerged in Lagos and how analogous or manual human management of the taxi

workforce was increasingly replaced and adapted by platforms using algorithms to manage the

labour process.

4.9.1. Uber Arrives in Lagos: Introducing the Hype and Rhetoric of a

Global Platform

In July 2014, Uber, a global platform company, landed in Lagos, Nigeria.

[

]

Uber,

valued at $18 billion at the time after raising $1.5 billion compared to their rival Lyft at $332

185

million, had launched in 205 cities in 45 countries (Hoge, 2014). The platform emerged in

Nigeria when the overall unemployment rose from 3.7% in 2013 to 4.5% in 2014 (World Bank,

2019b). On the other hand, youth unemployment was also high ranging from 9.8% in 2013 to

8.8% in 2014. It was becoming difficult for residents to acquire jobs after graduating from the

university. The former general manager of Uber, in an interview with John Rhoda, hinted at

creating jobs for potential Uber drivers.

[

]

As of August 2015, we created about 600 jobs, from when we started in 2014,

we are 17 months now. We are hoping that by the end of next year, that number would

have increased to 3,000 jobs. The aim is to create two financial opportunities in

Nigeria: the vehicle owner, the driver, and after a while, the driver can become a

vehicle owner as well because they now have the financial stability. (Ebi Atawodi,

2016)

Uber used its popular phrase "be your own boss", which was used globally in selling

the idea to drivers that are often unemployed or underemployed as a business to make extra

income without the control of a boss. The term indicates why workers should be free from a

traditional working environment to bolster their ability to define their labour conditions and

earn better salaries. Uber made the phenomenon of ride-hailing in Lagos attractive and the

adoption by drivers and usability by riders much easier. It classifies itself as a technology

company; it classifies its drivers as driver-partners; and passengers as riders.

Another of Uber's strategies was localising its operations by understanding the city of

Lagos through working with locals – a strategy its predecessor, Easy Taxi, failed to achieve in

previous years. According to the former general manager of Uber:

…as much as Uber is an international brand and founded in San Francisco, we

are very local as well. In every city we go into, we hire a local team who understands

the city. We think of our market as cities, not necessarily countries, Lagos, Joburg,

Cape Town, as much as you have similarities and differences in food and architecture,

even going from one city to the next. New York, for example, is different to San

Francisco, one thing that does not change is that people always want to move around…

(Ebi Atawodi, 2016)

186

When the Uber platform was launched in Lagos, they accepted saloon vehicles and

other luxurious cars manufactured from 2008, with a progressive increase in the vehicle

manufacturing date every year. This reflected its mobile capital and international influence in

the taxi industry, with more sophisticated vehicles and technology befitting a megacity like

Lagos. Uber also required smartphones with a good memory space to accommodate its

software application like its onboarding processes globally. Drivers then signed up online, after

which they were expected to answer 100 questions (FGD, November 2018). According to the

Uber drivers I interviewed, the exam tests drivers' intelligence quotient, literacy level,

knowledge of the city, people's habits, customer relations, and diplomacy. When probed about

the psychometric Uber test, Junior, a bachelor’s degree holder who has been driving for two

years, stated:

They want to test your I.Q. They want to test the driver's intellect, how smart

are you, and how can you relate with people and your environment. Then being able to

resolve issues. They are aware that people are complex, so when it results to cases,

they want to know how well you would resolve such solutions. (Interviewee Junior,

August 2018)

If potential drivers pass the test, they go through a vehicle inspection process and

comprehensive background checks where drivers present their valid driver’s licence, photo

identity, vehicle registration number, vehicle type, guarantors, roadworthiness, vehicle

insurance and hackney permit.

[

]

When the check is complete, drivers pass through a training

process about how to interact with riders, including ratings, maximise their time on the app by

making money, and how to resolve technical issues or conflicts with riders. Topics relating to

politics, football and relationships are discouraged in training sessions because it is perceived

that these often lead to disagreements and could cause conflicts, ultimately affecting ratings

(Interviewee Jonathan, November 2018). Drivers can start working within 2-5 days of the

registration. In a situation where a driver fails the psychometric test, they can retake the test.

Driving and using Uber is a prestigious experience; even musicians have incorporated the

brand into songs, demonstrating its capture of the market. However, this created an opening

for other platforms, including indigenous platforms, to emerge in Lagos.

4.9.2. Bolt (Taxify): Competition for Ride-hailing Platforms

187

In 2016, Taxify (now Bolt) emerged in Lagos. A 19-year-old founded the Estonian ride-

hailing company, Markus Villig, funded by Estonian and Finnish investors in 2013, and later

by Didi Chuxing, a China-based ride-hailing platform in 2017 and Daimler in 2018 (Russell,

2017). Nigeria experienced a recession, with GDP declining from 6.7% in 2013 to -2.06% in

the second quarter of 2016 (National Bureau of Statistics, 2016). The overall unemployment

rate rose from 4.6% in 2014 to 7.1% in 2016, and an increase in youth unemployment from

8.1% to 12.4% in the same period (World Bank, 2019 b). Similarly, the Bolt platform marketed

its service to desperate unemployed and underemployed residents, including traditional taxi

drivers. According to the Bolt platform:

No schedule, no boss, good pay. Are you in it to earn extra money besides your

9-5 or make money driving a few hours on a weekly basis? Or maybe you want to

become a full-time Bolt driver earning approximately N120,000 (£463) a week? … On

average, Bolt drivers in Nigeria earn up to N400,000 (£863) each month

. [

]

Again, the registration and requirements for being a Bolt driver were similar (see Table

19), except drivers were not required to take a psychometric test as they were for Uber, making

the process faster for drivers but less secure for riders.

A typical strategy for Bolt is to emerge in markets where Uber exists where the platform

aims to offer better service, which creates competition based on cheaper commission rates for

drivers (15%) and lower fares for riders (Petzinger, 2018; Satariano, 2019). Although this was

competitive from Bolt, it further legitimises the process of uberisation by adopting the business

model of Uber, channelling flexibility and autonomy of labour in the gig economy (Radu and

Psaila, 2017). The emergence of Bolt and other platforms in Nigeria is critical for the

realisation of network effects because, according to Knee (2018, p. 18), "every new participant

increases the value of the network to existing participants, attracts more new users, and makes

the prospect of a successful competitive attack ever more remote – thereby bolstering the

relative attractiveness of the business". Bolt targeted Uber drivers in Lagos using lower or no

commissions, radical incentivisation and enabling drivers to lure riders to the platform

physically. According to interviewee Koffi, a Higher National Diploma (HND) graduate in

Marine Engineering:

In the beginning of Taxify's entrance, they took commissions, but at the end of

the week, everything would be returned to you. You might just be working and see that

188

you are credited with about N50,000 (£108) – N60,000 (£130), and you do not know

where it is coming from. This made drivers troop in. All this happened in the first 6

months. (Interviewee Koffi, September 2019)

From field observations, a common strategy for new platform entrants was to approach

local taxi drivers and platform drivers on Uber or Bolt, making them ambassadors by giving

selected drivers branded T-shirts and flyers to place in their vehicles and to share with other

drivers to create awareness. In some cases, drivers market the new platform on SMCN groups

by posting earnings and benefits offered by those new platforms to facilitate more drivers to

adopt it. This became a regular practice for other ride-hailing platforms, which emerged in

Nigeria after Bolt. An interview with a former taxi and now platform driver George (September

2019), further highlighted how Bolt distributed flyers to drivers to facilitate riders to download

the platform app. It was not difficult to convince drivers to relocate from Uber because the fuel

price had increased from N97 per litre to N145 per litre, and Uber was still collecting a 25%

commission while reducing fares by a 40% margin. Both Koffi and George (September 2019)

acknowledged that platforms such as Bolt used drivers to build their network.

There is a new platform known as In-driver. They invited me to their office that

we should work hard to bring them in. I told them that if we bring you in, if we say that

you are going to have 500 drivers within three weeks or one month, it is certain. As we

are migrating, that is how we migrate with the riders along with us… we did it for

Taxify. (Koffi and George, September 2019)

This recruitment strategy by platforms became prevalent with other new taxi entrants,

both international and indigenous platforms. Uber and Bolt remained the dominant platforms

from field observation and interviews with all the drivers in this study, incessant

advertisements, progressive incentivisation and clarity in daily earnings. However, the

international platforms lacked a complete immersion in the culture of Lagos driving, its people,

and the everyday experiences of drivers.

4.9.3. Oga-Taxi: Integrating the City Culture

The Oga-taxi app was founded in 2015 but was active in 2016, the year of the

emergence of Bolt. The founders and serial entrepreneurs, Loko Ameh Udoko and Michael

Nnamadim, both graduates of the University of Massachusetts, abandoned their agribusiness

to launch a new indigenous platform to compete against Uber and Bolt. The name of the

189

company reflects the localisation of slang from the Nigerian Yoruba tribe. For example, the

lexical item ògá from the Yoruba tribe, according to Akere (1982), means boss or leader

(Akere, 1982; Manfredi, 2013). The name ògá has been used to address upper-class people in

Nigeria. Ògá includes but is not limited to top bureaucrats, ministers, agency heads, company

chief executives and individuals associated with people of power (Page, 2020). Table 19 shows

a summary of the requirements of these platforms in Lagos.

Table 19: Summary of Requirements for Uber, Bolt, and Oga-taxi

Driver and

Vehicle

Requirements

Uber

Bolt

Oga-Taxi

Sign-up

Online

Psychometric

test

Yes

Document

upload

-Valid Nigerian driver’s

licence

-Valid LASDRI photo

(recently added)

- Safety screening

-Valid Nigerian driver’s

licence

- Additional document: Valid

LASDRI card, driver badge,

LASRRA (recently added)

Vehicle

inspection

Yes (Car45 inspection)

Yes (Autogenius vehicle

inspection)

Yes

Paperwork

provision

-Vehicle inspection report

- Vehicle insurance

- Roadworthiness

-Hackney permit (recently

added)

Vehicle inspection report

-Vehicle insurance

- Roadworthiness

-Hackney permit (recently

added)

Training

Yes

Vehicle year

4-door vehicle from 2000

upwards

4-door vehicle between 2002

and 2004 upwards

4-door

vehicles

from 2005

upwards

190

Smartphone

requirement

- Minimum Android 6.0 or

IOS 8.0)

- 2GB RAM minimum

- Minimum Android 6.0 or

IOS 8.0)

- 1.5GB minimum

Source: Author's fieldwork (2018 and 2019)

The idea for Oga-taxi was to rival Uber and Bolt, which came shortly after, by

advertising lower commission rates of 15% and luring drivers and unemployed residents with

phrases like "our full-time partners make as much as N350,000 (£755) to N400,000 (£863) a

month" – earnings reflective of being an ògá at the time.

[

]

Unlike well-funded platforms like

Uber and Taxify (now Bolt), these entrepreneurs funded and managed the process of digitising

their services by convincing local taxi drivers of the advantages of being your own ògá. In an

interview with TV360 in 2019, Loko Udoko highlighted how he approached drivers.

According to the Loko Udoko: the target was visiting popular locations like taxi

ranks and hotels where people had car-hire services like the Eko-hotel and suites in

Victoria Island. He was convincing drivers to forfeit traditional taxi driving to drive

for his profitable platform in a nutshell. The reason was that in 2015, drivers making

between N7,000 (£15) to N8,000 (£17) for a trip between Victoria Island and the

airport, the hotel will often keep over 30% commission.

One driver who lost his job at an insurance company decided to assist the Oga-taxi

founder by convincing other traditional taxi drivers to adopt the indigenous hailing platform.

Unlike other ride-hailing platforms, the Oga-taxi platform introduced over fifty brand new

Hyundai Elantra vehicles with the possibility of drivers acquiring them on rental or hire-

purchase plans. The registration process for drivers was seamless and like Uber but without

psychometric tests. The Oga-taxi platform incorporated a Kabu-kabu feature, targeting

traditional taxi drivers and private vehicles, encouraging them to share a ride and fare with

colleagues going to a similar destination with Facebook connecting mutual friends to prevent

trust issues.

[

]

Similar Kabu-kabu types of services like Jekalo and Gomyway launched the

same year but became defunct due to lack of funding, users’ distrust and circumvention of the

app by riders. In an interview with a venture capitalist Abiola (September 2018):

191

…it failed because people started to circumvent it. For instance, if I found

someone going my way regularly, the passenger may no longer need the app again. I

would rather just follow someone I have trusted instead of meeting someone new every

time. The company started finding it difficult to monetise. They shut it down after a year

or two.

During field visits, discussions with drivers indicated the platform was passive and

dormant, i.e., not working hard enough to rival international platforms or keep drivers. Only

two drivers out of 25 mentioned using Oga-taxi as a supplementary platform. I also visited the

office, but it was no longer operational, and their contact number was not responsive.

[

]

Figure 16: Mapping the Emergence of Ride-hailing Platforms in Lagos Since Uber’s Arrival

Source: Author's fieldwork (2019)

Uber, Oga-taxi, and Bolt served as critical turning points in the taxi industry despite

earlier attempts from Easy Taxi and Afrocabs. While the dominant platforms since the

fieldwork experience have been Uber and Bolt, several other ride-hailing platforms, such as

2014

•International platforms: Uber, Taxi Pixi

•Indigenous platforms: Afrocabs (relaunch), Saytaxi (defunct)

2015

•International platforms: None

•Indigenous platforms: Oga-taxi

2016

•International Platforms: Taxify (now Bolt)

•Indigenous platforms: Holy Cabs (defunct), Soole Cabs (defunct)

2017

•International platforms: None

•Indigenous platforms: Oga-taxi (relaunch) Alpha1rides taxi, gudride, jetride, pro-taxi,

smartcab/taxi, GLT, Holla cabs

2018

•International platforms: Taxigo

•Indigenous platforms: Carxie, rideme, Zykte, Tripz, Go247

2019

•International platforms: In-driver, Opay (OCar)

•Indigenous platforms: Gidicabs, Yougo, Intercab

192

In-driver, mentioned above, have emerged in Nigeria as a competitor and a solution to reduce

unemployment

[

]

. As the forerunner of ride-hailing platforms globally, Uber presents itself as

another GN neoliberal ideology that has emerged in GS countries like Nigeria (Kaye-Essien,

2020), creating a spiralling effect for domestic platforms but too powerful to allow indigenous

platforms to grow. In the field, a total of four out of 25 drivers mentioned driving for an

indigenous platform such as Oga-taxi, Gidicabs or Clickcabs (Figure 16). The consensus from

interviews indicates that indigenous platforms are not well organised and do not advertise

adequately to attract drivers and riders. While the emergence of international platforms defines

our techno-utopian futures in the gig economy, it is critical to understand how their emergence

has changed the analogous labour processes to digital processes effectively managed by

algorithms.

4.10. Formalising the Informal: Uberising Gig Work in Lagos

The emergence of ride-hailing platforms in Lagos indicated an abrupt change in the taxi

industry, but there was also a development in how workers were managed and controlled.

Following the different classifications of informal transport, such as third world transport,

paratransit, and intermediate technologies, Cervero (2000) adopts the term informal transport.

According to Cervero (2000, pp.3-4), characteristics of informal transport

modes include “lacking necessary permits or registration for market entry; failure to

meet certification requirements for common-carrier vehicles (e.g., maximum vehicle

size, vehicle fitness standards); lack of liability insurance; substandard vehicles; low

fares; and several other factors”.

While these classifications typify the danfo and Okadas in Lagos, it is evident that the

traditional taxi industry blurs the lines of formality and informality based on other delineations

of the Kabu-Kabu and substandard yellow taxis discussed above. More so, the traditional taxi

industry in Lagos typifies other characteristics of informality which include weak regulatory

frameworks and existing regulatory frameworks without implementations or enforcements,

which facilitate illegality, evasion of taxes and fees, erratic scheduling and service, lack of

accountability and overall substandard quality of labour (Cervero and Golub, 2007; Ehebrecht

193

et al., 2018). The term gig, gig worker or gig employment is not defined or part of the

employment category in the Nigerian Labour Act (Enwukwe, 2021).

The labour laws in Nigeria were adapted from Great Britain during the colonial era

(Nwokpoku et al., 2018). While the legislative power has developed the labour constitution,

the theoretical aspect of these laws does not match the implementation and enforcement

processes in Nigeria. For example, following the centralisation of rules at the Federal level, the

recent implementation of the national minimum wage was not operationalised in different

states based on the inability to pay (Nwokpoku et al., 2018). Another example is the Kaduna

State governor Nasir El-Rufai, who sacked 30,000 workers following a competency check to

ascertain the professionalism and education of teachers (Abuja, 2017). According to

Nwokpoku et al. (2018), this was a violation and non-compliance with the Labour Decree No

21 of 1974, which aims to protect workers against unfair sanctions, including contracts and

terms of employment, wrongful dismissal, issues of wages and other factors. According to

Enwukwe (2021, p. 8):

“The dearth of the definition of the status of gig workers and the legal

framework regulating the terms of conditions of their employment and protection

explains the motivating factor for the increasing use of workers by employers and why

this category of workers is exploited by gig work employers who engage them”

This irregularity in the taxi industry is interrelated with its semi-formal or informal

labour standards and the business model of ride-hailing platforms, which formalises gig work.

In other words, the informality of the taxi industry is now digitised by formalising the

traditional self-employment models for ride-hailing platform drivers in Lagos (Lakemann and

Lay, 2019). While Lakemann and Lay (2019) argue that the uberisation of gig work provides

new opportunities for informal entrepreneurs in African cities, the findings in this thesis show

that the opportunities afforded to ride-hailing drivers specifically are extortionate, precarious,

and burdensome. Rekhviashvili and Sgibnev (2019) argue that the business model by platforms

in the GS are formal, while its labour relations remain informal. In Lagos, however, considering

that taxis have always been gigs, the formal business model has formalised the informal nature

of driving taxis with improved technology through the ease of registration, provision of

independent contracts, real-time pairing and management and trip predictability (see Table 20).

This phenomenon moves away from the idea of driving as a profession to driving as a skill

because it incorporates both professional taxi drivers and everyday citizens with the ideology

of making money as a part-time gig or as a full-time gig. The ride-hailing platform Uber, as

194

the forerunner, signals colonial tales under the guise of technology, such as liberating the

bottom billion, banking the unbanked and connecting the unconnected (Birhane, 2020).

Table 20: The Difference between Traditional Taxis and Ride-hailing Platform Gig Work

Gig Work in Lagos

Taxis

Platforms

Factors of Control: The State transport

ministry, unions, taxi parks, law enforcement

agents

Factors of Control: The smartphone app,

algorithmic management

Informal contracts with vehicle owners and

registration with the State transport ministry

(except operating illegally)

Independent contracts offered online and in-

person

Subject to state regulatory policies

Evades State policies because it is a technology

and not a taxi company, leaving drivers to take

on the risks

Assigned to motor parks

No assignment of parks

Union or association membership

Designed against unionisation

Union dues monthly payment and meetings

Not required as a ride-hailing driver

Passenger access is by hailing a cab or

receiving calls

A push of a button on a smartphone; trips are

assigned based on ratings and as a function of

demand and supply

Fare calculation is based on the driver’s

discretion, and payment is often in cash

Calculated automatically by algorithms and

paid in cash or via the app

Managing the labour process outside a taxi

park is based on the driver’s discretion and

knowledge of the city

Algorithmically managed with integrated maps

and GPS tracking

Source: Author's fieldwork (2018 and 2019)

Ride-hailing platforms exacerbated this notion by propagating utopian phrases such as

'be your own boss', 'no schedule, no boss, good pay’, 'make Oga kind of money', all of which

were timely in entering the system with rising unemployment levels. For these platforms, the

experience of driving feels prestigious, like a white-collar job with the need to dress properly,

195

keep the vehicle clean and respect riders (Rosenblat and Stark, 2016), a narrative observed

from Bolt driver training in Lagos. Platform’s classification of drivers as 'driver-partners’,

'entrepreneurs', 'bosses' reinforces that they are autonomous, independent contractors – a

company partner out of their free will. Drivers internalised such classifications as a prestigious

thing, a clear delineation from traditional taxi drivers and a reflection of what riders should see

them as by calling themselves 'pilots' and 'comrades' (Arubayi 2021a).

[

]

Before now, and from interviews, most of the taxis in Lagos had operated illegally

under the guise of independent unions. With traditional taxis, control processes were linked

with formal and informal political structures such as the state, unions, law enforcement agents,

and other facilitators of control within taxi parks and other contested spaces throughout the city

(Albert, 2007; Agbiboa, 2017). The labour process for taxi drivers was fragmented from how

workers were assigned to a union and a taxi park where they self-organised by creating a

physical list to facilitate trip assignments on a turn-by-turn basis. Like most cities globally,

platforms started their business without owning vehicles while enabling drivers, employees,

and entrepreneurs to make money without being part of a union, paying dues, not being

completely controlled by state regulation and freedom in creating working schedules (Purcell

and Brook, 2020). It became a case of formalising flexibility and autonomy within the industry

in the guise of entrepreneurialism with the capacity to manage drivers without political

structures from the state or control from conventional employers of labour (Jamil and Noiseux,

2018; Purcell and Brook, 2020).

In developed nations, these structures for traditional taxis are more formalised and

regulated by the state with adequate accountability and the ability to manage taxi labour. For

example, in London, Transport for London (TfL) ensures that black taxis and PHVs adhere to

its policies by instituting clear rules and regulations regarding access and payments (Dudley et

al., 2017). Accordingly, PHVs breaking the rules by picking up passengers without passing

through the system are not covered by insurance and could be liable to fines based on

surveillance apparatuses from the dispatch system or CCTV cameras throughout the city. In

Ghana, formalising the informal, as discussed by Kaye-Essien (2020), is based on the

subjectification of the State, considering how the Ghana Private Road Transport Union

(GPRTU) signed a statement of understanding with the Ministry of Transport, giving Uber the

power to create laws that outline examples of what constitutes a utopian mobility system. This

signals how the uberisation of the taxi industry is deep-rooted in our dependence on GN

contexts drawing from their influence on our media consumption, fashion trends, and

technological influence. The technological influence is evident in the reliance on VCs from

196

GN contexts and overall innovative ideologies and infrastructures. In addition, this can be

linked to efforts of the Lagos government to adapt neoliberal ideologies of deregulation and

value appreciation of the taxi license, which in reality does not fit into the Lagos context.

Instead, it is impinging on the weak regulatory environment based (Thelan, 2018), bolstering

informality of labour and impoverishing drivers while establishing a solid customer base across

Nigeria.

There are limitations to the innovation of local products because of the reliance on GN

contexts such as the US on technology and AI tools, which lack transferability and context

specificity (Birhane 2020). For example, according to the director-general of the National

Office and Technology Acquisition and Promotion, the low patent culture for innovations in

Nigeria has led to over 90% of imported technologies driving the Nigerian economy instead of

emerging from universities, research establishments and colleges of education (Ogunfuwa,

2017; Oyewale et al., 2018). These also apply to Uber and Bolt, which remain dominant but

truncate context specificity and limit local ride-hailing platforms like Oga-taxi from growing

due to their superior technology and better access to funding. According to Amorim and Moda

(2020), the real infrastructure of control of the labour process is data or what they call the

technological means of production cast via smartphone apps, with the capability of radicalising

historical processes of work subsumption. Drivers only possess the vehicle, in some cases

based on subcontracted arrangements, trickling down the ownership of some means of

production as suggested by Amorim and Moda (2020) to just the smartphone. This leads the

analysis into the next chapter, where I examine intricately how platform emergence has

integrated algorithmic management in Lagos and its impacts on ride-hailing platform drivers.

4.11. Conclusion

This chapter has examined the emergence of ride-hailing platforms in Lagos, and how

the technology has reconfigured the taxi industry. It outlines how taxis and ride-hailing

platforms emerged in GN and South cities and the differences that exist between them. More

critically, the emergence of international ride-hailing platforms such as Uber and Bolt has

improved managing drivers digitally. In previous years, despite efforts to formalise taxi

operations, taxis such as the yellow taxis have operated informally because of weak regulatory

frameworks, an elderly and uneducated workforce and power-driven unions that deprioritise

the drivers' needs. The analogous management of the taxi industry made it difficult for

197

companies to monitor and evaluate the driving process leading to shortcomings for taxi

companies. Managing drivers in atomised units of motor parks gave unions and agberos

extortionate power, which created some form of control over drivers by ensuring payment of

dues, the physical arrangement of trips and an enclosed environment for accessing passengers.

Metro taxis, Easy Taxi and Afrocabs, which attempted to use technology in managing the

workforce, also failed because of the challenges in adopting drivers and managing the process

– a challenge prevalent in yellow taxis. However, this has changed with the emergence of ride-

hailing platforms.

Uberisation in Lagos since 2014 has led to the emergence of other international

platforms and local platforms such as Oga-taxi, Bolt, In-driver and others. It has also fostered

entrepreneurialism or self-employment through independent contracts for Lagos's unemployed

and underemployed people. To an extent, platforms have formalised the taxi industry by

providing clear registration processes and the ease of monitoring ride-hailing gigs and

managing drivers. The digitisation of the taxi industry and the labour process have created an

avenue for platform companies through algorithms to assign work, monitor, evaluate and

sanction drivers through gamic elements such as rating metrics, trip acceptance and

cancellation metrics and an incentivised system. It has become possible to measure and manage

the labour process of drivers on these platforms using algorithms through surveillant

assemblage (Haggerty and Ericson, 2000) from a God’s-eye view held by platform companies

and due to customer surveillance (Newlands, 2020). However, "as technological solutions are

increasingly deployed and integrated into social, economic, and political spheres, so are the

problems that arise with the digitisation and automation of everyday life" (Birhane, 2020, p.

400). In other words, the improvement in managing and controlling ride-hailing gig work has

created new problems for ride-hailing drivers and reinforced pre-existing problems from the

taxi regime, which are examined in the next chapter.

198

5. Chapter Five: Examining the Impacts of Algorithmic

Management in Ride-hailing Gig Work in Lagos

5.1. Introduction

It is another day, and ride-hailing drivers are roaming the streets of Lagos, waiting for

a trip assignment from platform algorithms (see Figure 17). This experience has become the

reality of platform drivers globally, where ride-hailing platforms, through algorithms, can

control and manage the labour process under the guise of autonomy and flexibility for their

workers (Arubayi 2021a).

Figure 17: A Driver Roaming the Streets of Victoria Island Lagos for Trips

Source: Author’s fieldwork, (November 2018); Arubayi, (2020)

Min Kyung Lee and colleagues, who coined the term algorithmic management in 2015,

define it as software algorithms that rely on big data and surveillance to manage and control

gig workers' labour process (Lee et al., 2015; Arubayi, 2021). Here, algorithmic management

has replaced the traditional confinements of managing the labour process faced by conventional

199

taxis in Lagos, reinforced by a surveillant assemblage (Haggerty and Ericson 2000) and

external modalities which facilitate task assignments, monitoring and surveillance,

performance evaluation, payment and rewards, and deactivations and sanctions (Rosenblat and

Stark, 2016; Möhlmann and Zalmanson, 2017; Kellogg et al., 2020). The chapter examines

how drivers are managed by algorithms and how the hidden processes of algorithmic

management impact drivers and the labour process.

By so doing, I argue that algorithms are characterised by opacity, biases, and

information asymmetries which facilitate impacts on drivers (Pasquale, 2015; Rosenblat and

Stark, 2016). Defining the meanings of opacities, bias, and information asymmetries in

subsequent sections determines how these interact to impact drivers in Lagos. Ontologically,

and as an output in this chapter, I classify these underlying characteristics facilitating impacts

as algorithmic burdens, where these burdens are reinforced by the mismatched realities

between data and drivers based on contextual realities, everyday experiences, and the culture

of everyday driving in Lagos. While algorithms assign work and effectively manage the labour

process using performance metrics and gamified tools such as ratings, it also comes with

burdens of labour which exacerbates impacts on drivers such as arbitrary deactivations, and

manipulative incentives, which can lead to self-exploitation, and others discussed in this

chapter.

Because algorithmic management possesses assemblages and modalities, it is expected

that drivers should not experience these burdens (Haggerty and Ericson, 2000; Newlands,

2020). However, findings show that this is not the case in Lagos. It is also interesting to note

that based on the hidden nature and mystery of platform algorithms, I classify algorithmic

management reinforced by surveillance capabilities as weapons of the dominant, an ontological

rethinking of weapons of the weak by James Scott (1985; 1989), discussed in the subsequent

chapter. In this case, drivers are aware that the labour process is being managed and watched

through processes of the assemblage (e.g., gamic elements such as ratings) but are not entirely

aware of how algorithms operate, limiting their control over the labour process and decisions

affecting them (Arubayi, 2021). This chapter sets up the subsequent chapter to examine how

drivers resist the impacts of algorithmic management based on how they navigate through

decisions that affect the labour process.

200

5.2. Algorithmic Management as a Weapon of the Dominant: Effective

Control of Ride-hailing Gig Work in Lagos

The introduction of ride-hailing platforms in Lagos has made it possible for dominant

platforms like Uber and Bolt to manage the labour process because of data and algorithms

seamlessly. Before examining the impacts of algorithms in Lagos, this section outlines some

of the characteristics of algorithmic management adapted from Rosenblat and Stark (2016),

Möhlmann and Zalmanson (2017), and Katzenbach (2019) as positive contributions to the

mobility sector and the future of labour. As highlighted in the previous chapter, ride-hailing

platforms have digitised and formalised the labour process of driving in Lagos, which many

scholars fail to acknowledge. I classify algorithmic management as one of the non-intrusive

weapons of the dominant based on Scott’s weapons of the weak, i.e., the hidden and public

resistances of subordinate groups.

Weapons of the dominant here include algorithmic management, reinforced by

surveillance capabilities, which will be examined further in the subsequent chapter. This

section outlines the key components of algorithmic management in ride-hailing gig work,

drawing from Lee et al. (2015), Möhlmann and Zalmanson (2017), Katzenback and Ulbricht

(2019), and Kellogg et al. (2020). While Kellogg et al. (2020) conceptualise algorithmic

management as algorithmic directions, algorithmic evaluations, and algorithmic discipline, this

section simplifies these based on its technicalities in the field, such as task assignment, and

monitoring and surveillance, performance evaluation, payment and rewards and sanctions and

deactivations. Because these layers operate simultaneously, the following section attempts to

outline and analyse how algorithmic management works in Lagos. Subsequent sections

critically examine the impacts drivers experience based on opacities, information asymmetries,

and biases embedded in the process of algorithmic management.

5.2.1. Big Data and Surveillance

With the help of big data, ride-hailing platforms possess the power to store and manage

a larger quantity of information, which is also integral in the surveillance of labour and drivers’

behaviours (Andrejevic and Gates, 2014; Zuboff, 2015). In Lagos, it is impossible to track

drivers' behaviours through technological apparatus such as CCTV footage or dash cameras

compared to developed cities like London.

201

Figure 18: Conceptualising the Surveillant Assemblage and External Modalities of a Lagos Driver

Source: Author's fieldwork (2018 and 2019)

Platform Surveillance (god’s-

view). Platform companies

provide the Uber and Bolt

platform and monitor the overall

process in Lagos. They possess

data which are not disclosed.

Customer Surveillance. Riders

are responsible for rating

drivers while algorithmic

managers judge based on 4.5

(for Bolt) and 4.6 for Uber if

drivers are fit to work.

Algorithmic Surveillance. Algorithmic

management is responsible for the

following: assigning jobs to drivers;

ensuring real-time pairing through GPS

between riders and drivers; setting prices;

judging appraisals; assigning rewards;

terminating contracts.

Vehicle owner surveillance. Car

owners or leasing companies that

lease cars to drivers for investment

purposes and appraise drivers’ weekly

rental payments. Multiple defaults

could see drivers contracts

terminated.

Law Enforcement Surveillance.

These include LASTMA officials,

VIO, and the police who

reinforce the need for drivers to

follow the laws of driving in

Lagos. The State supervises the

whole process of e-hailing and

learns from platform companies.

It implements law and modifies

rules to accommodate platform

companies and its workers.

Self-Surveillance. Drivers are expected to

analyse daily decision-making based on

weekly earnings, labour conditions, and

are offered the freedom of choosing when

to work through the app. This can be

reinforced by external bodies such as

vehicle owners or law enforcement agents

which track and may punish behaviour

shortcomings.

Assemblages

External Modalities

202

Figure 18 shows external modalities drawing from the surveillant assemblage in

understanding platforms, which include god’s-view surveillance, algorithmic surveillance,

rider surveillance, vehicle owner surveillance and self-surveillance, all subject to algorithmic

management (Haggerty and Ericson, 2000; Newlands, 2020). With the ride-hailing apps

embedded in smartphones, algorithms are always watching, and with their unverifiable nature

(Foucault, 1995; 2008), tracking and managing drivers is a possibility. The figure above shows

how surveillance at different levels reinforces control of the driver. As a central point, the app

enables algorithmic surveillance based on the factors of algorithmic management; customer

surveillance based on ratings; platform surveillance based on data analytics; and self-

surveillance based on a combination of all the above and external modalities. Algorithms, as

an underlying factor, facilitate riders to rate drivers after trips. This reinforces the need for

drivers to drive properly to ensure that riders input 5-star ratings, which feedback into the

system and are stored on the platform company databases for future analysis.

This self-surveillance can also be influenced by external factors such as vehicle owners

or law enforcement agents who ensure that drivers adhere to traffic laws or contractual

agreements based on weekly payments. Because drivers are expected to pay weekly payments,

as discussed in the previous chapter, they are expected to maximise their efforts to make profits,

including maintaining the vehicle. While algorithms control the process during trips, vehicle

owners still possess control over drivers because of the expectation from weekly outcomes and

supervision in terms of how these vehicles are being utilised. Some vehicle owners even install

vehicle trackers to ensure that drivers are working and to prevent theft of their vehicles.

[

]

The

inability to interview Uber and Bolt platform core managers made it difficult to understand

how god’s-view surveillance works in reality. All indication from news outlets concerning the

use of Uber's god’s-view to spy on politicians and public figures (e.g., Hern, 2016) indicates

the presence of these infrastructures. In Lagos, however, one interview with a bike-hailing

company revealed some of the traditions of algorithmic management and surveillance

infrastructures. An interview with a key representative of the Max.ng bike-hailing platform

revealed:

We have what we call mission control, where we have data for every rider after

the verification and profiling. We also utilise god’s-view; it is like a dashboard where

we can track every rider at every point in time. It can help track how these drivers move

around, how often they apply their brakes, the kind of routes they take. We have those

analytics that helps us manage the data properly. These analytics tell us how many

203

rides they make a day and how many requests they get daily… (Benjamin, November

2018)

According to interviewee Benjamin (November 2018), the app is a surveillance

apparatus that also extends self-surveillance to riders by ensuring that they drive efficiently,

follow map directions and are cautious in applying brakes, amongst other things. This also

indicates that drivers can be trackable and traceable across the city of Lagos, particularly in

challenging situations, unless there is a disconnect from the platform. However, based on the

model, the bikes also possess trackers in scenarios of platform disconnection. This improves

drivers' safety because of the availability of data and the possibility of always knowing drivers’

locations. Based on this realisation, people’s experiences of ride-hailing platforms such as Uber

are quite similar; the caveat is the lack of knowledge and transparency from top to bottom,

which drives more control and power for platforms. Unlike the taxi regime, drivers can estimate

trip durations and pinpoint the locations of potential riders. With the help of Google maps

integrated within the Uber and Bolt apps, drivers can navigate the city seamlessly, although

first-hand knowledge of the city is critical, especially in unidentifiable places in the city. Digital

maps can provide drivers with routes in intense traffic congestion or blocked routes.

This accumulation of data for surveillance is integral for keeping records of their

weekly labour, central to payments, rewards and deactivations. Algorithms enable seamless

weekly payments to drivers recorded in the ride-hailing app (see Figure 19). This also makes

it easy for both drivers and vehicle owners to examine weekly payments from the platforms.

204

Figure 19: Display of Weekly Earnings on Uber and Hours Worked on Bolt

Source: Author's fieldwork (2018)

205

Similarly, the platformisation of the taxi industry made it possible for drivers to pay

taxes, which was challenging to track and manage in previous years for the Lagos State

Ministry of Transport. In 2016, Uber, in collaboration with PricewaterhouseCoopers (PWC),

Federal Inland Revenue (FIRS) and Lagos State Internal Revenue (LIRS), formalised the

process of tax payment for drivers, making it compulsory for drivers to provide a Tax

Identification Number (TIN) as a requirement for being a driver-partner (Uber Africa

Newsroom, 2016). Payment for trips with Uber and Bolt is integrated and managed by the app

with card or cash payment, and earnings summaries are displayed weekly within the app (see

Figure 20)

[

]

. During this study, 25% and 15% commissions were collected on Uber and Bolt,

respectively, after each trip. For Bolt, drivers receiving cash payments could owe the platform

N10,000 (£22) per week. If drivers exceed N20,000 (£43) without paying into a Bolt bank

account, trip assignments depreciate, and payments become limited to card trips only with a

risk of deactivation. On the other hand, Uber deducts its commissions before displaying the

final fare, as one driver noted.

5.2.2. Gig Assignments and Performance Evaluation

In attending a Bolt drivers' training exercise (November 2018) at their headquarters in

Lekki Lagos, I observed the rationale behind the algorithmic trip assignment. On Bolt, it is

advised that trips are kept within a 2km radius to prevent long trips to reach the next passenger.

Drivers can receive trips outside the radius but are not penalised if they reject them. On Uber,

the pick-up radius is hidden and is often related to a passenger's availability and the driver's

initiative. When algorithms assign a trip to drivers, they have 15 and 25 seconds on Uber and

Bolt to accept or reject a trip. The rate of drivers accepting or rejecting trips is defined as

acceptance rates and cancellation rates in the app dashboard. While the intention of autonomy

is clear, accepting or rejecting trips impacts how quickly algorithms assign trips in the case of

Uber and the activity score on Bolt. The activity score for Bolt indicates that the more active a

driver is, the more trip assignments they receive from algorithms.

[

]

For instance, if a driver

receives ten trips and cancels 5, the activity score would be 50%. If two drivers, A and B, are

in a location where A has 70% on his activity score and B has 99%, the algorithm assigns to

driver B first based on consistency in acceptance over time, even if driver A had more trips in

a day (see Figure 20).

206

Figure 20: Activity Dashboard and Activity Score Page for Bolt

Source: Author's fieldwork – Platform drivers' screenshot (December 2018)

The most critical metric is the ratings on a platform because it determines if a driver

continues to work on it (Rosenblat and Stark, 2016). This is because it is related to the number

of trips a driver gets and, in many cases, the bonuses involved. On Uber, for instance, ratings

are calculated by adding your individual trip ratings (1-5 stars) and dividing by the total number

of ratings usually received up to 500 most recent trips.

[

]

The Bolt platform also uses a similar

rating metric. However, the terms and conditions on Bolt clearly state in Article 8.2 that ratings

and activity scores are interlinked, such that it is measured based on the driver's activity in

relation to declining, accepting, not responding, and completing transportation service

requests.

[

]

Finally, these performance measurement metrics influence the incentivisation of gig

work using bonuses or surge pricing, where ride-hailing drivers qualify for bonuses from Bolt

and Uber for the hours of labour on these platforms. It is like a reward for working hard. As

mentioned in the previous chapter, Bolt utilised this as a strategy in recruiting drivers to its

platform six months after its emergence in Lagos. Despite the positive integration of platforms

207

in gig work in Lagos, the very elements that permit seamless workforce management are

creating burdens for drivers that impact the labour process and all-around productivity. The

following section shifts the discussion to examine how algorithmic management causes may

impact drivers, which are inherently driven by informational irregularities such as opacities,

informational asymmetry, and biases proliferated by a mismatch in the everyday realities of

drivers.

5.3. Characterising Algorithmic Impacts: A Mismatch of Data and Ride-

hailing Drivers’ Realities

Platforms promise the marketisation of transparency which breeds information

symmetries and the autonomy and flexibility of labour. However, as this chapter shows, ride-

hailing platform drivers are experiencing the reverse in Lagos, a phenomenon that is also

prevalent globally. Beer (2017, p. 3) argues that the "uncertainty about algorithms leads to the

misjudgement of its power by overemphasising its importance or misconceiving algorithms as

a lone detached actor and how power might be deployed through technologies”. The

dependency on data based on hidden coded values creates what Pasquale (2015) identifies as a

black box society. Platforms utilise this hidden management to safeguard their data and enable

more control and power over drivers' bodies. Here I argue that these impacts are driven by three

characteristics which include opacity, informational asymmetries, and biases, as discussed in

chapter two. These three interact interchangeably to create burdensome labour for drivers based

on the risks they inherit as a result of poor information sharing from platforms.

Opacities

The hidden nature of algorithmic management is such that information is limited to

ride-hailing drivers in Lagos. Based on real secrecy, legal secrecy, and obfuscation levels of

opacity, as highlighted in chapter two (Pasquale, 2015), platform algorithms facilitate opacities

that limit both drivers, researchers, regulatory bodies, and other gig economy advocates from

fully comprehending why and how certain decisions are taken. These levels of secrecy facilitate

opacities by platforms like Uber because it limits researchers, regulatory bodies, and gig

workers from fully comprehending why and how certain decisions are actualised. For example,

this chapter will show that drivers in Lagos often do not know why they were

blocked/deactivated, restricted to fewer trips on the app, or prevented from actualising a

208

promotional trip because they do not have access to how algorithms determine decisions behind

the scenes. This is why drivers continually try to anticipate decisions and counter these actions

by learning from SMCN (discussed in the next chapter).

A critical challenge of opacity for drivers in Lagos is how the payment and

commissions are determined, i.e., if they are consistently accurate weekly. For example, in

talking about the changes in payment and commissions for Bolt, Koffi stated:

“Formerly on Taxify, if a trip is N5,000 (£10), and they are taking 15%, that

means they would take N750 (£1.50). Driver will go home with N4,250. Now they have

increased price by 10%, that is N5,500 (£11). If they take 20% out of that, it would be

N1,100 (£2.30), and the driver will go with N4,400 (£9). Which is an additional N150

for the driver, but on their side, they have an additional N350 for themselves. Nobody

did these calculations. Drivers are still in slavery

This quote is based on Koffi’s personal observations and calculations, which are limited

by the lack of transparency in making these decisions that affect their labour. Jarrahi et al.

(2021) argue that most gig platforms possess the power to tweak their algorithms because these

are developed internally, hence enabling more opacity. This indicates that ride-hailing

platforms do have any accountability because they continue to obfuscate and hide their inner

workings, making it complex for workers to comprehend (Diakopoulos, 2015). Uber and Bolt

possess the power to manipulate information that facilitates burdensome labour, which is often

detrimental to their earnings, wellbeing, and work ethic.

Information Asymmetry

As discussed in chapter two, information asymmetry builds on the opacity of

information, i.e., drivers lacking sufficient knowledge about how the platform works, including

its gamic elements of engagement tools such as blind passenger acceptance (Rosenblat and

Stark, 2016). In other words, drivers appear to lack informational power compared to riders.

Through its surveillant assemblage, ride-hailing platforms have more information about drivers

and riders, but drivers have less power in bargaining about core decisions that affect their work

because of the lack of information to make informed decisions. This indicates that platforms

like Uber, through its algorithms, provide hierarchical informational flows which decide the

types of content and means of disclosing information to drivers, all of which are exacerbated

209

by the automated communications and algorithmic features, including ratings which act as

barriers between drivers and the platforms (Bokànyi and Hannàk, 2020).

Information asymmetries, as demonstrated in the subsequent section, are evident in how

drivers are unaware of the destination of potential riders but are coerced to accept trips even

though these may impact the safety and security of drivers. As a building block of the opacity,

these algorithms also create information asymmetries that limit drivers’ knowledge in rider

rating decisions, registration processes, promotional trips, and conflict resolution and appeal

processes.

Bias

A worker or group of people who are already privileged are shown preferential

treatment by algorithms while discriminating against vulnerable or marginalised groups based

on historical data sets of human behaviours and experiences (Noble 2018; Katzenbach and

Ulbricht 2019). Chapter two highlights different biases observed in a conventional workplace,

such as racial or age biases (Rosette et al., 2008; Ajunwa, 2019) and increasingly becoming

more evident within digital technology platforms, such as language and appearance bias

(Hanrahan et al., 2018). This phenomenon builds on opacity and asymmetrical information

platform drivers' experience daily. The platform algorithm subsequently favours decisions that

are beneficial to the platform business, irrespective of workers being disadvantaged. As the

subsequent section below will show, the main bias that was evident in Lagos is what I classify

as rider decision bias. Because the system is opaque, the algorithm possesses the power to make

asymmetrical decisions which in most cases are biased towards riders. For example, in

scenarios of conflict between a rider and a driver, drivers in this study argue that the algorithms

are always in favour of riders leading to arbitrary deactivations, as discussed in the next section.

Algorithmic management in Lagos, as this chapter shows, is an effective weapon of

control for drivers. This has brought about positives through the capacity for big data and

surveillance, which facilitate gig assignments and performance evaluations which were

previously not conceivable in Lagos. However, despite the positives to the transport industry,

algorithmic management possesses a conflation of opacities, information asymmetries and

biases which impact the everyday labour of drivers.

It is evident that the opacities, information asymmetries and rider bias are interactive

processes of secrecy based on the business models and design of Uber and Bolt platforms in

Lagos. As a blanket term, I propose to classify these three elements as algorithmic burdens of

210

labour because of the risks and toil of responsibility that are outsourced to drivers. Based on

these characteristics, the sections below show how these factors interplay to prevent drivers

from making informed decisions while possessing more information about them that is

disproportionately shared between the platform, riders, and drivers.

5.3.1. Price Mechanism and Precarious Vehicle Arrangements:

Proliferators of Overworking

During the field exercise, all the drivers complained about the unfairness of ride-hailing

platforms' pricing algorithm and how it does not change with the times. In transitioning from

traditional employment for over 13 years, Okoro, who has been driving for both Uber and Bolt

since 2018, expressed why the payment is not commensurate with labour.

The charges do not commensurate with the work one puts in. Assuming you

work from morning to 5 pm, by the time you subtract fuel money, commission and other

expenses, most times on an average day, the driver is left with only N5,000 (£10.6). The

return is not as rewarding as it should be. (Interviewee Okoro, August 2018)

The expression from Okoro was the reality for all the 25 drivers in this study,

particularly since 2017. In 2016, the base fare for Uber was N400 (£0.84), N90 (£0.18) per

kilometre and N9 (£0.016) per minute on Uber.

[

]

Even though it was significantly less than

the unstable charges of traditional taxis, it was substantial for drivers working on the platform

from inception. Also, the fuel price in 2016 was about N97 (£0.20) per litre, indicating that

drivers spent less money on fuel and saved more income. For example, a trip from Lekki Phase

One on the Island to the International airport on the Mainland, Ikeja, costs between N4,000

(£8.4) and N5,000 (£10.5) (Interviewee George, Sept. 2019). In this study, 6 out of 25 drivers

working for Uber since 2017 highlighted the depreciation of payments since the emergence of

Bolt. Drivers perceived that the occurrence led to competition in Lagos because the Bolt

platform introduced lower fares and lower models of saloon vehicles, forcing Uber to reduce

fares and vehicle models. Bolt proposed lower fares on its journeys which boosted drivers’

earnings because of higher demand and a 15% commission compared to Uber's 25%.

Consequently, for Uber to compete with Bolt, Uber slashed its trip fares by 40%, signalling to

211

drivers that cheaper fares would attract a higher demand and ultimately improve earnings

(Uber.com.ng). This change resulted in the first significant protest by Uber drivers discussed

in the subsequent chapter. Several drivers felt exploited by the company; some drivers like

Jacob (August 2019) and Koffi (September 2019) completely stopped working for Uber, while

others switched to Bolt and worked on Uber when convenient.

Implementing platform regulations and managing the labour processes should

recognise the Nigerian economic landscape changes that are likely to impact drivers' earnings.

According to World Bank (2018), the inflation rate increase was high in 2016 and 2017 at

15.7% and 16.5%, respectively. As much as the recession affected businesses, it also rendered

the Naira weaker than the Dollar, meaning the cost of basic amenities like food prices

increased. Essential car maintenance also increased over time. For example, the cost of

pumping a tyre was N50 (£0.10) between 2015 and 2016; it is currently N200 (£0.42) per tyre,

indicating more money assigned to maintenance costs (Interviewee George, Sept. 2019).

Platform algorithms do not consider these indicators in calculating driver earnings, which, in

most cases, are based on previous economic realities – expecting drivers to cope with pay that

does not reflect the current economic realities.

At the time of Bolt's emergence in 2016, the low commission rates and ample bonuses

outlined previously were beneficial to early adopters in this study, such as Koffi, Efe, Charles,

Osahon, George and drivers throughout the ride-hailing sector in Lagos. However, this was a

ploy to enable Bolt to capture most drivers and increase the commission rate to 20% in 2019.

Taxify (Bolt) has increased price by 10% per N60km. If you increase the price

by 10%, it should be about N66 (£0.14), but it is N65 (£0.13) on the platform. Formerly,

15% of N5,000 (£10.5) is N750 (£1.57). The driver goes home with N4,250 (£8.9). Now,

the price has increased by 10%, that is N5,500 (£11.6). 20% from that is N1,100 (£2.3),

and the driver goes home with N4,400 (£9.3), which means additional N150 (£0.32) for

the driver and an additional N350 for them. Nobody is asking them questions. We are

still in slavery. (Interviewee Koffi, September 2019)

Drivers are caught in the middle of the competition between Uber and Bolt, which

further exposes them to unfair decisions affecting the labour process. Instead, drivers are left

in the hands of the dynamic pricing algorithms, which facilitate long working hours but with

even lesser remuneration in addition to evolving maintenance costs. The precarious

arrangements with vehicle owners further exacerbate this. Only 7 out of 25 drivers owned their

212

vehicles in this study, while the other drivers paid weekly to independent vehicle owners or

rental companies. In many instances, vehicle owners purchase a fairly used vehicle (popularly

known as a Tokunbo car) which ranges from N1.2m (£2,526) to N1.9m (£4,000).

[

]

Vehicle

owners' contract potential drivers, demanding profits of between N300 000 (£632) and

N1,000,000 (£2,105) at the end of the contract. Sometimes, it is a fixed contract between 24

and 36 months, depending on the drivers' ability to meet the target. Many platform drivers

attempt to complete the contract as early as possible, leading to self-exploitation – pushing

themselves to acquire the vehicle.

In London, over 83 drivers highlighted how Uber's business model limits the chances

of making minimum wage despite working long hours and paying for maintenance costs

(Lawrence, 2016). In response, Uber claimed that drivers make an average of £16 an hour after

commission rate deductions, but not including maintenance costs and other expenses on the

job. Parliament member Frank Field compared the labour practice to sweated labour in the

Victorian era when drivers worked long hours with little pay, high risks, and insecurities

(Lawrence, 2016). In South Africa, Carmody and Fortuin (2019) observe a similar trend with

working hours of about 64 hours weekly, equating to 13 hours a day over five days or 11 hours

a day over six days. However, many drivers work between 70 to 90 hours a week in Lagos, all

trying to meet their weekly targets for their respective vehicle owners. According to Charles

and Jude's (November 2018) interviews, drivers who own their vehicle possess more freedom

to work when they choose, although they often work an average of 50 to 60 hours a week,

depending on the speed at which they make at least N15,000 (£32) daily. At the time of this

study, Uber implemented a 12-hour working cap, followed by a 6-hour break from the app to

prevent overworking. This action did not restore the underlying problem because algorithms

still decided the prices, and Uber still collected a 25% commission. On the other hand, Bolt

does not possess a working cap, as drivers can switch to Bolt for longer hours when Uber

restricts their app (Interviewee Jude, November 2018). It is important to note that these time

caps did not include the drivers roaming for trips, but only active driving hours, which enables

drivers to work beyond a 12-hour cap.

On average, weekly remittance to vehicle owners for all the drivers in this study ranges

between N25,000 (£52.6) and N35,000 (£73.7), with the former being more affordable for most

drivers (FGD, November 2018). In South Africa, weekly remittance is equivalent to $190 to

$250 a week, which is high for drivers because they are expected to pay rental fees or, in some

cases, receive earnings directly from partners registered on Uber in addition to daily and weekly

expenses (de Greaf, 2018). The problem is that platforms outsource these burdens to drivers

213

by facilitating vehicle owners and rental companies as a sub-layer of surveillance and control,

leading to the self-exploitation of drivers. According to the interview with Charles (November

2018), some vehicle owners install trackers in their vehicles as a safety practice to prevent

drivers from stealing them. This further reinforces self-surveillance because the driver is

appraising himself against the algorithm, customers, and vehicle owners to demonstrate that

excellent service is rendered, and the vehicle is utilised for its sole purpose – driving.

Interviewee Samuel (August 2019), an HND graduate driver who has been driving for Uber

for over eight months, outlined the issue with vehicle owners as an extended facet of Uber's

business model – an entrapment of drivers’ labour based on the modality of surveillance

(Haggerty, 2000). According to Samuel:

… Some car owners believe that when you are driving, you are making 1 million

Naira (£ 2,105). It was before, and I can guarantee that a driver can drive in a week

and make about N150,000 (£316); deduct fuel N30,000 (£63.4), deduct food another

N30,000 (£63.4) — the rest money he can pay the partner N30,000 (£63.4) or N35,000

(£73.7) and still go home with that much. But now we are struggling with N20,000 -

N25,000 (£42 - £53) … (Interviewee Samuel, August 2019)

Another driver (Akpos, August 2019), who has been working for Uber, Bolt and

Gidicabs since 2017, also complained about the reality of these precarious arrangements

considering that drivers bear most of the maintenance costs without assistance from platforms:

… 90% of partners will tell you N9,000 (£19) below will be handled by

the driver while the partner will handle N10,000 (£21) and above. If you look at it

logically, how often does a vehicle incur damages that will cost you more than N10,000

(£21) monthly? Most often, servicing will cost you about N6,000 -N7,000 (£12.6 –

£14.7). Change parts about N2,000 (£4.2). (Interviewee Akpos, August 2019)

When prices are determined by platforms and controlled by algorithms, it becomes difficult for

drivers to achieve targets considering the entanglements in precarious vehicle arrangements. It

is safe to argue that if ride-hailing platform drivers are independent, contractors' algorithms

should not be allowed to manage the labour process nested under layers of surveillance

modalities in vehicle ownerships adopted from conventional taxis. However, algorithms are

responsible for workers and integral for weekly remuneration to vehicle owners or rental

214

companies. This creates an avenue for self-exploitation, such as when workers overwork to

meet a target, and exploitation, such as dynamic price changes due to platform competition

without consulting drivers and their realities.

5.3.2. Data Misrepresentation: Exposing Drivers to Contextual Risks

The recruitment process for drivers on ride-hailing platforms through digitalised means

and proper database management improves the safety and security of riders because algorithms

possess the power to remove deviant drivers from platforms based on behaviour or low ratings.

However, this is not often the case because algorithms cannot resolve problems when there is

a low barrier for riders with inaccurate or no information about their existence. During the

registration process for both Uber and Bolt, drivers provide sensitive information, including

home addresses (which can be verified by providing guarantors), vehicle details, driver’s

licence, and two guarantors. Consequently, riders provide contact numbers, bank card details

(optional), email addresses and home addresses (which cannot be verified because of the

flawed addressing system in Nigeria). The Nigerian Postal Service (NIPOST) revealed that

only 20% of the Nigerian population possesses a correct address capable of receiving a mail at

their residence (Adeputun, 2017), confirming the difficulties in verifying riders' information.

Thus, riders possess more power because while the data sparsely verifies their existence,

algorithms are automatically biased towards them, leading to a misrepresentation of drivers.

Biased in the sense that drivers are always at the brunt of the problem. In scenarios where a

rider is wrong, and they report a misunderstanding, or there's a conflict with the driver, the

driver is often sanctioned or blocked first, as observed in Lagos. Drivers feel unequally treated

because the consequences of data misrepresentation are evident in Lagos.

To this present time, we have 30 – 35 of drivers been killed by riders because

they do not profile them well, many riders do not use their correct names and do not

put the correct information, and they are collecting cars and killing people, and that is

why we need the government to regulate things. (Platform driver and Union president

Dipo, January 2019)

A rider can register on a platform several times to avoid app sanctions or search for

promotional trips. Conversely, this occurrence makes it challenging to track riders who

215

defaulted because of the low entry barrier. For instance, it takes less than a 4.6/5 and a 4.5/5-

star rating on Uber and Bolt, respectively, for a driver to be blocked. In contrast, there are no

clear thresholds that permanently restrict riders from the system in places like Lagos because

riders can register with multiple accounts, which can be detrimental to the business model and

the safety of drivers. Although in places like San Francisco, Uber has particularly expressed

that riders with poor ratings can be restricted from utilising the app (Paul, 2019), during this

study, there was no indication that riders who perpetrate bad behaviours are blocked,

particularly on the Bolt platform.

An underlying causal factor for this problem is the porous identification system in

Nigeria. The National Identity Management Commission (NIMC) established an Act in 2007

to enrol the citizens of Nigeria with National Identity Numbers (NIN) (NIMC, 2014).

However, according to the Director-General Aliyu Aziz, only 36 million Nigerians out of

approximately 200 million Nigerians are registered (Okere, 2019; Olurounbi, 2019), indicating

the challenge in revealing the true identities of residents. It is interesting to note that this

anomaly is a function of the large population size and the incompetence of the commission to

issue NIN when due.

[

]

Accordingly, ride-hailing platforms provide some form of digital

identity for drivers through the app. However, with an imbalance in information, riders are not

accountable to the platforms, thereby affecting the efficiency and productivity of labour

discussed later in the chapter. Arguably, passengers were equally not accountable for their

actions within the taxi regime compared to platforms. With data and algorithmic management,

it is expected that riders like drivers should be more accountable, but the algorithm shifts these

risks onto the drivers creating a power imbalance.

In cities globally, Uber has faced criticisms about its vetting process for drivers,

expressing that it does not conduct proper background checks compared to traditional taxi

companies. In 2014, over three states in the US, such as Oregon, Illinois, and California, filed

suits describing Uber as an illegal transport company that does not rigorously vet drivers on

the platform because of the exploitation of riders on the platform (Isaac, 2014). According to

a CNN report, about 103 Uber drivers assaulted riders between 2014 and 2018 in the US

(O'Brien et al., 2018). In such contexts, it is relatively easy to investigate these cases because

of the availability of information and true identities such as the social security number in the

US and the availability of CCTV footage, an additional surveillance infrastructure integral for

dispute resolution and solving crimes. The interviewee Osahon (September 2018), a former

hotel accountant and now a driver for Uber and Bolt, narrates how the imbalance in information

creates difficulties for drivers. To summarise, according to Osahon:

216

A woman threatened to kill him because he refused to drive into her estate due

to the bad state of the road and flooding in the area. One of the neighbours told him

that the woman had exhibited this behaviour with several drivers before him.

(Interviewee Osahon, September 2018)

This observation indicates that despite the incessant complaints from drivers about this

rider, there was no sanction on the rider's account, and the rider still utilised the platform. These

information asymmetries are compounded by time limitations in accepting trips on the

platforms, which are 15 and 25 seconds on Uber and Bolt, respectively (Rosenblat and Stark

2016). The opacity from algorithms prevents drivers from detecting a rider’s destination and

prevents intense screening of a rider’s character, behaviours and, more critically, their ratings.

While the ratings sometimes assist six out of 25 drivers in this study in deciphering the

behaviours of potential riders, this is not adequate for determining their identity and history of

deviant behaviours. To summarise the interview with Akin (November 2018):

A rider he once carried had a low rating. During the trip, he was curious as to

why he had a low rating despite the excellent behaviour of the rider. The rider explained

that different people, such as friends and family, utilise the account and possibly

misbehave on a trip which attracted low ratings.

This observation implies that a good or bad rider rating on the platform may not be the

same in reality. Since drivers, in this case, act as the intermediary, they can invalidate the

judgement on the platform. However, this can be restricted if platforms adequately assess

riders' information with similar accountability metrics as drivers. In addition, because Uber and

Bolt have taken advantage of the unemployment situation in Nigeria, it is difficult for drivers

to reject trips from riders with low ratings because of the impact on their performance metrics

and the uncertainty of whether they will receive another trip from algorithms. The interviewee

Jonathan (November 2018), who has been an unemployed engineer driving for both Uber and

Bolt, said:

This unemployment thing. You have been waiting for request, and you see the

rider's rating as 4.2, you ask him where he is going, he says he is going from Ikotun to

Lekki, see trip. Will I look at the rating? I will go and collect my money. I feel that if

217

there is some kind of regulation… or check and probing properly as per proper

sanction. It will not be easy because the app companies would hide the info. But if

something can be done, riders too will sit up.

On platforms, rejecting a trip would affect the driver's activity score on the Bolt

platform or the acceptance ratings on the Uber platform, which is why drivers are quick to

accept a trip before any dialogue. This phenomenon indicates that the implied freedom of

choice or flexibility portrayed by platform companies only exists before a driver accepts a trip.

The interviewee Thomas (August 2019), an Uber and Bolt driver who started driving in 2018,

also highlights that:

Uber will not allow you to pick the next trip if you do not rate the rider, but the

rider can book 20 trips without rating the driver, and it should not be so. They are

giving priority to the rider. It should not be so. (Thomas, August 2019)

This shows an imbalance in the informational access between drivers and riders. The

algorithmic burden based on all three elements exposes drivers to contextual risks that are

unexplainable by mere data and detrimental to their lives. This occurrence reduces the power

and control of drivers, which is an irony when promoting flexibility and autonomy and being

your own boss. The burden to decipher deviant riders’ behaviours before a trip is shifted onto

drivers who lack the power to change algorithmic decisions without being negatively affected.

The following section further examines further impacts of algorithmic management due to

algorithmic burdens, demonstrating further the mismatched realities of data and the everyday

experiences of drivers.

5.3.3. Arbitrary Discipline and Punishment: Unclear Terms of

Platform Deactivations

The realities of algorithms result from data that captures drivers' behaviours, which is

more sophisticated than what Haggerty and Ericson (2000) regard as data doubles because

algorithms possess the power to analyse and present records that inform their decisions in real-

time. Algorithmic assemblages, which are based on data from driver behaviours, ratings by

passengers, activities directly by the algorithms, and external modalities extended to vehicle

218

owners and law enforcement agents, create informational flows that facilitate control of drivers

(Haggerty and Ericson, 2000; Newlands, 2020). This also indicates that platforms are always

aware of the everyday occurrences of drivers through the algorithmic management assemblage

and external modalities. In Lagos, several scenarios were examined around the

interconnectedness of different aspects of the assemblage designed to get information about

the drivers’ working patterns.

Kellogg et al. (2020) argue that the discipline of gig workers can be likened to technical

and bureaucratic control used by employers in the 20

century. For technical control, workers

who do not follow their employers' directives are replaced by the pool of secondary workers

who are ready to take over. With bureaucratic control, workers with good working behaviours

are rewarded with promotions, more tasks, higher pay, and other incentives. In contrast,

workers with bad working behaviours are penalised or fired for not obeying rules and policies

(e.g., Mcloughlin et al., 2005). However, with algorithmic management, workers are

automatically removed (permanently or temporarily) from the platform based on low

evaluation scores, conflict with a rider (whether at fault or not), and often due to unclear terms

of service. In online gig work while workers are not deactivated from the platform per se, the

algorithm filters work away from gig workers whose scores fall below a certain threshold such

that they must work hard to gain a new client (Wood et al., 2019). In Lagos, the analysis is

based broadly on three scenarios analysed here, including driver deactivations without warning

or clarity; deactivations based on low ratings or rider conflicts; and discrepancies in the app

payment infrastructures.

219

Figure 21: Screenshot Example of an Arbitrary Deactivation on Bolt

Source: Author's fieldwork (September 2019)

As highlighted previously, the threshold for sanctioning or blocking drivers on

platforms is opaque and does not reflect what leads to app restrictions. Some of the ratings may

reflect a conflict with a rider, wherein drivers in this study argued that riders were often at fault

(see Figure 21). Riders aware of the impact of low ratings for drivers often threaten to report

them to platform companies or threaten to give them 1-star ratings. Nevertheless, drivers Alex

and Gabriel from the FGD meeting (October 2018) highlighted how the Bolt platform coerces

drivers to rate riders highly.

I give riders 5-star, except the rider gives problems. It does not really matter

for me. On Taxify (Bolt), for instance, once you end trip, it is automatically on 5-star;

so sometimes, I could mistakenly rate riders 5-star by swiping, unlike Uber that you

have a choice to make. (FGD, Gabriel, October 2018)

According to Chan (2019), in normal circumstances, riders perceive 3/5-star ratings or

4/5star ratings as good ratings for drivers. However, receiving over 100 four-star rated trips

could indicate a temporary deactivation from the platform. According to Rosenblat and Stark

(2016), drivers were unsure about how low ratings are defined despite exhibiting good

220

behaviours to riders; it is difficult to challenge the reality constructed by algorithms. The

interviewee Okoro (August 2019) narrated a story between himself and a rider which led to a

temporary deactivation:

He politely explained to a woman about ending a trip because the destination

was awkward, and it was late, and he refused to collect any fare. She insulted and

reported him (also cited in Arubayi, 2021). In his words:

The next day, they just blocked me due to low ratings. Most times, they will not

tell you why they are blocking you; they just said low rating and blocked me for 24

hours. (Okoro, August 2019)

Okoro’s experience indicates how drivers lack the freedom and control to challenge

decisions affecting the labour process. This is unfair for drivers because algorithms cannot

determine the actual realities in conflicts between drivers and riders. From the Bolt experience,

a compulsory training exercise is required after the first deactivation occurrence of 24 hours.

The second and third time the ban is for two weeks and ten years, respectively, because

according to the Bolt instructor, such drivers cost the company resources due to poor customer

experiences for riders (Bolt Drivers Training, November 2018).

While this study has highlighted how trip cancellations and acceptance rates affect the

ability to attract more trips and bonuses, it also observed on online Facebook groups that high

cancellation rates lead to a temporary and subsequently permanent deactivation from the app.

Other examples I found were the app blocking drivers for cancelling trips cumulatively even

when it was the rider's fault; if drivers reject credit card payments, amongst other practices

observed on SMCN. Figure 22 shows drivers getting blocked for high cancellation rates. It

does not clearly explain what led to those cancellations, leaving drivers to question why riders’

decisions should be imposed on them. This is a clear algorithmic burden that limits the

information for drivers to comprehend the reasons behind the deactivation. According to Akin

and Junior's interviews (November 2018), cancellation or not accepting a trip could be due to

a dangerous location, inconvenient travel distance, traffic congestion, riders' bad behaviour,

and other limited experiences beyond the scope of algorithmic management.

221

Figure 22: Example of Blocked Drivers with High Cancellation Rates on Bolt and Uber

Source: Private Drivers' Facebook Group (September 2019)

There are circumstances when the app underpays or overpays drivers due to technical

difficulties such as a mobile network problem or a low phone battery. With Bolt, if a driver is

on a trip, and the phone goes off, or there is a network problem, the fare becomes inaccurate

and often undercharges drivers (interviewee Akin, November 2018).

The way they block drivers is something else, although some of our drivers are

also funny too. They should look at the history of the driver. Possibly, the driver might

not be the issue, but you can call the driver and talk to the person. When you talk to the

person, you will know if he or she is telling the truth or not. Sometimes, GPS might

overshoot or undershoot. I had a situation when GPS underpaid me and also a situation

when it overpaid me. (Interviewee Akin, November 2018)

When overpayment or underpayment occurs, the Bolt app has a feature known as the

'fare review', which reviews trip fares and attempts to rectify the situation. If it is a cash trip,

compensation takes a long time for drivers to receive because the Bolt platform would urge

222

drivers to collect the money from riders, leading to further conflicts. According to the

interviewee Samuel (Aug 2019), a driver can drive for several kilometres, and the app may

record a meagre sum of N500 (£1), citing network or device problems (see Figure 23). Some

insensitive riders would pay that amount indicating to the driver that it is a bonus for them. If

it is a card trip, a fare review is implemented through another trip, but it takes a long time,

sometimes over two weeks, to rectify the problem. However, if it is an overpayment by the

Bolt app, the company will chase drivers to refund excess amounts or risk the algorithm

limiting access to trips on the app. Drivers complicate this phenomenon when they perform

several cash trips, especially on the Bolt platform. If you are a driver who makes too many

cash trips and owes the platform up to N30,000 (£61), your trips would be restricted even with

a high activity score until the money is paid back (FGD – Charles 2018).

Figure 23: Screenshot Example of Fare Underpayment for Bolt and Uber

Source: Platform Drivers Forum (November 2019)

On the other hand, drivers acknowledge that Uber responds better to drivers and

compensates adequately for lost kilometres in cases of underpayment (see Figure 26).

[

]

least, if an issue is unresolved, Uber will call drivers from the San Francisco headquarters or

regional headquarters in South Africa and give directives on the issue (Interviewee Abiodun,

Sept. 2019; FGD, November 2018). This raises a question about the lack of context specificity:

223

how would an Uber worker in San Francisco or South Africa understand the realities of drivers

in Lagos based on algorithmic data that could be inaccurate? As interviewees Dipo (January

2019) and George (September 2019) argue, the constitution and jurisdiction in places like the

US vary from Nigeria. However, interviewee Abiodun (Sept 2019), who started driving for

Uber in 2016, perceives that Uber is better at resolving such issues; if there is an unfortunate

network issue or a phone goes off, the fare remains the same because it is monitored via the

satellite. Bolt is managed locally, making these inaccuracies very frequent and difficult to

manage (Interviewees Abiodun, Sept 2019; George, Sept. 2019).

With these biases in favour of riders, drivers require a clear communication channel

and seamless appeal process when algorithms misinterpret their everyday realities. However,

communication with platform companies is only possible through software application emails

or automated calls (Rosenblat and Stark, 2016; Rosenblat, 2018) and, more recently, under-

publicised support centres. This support centre for Uber is called the greenlight hub, located in

Lagos's Maryland mall on the Mainland. The communication process remains complex on both

platforms. To make a complaint on Uber after a trip, drivers would have to target a subsequent

trip to report the issue because of the absence of an email address or contact number outside

working periods (interviewee Junior, November 2018). The platform makes it difficult for

drivers to defend themselves against opaque decisions or high algorithmic burdens. According

to Akin (November 2018), there is support for the perspective of Lee et al. (2015) in calling

for a complementary human manager to reduce the errors experienced by drivers due to

algorithms. Some scenarios that have led to deactivations could be avoided if algorithms

demonstrate humanness or if complementary human managers could conduct critical

investigations before blocking a driver. If the rules are flexible, drivers can demand an

explanation following an unfair restriction or deactivation from the app. Drivers should not be

assigned the burden of deciphering unclear deactivations, trip restrictions or poor

communications based on the judgement of the algorithm or ratings and narratives from drivers

(Arubayi, 2021b). It can be a complex process mediating between right and wrong in a conflict,

but with a human intermediary, the judgement process is fairer and more lenient for drivers.

Overall, it is essential to note that realities from data created by algorithms often vary

from drivers’ realities, i.e., algorithms misjudge drivers’ everyday experiences leading to a

mismatch in understanding how and why certain decisions are instituted. These create

algorithmic burdens that transfer the onus of deciphering decisions from platforms. It also

facilitates rider bias, such that the platform only recognises the narratives from riders as

absolute truth. While there are no absolute truths in reality; when drivers visit platform offices

224

to resolve burdensome issues, the god’s-eye attempts at resolution are based on asymmetrical

data from algorithms and riders (Rosenblat et al., 2017). In several cases, the platform is forced

to adhere to the deactivation timeline (if available), or the driver permanently remains blocked

on the app. Effective communication between platform companies and drivers should

complement algorithmic management to reduce the burdensome labour, which is impactful to

their lives and the entire labour process.

5.3.4. Motivation Paradox: Manipulating Drivers by Incentivising

Ride-hailing Gig Work

Since the introduction of Taylor's scientific management, the management of workers

has been concerned with achieving maximum value from labour by developing new ways of

motivating, controlling, and supervising labour (Woodcock and Johnson, 2018). For platforms

in Lagos and globally, the need to incentivise labour through bonuses and promotional trips is

an avenue to ensure drivers are compensated for driving more hours while extracting maximum

productivity. Woodcock and Johnson (2018, p. 547) term this phenomenon as gamification-

from-above, which "entails an elite (whether academic or managerial) that decides to impose

game elements into the lives of other people, purporting to improve their experiences without

genuine engagement or dialogue". This understanding applies here because platforms integrate

gamic elements such as bonuses and rewards to make the labour experience engaging for

workers and, in turn, beneficial for platforms. Here, algorithms transfer the onus on drivers via

in-app messages or what can be classified as digital nudges (Gregor and Lee-Archer, 2016;

Schieber, 2017; Gino, 2017; Birhane, 2020) – facilitating them to decide if chasing an incentive

is worth it regardless of the number of hours and the potential risk that may be involved in the

process. Digital nudges are enablers based on behavioural science, behavioural economics and

political theory using information technology and predictive analytics to achieve a social policy

outcome or behavioural modification to suit commercial interests (Gregor and Lee-Archer,

2016; Birhane, 2020). Like other global cities, incentives in Lagos are bonuses or promotional

trips or alternative rewards such as medical and maintenance benefits and surge fares.

The observation of how these algorithms nudge drivers is prevalent in places like the

US, according to Schieber (2017) – highlighting how platforms utilise a similar algorithm to

the Netflix feature, which nudges people to continue binge-watching following the automatic

display of new programmes. The algorithm in both Uber and Bolt sends drivers their next fare

225

before completing a trip, encouraging drivers not to log off because they are close to their

target. Schieber (2017) argues that Uber employs several social and data scientists to gamify

the driving experience with video game experiments on graphics and noncash rewards with a

low value which nudges drivers to work long hours and sometimes in less lucrative locations.

In San Francisco, some drivers express that the bonus system helps with about 25 – 30% of

their weekly earnings, but it often leads to long working hours even when drivers do not meet

trip bonuses (Hook, 2017). For many drivers, this includes accepting every trip at odd hours

and odd locations, including potentially dangerous ones such as drunk riders (Mason, 2018).

There are few or no studies to show the perception and experience of drivers concerning

incentives on ride-hailing platforms. However, all the indications from Lagos show that

platform companies use the bonus system as a hidden exploitative medium to encourage drivers

to keep working, with opacity in terms of how compensations are constructed. Platforms

intentionally keep the incentivising process a secret to prevent ratings inflation and

manipulations from workers (Kellogg et al., 2020).

Figure 24: Screenshot of a Bonus Trip on Bolt

Source: Author's fieldwork – Platform driver's screenshot (2019)

226

Incentives or bonuses for drivers are competitive tools between platforms like Uber and

Bolt in Lagos. As mentioned previously, when Bolt came to Lagos in 2016, its 15%

commission rate was not the only tool used to lure drivers; it also rewarded drivers with several

bonuses that could increase earnings. In response to the Bolt platform strategy, Uber introduced

a bonus scheme where drivers complete at least 15 trips at the standard 25% commission rate,

after which the rate drops to 3% for the rest of the week. The driver Samuel, who became a

fleet owner, i.e., started managing drivers for vehicle owners, explained how he took advantage

of bonuses:

There are times I skip church, and I had to work daily. There are times I

delivered double payments in a week. About N50,000 (£104.6) in a week. I was able to

make about N78,000 in a week on Uber. Uber was doing this 15 trips promo, and I

planned to finish those trips between Monday and Tuesday afternoon, then the rest of

the week is 3% commission. I work hard from Wednesday to Friday to get the owner’s

money. Then on Saturday and Sunday, I work harder to get my money, and double pay

him. I have made 14 trips in a day before, and I got home at 1 am. (Interviewee Samuel,

August 2019)

While these bonuses can benefit drivers, the process of incentivisation in Lagos often

comes with hidden terms and conditions, such as the use of debit or credit card details, ratings,

percentages of cancellations and rates of completion (see Figure 24). Despite the intention of

these incentives to motivate drivers, the opacity in algorithmic management becomes a burden

to drivers. These companies regularly send drivers in-app messages or digital nudges to

complete several trips to qualify for a bonus or for them to keep driving. For drivers like Obus

(September 2019), such in-app messages motivate him to keep driving to complete bonuses,

especially on weekends. The reality for 15 out of 25 drivers in this study was that the bonus

incentives are not transparent, especially on the Bolt platform, with drivers like Temi (August

2019) and Jonathan (August 2018) calling Bolt's bonuses a scam. When asked how drivers

respond to in-app bonuses, interviewee Jacob, an MSc holder driving for Bolt since 2017,

stated that:

Sometimes, when you are close to getting a bonus, they would flash you a

request, and the request will cancel, and they would drop your acceptance percentage.

They are not sincere with the drivers… The bonus is stressful. The time frame, for

227

instance, they will tell you to carry eight people at an acceptance rate of 95% and in

60 minutes, at the end of the day, when you are close to it, the time will elapse.

(Interviewee Jacob, August 2019)

Objectively, platforms might argue against this being the case. However, drivers'

experiences indicate their awareness of this opacity and information asymmetry limits their

access to incentives. Another driver, Akin (November 2018), highlights that Uber is more

transparent in terms of bonuses. Speaking about the lack of transparency on Bolt, Akin argued

that it is unfair if a driver completes at least 70% of bonus trips but receives no compensation

despite the effort invested in several hours of driving. Jonathan (August 2019) is aware that

incentives are designed not to work for all drivers. The interviewee Koffi (September 2019)

understands explicitly the intricacies of algorithms incentivising drivers, which prevents him

from chasing bonuses:

...For bonuses, Bolt is more stringent. If Bolt asks you to take 25 trips and you

take 24 trips and cancel the 25

for some reason, forget about your bonus for that day.

It would not only affect your activity score, but before you get that 25% back, you would

need to take at least 25 trips, that is, 1% for each trip. (Koffi, September 2019)

Koffi is part of the few drivers who critically understand the hidden nature of these

bonuses, and this is based on innovative awareness and driving experience. It is important to

note that driving for a longer period does not equate to more experience, particularly with the

information shared in SMCN discussed in the subsequent chapter.

Uber and Bolt utilise alternative rewards to coerce drivers to work even more hours by

integrating them into medical benefits, maintenance costs and driver of the month awards. In

the Bolt training session (November 2018), the instructor hinted that drivers who make over

200 rides in a month qualify for a medical check-up, including family members, in the

following month. Analysing this phenomenon indicates that drivers need to take at least 6.7

trips a day for 30 days, which would vary according to the number of days worked per week.

This, therefore, indicates that drivers who make 198 or 199 trips in a month will not qualify for

these benefits, but the platform would benefit from these drivers' labour, emphasising Akin's

point above about the unfairness of labour and lack of compensation. Another example is Uber,

which partnered with Germaine Autos (an automobile tune-up shop in Lagos), where drivers

would only be eligible for a N3,000 (£6) discount for maintenance with at least 50 trips weekly

228

(Uber.com 2017).

Also, drivers qualify for discounted vehicle servicing for every 5,000 km

travelled. This further enables several drivers to work long hours to gain such rewards. There

is no guarantee that a driver on 190 trips, for instance, will get these bonuses. There is also no

clarity about how many drivers can acquire these bonuses per month. Therefore, it is safe to

argue that they are developed for only a few drivers to benefit from these gamification and

motivation schemes.

The digital nudges cut across all these incentivisation schemes, including surge pricing,

the final type of incentive discussed. These in-app messages nudge drivers to keep driving to

reach surging fares in areas lacking ride-hailing platform vehicles. All the drivers in this study

admit to having chased surge pricing at some point, particularly new entrants like Akin. Drivers

from two FGD sessions (October 2018; November 2018) highlighted that surges are most

beneficial during social events, festive seasons like December and fuel scarcity periods with

fares exponentially increasing on both Uber and Bolt platforms. In other words, surge fares

have been beneficial to all the drivers during their driving experience in Lagos, especially

drivers like Obus (September 2019) and Michael (October 2018), who supplement Uber and

Bolt as a side hustle or part-time job. Scholars such as Hall et al. (2015) analyse the

effectiveness and consistency of drivers' responses to surge prices when demand exceeds

supply and vice versa on New Year's Eve in cities in the US. They highlight that efficiency

gains occur when there is a supply of drivers and riders who value rides the most. Rosenblat

and Stark (2016) argue that chasing the surge is an inefficient strategy because of how the

rider's geolocation determines the pricing of a journey.

229

Figure 25: Screenshots showing Surge Trips for Uber and Bolt in Lagos

Source: Author’s fieldwork, October 2018

Figure 25 shows surging areas throughout Lagos on both Uber and Bolt – indicating a

low driver supply. For instance, why would algorithms advertise surging areas to drivers

outside a 2km radius with in-app messages on the Bolt platform? This is a gamic strategy for

drivers to rush to surge locations searching for higher fares, highlighting their lack of

autonomy. In reality, though, it entraps drivers to stay active on the platform. On the other

hand, Uber does not clearly define its kilometre radius for drivers. The algorithms coerce

drivers with exponential rates, leaving the decision-making process to drivers. Drivers that

decide to chase the surge become subjected to the rules of the algorithms, which subsequently

impact their acceptance and cancellation rates at the expense of unforeseen obstacles such as

traffic congestion, road closures and accidents (FGD, November 2018). The interviewees Noah

and Alex (FGD, November 2018) again highlight that the surge fares are more transparent on

Uber than on Bolt because Uber gives drivers a vague idea of when a surge will end. With Bolt,

surging areas could disappear quickly, making drivers identify them as a fake surge. However,

the consciousness to withstand algorithmic gamic elements and nudges in terms of chasing

surges comes with the experience of working for these platforms, which is discussed in more

detail in the next chapter.

230

Incentives and surge fares are a paradox of motivation by algorithms based on

gamification ideologies to keep drivers working at the expense of their health and wellbeing.

Birhane (2020) argues against transmitting the decision-making for social outcomes to

algorithms developed by profit-driven corporations like Uber because it would rewrite morals

and the labour code of conduct by integrating corporate incentives driven by selfish interests.

If algorithms incentivise ride-hailing drivers, the process should be more transparent and less

exploitative, reducing the burdens on drivers and enabling decisions that improve monetary

gains on the job.

5.3.5. Digital Map Limitations for Driving in Lagos

A key point that cuts across algorithmic burdens, as mentioned previously, is the

mismatch of realities between data, the everyday experiences of drivers, the city and contextual

culture specificity, which impacts the productivity of gig work. When discussing this with

platform drivers, they noted that the maps help locate trip destinations, especially on the Island,

but not so much on the Mainland of Lagos because of the complex nature of routes and real-

time map representations (Interviewee Henry, July 2019). Drivers like Henry prefer to work

on the Island because of accurate map representations, which are properly networked based on

its affluence compared to the Mainland of Lagos. Xiao (2019) builds on the concept of enclaves

(fixed and bounded sites) and armatures (infrastructure channels and transit spaces) according

to Shane (2005), proposing two concepts in the study of everyday mobility in the Lagos

landscape, known as routes and sites. Accordingly, routes are road networks that cut across

different city corridors, while sites are places attached to routes or along routes, such as

buildings, informal and formal bus stops, and open spaces along the roadside. As Xiao (2019)

argues, these routes and sites transcend Shane's conceptualisation of bounded sites and transit

spaces. Instead, they are constantly evolving with the dynamics of developments around the

city and the mobilities of urban residents, which are fluid and effective in reproducing bounded

sites and spaces (Jensen, 2009). Again, this logic emphasises the mismatch of the realities of

drivers and algorithms during and after trip assignments. There are limits to the extent to which

algorithms interpret wrong locations or fake addresses, bad roads, single carriage roads (or

what drivers know as one-way), and traffic congestion on the measurement of performance and

productivity, which is a burden for drivers that can impact their evaluation metrics and lead to

conflicts with passengers.

The map feature is an integral facet in niche platforms like Uber because trip allocations

and real-time pairings thrive on the availability and functionality of the GPS signal. Digital

231

maps would function optimally in a well-built city like London because of a well-organised

environment with a proper address system. Poorly built cities like Lagos, with consistent urban

sprawl, make it difficult for digital maps to cope with everyday infrastructural complexities.

This phenomenon is inherent in the city's materiality; the poor interlinking of roads and

symbols constantly evolves without proper integration within the digital map, which are

infrastructural defects that impact route and site decisions in Lagos (Xiao, 2019). Thus, it is

difficult to determine how to navigate familiar routes affected by road construction, particularly

for drivers unfamiliar with a new route. When this happens, it affects the driver's arrival time

or departure time, which creates a false impression for riders and could affect the ratings of

these drivers. For example, platform driver Henry (October 2018) does not trust Google maps

because it can be misleading on the Mainland in Lagos, particularly when it does not indicate

when a route is a single-carriage way or 'one-way'. In some instances, this could lead to a road

traffic offence with Lagos Metropolitan Area Transport Authority (LASTMA) officials, which

could affect drivers' earnings if they have to pay traffic offence fees that could be illegally

inflated, or the vehicles of these drivers could be impounded. Henry prefers driving on the

Island and is more likely to reject trips to the Mainland of Lagos, especially at night because,

even when driving to an unknown destination, it is easier to navigate with the help of familiar

landmarks beyond the assistance of digital maps. How can algorithms then penalise drivers due

to circumstances beyond their control? Another driver, Jonathan, explains how the landscape

and design can differ from the map interpretation in the app.

With the map, there might be a road closure, for instance. In Lekki, there are so

many housing estates, and you cannot cross these estates like that. It is a big issue for

us. The map leads us there because it knows there is a road there. However, a gate

might just have been installed, which could affect your arrival time or pick-up time.

The passenger may wonder if the driver wants to increase the fare. (Interviewee

Jonathan, August 2019)

The flawed address system compounds this revelation in Lagos, which often affects the

pick-up times of riders. In Nigeria, describing a location often involves referring to a nearby

landmark, making it difficult for maps to determine exact locations. In many instances, the map

could misdirect drivers via a longer route, affecting the pick-up or drop-off times and leading

to potential conflicts with riders. This indicates that riders can be another source of map

misinterpretation for drivers. Riders input a false origin when requesting a trip via the Uber or

232

Bolt app. This act affects the driver's arrival time because the rider is not at the actual origin,

and the waiting time of 5 minutes would not be activated until the driver has arrived.

Alex narrated a story of a woman who switched the location within the 1004

estate in Victoria Island, and it made him roam around the housing estate until he

cancelled the trip after 15 minutes (FGD, November. 2018)

This experience was on Bolt, and at the time, there was no cancellation compensation

policy like the N400 (£0.8) on Uber. Potential riders indulge in this act of time-wasting to avoid

the cancellation fine by platforms. The Bolt platform did not charge riders for time-wasting

during the field study.

Figure 26: Traffic Congestion on the Muritala Mohammed Airport Road in Lagos

Source: Author’s fieldwork (July 2019)

The presence of contextual realities such as traffic congestion, bad roads, and single

carriage roads contributes to drivers’ urban experience, which impacts performance evaluation

by the platform algorithms (see Figure 26). In 2018, the National Bureau of Statistics (NBS)

recorded about 5 million vehicles and over 200 thousand commercial vehicles on the road per

day (Oshodi, 2015). Compared to the national average of 11 vehicles per kilometre, Lagos

233

averages over 227 vehicles per km (Zaccheus, 2017).

Lagos has 3,067 persons per kilometre

with a total road length of 7,598 km, compared to Tokyo, another megacity with about 544

persons per km with a total road length of 24,431 kilometres (Oshodi, 2015). This phenomenon

indicates that the road network density in comparison to the size of the population is low. All

the drivers in this study complained about how traffic congestion affects decision-making

during the labour process. At the time of my data collection, Uber charged N11 per minute and

N60.01 per km; Bolt charged N10 per minute and N60 per km, indicating that long hours spent

in one location due to traffic do not increase earnings on the platform.

… the problem we have is traffic. For instance, the trip, you are meant to spend

an hour and make about N3,000 (£6) and go on another trip. A driver would spend five

hours on that trip. There is no way to compensate for all that time because it means

riders would pay more. Because of bad roads and traffic, many drivers sometimes

cancel trips. (Interviewee Junior, November 2018)

Particularly on the Bolt platform, interviewee Samuel (August 2019) explains that it is

worse because often, the app does not start until at least 10 minutes into the trip compared to

Uber. Therefore, a 15-minute trip can be more profitable than being stuck in traffic for 45

minutes and over. Also, an HND graduate from statistics and former Bolt driver Temi (August

2019) highlighted how it took 45 minutes to make a U-turn to a rider's origin, which usually

takes just over 5 minutes without congestion. Again, trip cancellation would affect the ability

to receive jobs from the algorithms because of the low percentage of trip acceptances.

Platforms are aware of some of these problems raised by drivers. However, these

challenges that impact a driver's performance metrics should be integrated into the data used

in managing drivers to reduce the rigidity and lack of emotions from algorithms. If routes and

sites are consistently evolving due to infrastructural developments and the everyday mobilities

of urban residents, algorithms should be programmed to interpret such actions based on a

driver's reproduction of space in Lagos.

5.4. Algorithmic Management and the Burdens of Labour.

The section above has shown how drivers are impacted by a lack of information based

on the opacities, information asymmetries and rider decision biases that are unequally balanced

234

towards drivers (see table 21 below). Through the surveillant assemblage, platforms possess

the power to acquire information about drivers, which is integral for performance evaluations

from riders and other outsourced forms of surveillance and external modalities that control

their labour. However, drivers possess little information about core aspects of the labour

process, which puts them at a disadvantage. For example, these risks that drivers encounter are

inherent in experiencing a robbery attack in a rider’s destination because of the opacity

embedded in the app; as a result, withholding the information before the driver accepts a trip.

Secondly, riders with fake registration identities are allowed access to the platform but possess

the power to rate a driver, which in scenarios of conflict can increasingly and instantly impact

their ratings depending on why the driver was reported.

In GN cities, these burdens are less impactful because of robust existing regulatory

frameworks, innovative environment apparatuses (e.g., CCTV cameras), and an enabling

environment (Kwet, 2019; Birhane, 2020). For example, in the previous chapter, the failures

of Easy taxi and Afrocabs indicated the lack of an enabling environment compared to

California and the lack of innovative awareness and education among drivers. Although the

burdens were less compared to now, traditional taxi drivers complained about the difficulties

of adopting taximeters which prevented further penetration. In London, these were instituted

policies and rules that boosted drivers' awareness and education by learning how to use paper

maps, and digital maps and passing the knowledge test (Dobraszczyk, 2008; Chen, 2013;

Dudley et al., 2017). These attributes of a progressive society propagated the adoption of ride-

hailing platforms in GN cities. This does not imply that drivers in the North did not experience

burdensome labour processes, but drivers in the GS experienced worse with fewer capabilities

to withstand the impacts because of the lack of contextual integration, further undermining the

notions of autonomy and flexibility of labour which these platforms propagate.

235

Table 21: Summarising the Algorithmic Burdens and Impacts in Lagos

Burdens

Definition

Impacts

Opacities

Drivers not knowing how certain

decisions are made and how they may

impact them based on limited

information from the platform.

- It can lead to drivers overworking or not working because of a lack of clarity in the

algorithm. For example, manipulative incentives include irregularities in the price

mechanism and commission calculation.

- Manipulative incentives nudge drivers to keep working, which can lead to fatigue,

loss of income, accidents and an overall impact on the health and wellbeing of

drivers. It can also impact your activity score or other metrics if drivers cancel a false

promotional trip.

Information

asymmetries

This is also entwined with the opacity of

the algorithm because drivers possess

less information than riders concerning

certain decisions affecting their work.

- Based on the opacities and informational asymmetries drivers experience, deviant

riders who may have been deactivated can re-enter the platform to harm drivers.

- Also entwined with opacities is drivers not knowing the destination of drivers on

accepting a trip, which can lead them to encounter deviant riders or dangerous

locations, which can impact the evaluation metrics or lead to assault or death.

- With less information about riders, there are often inaccuracies with the accuracy of

the map reinforced by a bad address system etc. This can impact the driver's

evaluation metrics if a rider cancels or reports the situation. Cumulatively can lead

to temporary or permanent deactivation.

236

Bias

This occurs when the algorithm favours

riders in critical situations without proper

investigation. I recognise this as rider

decision bias.

- Drivers experience deactivations or impacts on their evaluation metrics, often due to

rider decision biases, i.e., without due process or investigation on the platform. Due

to opacities, there is no straightforward process for drivers to appeal the decision by

the platform. It ultimately affects their livelihoods because they are unable to work

Source: Designed by Author, 2019

237

These burdens of labour deny drivers the so-called flexibility and autonomy that

platforms advertise as core aspects of being an independent contractor. Instead, autonomy and

flexibility are paradoxical because of the lack of information to determine when and how

decisions are constructed and how these affect the labour process. According to Mazmanian et

al. (2013), the autonomy paradox from observations in the workplace showed that smartphone

devices that help professionals with seamless communications and improved team

management also increase their commitment levels and make it difficult to disconnect from

work. In a ride-hailing workplace, even the idea that ride-hailing gig work is flexible and

autonomous is misleading because the embedded surveillant assemblage ensures that drivers

are always working, and the burdens of labour limit information on the decision-making

process, which further prevents knowledge on how to request better working conditions.

Algorithms prioritise rules related only to platforms, downplaying the contextual

realities such as riders’ behaviours and the nature of the city (e.g., social vices, traffic

congestion, bad roads, gated communities, and others). This phenomenon often makes it

difficult for drivers to determine their working time because they have no control over trip

assignments, highlighting similar findings to Wood et al. (2019), who observed how freelance

platforms create a flexibility myth in online gig work. In summary, the impacts discussed in

the previous section hinge on algorithmic burdens of labour which is a conflation of opacities,

information asymmetries and rider biases. These manifest from the mismatch between the data

algorithms process and the contextual realities of the city, including the culture of driving and

the everyday experiences of drivers.

5.5. Conclusion

The implications of informal labour contracts and misclassification of drivers as

independent contractors have transferred burdens of risk and unfair decision-making to drivers.

This contributes to the ongoing discussion on algorithmic management and resistance from

scholars across different sectors (cf. Rosenblat and Stark, 2016; Möhlmann and Zalmanson,

2017; Cheng and Foley, 2019). However, this varies by first identifying that algorithmic

burdens are the underlying factors facilitating impacts, and these burdens vary across GS and

North contexts. Weapons of the dominant through algorithmic management facilitate the

continuous surveillance of drivers, which is critical in establishing algorithmic burdens based

on information asymmetries, opacities and biases (Haggerty and Ericson, 2000; Newlands,

238

2020). As such, while the surveillance capabilities of the platform should reduce algorithmic

burdens on drivers, they limit and withdraw the information in conflicting circumstances, thus

pitting drivers against data that are not reflective of contextual realities and the experiences of

driving.

Practically, this one-sided management does not adequately integrate the realities and

uncertainties of ride-hailing labour embedded in the different contextual realities, creating

more burdens for drivers. This is compounded by information asymmetries which create unfair

judgement in the face of conflicts or problems which impact drivers’ performance metrics and,

in some cases, lead to arbitrary discipline and punishment (Rosenblat and Stark, 2016;

Foucault, 2008). Drivers are blocked from the app without a fair process of appeal or

contestation when, in reality, some decisions by algorithms do not integrate the unpredictability

of ride-hailing gig work in a city like Lagos. Even if algorithms integrate the unpredictability

in the field, it is unclear why the data is opaque, and drivers still suffer the impacts of biased

decision-making. It is revolutionary that platforms gamify what is traditionally known as taxi

driving (Woodcock and Johnson, 2018). However, they have done so without properly

integrating the nuances embedded in different contexts, which contributes to redefining the

everyday experiences of drivers by simply assigning the process to algorithms.

The consequences of bad algorithmic management have created what Mazmanian et al.

(2013) and Wood et al. (2019) identify as an autonomy-paradox and flexibility myth because

ride-hailing gig work is micro-managed, particularly by algorithms ensuring that drivers follow

the rules of platforms. The autonomy-paradox contributes to workers' exploitation and self-

exploitation, which was not prevalent in the taxi regime because drivers were not operating in

a gamified environment. Creating a utopian experience from the marketisation of being your

own boss is an entrapment for drivers to acknowledge that freedom and flexibility are the new

normal. However, the findings in Lagos show that the hidden power of algorithmic

management, which is reinforced by the surveillant assemblage and external modalities,

ensures that drivers are always subject to riders and algorithms, which are burdensome to the

labour process based on information asymmetries, opacities, and rider biases. If drivers escape

digital nudges to continue driving, they face rider ratings, trip cancellations or acceptance and

vehicle owners’ conditions.

If ride-hailing gig work was genuinely autonomous and flexible, drivers should have

the ability to challenge unfair decisions considering that they do not contribute to the

management and control of the labour process. Algorithmic burdens, therefore, establish a gap

239

between drivers and platforms in deciphering decisions that impact the labour process. In the

subsequent chapter, I unpack how drivers resist algorithmic management and control in Lagos.

240

6. Chapter Six: Everyday Resistance: Exploring the Hidden and

Public Practices of Platform Drivers in Lagos

6.1. Introduction

The hidden nature of ride-hailing platforms through algorithmic management facilitates

drivers' hidden resistances in Lagos. As a conceptual model, pairing the everyday resistance

concept by Scott (1989, 1990) with the contemporary notion of algorithmic management, this

chapter critically analyses the hidden and public resistance practices against ride-hailing

platforms in Lagos. The city of Lagos presents unique resistance practices that represent the

GS, which lack critical scholarship in terms of how gig workers subvert the power of

algorithmic management, which imposes algorithmic burdens on drivers. Platform drivers

react because most of the decisions during the labour process are beyond their control and, in

several cases, are unfair in relation to their perceptions of a suitable working environment.

Regaining control by challenging the platform business model and deciphering elements of

algorithmic burdens does not inherently reduce the power of algorithmic management. If

anything, it increases the ability of the system to adapt to resistance practices based on

surveillance capabilities and data analytics, thereby leading to a power tussle across digital or

virtual spaces.

This chapter contributes by emphasising the spatiality and temporality of workers’

resistances in the gig economy, which is critical in unpacking how and when everyday

resistances occur and evolve. Further, this leads to examining how public and hidden

resistances are interconnected based on the previous improvements of everyday resistance

concepts by Adas (1986), Adnan (2007) and Shalhoub-Kevorkian (2012) due to the presence

of algorithms. Therefore, it examines and rethinks the so-called weapons of the weak as part of

what I call weapons of the dominant because algorithmic management, which is burdensome

to the labour process, also facilitates resistance practices. In Lagos, I argue that developing

algorithmic literacies through Social Media and Communication Networks (SMCN) is central

to developing public and other hidden practices which are critical in reducing the negative

impacts of algorithmic burdens. For instance, the mobilisation of workers for public offline

and online protests is firstly coordinated through social media and communication networks

(e.g., WhatsApp, Telegram). These observations were slightly similar to the experience of

Deliveroo riders in the study of both Briziarelli (2019) and Cant (2020), where "log-in" and

241

"log-out" functions were critical for collapsing physical and virtual spaces because of the

absence of traditional workplaces or actual headquarters for confrontations.

Investigating the forms of resistance is integral to understand the complexity of the

characteristics of power relations relevant to platform drivers and ride-hailing platforms in

Lagos (Foucault, 1980;1995). These struggles exist against the privileges of knowledge and

oppose deformation, secrecy and bewildering representations imposed on people. The

following sections discuss the public and hidden practices in Lagos with contextual examples.

I conclude by presenting the preliminary contributions to knowledge by acknowledging the

improvement of the everyday resistance concept, introducing weapons of the dominant as

reinforcing weapons of the weak and considering everyday resistances as everyday digital

resistances which can occur across contexts. This sets the context for expanding on these

contributions in the subsequent chapter.

6.2. Public Resistance of Platform Drivers in Lagos

In Lagos, field observations showed an open way to engage platforms by conducting

protests discussed in more detail below. Protests in the UK, Italy, and other GN cities across

platforms such as Deliveroo, Foodora and other platforms indicate public visibility of workers’

solidarity and collective actions (cf Brizarialli, 2019; Cant, 2019; Cant and Woodcock, 2020;

Tassinari and Maccarone, 2020). In Scott's understanding, these public forms of resistance

practice are often recorded and intentional against dominant groups; in this case, platform

companies managed by invisible algorithms. In this sense, workers are challenging platform

companies to review their labour conditions surrounding pay, terms of service, representations

and, most critical to this thesis, algorithmic burdens facilitated by algorithmic management.

This first part of the chapter outlines what was considered a public and visible practice

in challenging ride-hailing platforms in Lagos compared to other cities in the GN. These public

resistance practices are relevant in understanding why workers deploy hidden practices in the

subsequent part of this chapter.

242

6.2.1. Platform Drivers’ Solidarity: Online Strikes and Offline Protests

In the resistance discourses, protests, demonstrations, public dissidence, riots and

boycotts are forms of public practices that collective groups, unions, social movements and

associations have employed in challenging dominant groups (Scott, 1985; 1989). In Lagos, one

outright public practice I identified was by protesting in the form of physical visits to Uber and

Bolt offices, displaying placards by the Uber and Bolt offices or collectively boycotting the

app by refusing to work on these platforms during certain hours or days.

As a medium to voice the opinion of drivers in Lagos, the National Union of

Professional E-hailing Driver Partners (NUPEDP), in affiliation with the Trade Union

Congress of Nigeria, was formed in Lagos in 2017. In the interview with the president and

founder of the union Dipo (January 2019), he posited that the NUPEDP is a legitimate tool to

improve the bargaining power of platform drivers throughout Nigeria.

I decided to start the union because of the way our operation is going in Nigeria.

When I first heard of Uber and invested in it, the job was good; the return of our

investment was good until Taxify (Bolt) came, and they slashed prices by 40%. It

affected our business because they did not inform us. I noticed a gap between us and

them, and what are the things that can bridge that lacuna is a pressure group, which is

how the union came. If you have a pressure group such as a union that abides by the

law of the country, it will give us the power to challenge the platform policy.

Unlike the Rideshare Drivers United (RDU) in the US, different platform unions in

Lagos are conflicting because of selfish interests and the already porous institutional

framework in Lagos. These unions share a similar objective of building an alliance that

improves their position of power. Some of these other unions include the Professional E-hailing

Driver-partners Association (PEDPA), Association of Professional and Commercial E-hailing

Drivers (APCED), Moving Train Pilots Association (MOTPA); Union Leader Council (ULT),

and Amalgamation of E-hailing Stakeholders (AES). One of the leaders of PEDPA,

interviewee Koffi, describes how all other unions emerged from the NUPEDP and ULT

because of the perception of unsolicited monthly dues and tribalism (selecting only Lagosians

and Yorubas). Moreover, the nature of platforms through internationalised product markets and

corporate ownership have diminished collective bargaining and multi-employer coordination

243

on a national level (Wright and Brown, 2013). This phenomenon reduces the effectiveness of

unions in the gig economy.

In GN contexts, protests yield more tangible results compared to cities like Lagos

because of robust institutional frameworks and labour regulations that respond to the demands

of recognised labour unions and pressure groups. For example, in cities in the US and UK,

unions such as the Rideshare Drivers United (RDU), Independent Workers Union of Great

Britain (IWGB) and other unions have challenged platform companies about the

misclassification of workers as independent workers, which prevents them from contributing

to decision-making in the labour process and achieving genuine autonomy and flexibility. In

San Francisco, California, incessant public protests and strikes have led to Assembly Bill No.5

(AB5) institutionalisation, which aims to protect the rights and livelihoods of all gig workers

in America (Paul, 2019; McNicholas and Poydock, 2019). In a nutshell, this law advocates for

the reclassification of independent contractors to employee statuses with the relevant

employment benefits for workers

. While this is specific to places like California, pressures

from workers’ solidarity in lobbying regulatory bodies are evident in other cities such as

London, where Uber, for instance, experiences partial bans because of collective action both

from platform unions and traditional taxi unions. According to Scott (1985), he recognises

behaviours similar to strikes when farm labourers, through a leader, 'let it be known (cara

sembunyi tau)' by informing the farmer that harvesting will take a longer time, thereby

indirectly requesting higher pay. The farmer may 'let it be known' that the price is fair or that

they would be willing to increase labourers’ pay. However, according to Anwar and Graham

(2019), unionised struggles are pitted against a highly mobile capital. The choice for labourers

to resist working conditions openly may improve or aggravate working conditions depending

on the platforms' response.

Strike action on platforms like Deliveroo has led to permanent deactivations of known

strikers in places like London. Cant (2020) observes how consistent strike action by Deliveroo

riders as a result of collective efforts and mobilisations through a monthly bulletin called Rebel

Roo led to positive responses for workers and unions. The IWGB and other workers demanded

an increase in the piece-rate payment to £5 per drop, no victimisation of unionised riders and

a recruitment freeze (Cant, 2020). At the time, the Deliveroo platform responded to strike

actions in Brighton and Leeds by not victimising any strikers and temporarily freezing

recruitments of workers. However, at least at the time of the data collection exercise in Lagos,

protests remain insignificant as a public resistance practice in challenging ride-hailing

platforms.

244

Similarly, unions in Lagos also believe in the strength of numbers. Dipo, George and

Koffi highlighted plans to develop large followership of between 4,000 and 15,000 drivers per

union in three separate interviews. Dipo (January 2019) highlights building a union database

that encompasses platform drivers and traditional drivers, supervised by the State Ministry of

Transport in Lagos. These unions were created to challenge platform companies by lobbying

Uber and Bolt to redefine the terms of labour, classification of workers and algorithms, which

do not represent the realities of drivers. For example, Koffi (September 2019) discussed how

recruiting drivers was integral to forming a driver-led ride-hailing platform that recognises

drivers' needs and contextual realities, including providing safety protection benefits in Lagos.

In May 2017, the first significant protest in Lagos occurred when Uber slashed drivers'

fares by 40% in response to Bolt's lower fares (Interviewee Dipo, August 2019). Due to its

market dominance, there was a similar response from Uber following Bolt’s emergence in

Kenya and South Africa (Carmody and Fortuin, 2019). In this case, algorithmic burdens are

embedded in the opacities and information asymmetries of payment and incentives and how

these do not reflect the current economic standings of the country is apparent. In Lagos, the

main complaint from platform drivers is their lack of power to contribute to critical decisions

which affect their everyday labour. However, an autonomous response from platform

companies lies in classifying drivers as independent workers or driver-partners. Figure 27

summarises the essential complaints of drivers in Lagos. Although algorithms are central to

controlling workers’ labour in Lagos, they cannot directly voice their opinions to the app. These

protests serve as a public practice to ensure that platforms improve their labour conditions and,

most importantly, enable algorithms to fairly manage the labour process. After all, algorithms

result from programming that reinforces historical labour biases and unfair treatments

(Ajunwa, 2019), which are replicated through platforms.

245

Figure 27: The First Significant Protest Against Platforms in Lagos

Source: Gbadeyanka (2017)

… we buy vehicles and register under the company. That does not make them

our boss. We have over 80% of the investment on these platforms, but they are the only

ones that decide the price mechanism; meanwhile, they do not know what drivers are

facing on the road … Why is the company treating drivers like slaves? (Interviewee

Dipo, January 2019)

While the protests were not as demonstrative as in places like South Africa, unions were

relevant to practice collective action and solidarity despite it being a disadvantage in gig work.

In 2018, at Green Point, Cape Town, in South Africa, drivers engaged in protests and

demonstrations by hooting, driving slowly and creating roadblocks and traffic congestion

(Saal, 2018). Despite reward systems in Lagos and Southern African cities to disguise social

protection schemes, platform drivers called for a more transparent and inclusive process in

making critical decisions that impact their labour.

246

Figure 28: A Union Leader Advocating for Drivers at the Bolt Head Office

Author's fieldwork (September 2019)

Figure 29: Protest Placards Against Bolt’s Commission Increment

Source: Author's fieldwork (September 2019)

Another protest I observed through discussions with participants was the global

#appsoff, boycotting the app on 8

May 2019, led by the RDU and other unions across the GN

cities (Sainato and Paul, 2019). The NUPEDP union and other worker groups, such as

interviewees in this study (Dipo, Junior, Jonathan, Samuel, Phil), participated in the global

#Appsoff protest by switching off their phones for 24 hours without engaging in any trips. The

247

level of participation and coordination compared to RDU members in the US was significantly

lower because of certain pre-existing circumstances for platform drivers in Lagos. Interviewee

Samuel (August 2019) indicated that people did not respond to the call to boycott the app,

stating that gig workers in GN cities were more responsive and organised. Also, considering

that ride-hailing platforms facilitate entrepreneurship for independent vehicle owners and

rental companies, as mentioned previously, it was difficult for drivers to switch off their apps

for 24 hours and not make money (Interviewee Jude, November 2019; Interviewee Abel,

September 2019). This is due to the surveillant modality of vehicle owners or rental companies,

which reinforces the self-surveillance of drivers, facilitating continuous working arrangements

that prevent them from strike action. Drivers who owned their vehicles possessed more

autonomy and flexibility to engage actively in protests because they were not under pressure

to pay the rental company or vehicle owner weekly remittances; they worked for themselves.

Another reason for drivers' low turnout was the preconception of violent misconduct and

corrupt practices of traditional unions such as the NURTW and RTEAN in Lagos. Drivers such

as Obus, Jonathan, Henry and Jude perceive that platform unions could morph into traditional

taxi unions by collecting levies without improving the welfare of drivers. Finally, during the

global #Appsoff protest, interviewee Koffi, a prominent member of PEDPA, admitted to not

participating in the strike.

We did not partake in that. We told them not to do it that way, but we should all

come together. Our association opens doors to other associations, even if you have

another group under you. So that when we want to do something, they would see that

we have one mind. (Interviewee, Koffi September 2019)

Koffi's union colleague, Akpos (August 2019), who drives for Bolt, Uber and Gidi cabs,

also did not participate because he believes only the government can intervene. Later in

September 2019, NUPEDP and other drivers conducted another significant protest which was

more coordinated (see Figures 28 and 29). It is important to note that my interaction with

interviewee George led to a collaboration with comrade Dipo before the protest morning.

Previously, George (September 2019) was also sceptical about the rise of platform unions,

particularly the NUPEDP union, because of one of the mission statements (which he did not

disclose) that a member posted in a private WhatsApp group chat. The focus of this protest was

on the Bolt platform, particularly because of a sudden commission increment from 15% to 20%

without consulting with platform drivers or considering the depreciation of the economy. The

248

country was still recovering from the 2016 recession, and the growth rate increased to 2.2%

before Covid-19 (World Bank, 2020). Inflation rates also increased from 11.4% in 2019 to

about 12.8%; the increase in electricity tariffs, food prices and the removal of fuel subsidies

piled pressure on workers (NBS, 2019; AFDB, 2021). Over 40% of the population (82.9

million people) were below the poverty line, while another 25% (53 million) were vulnerable

– they could fall below the poverty line if the debilitating economic situation continues (NBS,

2019). However, according to updates on the Bolt website, commission rates are calculated

according to the market.

During the protest, the NUPEDP leader was granted an audience by some of the Bolt

staff members in a small room but took the protests outside the building because Bolt staff

members granted entry to only a few drivers. According to video footage sent by interviewee

George (September 2019), drivers refused to engage in dialogue in the Bolt auditorium because

Bolt representatives chose to dialogue with few drivers. This led drivers to protest outside the

building with placards and megaphones to voice out their concerns (see Figures 28 and 29).

The regularity of protests is indicative of the disgruntled perceptions of platform drivers and

how they perceive that the power in numbers can change their precarious condition.

6.2.2. Public Dissent through Media Engagements

Mainstream media engagements are minor aspects of public resistance that enable

workers to express their rights, exposing the reality of ride-hailing labour conditions in Lagos.

In reviewing a few online resources, I uncovered popular resources such as press interviews on

YouTube, interviews with popular online news outlets like The Guardian, Vanguard, and

Quartz, and critically popular radio shows such as the 96.1fm – Traffic radio Lagos.

… We have been able to achieve many press meetings—both in the press

of Lagos and Abuja. We even have a YouTube channel. We are very transparent. This

year alone (2019), we have many programmes planned out. We can invite you to one

of our meetings. (Interviewee Dipo, January 2019)

I tuned into two radio shows on the traffic radio in January 2019. The traffic radio

(96.1fm) is a critical state media resource because it announces traffic-congested areas in real-

time throughout Lagos with the help of everyday commuters who call into the show to report

on the situation. However, union representatives such as comrades Dipo and Omolawal

249

(January 2019) had multiple interviews via these media platforms with large audiences. The

NUPEDP union was not the only union observed on the show. A representative from the

Amalgamation of E-Hailing Stakeholders (AES) was also interviewed. Observations from both

interviews centred around the labour conditions of ride-hailing gig workers, driver experiences,

challenges and a request for all drivers across the State to be united under one union. In essence,

these engagements created more insights for this research in understanding their core

challenges. These were also useful for workers in transmitting information concerning their

needs to lobby the State government to establish strict policies that regulate ride-hailing gig

work. The interviews also target existing and upcoming ride-hailing platforms with demands

to treat drivers fairly.

A central argument that summarises these media protests concerns how flexibility and

autonomy are myths in ride-hailing gig work in Lagos. Workers, through unions, request more

power within platforms, especially as partners. In one of the radio shows, Omolawal (January

2019), a NUPEDP union member, expressed:

All we do is provide the cars and do the job. There are some rules that should

involve us. Take, for instance, security. We have recorded over twenty drivers killed

over the last three years and even theft….

Similarly, two representatives of AES stated categorically that:

We are not saying increase the price; we say have a meeting with us and take

into consideration and fix the price. Do not sit in your office and do the calculation by

yourself. Do not wake up one morning and slash prices at the comfort of your office

because of competition. (Kevwe, Radio interview January 2019)

We are calling for the government to review their transportation framework and

include the E-hailing sector. To every transportation service in the industry, there is

one regulation or the other. Except us. (Jolomi, Radio interview January 2019)

These are typical quotes highlighting radio interviews in Lagos, which are interactive

because both drivers and riders call in to ask questions or proffer solutions. The interviewee

George (September 2019), admitted that he was infuriated with some of these radio interviews

that gave false information, such as drivers’ weekly remittances to vehicle owners. The

250

problem is that it gives vehicle owners and platform companies a false reality of their labour.

Hence, he advocated for such information to be withdrawn from public spaces and a clear

indication of the kinds of information that should be public and private. This public dissent is

against platform ideologies through the media, with selected worker representatives exposing

details that were not publicly known, such as the realities of their everyday experiences,

including the loss of life of workers.

It is important to note that, in Lagos, these media interviews also serve as a platform

for representatives like Kevwe and Dipo to publicise their unions for potential union members

and to grow their members by facilitating more coordinated protests and app boycotting in

Lagos and throughout Nigeria. However, key informants such as Dipo and George argue that

Uber and Bolt adopt hidden practices by using ambassadors to disrupt the formation of unions,

such as speaking against unions on live radio events. Tuning into Dipo's radio session in

January 2019, I witnessed when a driver called into the radio show to antagonise and probe

their analysis of ride-hailing gig work as inaccurate and going as far as questioning the

legitimacy of their union. From a researcher's perspective, this presented itself as union

disagreements because the caller was a representative of another union. However, in the FGD

(August 2019) with Dipo and his colleagues, it was apparent that this was beyond union

antagonism but an invisible operation by platforms. While this section analyses public

practices, it is essential to highlight how dominant platforms in Lagos facilitate hidden or

backstage performances to dismantle union formation and activism.

…We call them super ambassador and agent. These are the guys they will come

and scatter the driver-partners, and they will collect money every week, and they will

pretend that they are driving car like us. And they are suffering like us. Meanwhile, it

is a lie. Bring an idea, and they will scatter the idea. (Dipo, FGD, August 2019)

While these allegations remain difficult to validate with platform companies, further

confirmation from drivers like George, Thomas, and Koffi highlights these allegations as part

of the hidden transcripts from platforms in Lagos. While the goal of public resistance is to

ensure platforms recognise the needs of workers, Uber and Bolt, by design, do not recognise

unions and collective actions, which is stated in their terms and conditions. Considering the

lack of coordination amongst unions, drivers realise that regaining control of the platform

would require tact and hidden practices focused on evading the entrapment of algorithmic

burdens based on poor algorithmic management. Therefore, this sets the context for unpacking

251

what this research identifies as workers' hidden practices and how they evolve across space and

possess high temporality.

6.3. Hidden Practices of Everyday Resistance in Lagos: Challenging the

Weapons of the Dominant

This section focuses on the hidden practices of platform drivers in Lagos, what Anwar

and Graham (2019) classify as 'the hidden transcripts' of the gig economy because of their

anonymity and sometimes collective and cumulative acts that disrupt platform companies by

resisting the power of algorithmic management. These hidden resistances are inconsistent

because they transition through digital and physical spaces because of social media and

communication networks (SMCN). The practices further emphasise temporality because a

hidden resistance today may become public and vice versa because of algorithmic learning

capabilities that may expose them to the public until drivers discover new hidden resistances.

This spatiality and temporality typify the digital world today with continuous changes by

simply optimising data codes that reflect user experiences that can be implemented through

software updates. As mentioned in the previous chapter, gamic elements are central in

managing drivers in Lagos, which Woodcock and Johnson (2018) would classify as

gamification-from-above. Likewise, hidden resistances here take on what they classify as

'gamification-from-below' which is "a set of politics around the subversion, undermining, and

even mockery, of serious life, through its reduction to the non-instrumentality and therefore

pointless, under neoliberalism, of play" (p. 549). Accordingly, this means ride-hailing drivers

undermine algorithmic management through gamic elements and literacies to regain control of

their work.

In the first instance, this section argues that although algorithmic literacy or online

sensemaking are widely hidden in nature amongst drivers, the shared and lived experiences on

SMCN are integral to the interconnectedness of hidden and public resistances in Lagos.

252

6.3.1. Algorithmic Literacy: Making Sense of Ride-hailing Gig Work

The isolating nature of ride-hailing work in Lagos and more generally creates an avenue

for the mobilisation of platform drivers as a collective group either through social media

platforms (e.g., Facebook), communication platforms (e.g., WhatsApp and Telegram) (Kinder

et al., 2019) or physical meeting points. For instance, scholars recognise online forums as a

medium for developing algorithmic competency (Jarrahi and Sutherland, 2018) and

sensemaking (Möhlman and Zalmanson, 2017; Cheng and Foley, 2019) which empowers

platform workers with information to understand algorithmic assignments.

In studying online gig workers (e.g., Upwork), Jarrahi and Sutherland (2018) coined

the notion ‘algorithmic competency’ in the form of data infrastructure literacy or infrastructural

competency (Strauss and Corbin 1990; Gray et al., 2018), which they conceptualise as

sensemaking, circumvention and manipulative practices. Accordingly, "algorithmic

competency refers to workers' understanding of algorithms that assign and assess work

conducted on gig platforms and learning how to work with and around those algorithms"

(Jarrahi and Sutherland, 2018, p. 9). While this term is plausible and adaptable to this research,

I would argue that online sensemaking practices are central to competently subverting the

power of algorithms in the gig economy. SMCNs were central resources to decipher

algorithmic opacity for Airbnb hosts to experiment, manipulate and resist unfair terms of

service (Cheng and Foley, 2019). Therefore, I propose the term algorithmic literacy in the place

of sensemaking in reference to the information gained through gig workers shared or lived

experiences online and offline for regaining genuine autonomy and control of their labour.

Making sense of algorithmic burdens equips drivers with the competency that facilitates

mobilisation processes central to hidden and public resistances. Information shared on these

platforms is a "set of socially and culturally established ways to identify, see, use and share the

information available in various sources such as television, newspaper, and the internet”

(Savolainen, 2008, p. 2). The confinement of information sharing within a cognitive paradigm

encapsulates the social and cultural conditions that produce information and information

behaviour (Lloyds, 2014).

The physical form of mobilisation for many platform drivers in Lagos is the training

session for ride-hailing work. Uber and Bolt conduct regular training and onboarding sessions

for successful platform drivers. Each of these sessions admits around 100 or more people from

various backgrounds. Drivers also interact in vacant spaces in the city, such as banks, eateries

253

and hotel car parks. Identifying each other is built on symbolic appearances and habits amongst

platform drivers. For instance, the popular vehicles for ride-hailing platform drivers were the

Toyota Corolla 2005 to 2007 models, Toyota Camry 2004 and Honda Accord. Drivers can also

recognise each other from their colleagues placing smartphone devices on their legs or a

dashboard during a trip or when parked in a vacant car space. Other habits include the

availability of refreshments and charging cables or different smartphone devices in vehicles.

These meetings are often interlinked with drivers exchanging SMCN contact information and

joining these groups. For instance, it was evident to observe from a Bolt training session in

November 2018 how platform drivers exchanged contact information for the potential to form

or join SMCNs subsequently.

Virtual forms of mobilisation and algorithmic literacy are through SMCN groups.

Accordingly, the SMCN serves as a platform for driver sousveillance (Mann, 2004) based on

how they post their everyday experiences captured through smartphone devices. Unlike

evidence from the literature that depicts social networks as a tool that facilitates activism and

social movements (e.g., Harlow, 2011; Carty, 2015; Mundt et al., 2018), the initial intention

for platform drivers in Lagos was not for activism but to mutually learn about their experiences,

considering the digital nature of ride-hailing gig work in Lagos. However, this has become a

basis to demand better labour conditions through different strategies of activism. During my

fieldwork exercise, most of the Facebook groups for platform drivers in Lagos were private.

To access the online environment, the incoming platform driver would need to show evidence

such as Uber or Bolt dashboards, which indicate weekly earnings, in-app identity, ratings and

other metrics. However, despite its hidden nature, these groups comprise vehicle owners

(partners) or rental companies that advertise to unemployed platform drivers. It is an avenue

for experienced platform drivers to boost their learning capabilities by sharing information on

Facebook groups, and less experienced drivers contribute by asking questions about vague

messages on their platforms (Interviewee Abel, September 2019). Some of these pieces of

information include but are not limited to requirements for applying for Uber or Taxify (Bolt)

gigs, choosing a vehicle owner, the kind of weekly payments for rentals or hire purchase,

posting screenshots and asking about their meaning and such other practices.

As a result of different categories of people on Facebook groups, some drivers further

choose more secure data-protected platforms like WhatsApp and Telegram to verify the

identity of potential group members. In these private groups, platform drivers self-organise by

assigning roles such as 'head admin' and 'assistant admin' on private groups. Most of these

groups possess specific requirements to join. For instance, Charles (November 2018), a retired

254

platform driver who is a lead administrator of a WhatsApp group, requires evidence such as

the data page of any platform company and a driver's licence from platform drivers before they

can join his group. The core reason for this is to prevent spies from Uber or Bolt from joining

the group (Interviewee Charles, November 2018). Additionally, these groups help prevent the

spread of false information amongst drivers, which are internal enablers of information

asymmetries. They also equip drivers with knowledge on avoiding law enforcement agents,

who exercise extortionate power based on the identities of the kind of vehicle utilised, as

mentioned above, and behavioural actions such as putting the smartphone devices on their

dashboard or legs. The interviewee Abel (September 2019) explained that:

The essence of drivers' forums and WhatsApp groups is to learn from the

experiences of others. It is like having a commentary on daily activities in the industry.

Because the industry is online, it is easier to understand. Perhaps, should there be a

problem in a destination you are going to, somebody who is already there could advise

that the area is blocked or locked down.

In this study, 95% of the platform drivers highlighted Abel's testimony as a reality for

surviving ride-hailing work. For instance, platform drivers who have experienced terrible

events due to fraudulent riders in unsafe areas share their bio-information (e.g., names, pictures,

phone numbers, receipts) via communication platforms to prevent other drivers from falling

prey to them or to mobilise drivers to seek justice. For example:

Interviewee Samuel (August 2019) narrated a story when a female passenger

refused to pay for a trip in Sangotedo Lagos, after which she mobilised thugs to harass

the Uber driver requesting his money. The Bolt driver posted the information on his

WhatsApp group, mobilising 27 platform drivers to apprehend the lady, seizing her

iPhone, and collecting N15 000 (£31) for the fare and damages.

In a situation like this, if the female rider reported to the Bolt platform, the driver could

be unduly deactivated, as is commonly the case. One of the three broad conceptualisations of

competencies by Sandberg and Pinnington (2009) argues for competence as a capability or

practical accomplishment (Gherardi, 2000) to realise specific work tasks. For Gherardi (2000,

p.220), “practice is the figure of discourse that allows the processes of ‘knowing’ at work and

in organising to be articulated as historical processes, material and indeterminate”. While

255

algorithmic literacy in itself is not a hidden resistance practice, it is central to the competencies

Jarrahi and Sutherland (2018) highlight, which, in many instances, are integral to both hidden

and public resistance in Lagos.

6.3.2. False Compliance of Platform Drivers: Gaming Spaces for

Rewards

Drawing from Scott's notion of everyday resistance, falsely complying with instructions

was one practice subordinate groups used to bargain for better labour conditions from dominant

groups in South Asia (Scott, 1985; 1989). In this case, ride-hailing drivers evade, bend and

game the instructions of algorithms for personal benefits, which include surge prices,

promotional trips or bonuses, and subsequently instigating offline trips. In Lagos, observations

of practices that subvert algorithmic managerial assignments were made, including creating or

gaming surge prices and promotional trips. Jonathan, an experienced driver and IT developer,

highlights how chasing surges and bonuses is not worth it.

You will see people driving all the way to go and catch surge. Surge is

programming too. I have been with many programmers. Sometimes you will be seeing

surge on the Atlantic Ocean…There is no need to slave yourself in this work. What if

you are trying to meet a target and something happens, such as speeding past a red

light; LASTMA now arrests you. (Jonathan, August 2018)

In Lagos, interviewee Osahon (September 2018) acknowledged that targeting surge

prices was a core practice for most drivers. However, globally, drivers are increasingly

becoming aware of some of these surges' opacity and are creating artificial surges or gaming

potential surging spaces based on their knowledge of the city. At the Ronald Reagan Airport

in Virginia, 120 to 150 Uber drivers admitted to switching off their apps five minutes before

flights arrived (Sweeney, 2019). These drivers possess third-party flight apps for monitoring

arrival times which enables efficient gaming for surge fares. Automatically, the area surges

with fares ranging between $10 (£7.44) and $19 (£14.13), after which drivers switch on their

apps. As a response to unreliable surge prices, Lagos' platform drivers learned to create surge

areas in Lagos, following examples in developed cities. In November 2018, interviewee

256

Charles highlighted using a third-party app known as flight stats to target trips from the Murtala

International Airport in Lagos. At the time, it was difficult for drivers to work in unison to

create surges. However, interviewee Dipo (August 2019) admitted plans to emulate these

evasive practices from the US by collectively creating surging areas.

What we want to do is that we want to predict how many riders are moving in

the morning and afternoon. So that whenever a rider calls, it would be a pool of drivers

that are there … Ten drivers can switch off their app in one location; one of them will

be controlling it. I can say, "comrade Taiwo, switch on your app". It is only one driver

who would interface with Uber at a time, with about 100 rider requests. Uber will not

have any choice but to increase the surge. Ten people will make correct money from it.

(FGD – Dipo, August 2019)

These are algorithmic literacies gained across digital spaces via SMCN, although

drivers lack the unity to initiate such practices in Lagos. Instead of waiting for the algorithm to

suggest surging areas, drivers utilise their knowledge of the city to predict potential surge

spaces. For example, I observed how drivers slept overnight in vacant parking spaces in Lekki

Phase One area between 10 pm and 5 am daily to potentially target high demands in the

morning because of the high demands from workers in locations like Victoria Garden City,

Ajah and Chevron locations in Lagos (FGD, 2018). Later in the day, drivers target the Central

Business District (CBD) – Victoria Island towards the close of work using private SMCNs as

real-time apparatuses to gauge the movement of people (Interviewee Charles, November

2018).

257

Figure 30: A Driver Soliciting for Trip Requests from Colleagues (Before and After)

Source: Uber and Taxify Platform Driver Facebook group (closed group) (November 2019)

Similarly, platform drivers have increasingly become aware of the opacities and

information asymmetries involved in a bonus or incentivised trips. Some drivers, such as

Jonathan (August 2018), depict promotional trips as unclear work assignments because of his

profession as an IT developer; he states, "It is a scam…". For example, promotional trips such

as Figure 30 may nudge drivers to make a certain number of trips, but drivers often narrate that

Uber and often Bolt may flash a fake trip request or issue long trips, preventing them from

meeting their targets. Platform drivers installed an Uber or Bolt rider application on their

smartphone devices to initiate self-requests to subvert this opacity.

258

… If I have done 24 out of 25 trips and I need one more request to complete it.

I would rather use my rider app to request myself, then drive like 5 minutes, then pay

N500 (£0.9) and collect the remaining money. (Interviewee Koffi, September 2019)

This practice becomes fraudulent after updating the app when the platform detects the

smartphone's IMEI number and the driver's user identification number (interviewee Jude,

November 2018). However, platform drivers have developed a counteraction by requesting

each other with a rider application. For example, on 9

January 2019, my fieldwork

observations between two platform drivers showed platform driver Efe calling a colleague,

Henry, to initiate a request to achieve the day's bonus.

These rewards, such as bonuses or surge fares, keep drivers on the road, generating

profits for platform companies. Even with drivers' gaming surges and bonuses, profit gains are

small, indicating that platform companies are still exploiting drivers. The Uber platform might

argue that drivers have 12-capped working hours, but this only includes time spent on a trip,

not waiting times. Apart from this being a competitive strategy amongst platforms like Uber

and Bolt, it is a deliberate action to hide the fact that drivers are paid unfairly. The algorithm

can be programmed to reflect decent payments with a retrofitted commission rate for Lagos,

but platforms choose to propagate precarity.

6.3.2.1. Escaping the App: Gaming for Offline Trips

The decision to circumvent the algorithms managing work for offline trips is one of the

few hidden practices platform drivers in Lagos utilise to maximise personal benefits. 'Offline'

as the name implies, means that trips transcend the confines of the platform – an act of

circumventing algorithmic management rules by switching off the app. This allows the driver

time to negotiate higher fares without ratings or other evaluation metrics. There are three types

of offline trips platform drivers utilise in Lagos, which include: intra-state offline trips, inter-

state offline trips and offline courier trips. While these existed within the traditional taxi

regime, online integrates aspects of surveillant assemblages and external modalities absent with

taxis. Instead of taking trip assignments directly from the algorithms, ride-hailing drivers

circumvent the app based on developed relationships with previous riders. Drivers inherently

practice some form of disintermediation by eliminating the riders from the app (Rosenbloom,

2007; Graham et al., 2017), although temporally. Hence the need to go offline to hide from the

259

algorithmic gaze. Platform drivers actively game or passively game for these three categories

of offline trips.

According to interviewee Okoro (August 2018), offline trips are more rewarding than

online trips despite the risks. For intra-state trips, 100% of the drivers in this study

acknowledged indulging in the practice for higher income. The passive method for recruiting

potential offline riders is through networking or honest service during Uber or Bolt trips.

… Offline trips save us more on this job. You could be so nice to some riders,

and they would make you their personal Uber driver; when they want to go anywhere,

they would call you… (Interviewee Junior, November 2018)

Other drivers like Charles (November 2018) testified about this method of recruiting

offline riders. Acts of geniality can be in the form of symbolic gestures, which include vehicle

refreshments like sweets, bottled water, phone chargers (Rosenblat and Stark, 2016; Rosenblat,

2018), opening the vehicle door, putting on the air conditioner and giving the rider the freedom

to choose different songs during a trip. Sociologist Arlie Hochschild (1983) classified such

behaviours as emotional labour – in this case, platform drivers present their best service

regardless of the rider's emotional state for beneficial purposes. Blackmailing riders with the

issue of traffic or unsafe areas is an active method of recruitment for intra-state offline trips. In

this case, platform drivers express fear and uncertainty about a location because of the traffic,

the distance (e.g., from Lekki to Ikorodu at night) and security.

… I was to pick up somebody, and I anticipated a traffic gridlock at his

destination. If I take him and return to town, it would be heavy traffic. When I started

the trip … I just said I would not be able to go. If he is willing to go, it would have to

be on an offline basis, but on the app, I cannot go… (Interviewee, Okoro August 2018)

When a platform driver cannot accept a rider's trip and is unwilling to bargain for an

offline trip, the driver negotiates with the rider to cancel the trip to avoid a trip cancellation,

which would impact future algorithmic assignments. Negotiating with riders is a practice to

resist trips imposed on the driver by the algorithms. In the stated example of trip cancellations

by Uber drivers in Mölhmann and Zalmanson’s (2017) study, although the practice analysed

is slightly different, they did not state if drivers are sanctioned or how they avoid sanctions

from algorithms. Negotiating trip cancellations with riders in Lagos is one trivial practice

260

platform drivers employ to resist algorithmic control. In a typical situation, because drivers are

aware that trip cancellations can negatively impact their acceptance rates, platform drivers

transfer the onus to cancel a trip to the rider. In some instances, such as delaying more than

five minutes on Uber or four minutes on Bolt, platform drivers benefit from a cancellation fee

of N400 (£0.83).

Inter-city offline trips are more strategic than intra-city offline trips, even though some

passive recruitment methods, such as networking, are similar. Although the practice of inter-

city offline trips is not new in Lagos, evading algorithmic management and surveillance on

platforms is a resistant behaviour in one instance and a survival tactic in another instance for

platform drivers. Traditional taxi driver and BSc graduate Adedayo (September 2019), who

has been driving for over eight years, explained in an interview how taxi drivers, particularly

'for-hire’ taxis, facilitate inter-state trips for higher-income with customers based on trust. I

also observed this at taxi ranks. More recently, interviewee Adedayo (September 2019)

purchased a Toyota Space bus for inter-state trips due to a dearth of intra-city trips since the

emergence of ride-hailing platforms. At least for these taxis, there is no form of monitoring

except for the coloured taxis (e.g., yellow taxis), which are more symbolic and confined to the

State. For-hire taxis experience better freedom of movement because they appear as private

vehicles with no designated representation colours. Regardless, the network for facilitating

these inter-city trips for traditional taxi drivers is limited compared to platform drivers. The

advantage for platform drivers is the presence of both the platform app and the SMCN, which

supplement passive methods for recruiting potential offline riders.

I have my personal trips to Ogun State, Ibadan, Benin. Sometimes you get to the

pick-up location, and the person can tell you to switch off your app that he is going far

away. Last week, I got a request, the person said: "Oga I requested you, I am going to

Ibadan, will you go?" I said, “Yes sir.” He said I should cancel the trip. I charged him

N30,000 to Ibadan and back. (Interviewee Jacob, August 2019)

This narrative also indicates riders' awareness of offline trips as a hidden resistance

practice. These longer trips are beneficial to drivers because they gain more money outside the

app and without commission deductions. For riders, they save more money because of the

absence of algorithms that will calculate long wait times, which are not equally beneficial to

drivers as well. Some riders know that the fare from dynamic pricing in Lagos is exploitative

for platform drivers. For example, interviewee Koffi (September 2019) expressed how an

261

American rider he transported from Murtala International Airport to Abraham Adesanya on

Lagos Island was surprised at the fare of N5,200 (£10.8) for 54km. The distance from the rider's

residence to the airport in America is about 23km, with a fare of $60 (£45) – $70 (£52).

Although the economies between America and Nigeria are different, these interactions

reinforce self-awareness of unfair payments, facilitating intentions to game the system. Using

SMCN, platform drivers actively share or search for offline trip opportunities or seek advice

on an offline trip proposal amongst other platform drivers. For example, one Facebook post I

observed of a driver in July 2019 stated (paraphrased):

Professional advice needed for an offline trip to take someone from Lekki to Ile-

Ife for four days, stay there for two days, then go to Ibadan on the third day. On the

fourth day, back to Lagos for N52,000 (£109). Is this a good deal? (Facebook user, July

2019)

There were 58 replies from colleagues advising the driver on whether it was a good

deal and other drivers telling the poster to be careful. This practice is also prevalent on intra-

city trips but requires less planning and fewer deliberations. On accepting an offline inter-state

trip, platform drivers post screenshots of the rider's information in WhatsApp groups as a form

of surveillance from their colleagues for security reasons (Interviewee Jacob, August 2019).

This shows drivers using information from algorithms to their benefit and controlling the

output with colleagues on SMCNs. The power of screenshots not only prevents conflicts in

friendships, according to Jaynes (2019), but it is also evidence by which to monitor and ensure

the safety of platform drivers during offline trips in Lagos.

Finally, offline courier trips are creative practices platform drivers utilise for similar

reasons, as mentioned above. One Facebook user, who could not get trips posted about a

delivery service for a ram, stated:

If people refuse to request for a ride, no wahala (no problem). I have a trip to

deliver a ram from Obadeyi to Yaba on Wednesday night for N8,000 and two rams for

N15,000. God bless our hustle. (Facebook user, August 2019)

The items for delivery could vary, such as animals, heavy equipment or food items like

cakes. Interviewee Samuel (August 2019), a platform fleet manager, reaffirmed this practice

when one of his drivers remunerated his weekly payments before the due date.

262

In conclusion, offline trips are hidden everyday practices platform drivers employ to

subvert the control of algorithmic management by taking advantage of a rider's circumstances

or vice versa for the advancement of personal benefits such as a higher payment without

commission impingement from algorithms. The phenomenon temporarily subjugates the

power of algorithms and facilitates faster weekly remunerations to partners by granting

platform drivers real autonomy to bargain for more income and evade algorithmic rules.

6.3.3. Sabotage: Manipulating Algorithms to Regain Control

The act of sabotage has been conceptualised in resistance literature within plantations,

mines and more traditional working environments as deliberately damaging, disrupting, or

obstructing labour with discreet practices that are targeted at the reputation and properties of

dominant groups as a way of demanding fair treatments (Scott, 1989; Gupta, 2001; Jordan,

2003; Mutekwe, 2019). Using Scott's weapons of the weak, Mutekwe (2019) argues that

Zimplat miners in Zimbabwe engage in everyday hidden workplace resistances. He classifies

these as absenteeism and Kukanda (sick notes), calling and unique language, desertions, and

resignations, and, critical to this section, sabotage. Miners at Zimplats engaged in sabotage by

damaging machines, which would cause them to waste time on simple tasks, as revenge against

managerial repression, poor working conditions and low wages. The difference with sabotage

examined by Scott (1989; Mutekwe, 2019) is that it identifies physical attributes such as

farmlands or miners’ equipment that workers can target or behavioural traits that focus on a

farm owner, job owners and other dominant representations of the powerful. However, in a

context like ride-hailing, where the workplace is embedded in digital space, it becomes difficult

for drivers to destroy their phones, for instance, or their vehicles. Although public practices

can target platform offices, their workplaces remain with algorithms as their managers via the

app.

Manipulating algorithmic management is the most strategic and hidden practice by

platform drivers in Lagos. Elsewhere, Jarrahi and Sutherland (2018) observed manipulating

the Upwork platform algorithm as a practice to alter, observe and improve workers' working

conditions during labour. This section, however, conceptualises algorithm manipulation as

hidden practices for ride-hailing drivers facilitated by lived experiences via digital spaces such

as private SMCN groups. It was found that there are technologically savvy drivers aware of

the loopholes on ride-hailing platforms, which are prevalent because of poor urban

263

infrastructure, traffic jams, and weak regulatory frameworks. This means that for the realities

that algorithms fail to integrate, drivers utilise them to their advantage in closing the gap against

algorithmic burdens. These weaknesses, such as an inaccurate representation of satellite maps

in specific locations and absent-mindedness of riders, help drivers manipulate the platform

during or before a trip and later share these experiences on private SMCN groups.

In China, Chen (2017) found that about 40% of drivers use bot apps on the Didi

Chuxing platform and register accounts on different devices to manipulate the app. These

actions enable drivers to select trips with the highest fares while rejecting unfavourable trips

without impacting their evaluation metrics or receiving sanctions. Möhlmann and Zalmanson

(2017) outline some manipulative practices they observed via online forums in New York and

London. For instance, to avoid angry riders who would rate them poorly, drivers usually cancel

trips first to avoid negative ratings that impact their labour. In Lagos, I examined four dominant

practices of manipulating ride-hailing apps in Lagos. These involve using third-party apps and

techniques, which drivers identified as Lockito; fraudulent transaction deductions; using

embedded GPS functions on smartphone devices; and using platform-forbidden devices (e.g.,

Phone 6) – a practice collectively known as 'Sakamanje'.

Figure 31: Fake GPS Trip Data Using Lockito Verus Genuine GPS Trip Data

Source: Nairaland Forum (2017); Arubayi (2020)

According to 95% of the drivers in this study, the pricing was fair before Bolt came

into Lagos in 2017. To recap on their emergence detailed in Chapter four, drivers perceived

that the occurrence led to competition in Lagos because the Bolt platform introduced lower

264

fares and lower saloon car models for driving, forcing Uber to do the same. The competitive

battle between Uber and Bolt in 2017 led to a 40% fare cut for Uber while retaining its 25%

commission rates. Apart from this creating a public response through the first significant

protest in 2017, specific platform drivers discovered Lockito, a third-party Android app

designed for developers to test geofencing-based apps or test the app in different locations.

Platform drivers use the app alongside Bolt or Uber apps, especially on Android devices.

Before a trip, as in figure 31, drivers input the location in the Lockito app in the background

such that the smartphone receives a wrong GPS signal which increases the length of the journey

(Nairaland 2017). As the distance increases, the kilometres travelled increase, which signifies

increased fares. An original trip fare of N5,000 (£9.4) can increase to N10,000 (£19) because

of the Lockito effect (FGD, Charles, November 2018). Besides the interviewee, Kene (FGD,

October 2018), the other 24 drivers in this study denied using this manipulative technique, even

though their knowledge of these practices was robust. Platform drivers were also surprised

about my questions on this technique – signalling a covert action only amongst themselves.

The second category is a response to platform updates against Lockito manipulation.

According to platform drivers, 'Phone 6' is a device that possesses embedded functions that

manipulate the GPS signal during a trip, thereby increasing kilometres travelled (FGD, Charles,

November 2018). The FGD suggested that the IMEI number can be changed with these

devices, which contributes to evading app updates, analysed in the subsequent section. This

device, however, falls into what platform drivers call 'Sakamanje'.

Sakamanje is not just an app; it is also a phone. If it happened to be one of the

phones not recommended by Uber/Bolt, such as phones low in size, RAM, or the patent

is not up to standard. The app will automatically be malfunctioning. It would be moving

from the front to the back and adding kilometres. (Interviewee Koffi, September 2019)

Although the developers of these devices were untraceable, platform driver Charles

(FGD, November 2018) identified their name as 'Uber boys' – a codename for tech drivers

specialising in discovering manipulative practices. According to Charles, these 'Uber boys'

secretly sold these devices to drivers at the Bolt head office in Lekki after training sessions.

This device, however, falls into a broader and loosely used term platform drivers call

'Sakamanje (Saka for short)'. This codename was used to anonymise the identities of these so-

called Uber boys. I attempted to investigate these Uber boys in the field, including waiting at

the Bolt office to observe and ask questions, but they were evasive.

265

Preceding this practice by Uber boys is one I refer to as fraudulent transaction

deductions. These transactions involve drivers using information from at least one stolen debit

card of a rider to initiate multiple distant trips by requesting themselves. According to Charles

(FGD, November 2018):

Uber boys, na research them dey do. Uber boys no dey go work, there fit use

one ATM card initiate multiple trip requests to places like EPE among 5 to 6 of them.

Each person go collect like 20k each. I know one guy wey die as a result of these things.

[

]

This implies that these Uber boys manipulate the algorithm by initiating long-distance

trips amongst a group of drivers for income gains. In reflecting on this, it was unclear if these

debit cards were stolen rider cards or a single debit card with multiple drivers sharing the

profits. The latter appears counterproductive because drivers paying themselves involves

taking money from each other. The former, however, involves using rider card details to gain

profits. Even if these card accounts possess no money, it would be deducted when money enters

the account unless the rider reports it as fraudulent. However, this typifies the hidden nature of

this practice based on the ambiguity from drivers.

The fourth category is the 'battery savings' and 'location' practice. Compared to the

previous techniques, the battery-saving technique targets the 'minutes per kilometre metric'

during a trip and further increases fares. Platform drivers discovered that putting the battery on

'power saving' mode as an embedded function distorts the efficiency and accuracy of the GPS

signal. This technique's accuracy is most beneficial when a driver is in traffic; otherwise, it

reduces the driver's fare when in motion (FGD, Jude, October 2018). This practice involves

wider knowledge of the city, such as peak times and traffic-congested zones on different days.

This phenomenon was a surprise to the driver Akin during one of my FGDs in October 2018.

The driver was unaware of how to deploy this technique, which negatively impacted his

earnings. Similarly, platform drivers switch off the location icon in their phone settings,

preventing the app from recognising the actual travel distance, an act known as 'location'

(Interviewee Koffi, September 2019). Getting to the end of trips, platform drivers switch on

the location setting, which most often creates an increased fare that may benefit them. In

Australia, to create unlimited time for orders on food delivery platforms, one driver claimed to

hack the app's auto-acceptance function and utilise a location-masking tool (Veen et al., 2019).

266

It is important to note that exposure to these actions affects platform companies'

reputations because some riders may choose cash preferences due to trust issues. For example,

attending a workshop on promoting road safety in a Smart City organised by the Ministry of

Transport ALAUSA, Ikeja Lagos, the former commissioner of transport, explained that he had

to delete his Uber account because of continuous unknown debit alerts from his bank account.

[

]

Reflecting on this, I linked it to Charles’ account of how drivers initiate multiple requests

with a single ATM card. However, these were difficult to validate despite a clear linkage

between them. In addition, Uber had discovered this practice during my fieldwork, making it

difficult to investigate.

6.3.3.1. Rejecting App Updates: Prolonging Algorithm Manipulation

As a way of circumventing new regulations that come with app updates, drivers delay

updating their accounts. This behaviour is central to practices that depend on the app for

manipulation. Riders who report these hidden practices through the app create a reference

record for platforms to enforce sanctions. For example, in 2017, the Lockito technique was

reported via 96.1fm as a fraudulent practice. This feedback loop to platform companies enables

mutual learning capabilities that train the algorithms to detect these practices as fraudulent

activities with the penalty of immediate deactivation or flagging of drivers' accounts. Following

this, platforms introduce app updates, not just bug fixes, but to close these loopholes. It is

important to note that, while some practices such as Lockito began on the Uber platform, the

Bolt platform possesses more loopholes that platform drivers exploit to develop more

techniques (Interviewee Abiodun, September 2019). Platform driver Abiodun (September

2019), a developer, recounts the difficulty in cracking the Uber platform because of its robust

security level compared to the Bolt platform. Using different third-party apps and techniques,

they have successfully hacked the Bolt app in several ways. However, closed loopholes enable

drivers to discover new hidden practices that facilitate profit gains.

If possible, drivers that are experienced in the field delay updating their apps. The first

time I recorded this practice was during a FGD in October (2018) when Charles advised other

drivers in the session not to update their Bolt or Uber apps. At the time, I could not understand

what it meant. However, in September 2019, platform driver Koffi explained it as a way for

drivers to counter app updates that detect manipulative techniques as fraudulent. Using the

older version of these apps indirectly indicates that platform drivers are not subject to the

updated rules on a newer version and further facilitates manipulative techniques. This

267

phenomenon creates a market for platform drivers with older versions of these apps or with

multiple accounts because they sell the accounts to desperate drivers for between N5,000 (£10)

and N40,000 (£42) (Interviewee Koffi, September 2019).

… Some people that have family people in the village can bring about 5 to 6 of

them to Lagos and open accounts on these platforms for them. They will go back, while

you will be selling those accounts… (Interviewee Koffi, September 2019)

As I observed on the Facebook group, drivers advertise these accounts with high-end

ratings of 4.7 to 4.8-star ratings to entice potential buyers with low ratings; drivers blocked

from platforms and drivers who resell them or use them to discover hidden practices. Anwar

and Graham (2019) identify the 'creation of multiple accounts' and 'buying and selling’ of

accounts as a hidden practice of resistance, but only in terms of Katz’s (2004) classification of

'reworking'. According to Katz (2004, p. 247), reworking practices “alter peoples' conditions

to enable more workable lives and create more viable terrains of practice”. This is associated

with how people are redirecting available resources and redeveloping themselves as social

actors and political subjects to improve their material stances. While this was used to increase

jobs for Upwork workers in Anwar and Graham's study, I would also add that it is a deliberate

action to resist the rules and regulation of algorithms for buyers while sellers expand material

benefits and monetise their digital identities within their scope.

The intention for these deliberate acts of resistance against platforms was evident in the

interviews with platform drivers George and Koffi (September 2019); they admitted that IT

professionals amongst drivers would continually develop techniques to disrupt the system until

platforms treated drivers fairly. Admittedly, one platform driver, Thomas, a union member,

read out a statement from one of the IT developers on his WhatsApp group in a focus group

discussion (August 2019). In summary:

The driver claimed to be the developer of the first and second manipulative

techniques. As an IT professional, his motivation was for drivers to achieve their goals

since the price slash in 2017, which was no longer commensurate with their labour. He

highlighted that these manipulative techniques, such as Sakamanje, helped him acquire

his first vehicle on hire-purchase. He vowed to keep developing more apps until app

companies treat drivers fairly by adjusting prices that suit the market.

268

The above statement makes it evident that there is an ongoing silent battle between

drivers and the consequences of algorithmic burdens. Drivers develop algorithmic literacies

that boost competency levels that lead to the subversion of the power of algorithmic

management. Therefore, it is evident that there is a silent battle against platforms because of

unequal power relations. What is fraudulent according to platform companies is resistance and

agency according to platform drivers, which empowers them to develop algorithmic literacies

and competencies to evade the burdens of labour which are the consequences of bad

algorithmic management.

6.4. Conclusion

Building on the previous chapters, this chapter has emphasised how international

platforms such as Uber and Bolt emerge in GS cities like Lagos. These international platforms

have formalised the informal processes of driving and management which existed in traditional

taxi regimes. In other words, platforms have exacerbated the deregulation of the taxi industry,

leading to disguised informal working conditions which further isolate the drivers in Lagos,

deny drivers access to safety nets, and ultimately prevent them from unionising and collectively

bargaining against platforms. The algorithm as a managerial weapon for these dominant

platforms has created algorithmic burdens based on a combination of information asymmetries,

opacities, and bias, limiting drivers' autonomy, flexibility, and decision-making. This

propagates control over drivers and, by design, destabilises collectivism which develops from

the gradual mobilisation of platform unions in Lagos.

In integrating Scott’s concept of everyday resistances, this chapter examines resistances

against ride-hailing platforms based on his notion of public and hidden transcripts (Scott, 1985;

1989; 1990). Surveillance assemblages and external modalities, which are integral for

algorithmic management, are critical for what I call weapons of the dominant, which are

instrumental in enabling control over the body of drivers. However, drivers devise public and

hidden practices to regain control. The analysis of public resistance discusses how platform

drivers initiate public practices based on building platform unions, engaging in offline protests

and online app boycotts and public dissent in media engagements against platforms. These

practices are similar to those occurring in other GS and North cities (Carmody and Fortuin,

2019; Cant, 2020; Woodcock, 2020). However, in the Lagos case, collective actions are not

substantial in redefining the nature of their work due to porous regulatory frameworks and

perceptions of predatory traditional taxi unions, which also prevents union members from

269

working as a unit (Agbiboa, 2020). For instance, this is unlike the positive results from protests

and court cases in GN cities such as the UK, which have led to the reclassification of workers

as employees.

Consequently, ride-hailing drivers in Lagos utilise more hidden practices based on

developing algorithmic literacies through SMCNs, to subvert the power of algorithmic

management. Figure 32 highlights a conceptual framing of algorithmic management and

everyday resistances in Lagos. Practices identified aim to manipulate, circumvent and gamify

burdensome rules that impact their everyday working lives. These are also intentional in

ensuring that platform companies treat ride-hailing drivers fairly. However, a key finding

identifies that hidden resistances are not static due to the temporalisation of labour, the ability

of algorithms to learn resistant practices, and drivers' ability to discover new ones because of

the machine learning by algorithms. Hidden practices assume the status of public practices

when platform companies become aware by further creating app updates and awareness for

riders. However, the continuous process of developing more hidden gamic or manipulative

practices creates an inconsistent loop between the hidden and the public realms of resistance.

This further reinforces the collapsing of in-between spaces based on spatial mediation because

public and hidden resistance, in this case, is entwined in multiple conjunctions of code,

space/place, and content facilitated by the practices of drivers’ everyday life (Leszczynski,

2015). The in-between space, as a site of the incessant battle with platforms, cannot be reduced

to only occurring in free spaces (Polletta 1999), but continuous collapsing of the in-between

space based on the complexities existing between drivers operating the device and algorithms

managing the process.

This inconsistency creates an entwined interrelationship between the hidden resistances

and public resistances (Johansson and Vinthagen, 2016), indicating the fragility of spatial and

temporal dimensions made possible by drivers' algorithmic competencies (Jarrahi and

Sutherland, 2018) in subverting the power of algorithms in Lagos. It was also critical for the

algorithmic management concept to acknowledge that just as gig workers gain competencies

in subverting their power, the system also learns and restricts practices of resistance.

270

Algorithmic Management

• Task assignment

• Monitoring and

surveillance

• Performance

evaluation

• Payments & Rewards

• Sanctions and

Deactivations

Algorithmic burdens

• Opacities

• Information

asymmetries

• Biases

Impacts on Gig workers

• Overworking

• Data misrepresentation

• Arbitrary deactivation

and punishment

• Manipulative Incentives

• Digital Map Inaccuracy

Gig workers

Everyday

Resistances

• Public

resistances

• Hidden

Resistances

Regaining Control

Enabling Control

Mutual Learning capabilities

271

Figure 32: Conceptual model of algorithmic management and the everyday resistances in

Lagos.

This chapter improves the understanding of everyday resistances, which not only builds

on Scott but also on Adas (1986), Adnan (2007) and Shalhoub-Kevorkian (2012), whose

conceptual refinement is based on technology advancements that emphasise the

temporalisation and spatialisation of everyday resistances. This also indicates dominant

platforms with practices that are part of the hidden transcript without physical offstage

performances as in Scott (1989) but utilising weapons of the dominant to restore control over

ride-hailing workers. While it is plausible in recognised actions that originate from digital

spaces as hidden resistances, it is worth noting that these spaces quickly merge based on the

co-constitution of humans and technology, i.e., based on mutual reinforcing tendencies as a

result of the interdependencies between real and virtual places (Kinsley, 2014). Therefore, it is

safe to classify actions hidden in digital spaces as everyday digital resistances, which

encapsulate the changes in strategies for both drivers and platforms, which I expand upon in

the subsequent chapter.

272

7. Chapter Seven: Conclusion and Recommendations

7.1. Introduction

In understanding how ride-hailing platforms emerge in a GS context like Lagos,

Nigeria, this thesis has further critically examined and unpacked the impacts of ride-hailing

platforms through algorithmic management on drivers. Furthermore, it outlines how drivers

are choosing to subvert algorithmic control by inventing everyday resistances, which are

critical for survival in the ride-hailing sector in Lagos. This chapter synthesises all the findings

that inform methodological and theoretical contributions relevant to GS cities and other

contexts within the gig economy. This thesis also provides valuable recommendations for an

overall improvement in researching gig work and adapting algorithmic management as a

concept. These recommendations apply to researchers, policymakers, gig workers, platform

companies, and think tanks within GS contexts.

7.2. Research Question 1: How do Ride-hailing Platforms Emerge in a

Global South Context like Lagos?

First, this thesis agrees with Kaye Essien's (2020) assertion that ride-hailing platforms

are another neoliberal project that has been exported to GS contexts. As discussed in chapter

two, neoliberal policies such as SAPS emerged following the debt crisis in the 1980s, which

were based on privatisation, deregulation, free market, and free trade improvement (Bryseson

and Potts, 2006). These policies are often unilateral, i.e., one-size-fits-all policies that do not

consider the contextual differences. In other words, these westernised concepts have not

succeeded in places like Nigeria. Instead, they have increased unemployment rates, increased

income inequality, and diminished the bargaining power of unions (Potts, 2008; Obeng-

Odoom, 2012; Lebaron and Ayers, 2013). Overall, this phenomenon has depreciated formal

labour and proliferated more individuals to adopt informal labour practices with minimal or no

safety protections and a decent wage. These informalities exist due to deregulation or no

regulation in most sectors and often with the deliberate complicity of the state.

In answering this question, this research sought to understand how ride-hailing

platforms such as Uber first emerged in GN contexts such as the US before emerging in GS

contexts like Lagos, Nigeria. The mobility sector, specifically the taxi industry, is one of the

273

areas where there are blurred lines between formal and informal labour, further leading to

precarious working conditions for drivers and a lack of safety nets. In comparing the taxi

industry between GN and GS contexts, this thesis highlights three factors: the regulatory

environment, dynamics of trade unionism, and an existing digitised environment. In a nutshell,

I argue that GN contexts such as the US possess a strong regulatory environment (e.g., robust

policy frameworks), structured trade union dynamics, and an existing digitised environment

that is fully equipped to manage the affairs of taxi driving compared to GS contexts. The lack

of a strong regulatory environment, structured trade union dynamic and digitised environment

in the taxi industry in GS contexts proliferates informality and limits drivers' bargaining power

compared to GN contexts (Heery and Abbott, 2000; Lomas, 2015; Igihe, 2019; Fobosi, 2019).

The lack of these factors in GS contexts like Nigeria has exacerbated taxi driver struggles since

Uber's emergence. This is because taxi drivers and regulatory bodies have no proper basis to

withstand the dynamics of ride-hailing platforms in GS contexts compared to GN contexts.

While the taxi industry and regulatory bodies have experienced difficulties regulating ride-

hailing platforms and protecting taxi drivers, they have better structures regarding these three

factors.

In GN contexts, I outline and discuss four core factors that define the emergence of

global ride-hailing platforms like Uber. This analysis was critical in first understanding the

relevance of the US, specifically San Francisco, in the emergence journey of Uber. I argue that

the innovative environment based on the early experiment of ride-hailing platforms, regulatory

battles in San Francisco, ease of access to venture capital, and the hype and rhetoric

surrounding Uber was critical. The critical factor here was the access to venture capital because

it facilitated the expansion of Uber within the US and globally, especially into other GS

contexts in 2013 (Brail 2020; Adler and Florida 2021). However, it was also a combination of

the innovative environment in San Francisco, which attracts the most capital funding in the US

and globally. The relevance of San Francisco brings validity and reliability for start-ups

because of pre-existing successes like Apple, Alphabet and others (Florida, 2017). Also,

because the US is a free market economy, the government has minimal interference in

structuring and funding a digital platform like Uber. However, these were not without any

regulatory battles.

Venture capitalism and the hype and rhetoric of Uber from the GN were particularly

critical factors towards the expansion to Nigeria and other GS cities. The year Uber launched

in 2014, the company acquired $18 billion in venture funding and had built a global image in

205 cities (Hoge 2014). Lagos, the former capital of Nigeria with about 21 million inhabitants,

274

is the centre of manufacturing, banking, fashion, innovation, and overall centre for businesses

to thrive, also serving as a motivating factor for Uber’s arrival in Lagos (Bloch et al., 2019;

Pilling, 2018). In other words, it was equally beneficial to Uber and Bolt because Lagos has

grown to become the innovative hub of Nigeria, with several start-ups and tech talent migrating

to the city (Kazeem, 2017b; Olawoyin, 2021). While the news surrounding the platform was

not all positive, the hype of creating jobs and the utopia of ordering a taxi by touching a

smartphone screen appealed to potential drivers and passengers. As Zukin (2020, p. 960), the

aim of tech platforms “is not only to create jobs, especially highly paid jobs that mayors desire

but to remake the city for a new modernity”. This phenomenon fits into the ideals of making

Lagos a modern mega city (Olajide et al., 2018), with technologies such as Uber that indicate

the growth and development of its local economy. Consequently, this enabled ride-hailing

platforms to impose their initiatives and solutions to rising unemployment, notably by

proposing high wages that supersede the minimum wage monthly of N18,000 (£49) and

demonstrates the flexibility of labour.

The taxi industry has been lagging and deficient in three aspects: regulatory policies, a

digitised environment, and formidable unions that have the freedom to bargain collectively.

According to interviews with a transport regulator (October 2018), the Lagos State government

has previously attempted to modernise the taxi system by adopting the Medallion taxi

Franchise system, where the licenses have tradeable values in the stock market, with practical

examples like the Red cabs, Metrocabs, Corporate cabs and other taxis. This phenomenon was

part of the neoliberal ideology based on how the government aimed to privatise these taxi

companies' ownership and failed because it did not fit the context. The lack of an innovative

environment in Lagos and the willingness for drivers to succumb to aspects of digitisation (e.g.,

use of taximeters, CCTV, e-payments) was farfetched. This is because taxi companies lacked

the business acumen, drivers lacked the technical-know-how and were elderly. More critically,

the taxi industry lacked proper laws to protect the rights of informal workers and, recently, gig

workers. Uber developed self-regulation based on how trips can be accessed, vehicle

requirements, aptitude tests and surveillance based on how drivers can be effectively managed

using algorithms to curate trips, facilitate payments, and incentivise work.

Finally, Uber's emergence facilitated the other ride-hailing platforms in Lagos.

Although Afrocabs and Easy taxi came into Lagos before Uber, there was already a global shift

in the taxi space from 2010 to 2014, including regulatory challenges in New York and other

GN contexts (Flores and Rayle, 2017; Bonini and Capizzi, 2019). While Uber and Bolt were

the main competitors, local platforms could not compete because of the lack of venture capital,

275

critical mass of users, and extortionate regulatory policies based on the franchise system by the

Nigerian government. This further demonstrated Uber's lobbying power against Lagos State

because it remained unclear how Uber continued to evade the franchise policy while putting

its drivers at risk of police arrests. All indications from discussions with NUPEDP signalled

the complicity of the Lagos state government regarding the operationalisation of Uber and Bolt

(Meagher, 2018), while local platforms without the resources to meet the demands of the

government fade out of the system.

The emergence of ride-hailing platforms in Lagos is a re-emergence or improvement

of existing gig work which is argued based on informal mobility means and, more specifically,

taxi labour. Arguably, GN cities are catching up to the gig work model in the GS by formalising

or digitising informal labour processes (Rekhviashvili and Sgbinev, 2019; Kaye-Essien, 2020),

which instead of being accountable to specific country contexts, are now accountable to digital

codes and innovators in founding platform contexts. Therefore, this redefines old problems,

creates new ones, and further institutes the need for platform unions that can operate across

space for the greater good of ride-hailing drivers in Lagos.

7.3. Research Question 2: How is the algorithmic management in the ride-

hailing sector of the gig economy impacting platform drivers in Lagos?

In chapters two and four, ride-hailing platforms emerging in GS cities appear as another

one-size-fits-all neoliberal project (Kaye-Essien, 2020), and these come with new ways of

managing labour, such as through algorithms, which are not entirely context-specific. Using

algorithmic management as a concept, as coined by Lee et al. (2015), has further been

developed by a few scholars such as Möhlmann and Zalmanson (2017); Katzenbach (2019),

Kellogg et al. (2020); Wood (2021) as workers interact with a system without human

intervention, thereby possessing capabilities of constant surveillance based on big data,

assigning gig work, automated unilateral decision-making, performance evaluation and

sanctioning or deactivations. Möhlmann and Zalmanson (2017) characterised platforms using

algorithmic management as opaque socio-technical processes because platform companies do

not disclose the hidden nature of algorithms, which leads to unfair labour decisions against

drivers (Rosenblat and Stark, 2016; Möhlmann and Zalmanson, 2017). Algorithms are

responsible for assigning tasks and overall direction of the labour process, the evaluation, and

276

disciplining of drivers who do not comply with the rules and regulations on the platform

(Wood, 2021, Kellogg et al., 2020)

Building on the definition of algorithmic management in chapter two, I outline opacities

(e.g., concealed information based on real secrecy), information asymmetries (e.g., one-sided

information) and biases (e.g., rider biases, i.e., preferential treatments for certain groups of

users) (Pasquale, 2015; Diakopoulos, 2015; Rosenblat and Stark 2016; Mittelstadt, 2016) in

chapter five. These elements are responsible for the impacts on drivers’ work in Lagos. As a

blanket term and ontological contribution, I refer to these three characteristics as algorithmic

burdens because of the hidden nature and one-sided management of the labour process leading

to unfair decisions of drivers. Therefore, I define algorithmic burdens as the interplay of

information asymmetries, opacities, and biases that limit gig workers from fully understanding

the terms of labour which thus may lead to impacts that affect the labour process and gig

workers. Chapter five provides substantial evidence on how the interactions of these

characteristics lead to arbitrary deactivations, data misrepresentations, driver manipulations

based on opaque incentives and price mechanisms and map limitations. More scholars such as

Grohmann et al. (2022) are starting to document the inherent nature of algorithmic burdens that

facilitate platform manipulations.

Because platforms originate in GN contexts such as the US, there is already a basis to

aid the integration of ride-hailing platforms into their system compared to GS contexts like

Lagos. For example, fare calculations based on distance and digital maps already existed in

GN contexts before now but are now adopted through an app in Lagos. These burdens of labour

lead to mismatched contextual realities or data misrepresentations that are causal factors of

drivers experiencing assaults, harassment, and death. For example, embedded platform maps

that are limited in accurately displaying the landscape of Lagos become a source for impacts

on drivers’ evaluation metrics such as their ratings, cancellation/acceptance scores, and activity

scores, particularly on Bolt. It is critical to note that these are also reinforced by the factors that

facilitated informality in the taxi industry, such as weak regulatory frameworks, lack of

dynamic trade unions, and lack of a digitised environment before Uber’s arrival. This

phenomenon was observed when drivers complained about the low barrier of entry for

passengers, which often do not verify their identities based on their biases, the lack of proper

address systems and the porous identification system for residents. It is important to note that

algorithmic burdens determine the level of impact a driver may experience. In other words,

algorithmic burdens of labour are the characteristics of algorithmic management that form the

basis for impacting drivers in the real world. For example, if algorithmic management were

277

transparent, inclusive, and non-biased, a driver would not encounter a deviant passenger that

was deactivated from the platform.

In chapter five, I also draw from Haggerty’s surveillant assemblage to outline the

hidden infrastructures that enable algorithms to control and manipulate workers (Haggerty and

Ericson, 2000; Newlands, 2020). I argue these are based on an integrated experience through

different socio-technical aspects of the app and extending outside of the app, which facilitates

interconnected layers of surveillance. I classify aspects extending outside of the app as external

modalities, i.e., they can function offline but are often reinforced by the information embedded

in the app. There is a disconnection between the findings of what is grouped as assemblages

and external modalities. For example, vehicle owners passively or actively rely on assemblage

components such as dashboard data to crosscheck if drivers are meeting their weekly targets.

This helps the vehicle owner corroborate or negate excuses from drivers who claim to be

working. The surveillant assemblage consolidates control of workers, indicating that ride-

hailing gigs are not precisely flexible and autonomous (Mazmanian et al., 2013; Wood et al.,

2019). Instead, these different aspects of surveillance and external modalities help reinforce

the power of algorithms through digital spaces, particularly in regaining control from counter-

resistant hidden strategies discussed in chapter six.

Researchers must interrogate the possible mismatch between algorithms and the

contextual realities, gig worker experiences, and culture in platform research. This is a similar

contribution for platform companies to develop algorithms that reflect the lived realities of gig

workers in places like Lagos. In addition, it is essential to note that these nuances vary across

different contexts with varying specificities that emerge from within them. It further reinforces

the call for human-centred algorithmic management, according to Lee et al. (2015), by

interrogating the need for algorithms to possess less power and control over workers.

7.4. Research Question 3: What are platform drivers' resistance strategies

against algorithmic impacts in Lagos?

Algorithmic management already facilitates agency in gig workers, such that

algorithmic burdens enable workers to develop strategies or practices that subvert or regain

control of the labour process. In applying the concept of everyday resistances, the conceptual

model demonstrates spatialisation and temporalisation approaches to investigate how these are

part of the public or hidden transcripts (Scott 1985;1989; Johansson and Vinthagen, 2016). For

278

Adnan (2007, p. 185), the transformation from covert to overt resistance for subordinate groups

"goes beyond the problematique of everyday resistance, as articulated by Scott (1985, 1989),

which does not explicitly address the question of how, and under what conditions, such

transformation can come about". Spatial and temporal dynamics are critical because ride-

hailing platform drivers do not possess physical office locations. Workers' identities and work

locations are embedded within digital spaces entwined with physical spaces throughout Lagos,

such as online via the ride-hailing app and when drivers work offline (Polleta, 1999; Briziarelli,

2019). While the core contribution in this thesis is on hidden resistances, it was critical to

highlight public forms of resistance because of the highly spatialised and temporal mobile

working environment in ride-hailing gig work.

In Lagos, I categorise drivers' public resistance in the form of online strikes, offline

protests, and public dissent through media engagements as a way of shaming the unfair

treatment by platform companies. Elsewhere according to Can (2020), Briziarelli (2019), and

Woodcock and Johnson (2020), there is an in-between space or third space that drivers

appropriate when logging in and logging out of delivery apps to reclaim their collective power.

Space here is political and ideological – platforms possess a higher degree of access or control

over platform drivers with limited access to space (Johansson and Vinthagen, 2016). In GN

contexts, when done collectively, these create a supply deficit of gig workers, which can lead

to high surge fares on ride-hailing platforms. This phenomenon either forces the passengers to

switch between competitors or abandon a potential trip which indirectly affects the profits of

platforms. In Lagos, logging out as a protest did not precisely limit jobs on Uber or Bolt

compared to GN contexts. This phenomenon is because drivers lack coordination and

organisation due to multiple worker collective groups and multiple subcontracted layers of

vehicle ownerships that require drivers to meet weekly payments. This phenomenon also

affected physical or locational protests at platform head offices of Bolt and Uber. However,

drivers did not receive positive responses concerning issues of arbitrary deactivations, price

mechanisms, robberies, deaths, and general opacity on platforms. While the public resistance

identified in this thesis can be critical in regaining control, the informal nature of taxi driving

preceding platforms, weak regulatory frameworks, and a lack of a digitised and enabling

environment limit public actions' effectiveness compared to GN cities. Therefore, it is no

surprise that drivers in Lagos engage in everyday hidden forms of resistance in their mobile

workplaces, which undermines the overarching ethos of platform companies.

Gradually, scholars are starting to document the hidden transcripts in places globally,

such as in China, the UK, Australia, and the US (Chen, 2017; Sun, 2019; Anwar and Graham,

279

2019; Veen et al., 2020). Lagos documents perspectives from the GS and, specifically, African

cities, which further shows why and how algorithms misrepresent such contextual realities. In

this study, drivers were involved in circumvention or gaming trips (e.g., offline trips, gaming

for incentives), sensemaking or what I classified as algorithmic literacy (e.g., acquiring

knowledge via SMCN), and manipulation (e.g., using the platform app or third-party app to

disrupt algorithms; prolong app updates to resist platform rules). To show similarity between

Scott's hidden practices in a physical environment, I link the hidden resistance of ride-hailing

drivers to false compliance and sabotage because of drivers' deliberate action to ignore rules

from algorithms or disrupt the system from within. Central to both hidden and public

resistances is algorithmic literacy, which is based on the knowledge drivers acquire and share

within closed or open SMCN groups to subvert algorithmic control, generate organised justice

(e.g., mobilising drivers to respond to passenger payment refusals), and integrate added

surveillance during offline trips (Möhlmann and Zalmanson, 2017; Jarrahi and Sutherland,

2018; Cheng and Foley, 2019).

Algorithms through machine learning techniques also develop literacies of workers'

resistance practices. Because these practices occur within digital spaces, hidden practices

eventually evolve into public realms aided by components of the surveillant assemblage, such

as riders who report unusual practices to platforms, ratings, platform gods eye surveillance and

other external modalities, including vehicle owners using trackers to monitor the behaviour of

drivers. This leads to the exposure of these practices in mainstream media such as radio stations

and updating the platform app to detect such hidden practices of drivers in Lagos. This

documentation is also aided by algorithms mutually learning the actions of drivers based on

machine learning and riders that detect these practices and report them to platform companies

which expose these actions; again, in the mainstream media such as radio stations. Today, a

hidden resistance practice can become public subsequently, leading to a continuous discovery

of new hidden resistance practices by ride-hailing drivers, embracing the flexibility and

plurality of adopting everyday resistances that demonstrate interconnectedness and

interdependencies based on the collapsing of digital and physical spaces as highlighted by

scholars (e.g., Adnan, 2007; Shalhoub-Kevorkian, 2012; Kinsley 2014).

While Scott (1985;1985) has highlighted possibilities of counter resistance by dominant

groups, he has not critically examined it within the context of emerging trends of

platformisation of labour, especially within the African context, which consequently

demonstrates that platforms can be part of the hidden transcripts, often through an extension of

its algorithms. This phenomenon can be categorised as a weapon of the dominant as opposed

280

to Scott's weapon of the weak because the burdens algorithms impose on workers are equally

manipulative and strategic towards subverting their sense of autonomy and flexibility. In other

words, platforms deliberately conceal information from public spaces by restricting knowledge

about platforms' operations, the kinds of data collected, and the implication of this data on their

everyday working lives. Evidence from the thesis highlights the use of third-party apps for

manipulating the Uber and Bolt apps, but this weapon of the dominant, i.e., algorithms, will

enable even more disruptive and invisible practices that could involve hacking the app for

material gains (cf. Wazid, 2013; Najafabadi and Domanski, 2018).

This phenomenon indicates a continuum of resistance and counter-resistance practices

between platforms and drivers, which I ontologically frame as digital everyday resistances.

Kinsley (2014) and Leszczynski (2015) would argue that this in-between space of power tussle

is entwined in multiple conjunctions of code, space/place and content based on the practices

and spaces of drivers' everyday life. In this case, algorithms and drivers co-constitute each

other, drawing from repertoires of resistances and counter-resistances that aid in understanding

the burdens of labour and drivers' agency (Kinsley 2014; Johansson and Vinthagen, 2016).

The limitation to fully understanding the complexities behind how algorithms control

drivers remain opaque, particularly from platform companies, due to the refusal to share data

with external researchers and based on competition with other platforms. Therefore, this calls

for more research in unpacking the layers that constitute platform and algorithmic counter-

resistance from the platform perspective in Lagos and globally.

7.5. Research Question 4: What are the implications of ride-hailing

platforms, algorithmic management, and drivers’ resistances for the

future of labour in the gig economy?

While digital labour platforms such as Uber and Bolt have been beneficial to many

platform workers in GS cities through temporary self-employment, research findings in this

thesis also indicate that the business model is increasingly unsustainable for platform workers.

The fact that taxi mobility and the driving workforce experienced slightly more freedom and

flexibility does not indicate that things were better in the past. Platforms help create data

repositories that can revamp the taxi industry and the entire mobility system in Nigeria. This

Uberised model can be replicated within indigenous informal modes such as the danfo, Okada,

traditional taxis and water transport which can improve the overall outlook of driving; demand

281

accountability from unions; develop robust regulatory frameworks; and limit corruption

practices. In some ways, this is already occurring with bike-hailing platforms such as Max.ng

and Gokada and other platforms, as mentioned previously in Chapter four. While these have

not been fully adopted, it demonstrates the influence of Uber in the Nigerian context. However,

much of this remains a paradox in the Nigerian context because ride-hailing platforms and

algorithmic management misrepresent the realities of drivers, indirectly limiting union

formulation and collective organisation. By so doing, drivers inherit the challenges of

traditional taxi drivers and experience new problems that come with driving for ride-hailing

platforms, including robbery, kidnap, assault, and loss of life, all because of better vehicles and

smartphones used on the job.

As much as the regulation is a challenge for platforms in Lagos, the current franchise

system appears to be rather short-term and profit-driven, making it difficult for indigenous

platforms to thrive. This fosters opacity between the state government and platform unions

based on drivers being systematically excluded from decisions affecting their labour. Also,

there is no clarity on whether dominant platforms pay these licenses following assertions from

interviewee Dipo and the media trends in Lagos. Considering this opacity, if the Lagos State

government creates a state-owned platform or integrates algorithms within the mobility

industry, this could lead to a modern-day panopticon that breaches data protection laws based

on pre-existing management of the taxi industry, which was also opaque for drivers. Further,

as has occurred more than once, drivers are exposed to harassment from law enforcement

officers when the state is attempting to clamp down on platforms refusing to pay for these

licenses or whenever there are disagreements between the state government and platform

companies. On the positive side, the implication is that the state government is starting to

develop regulatory policies that can be perfected progressively as GS cities grapple with

regulatory difficulties. For example, the Lagos State Government released an official policy

document in 2020, classifying platforms as service entities under the taxi system model.

However, this policy fails to recognise the managerial side of things that significantly impacts

drivers.

This onus remains on platform unions which are lobbying platforms to treat workers

fairly, reduce algorithmic burdens by being more inclusive in decision-making processes and

the Lagos State Government to develop an adequate regulatory framework for dominant

platform companies in Lagos.

Beyond the mobility sector and other aspects of gig work, scholars (cf. Dugan et al.,

2019; Gal et al., 2020; Jarrahi et al., 2021) are beginning to examine the transferability of

282

algorithmic management within standard employment relations. Findings demonstrate how

algorithms are critical to human resource management by effectively filtering applications

(Dugan et al., 2019; Leicht-Deobald et., 2019), effectively monitoring and firing workers based

on their speed or inactivity (e.g., Amazon Warehouse) (Dzieza, 2020; Gal et al., 2020), and

generally foster more algorithmic burdens within the workplace. However, Jarrahi et al. (2021,

p.4) reckon that algorithmic management cannot be perfectly translatable to standard

employment contexts because of people's connectedness to a traditional employment

arrangement and organisational hierarchies reflective of conventional coordination

mechanisms. Considering the still analogous working environment in Nigeria and many

African countries, this is particularly relevant because it would require both organisational

leadership and employees to upskill. Upskilling for youths will not be a problem because over

40.1% of first-degree holders were without jobs which indicates that the supply exceeds the

demand for labour (NBS, 2020; Nwokoma, 2021). Since the pandemic, overall unemployment

has increased to 33.3% or 23.2 million of about 70 million people in the labour force (NBS,

2020). Instead of permeating standard working environments, there would be an increase in

gig work platforms or platformisation of labour in Nigeria through adopting both location-

based gig work, and online-based gig work.

This can spiral into the terrains of algorithmic governance when cities adopt the

platform model using algorithms to rate and predict behaviours and choices of citizens for the

greater good (Katzenbach and Ulbricht, 2019). However, this is risky because many African

cities' political landscape and governance are primarily opaque and driven by one-sided

decision-making, forcing citizens to resist, adapt, and survive. If this power is transferred to

the governance of Lagos State, there would be limitations on improving transparency and

accountability. This is because, ultimately, algorithms would inherit pre-existing historical

burdens that would impact citizens' everyday lives because of unethical mass surveillance

(Norris and Armstrong 1999; Gal et al., 2021) of their livelihoods and well-being. This beckons

on why pioneer users of algorithmic management such as platforms need to ensure that

algorithmic burdens are not entirely transferred to gig workers but are shared by

acknowledging their voices in the decision-making process.

283

7.6. Methodological and Fieldwork Contributions

The opacity of algorithmic management in platforms reflects the difficulties in studying

them. Unlike traditional taxi services, researching platforms requires creativity from

researchers, as extensively discussed in the methodological chapter of this thesis. Below are

contributions to effectively study platforms, particularly within a GS context.

A. The importance of online observations: SMCN as a resource

This research contributes to the importance of social media and communication

networks (SMCN) as critical components in studying gig work. In the book Uberland,

Rosenblat (2018) outlined that joining social media communities was critical to the

operationalisation of platforms and the lives of ride-hailing drivers. Similarly, SMCN,

particularly Facebook platform groups, were critical for the recruitment of drivers and

understanding of the ride-hailing environment in Lagos. This contributes to the

constructivist paradigm of learning about different realities of ride-hailing drivers and

other participants within these shared spaces that contribute to platform discussions.

Besides joining platform groups, commenting, liking, and posting improve familiarity

and builds drivers' trust.

SMCN can serve as critical resources for researching black box society

(Pasquale, 2015), considering their hidden nature and the impacts of the Covid-19

pandemic on travelling and physical contact. Researchers can choose to recruit gig

workers outside of their geographical location. However, the researcher must critically

verify the identities of gig workers. For ride-hailing platforms, this can be done by

ensuring that drivers show the researcher the app dashboards of their daily activities if

possible. There is a possibility that specific drivers are unwilling to share such

information based on personal data privacy. However, these four things contribute to

ensuring the validity of participant recruitment.

i. Click on the profile of the potential participant to ensure they are not robots and

are actual drivers. The profile states the date the potential participant joined the

platform and, in some cases, the nature and number of posts they have

contributed to the group.

284

ii. Use the search tool to search for the participant's name to deduce their

authenticity from the information and frequency of posts. Researchers can

choose to search based on yearly queries, which the Facebook algorithm

affords.

iii. Ensure to send participant information sheets and consent form copies to

participants that show interest, primarily via direct messages.

iv. Engage in an informal conversation by requesting their mobile numbers to

schedule interview dates. This informal conversation enables the researcher to

assess the knowledge of potential participants about the job.

While these are critical in validating the identities of potential participants, the

researcher should also rely on other data collection methods to triangulate information.

Alternatively, SMCN (e.g., WhatsApp and Telegram) can be created for a few

drivers (two to five) for short periods in addition to recording screenshots,

conversations, and the decision-making process of drivers. This would also give critical

insights on other hidden resistances drivers use in subverting algorithmic control.

B. Compensation as a Service.

Compensation has widely been successful as a motivating factor for research

discourses. However, participants' payment for research purposes has led to ethical

concerns and coercion of researchers to offer information for money or incentives

(Head, 2009; Grady, 2019). The difficulty in recruiting platform company workers or

representatives makes drivers critical resources based on their data. Drivers are

primarily entangled in precarious vehicle contracts that prevent them from making

minimum or decent living wages without impacting their well-being. Therefore, the

idea of adequately compensating drivers creates an entry for researchers by

compensating for their time and potential trip engagements.

This research contributes by acknowledging that posting recruitment messages

on SMCN driver groups motivates driver participants in GS cities. Researchers should

post recruitment messages stating that "participants would be duly compensated" and

not the exact sum in the first instance. If participants do not respond adequately, a

285

follow-up message should state the sum measured according to the minutes in the

interview. It may be necessary to cap the number of participants to prevent infeasibility

in recruitment and conducting interviews.

Findings from this study depict that interviewees in Lagos prefer cash payments

compared to incentives (Head, 2009) because of the freedom of usage. With mobile

banking apps, researchers can compensate drivers with detailed narratives as a hybrid

for cash and incentives in transport, data and feeding costs. Detailed narratives are

critical, invalidating the reason for these costs and the importance of research. It is

difficult to check if drivers utilise compensation for the above reasons. This highlights

the need for interviewees to possess more autonomy in utilising compensation costs as

they choose.

C. Gamification-from-within: Mobile methodologies

Gamification-from-within is indicative of the strategic role of this researcher in

selecting trips to observe drivers closely. Broadly, this contributes to virtual

ethnographies (Hine 2008; Murthy 2008), where the researcher digitally immerses

themselves within the context of the platform and the city to develop literacies in

understanding digital gig work. As highlighted in the methodological chapter, this

thesis contributes to research utilising mobile technologies as part of the new mobility

paradigm (Urry, 2006; Evans and Jones, 2011) and emerging mobility methodologies.

Rosenblat (2018) also identified ride-hailing platform vehicles as conduits for

interviewing and recruiting platform drivers. Aside from interviewing drivers on a trip,

observing their actions on trips is critical in unpacking how drivers make decisions and

respond to algorithmic management. The researcher deliberately becomes a tourist by

trying to understand the internal processes of ride-hailing drivers in Lagos. This could

develop into perceiving the city as a means to an end, whereby the researcher genuinely

creates an itinerary of places to visit using ride-hailing platforms. In addition,

researchers can replace everyday mobility within the city with ride-hailing platforms

instead of personal mobility or public transportation modes. This can be cost-intensive,

which is why researchers may need to cap the number of trips to a feasible number.

286

7.7. Recommendations

Based on the global advocacy for fair treatment of gig workers, developing robust

regulatory frameworks, and a call for ethical and transparent algorithmic management, I outline

recommendations for unions, government bodies (including transport authorities), and

platform companies. While these recommendations are deduced from the Lagos case, they also

apply to other GS cities and digital platforms outside the mobility sector.

7.7.1. Recommendations for Platform Companies in GS Cities

1. International platform companies emerging in Lagos possess the power to improve

transparency and accountability in the taxi industry, mobility sector and gig work in

general. They need to programme algorithms according to the specificity of the city,

the users and the culture of driving in Lagos. Platform companies must realise that

algorithmic management should be embedded in the city and other emerging realities,

thereby ensuring that the error margin for drivers is inherently fluid.

2. Besides learning from international platforms, indigenous platforms need to invest in

research and development that empowers them to improve the operationalisation of

algorithms. This would equip indigenous platforms with the preparedness to capitalise

on the loopholes of international platforms. Subsequently, this would attract funding

that can boost the competition against international platforms.

3. As the findings suggest, tailoring algorithms to reflect the nuances of the city is

particularly not successful for indigenous platforms if there is no radical investment in

marketing. Indigenous platforms must invest in marketing to improve visibility and

adoption by both users and drivers.

4. Platforms must integrate initiatives that boost the fairness of ride-hailing workers

contextually and globally. For example, the Fairwork Foundation utilises five

principles: Fair pay, Fair conditions, Fair contracts, Fair management, and Fair

representation to rate platforms over a score of ten, based on platform policies and

interviews with drivers, which is a useful initiative to adapt. Platform companies should

progressively aim for a perfect score to reflect their commitment to improving gig work

conditions.

287

5. Critically platforms need to recognise platform unions and collectivism as critical

aspects in shaping policies and decisions that affect the labour process, which is

reflexive of the fair representation principle by the Fairwork Foundation.

7.7.2. Recommendations for State Governments and Transport

Authorities

1. If platform companies evade data access, ministry representatives should invest in

studying and researching platforms. This can be done by developing a simulation of the

Uber model to understand the inner workings of algorithms. Participants in this study

should be duly compensated for the period of study. The result from generating this

data presents city governments with concrete evidence that would assist in developing

critical legal frameworks and policies that safeguard the affairs of ride-hailing drivers

and other gig workers.

2. Alternatively, the state can partner or contribute to research initiatives like the Fairwork

project to provide insights into the platformisation of labour in GS cities. This includes

providing networks that facilitate access to platform representatives and developing

workshops for knowledge creation and sharing.

3. Recent attempts to revamp traditional taxis using technology and further introduce

state-owned platforms or Ekocabs have slowly progressed. Instead of developing state-

owned platforms, which reduce trust from citizens, state ministries should support

indigenous platforms by creating an enabling environment for them to thrive. These

include:

I. Providing access to funding for indigenous platform companies.

II. Introducing accessible loan schemes for emerging platform companies that can

facilitate scalability and generate revenue for the state.

III. Introduce debt relief funds that can assist struggling indigenous platform

companies to recover from financial difficulties.

4. The introduction and improvement of Lagos State Drivers Institute (LASDRI) are

commendable in ensuring professional training of drivers. However, the state should

288

improve supervision and frequent training of drivers. This can help improve the

adoption of platforms by enabling drivers to understand and embrace technology.

5. State governments and policymakers should be more inclusive of platform unions in

decision-making processes and the development of regulations that affect ride-hailing

gig work.

7.7.3. Recommendations for Platform Unions.

1. Compared to GN cities with multiple platform unions that succeed in fighting platform

companies and facilitating regulations, GS cities do not possess formalised and robust

legal frameworks to facilitate similar results. Platform unions in Lagos need to work

together under one umbrella. This will improve their lobbying abilities to state transport

ministries and platform companies.

2. Besides learning from unions in GN cities, platform unions can contribute to initiatives

that advocate improving fairness in the gig economy by providing critical information

and attending workshops and conferences that boost algorithmic literacies.

3. Platform unions should develop a database that keeps records of union activities,

members, and discrepancies with platform companies. This creates another layer for

drivers to penetrate the opacity of algorithms by recalling records from past

experiences.

7.8. Future Research – Gig Work and Algorithmic Management in 30

Years

Following the Covid-19 pandemic, more people are starting to subscribe to the idea of

online or location-based gig work based on the utopia of autonomy and flexibility of labour.

This is a critical reason why dominant platforms such as Uber need to ensure that the codes

behind algorithmic management are less burdensome for gig workers by limiting their ability

to inherit historical biases, opacity, and information asymmetries. This is even more critical for

GS cities based on the disparity in innovative start-ups compared to GN cities. For instance,

289

considering Uber is seeking to integrate driverless vehicles in GN cities, many GS cities may

still be left behind. Even if cities such as Lagos develop robust policies to cope with the rise

in gig work and algorithmic management, they need to start developing platforms by and for

these contexts with codes that demonstrate the specificities of gig workers.

Speculatively, if Uber adopts driverless vehicles, they can fully embrace the framing of

being a technology platform and not a transport company. This will further proliferate opaque

algorithms and unfair practices. Also, with the announcement of the Metaverse, these ideals of

precarity could spread to traditional working environments by merging the digital and physical

worlds under the control of algorithms and surveillance assemblages.

This further indicates that researching these platforms becomes opaque and one-sided,

i.e., based on consumers' perspectives. Therefore, researchers should consider developing

simulation apps that investigate these platforms from within. Alternatively, apps that empower

drivers and gig workers alike to keep track of the labour process or practice sousveillance, with

open ethical data repositories facilitating ease in studying platforms. Whatever specific

technological developments come to pass, the platformisation of labour in GS cities is likely

to continue, potentially colonising all forms of informal and formal labour. These will impact

global development, sustainability, and livelihoods for workers based on disproportionate

policies between GS and North cities. These phenomena must not be neglected in GS cities

just because they are difficult to research.

290

References

Aarhaug, J. (2016). Taxis as a part of public transport, Deustsche Gessellschaft fur

Internationale Zusammenarbeit (GIZ). Bonn, Germany. (Online) Available at:

http://sutp.org/files/contents/documents/resources/B_Technical-

Documents/GIZ_SUTP_TD16_Taxi_EN.pdf [Accessed March 2, 2018]

Aarhaug, J. and Skollerud, K. (2014). Taxi: different solutions in different segments.

Transportation Research Procedia, 1(1), pp.276–283. [online]. Available from:

http://dx.doi.org/10.1016/j.trpro.2014.07.027.

Abuja, C. (2017). NLC faults sack of 30,000 Kaduna workers by El-Rufai | The Guardian

Nigeria News - Nigeria and World News. The Guardian Nigeria News - Nigeria and World

News. [online]. Available from: https://guardian.ng/news/nlc-faults-sack-of-30000-kaduna-

workers-by-el-rufai/ [Accessed January 5, 2020].

Adam, B. (2003). When time is money: contested rationalities of time in the theory and practice

of work. Theoria: A Journal of Social and Political Theory, 102, 94–125.

http://www.jstor.org/stable/41791393

Adas, M. (1986). From Foot dragging to Flight: The Evasive History of Peasant Avoidance

Protest in South and Southeast Asia. The Journal of Peasant Studies, 13(2), pp.64–86.

Adedun, E. (N.D). "CARRY GO”: A Sociolinguistic Study of The Language of Bus

Conductors in Lagos (Nigeria). University of Lagos [online]. Available from:

https://ir.unilag.edu.ng/handle/123456789/3065 [Accessed March 16, 2019].

Adepetun, A. (2017). Nigeria adopts new addressing system, NIPOST sets 2020 target. The

Guardian. [online]. Available from: https://guardian.ng/news/nigeria-adopts-new-addressing-

system-nipost-sets-2020-target/ [Accessed May 28, 2020].

Adler, P. and Florida, R. (2021). The rise of urban tech: how innovations for cities come from

cities. Regional Studies, 55(10-11), pp.1787 –1800.

291

Adnan, S. (2007). Departures from everyday resistance and flexible strategies of domination:

the making and unmaking of a poor peasant mobilization in Bangladesh. Journal of Agrarian

Change, 7(2), pp.183–224.

AFDB. (2021). Nigeria economic outlook. African Development Bank - Building today, a

better Africa tomorrow. [online]. Available from: https://www.afdb.org/en/countries-west-

africa-nigeria/nigeria-economic-outlook [Accessed February 27, 2021].

Agbiboa, D., 2020. Between cooperation and conflict: the national union of road transport

workers in Lagos, Nigeria. Crime, Law and Social Change, 73(5), pp.605-622.

Agbiboa, D.E. (2018). Conflict analysis in world class cities: urban renewal, informal transport

workers, and legal disputes in Lagos. Urban Forum, 29(1).

Agbiboa, D.E. (2017). Mobile bodies of meaning: city life and the horizons of possibility.

Journal of Modern African Studies, 55(3), pp.371–393.

Agbiboa, D.E. (2016). No condition is permanent’’: informal transport workers and labour

precarity in africa’s largest city. International Journal of Urban and Regional Research, 40(5),

pp.936 – 957.

Aglietta, M. (2000). A theory of capitalist regulation. London: Verso.

Ajunwa, I. (2019). Age discrimination by platforms. Berkeley Journal of Employment and

Labor Law, 40(1), pp.3 – 28.

Akere, F. (1982). Language use and language attitudes in a yoruba suburban town: a

sociolinguistic response to the factors of traditionalism and modernity. Anthropological

Linguistics, 24(3), pp.344-362.

292

Albert, I. O. (2007). NURTW and the politics of motor parks in Lagos and Ibadan. In L.

Fourchard (Ed.), Gouverner les villes d’Afrique. Etat, gouvernement local et acteurs privés.

Paris: Karthala.

Albert, I.O. (2005). Explaining godfatherism in Nigerian. , (2), pp.79 –105.

Aliu, M. (2015). Labour movements in the GS and the struggles against neoliberal

globalisation: the Congress of South African Trade Unions ( Cosatu ) Example. , 11(28),

pp.461–471.

Allmark, P. et al. (2009). Ethical issues in the use of in-depth interviews: literature review and

discussion. Research Ethics, 5(2), pp.48–54.

Aloni, E. (2016). Pluralizing the sharing economy. Washington Law Review, 91(4), pp.1398 –

1458.

Amorim, H. and Moda, F. (2020). Work by app: algorithmic management and working

conditions of uber drivers in Brazil. Work Organisation, Labour and Globalisation, 14(1),

pp.101–118.

Andrejevic, M. and Gates, K. (2014). Big data surveillance: Introduction. Surveillance and

Society, 12(2), pp.185–196.

Anwar, M.A. and Graham, M. (2019). Hidden transcripts of the gig economy: labour agency

and the new art of resistance among African gig workers, Environment and Planning A, 52(7),

pp. 1269–1291.

Aneesh, A. (2009). Global labor: algocratic modes of organization. Sociological Theory, 27(4),

pp.347–370.

Arubayi, D., (2021a). Be your own boss: examining algorithmic burdens on ride-hailing

platform drivers in Lagos. [Blog] The digital Inequality Group, Oxford Internet Institute,

University of Oxford, Available at: <https://dig.oii.ox.ac.uk/2021/05/18/be-your-own-boss-

293

examining-algorithmic-burdens-on-ride-hailing-platform-drivers-in-lagos/> [Accessed May

23, 2021].

Arubayi, D. (2021b). Documenting the Everyday Hidden Resistance of Ride-Hailing Platform

Drivers to Algorithmic Management in Lagos, Nigeria. South Atlantic Quarterly, 120(4),

pp.823-838.

Arubayi, D. (2020). Ride-hailing Platform Asymmetries between Riders and Driver-partners

in Lagos, Nigeria. ICTs for Development.

Atolagbe, M. (2016). South Africa’s zebra cabs has raised $21.5 million from futuregrowth to

expand its metered cab service | TechCabal. [online]. Available from:

https://techcabal.com/2016/11/29/sa-zebra-cabs-raised-21-5-million-from-futuregrowth/

[Accessed August 6, 2020].

Ayamga, E. (2021). Uber and Bolt drivers hit the streets to protest over poor fares. Pulse

Ghana. [online]. Available from: https://www.pulse.com.gh/news/local/uber-and-bolt-drivers-

hit-the-streets-to-protest-over-poor-fares/8nphkvg [Accessed July 29, 2021].

Bain, P. et al. (2002). Taylorism, targets and the pursuit of quantity and quality by call centre

management. New Technology, Work and Employment, 17(3), pp.170–185.

Bain, P. and Taylor, P. (2000). Entrapped by the ‘electronic panopticon’? worker resistance in

the call centre. New Technology, Work and Employment, 15(1), pp.2–18.

Ball, K. (2010). Workplace surveillance: an overview. Labor History, 51(1), pp.87–106.

Baker, L. (2006). Observation: a complex research method. Library Trends, 55(1), pp.171 –

189. [online]. Available from: https://muse.jhu.edu/article/202353 [Accessed July 19, 2020].

Barret, J. (2003). Organizing in the informal economy: a case study of the minibus taxi industry

in South Africa. Geneva: International Labour Organisation. [online]. Available from:

https://www.ilo.org/wcmsp5/groups/public/---ed_emp/---emp_ent/---

ifp_seed/documents/publication/wcms_117698.pdf [Accessed May 4, 2019].

294

Batty, M. and Miller, H. (2000). Representing and visualizing physical, virtual and hybrid

information spaces. In J. Donald & H. David, ed. Information, Place, and Cyberspace. Berlin,

Heidelberg: Springer, pp. 133 –146.

Bayat, A. (2000). From dangerous classes to quiet rebels. International Sociology, 15(3),

pp.533 – 557.

Bayat, A. (1997). Un-civil society: the politics of the informal people. Third World Quarterly,

18(1), pp.53–72.

BBC. (2019). Uber’s expansion into Africa. BBC News. [online]. Available from:

https://www.bbc.co.uk/news/av/business-47518342 [Accessed July 10, 2020].

Bigo, D. (2008). Terror, insecurity and liberty: illiberal practices of liberal regimes after 9/11.

1st ed. London: Routledge.

Bigo, D. (2006). Security, Exception, ban and Surveillance. In D. Lyon, ed. Theorising

surveillance: The panopticon and beyond. Portland: Willan Publishing, pp. 97 – 122.

Bellafante, G. (2018). A driver’s suicide reveals the dark side of the gig economy (Published

2018). Nytimes.com. [online]. Available from:

https://www.nytimes.com/2018/02/06/nyregion/livery-driver-taxi-uber.html [Accessed May

22, 2022].

Beneito-Montagut, R. (2011). Ethnography goes online: towards a user-centred methodology

to research interpersonal communication on the internet. Qualitative Research, 11(6), pp.716–

735.

Burrell, J. (2016). How the machine ‘thinks’: Understanding opacity in machine learning

algorithms. Big Data & Society, 3(1), p.205395171562251.

295

Beer, D. (2017) 'The social power of algorithms', Information Communication and Society.

Taylor & Francis, 20(1), pp. 1–13. doi: 10.1080/1369118X.2016.1216147.

Bergantino, A. and Longobardi, E. (2000). The drawbacks of deregulation in the taxi market:

evidence from the international scenario and the Italian experience. Urban Transport, pp.86-

94.

Birhane, A., (2020). Algorithmic Colonization of Africa. SCRIPT-ed, 17(2), pp.389 –409

Bhabha, H., (1994). The location of culture. London: Routledge.

Bernard, H. (2006). Research methods in anthropology. Lanham, Md.: AltaMira Press.

Bloch R., Fox S., Monroy J., and Ojo A. (2015). urbanisation and urban expansion in nigeria.

urbanisation research nigeria (URN) Research Report. London: ICF International. Creative

Commons Attribution-Non-Commercial-ShareAlike CC BY-NC-SA

Block, E.S. and Erskine, L. (2012). Interviewing by telephone: specific considerations,

opportunities, and challenges. International Journal of Qualitative Methods, 11(4), pp.428–

445.

Bokányi, E. and Hannák, A. (2020). Understanding inequalities in ride-hailing services through

simulations. Scientific Reports, 10(1).

Bonini, S. and Capizzi, V. (2019). The role of venture capital in the emerging entrepreneurial

finance ecosystem: future threats and opportunities. Venture Capital, 21(2-3), pp.137 – 175.

Bornstein, J. (2015). The great Uber fairness fallacy: as a driver, how do you bargain with an

app?. the Guardian. [online]. Available from:

https://www.theguardian.com/technology/2015/aug/24/uber-fairness-independent-

contractors-employees-rights [Accessed February 4, 2016].

296

Brail, S. (2020). World cities of ride-hailing. Urban Geography, 43(1), pp.12-33.

Branstetter, L., Glennon, B. and Jensen, J. (2018). The IT Revolution and the Globalization of

R & D. Innovation Policy and the Economy, 19, pp.1 –37.

Bray, M. and Underhill, E. (2009). Industry differences in the neoliberal transformation of

Australian industrial relations. Industrial Relations Journal, 40(5), pp.372–392.

Braun, V. and Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research

in Psychology, 3(2), pp.77–101.

Brenner, N. and Theodore, N. (2005). Neoliberalism and the urban condition. City, 9(1), pp.101

–107.

Briziarelli, M. (2019). 'Spatial politics in the digital realm: the logistics/precarity dialectics and

Deliveroo's tertiary space struggles', Cultural Studies. Taylor & Francis, 33(5), pp. 823–840.

doi: 10.1080/09502386.2018.1519583.

Briterbridges.com. (2018). eTransport Companies innovation maps 2018. Available at:

https://briterbridges.com/innovation-maps [Accessed July 5, 2019]

Bryceson, D. and Potts, D. (2006). Viability, vitality or vitiation?. In D. Bryceson & D. Potts,

ed. African Urban Economies. London: Palgrave Macmillan London.

Bruner, J. (2019). Acts of meaning. Cambridge, MA: Harvard University Press

Butcher, L. (2016). Taxi and private hire vehicle licensing in England and Wales. London:

House of Commons Library, pp.4–11. (Online). Available at:

http://researchbriefings.parliament.uk/ResearchBriefing/Summary/SN02005#fullreport

[Accessed Febraury 8, 2018]

Butler, S. (2021a). Uber drivers entitled to workers' rights, UK supreme court rules. the

Guardian. [online]. Available from:

297

https://www.theguardian.com/technology/2021/feb/19/uber-drivers-workers-uk-supreme-

court-rules-rights [Accessed May 19, 2022].

Butler, S., (2021b). Uber agrees union recognition deal with GMB. [online] the Guardian.

Available at: <https://www.theguardian.com/business/2021/may/26/uber-agrees-historic-deal-

allowing-drivers-to-join-gmb-union> [Accessed 30 August 2021].

Cahill, D. (2010). Actually existing neoliberalism and the global economic crisis. Labour &

Industry: a journal of the social and economic relations of work, 20(3), pp.298–316.

Cant, C. (2020). The warehouse without walls: a workers’ inquiry at deliveroo. Ephemera:

Theory & Politics in Organization, 20(August 2016), pp.131–161. [online]. Available from:

http://www.ephemerajournal.org/sites/default/files/pdfs/contribution/20-4Cant.pdf.

Cant, C. and Woodcock, J. (2020). Fast food shutdown: from disorganisation to action in the

service sector. Capital and Class, 44(4), pp.513–521.

Caliandro, A. (2018). Digital methods for ethnography: analytical concepts for ethnographers

exploring social media environments. Journal of Contemporary Ethnography, 47(5), pp.551–

578. [online]. Available from: https://doi.org/10.1177/0891241617702960.

Carmody, P. and Fortuin, A. (2019). “Ride-sharing”, virtual capital and impacts on labor in

Cape Town, South Africa. African Geographical Review, 38(3), pp.196 –208.

Carty, V. (2015). Social Movements and New Technology. 1st ed. New York: Routledge.

Clarke, R. (1988). Information technology and dataveillance. Communications of the ACM,

31(5), pp.498 –512.

Cervero, R. and Golub, A. (2007). Informal transport: A global perspective. Transport Policy,

14(6), pp.445–457.

Cervero, R. (2000). Informal transport in the developing world. Nairobi: United Nations

Centre for Human Settlements (Habitat).

298

Chan, N., (2019). The rating game: the discipline of uber's user-generated ratings. Surveillance

& Society, 17(1/2), pp.183 –190.

Chan, N.D. and Shaheen, S.A. (2012). Ridesharing in north America: past, present , and future.

, 1647(January), pp.37–41.

Chen, J., (2017). Thrown under the bus and outrunning it! the logic of didi and taxi drivers’

labour and activism in the on-demand economy, New Media & Society, 20(8), pp.2691-2711.

Chen, M. (2012a). The informal economy: definitions, theories and policies. Hks.harvard.edu.

[online]. Available from: https://www.hks.harvard.edu/publications/informal-economy-

definitions-theories-and-policies [Accessed July 30, 2019].

Chen, B. (2012b). A feisty start-up is met with regulatory snarl. Nytimes.com. [online].

Available from: https://www.nytimes.com/2012/12/03/technology/app-maker-uber-hits-

regulatory-snarl.html [Accessed April 14, 2022].

Cheng, M. and Foley, C. (2019). Algorithmic management: the case of airbnb. International

Journal of Hospitality Management, 83(April), pp.33–36.

Cheng, M. (2021a). The UK supreme court ruled that uber drivers are not self-

employed. Quartz. [online]. Available from: https://qz.com/1974912/uber-drivers-are-not-

self-employed-uk-supreme-court-rules/ [Accessed April 20, 2021].

Cheng, M. (2021b). Uber drivers in Africa are taking notes from their triumphant British

peers. Quartz. [online]. Available from: https://qz.com/africa/1997485/africa-uber-drivers-

take-note-of-wins-by-uk-drivers/ [Accessed June 4, 2021].

Chin, C.B.N. and Mittelman, J.H. (1997). Conceptualising resistance to globalisation. New

Political Economy, 2(1), pp.25–37.

299

Chin, M. (2008). Facebook for iPhone. [online] Facebook.com. Available at:

https://www.facebook.com/notes/facebook/facebook-for-iphone/22389032130/ [Accessed

March 12, 2018].

Chonghaile, C. (2015). Road safety initiative helps Rwanda's motorbike taxis stay on the right

track | Clar Ni Chonghaile. The Guardian. [online]. Available from:

https://www.theguardian.com/global-development/2015/dec/28/safemotos-rwanda-

motorbike-taxis-road-safety [Accessed October 5, 2017].

Clawson, D. and Clawson, M.A. (2017). IT Is Watching: workplace surveillance and worker

resistance. New Labor Forum, 26(2), pp.62–69. [online]. Available from:

https://doi.org/10.1177/1095796017699811.

Cohen, P. et al. (2016). Using big data to estimate consumer surplus: the case of Uber. [online].

Available from: http://www.nber.org/papers/w22627.pdf.

Coleman, D. (2019). Digital colonialism: the 21st century scramble for Africa. Michigan

Journal of Race and Law, 24. [online]. Available from:

https://repository.law.umich.edu/mjrl/vol24/iss2/6/.

Conner-Simons, A. (2017). How ridesharing can improve traffic, save money, and help the

environment. MIT News | Massachusetts Institute of Technology. [online]. Available from:

https://news.mit.edu/2016/how-ride-sharing-can-improve-traffic-save-money-and-help-

environment-0104 [Accessed May 16, 2022].

Constantiou, I. and Kallinikos, J., (2015). New games, new rules: big data and the changing

context of strategy. Journal of Information Technology, 30(1), pp.44 – 57.

Constine, J. (2012). Hail a fellow human, not a taxi with Sidecar - The new P2P Uber.

Techcrunch.com. [online]. Available from: https://techcrunch.com/2012/06/26/sidecar/

[Accessed June 27, 2022].

Conneran, L., Bridgett, H. and Randolph, M. (2017). Shared Rides – City and California Public

Utilities Commission Perspectives. In League of California Cities. California. [online].

300

Available from: https://www.cacities.org/Resources-Documents/Member-

Engagement/Professional-Departments/City-Attorneys/Library/2017/Spring-Conf-

2017/Conneran-SharedRides-CityAndCaliforniaPublicUtilit.aspx [Accessed March 18, 2018].

Corlett, S. and Mavin, S., (2018). Reflexivity and researcher positionality. The SAGE

handbook of qualitative business and management research methods, pp.377 –399.

Cramer, J. and Krueger, A. (2016). Disruptive change in the taxi business: the case of Uber.

American Economic Review, 106(5), pp.177 –182.

Creswell, J. (2009). Research design. 3rd ed. London: Sage.

Creswell, J. et al. (2007). Qualitative research designs: selection and implementation. The

Counseling Psychologist, 35(2), pp.236–264.

Crotty, M. (1998). The foundations of social research: meaning and perspective in the research

process. 1st ed. London: Sage publications Ltd.

Dahir, A. (2018). Uber’s innovations in Africa have helped shape its global operations. Quartz.

[online]. Available from: https://qz.com/africa/1368732/uber-in-africa-cash-payments-are-

shaping-global-operations/ [Accessed June 30, 2019].

David-West, O. and Evans, P. (2016). The rise of African platforms: A Regional Survey. The

Center for Global Enterprise.

Deleuze, G. (1992) 'Postscript on the societies of control: The MIT Press Stable URL:

https://www.jstor.org/stable/778828 Postscript on the Societies of Control *', 59, pp. 3–7.

Deleuze, G. and Guattari, F., 1987. A thousand plateaus: capitalism and Schizophrenia.

London: Athlone.

301

Denzin, N. and Lincoln, Y. (2021). The discipline and practice of qualitative research. In N.

Denzin & Y. Lincoln, ed. The Sage handbook of qualitative research. Sage publications, pp. 1

–21.

DFT. (2021). Taxi and private hire vehicle statistics, England: 2021.

Assets.publishing.service.gov.uk. [online]. Available from:

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data

/file/997793/taxi-and-private-hire-vehicle-statistics-2021.pdf [Accessed April 27, 2022].

de Certeau, M. (1984). The practice of everyday life. Berkeley: University of California Press.

de Greef, K. (2018). Driving for Uber when you can’t afford a car. The Atlantic. [online].

Available from: https://www.theatlantic.com/business/archive/2018/09/uber-south-

africa/567979/ [Accessed September 17, 2019].

de Haas, R. (1998). Vizcaino v. Microsoft. Berkeley Technology Law Journal, 13(1), pp.483 -

499.

de Souza Silva, L., de Andrade, M. and Alves Maia, M. (2018). How does the ride-hailing

systems demand affect individual transport regulation?. Research in Transportation

Economics, 69, pp.600 –606.

Leicht-Deobald, U. et al. (2019). The challenges of algorithm-based hr decision-making for

personal integrity. Journal of Business Ethics, 160(2), pp.377 –392.

Dickson-Swift, V. et al. (2008). Risk to researchers in qualitative research on sensitive topics:

issues and strategies. Qualitative Health Research, 18(1), pp.133 –144.

Dobraszczyk, P. (2008). Useful reading? Designing information for London’s Victorian cab

passengers, Journal of Design History, 21(2), pp. 121–141.

302

D'Onfro, J. (2014). Uber says its drivers are making $75,000 - $90,000 a year. Business Insider.

[online]. Available from: https://www.businessinsider.com/uber-drivers-salary-90000-2014-

5?r=US&IR=T [Accessed June 20, 2022].

Dudley, G., Banister, D. and Schwanen, T. (2017). The rise of Uber and regulating the

disruptive innovator, Political Quarterly, 88(3), pp. 492–499, (Online). Available at: doi:

10.1111/1467-923X.12373 [Accessed December 2, 2017]

Dubal, V., Collier, R. and Carter, C. (2018). Disrupting Regulation, Regulating Disruption:

The Politics of Uber in the United States. Perspectives on Politics, 16(4), pp.919-937.

Duggan, J. et al. (2020). Algorithmic management and app-work in the gig economy: a

research agenda for employment relations and HRM, Human Resource Management Journal,

30(1), pp. 114–132. doi: 10.1111/1748-8583.12258.

Dunn, K. (2005). Interviewing. In I. Hay, ed. Qualitative Research Methods in Human

Geography. Melborne: Oxford University Press, pp. 79 - 100.

Dzieza, J. (2020). Robots aren’t taking our jobs — they’re becoming our bosses. The Verge.

[online]. Available from: https://www.theverge.com/2020/2/27/21155254/automation-robots-

unemployment-jobs-vs-human-google-amazon [Accessed June 1, 2020].

Ebenezer, A. (2018). The many colours of the Nigerian Local Taxis. The Guardian. [online].

Available from: https://guardian.ng/life/the-many-coat-of-colours-of-the-nigerian-local-taxis/

[Accessed June 28, 2019].

Edelman, B. (2017). Uber can’t be fixed — it’s time for regulators to shut it down. Harvard

Business Review. [online]. Available from: https://hbr.org/2017/06/uber-cant-be-fixed-its-

time-for-regulators-to-shut-it-down [Accessed April 18, 2022].

303

Ehebrecht, D., Heinrichs, D. and Lenz, B. (2018). Motorcycle-taxis in sub-Saharan Africa:

current knowledge, implications for the debate on “informal” transport and research needs.

Journal of Transport Geography, 69(May), pp.242–256.

Ekanade, O. (2014). The dynamics of forced neoliberalism in Nigeria since the 1980s. The

Journal of Retracting Africa, 1(1), pp.1-24. [online]. Available from:

https://encompass.eku.edu/jora/vol1/iss1/1 [Accessed November 4, 2020].

Elebeke, E. (2016). Lagos leads in latest active voice, internet users' survey. [online] Vanguard

News. Available at: https://www.vanguardngr.com/2016/05/lagos-leads-latest-active-voice-

internet-users-survey/ [Accessed May 21, 2018].

Elliot, L. (2021). Membership of UK trade unions rises for fourth year in a row. The Guardian.

[online]. Available from: https://www.theguardian.com/politics/2021/may/27/membership-of-

uk-trade-unions-rises-for-fourth-year-in-a-row [Accessed March 27, 2022].

Enwukwe, N. (2021). The employment status of nigerian workers in the gig economy: using

Uber as a case study. Journal of Law, Policy and Globalization, 107, pp.1–10. [online].

Available from: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3818696 [Accessed June

5, 2021].

Escobar, A. et al. (1994). Welcome to cyberia: notes on the anthropology of

cyberculture. Current Anthropology, 35(3), pp.211-231. [online]. Available from:

https://www.jstor.org/stable/2744194 [Accessed February 3, 2019].

Evans, S. and Boyte, C. (1986). Free Spaces: The Sources of Democratic Change in America.

New York: Harper & Row.

Evans, P.C. and Gawer, A. (2016). The rise of the platform enterprise a global survey. The

Center for Global Enterprise, (January), pp.1–30.

Ezeoha, A. et al. (2019). Second-hand vehicle markets in west Africa: a source of regional

disintegration, trade informality and welfare losses. Business History, 61(1), pp.187–204.

304

Fern, E. (1982). The use of focus groups for idea generation: the effects of group size,

acquaintanceship, and moderator on response quantity and quality. Journal of Marketing

Research, 19(1), pp.1 –13.

Filani, M.O. (2012). The changing face of Lagos: from vision to reform and transformation.

Cities Alliance, Brussels. [online]. Available from:

http://www.citiesalliance.org/sites/citiesalliance.org/files/Lagos-reform-report-lowres.pdf

[Accessed March 16, 2019]

Finn, B., Arthur, B. and Gyamera, S. (2009). New regulatory framework for urban passenger

transport in Ghanaian cities. In 11th Conference on Competition and ownership in Land

Passenger Transport. Delft, pp. 2 –14.

Florida, R. (2017). Venture capital remains highly concentrated in just a few cities.

Bloomberg.com. [online]. Available from: https://www.bloomberg.com/news/articles/2017-

10-03/the-geographic-concentration-of-venture-capital [Accessed May 27, 2022].

Fobosi, S.C. (2019). Employment practices within the minibus taxi industry in Johannesburg.

African Socialogical Review, 23(2), pp.103–124. [online]. Available from:

https://www.jstor.org/stable/10.2307/26868086.

Foucault, M. (2008). Panopticism from discipline & punish: the birth of the prison.

Race/Ethnicity: Multidisciplinary Global Contexts, 2(1), pp.1–12.

Foucault, M. (1995). Discipline and punish: The birth of Prison. Translated by A. Sheridan.

2nd ed. New York: Random House Inc.

Foucault, M. (1980). Power/knowledge. a selected interviews and other writings 1972-77.

Translated by C. Gordon, L. Marshall, J. Mepham, and K. Soper. New York: Pantheon Books.

Fourchard, L. (2011). Lagos, koolhaas and partisan politics in Nigeria. International Journal

of Urban and Regional Research, 35(1), pp.40–56.

305

Flores, O. and Rayle, L. (2017). How cities use regulation for innovation: The case of Uber,

Lyft, and Sidecar in San Francisco, Transportation Research Procedia. Elsevier B.V., 25, pp.

3760–3772, (Online). Available at: doi: 10.1016/j.trpro.2017.05.232 [Accessed December

8,2017)

Friedman, G. (2014). Workers without employers: shadow corporations and the rise of the gig

economy. Review of Keynesian Economics, 2(2), pp.171–188.

Friedman, B. and Nissenbaum, H. (1996) Bias in computer systems, Computer Ethics, 14(3),

pp. 215–232. doi: 10.4324/9781315259697-23.

Gal, U., Jensen, T.B. and Stein, M.K. (2020). Breaking the vicious cycle of algorithmic

management: a virtue ethics approach to people analytics. Information and Organization,

30(2).

Galič, M., Timan, T. and Koops, B.J. (2017). Bentham, deleuze and beyond: an overview of

surveillance theories from the panopticon to participation. Philosophy and Technology, 30(1),

pp.9–37.

Galière, S. (2020). When food-delivery platform workers consent to algorithmic management:

a Foucauldian perspective. New Technology, Work and Employment, 35(3), pp.357–370.

Gandy, M. (2007). Planning, anti-planning, and the infrastructure crisis facing metropolitan

lagos. Cities in Contemporary Africa, 43(2), pp.247–264.

Garcia, A.C. et al. (2009). Ethnographic approaches to the internet and computer-mediated

communication. Journal of Contemporary Ethnography, 38(1), pp.52–84.

Garner, S. and Stokoe, G. (2000). ‘Taxi’. London: Frances Lincoln Limited

Gawer, A. (2014). Bridging differing perspectives on technological platforms: toward an

integrative framework. Research Policy, 43(7), pp.1239-1249. (online) Available at:

/dx.doi.org/10.1016/j.respol.2014.03.006 [Accessed June 2, 2018

306

Gbadeyanka, M. (2017). Uber drivers protest in Lagos. Business Post Nigeria. [online].

Available from: https://businesspost.ng/auto/uber-drivers-protest-lagos/ [Accessed March 10,

2019].

Gino, F. (2017). Uber shows how not to apply behavioral economics. [online] Harvard

Business Review. Available at: <https://hbr.org/2017/04/uber-shows-how-not-to-apply-

behavioral-economics> [Accessed July 29, 2021].

Gharaghooshi, F., Mcmeel, D. and Manfredini, M., (2014). Digital technologies and the

production of spaces of resistance, in international contemporary urban issues conference on

rethinking informality. Istanbul, pp.1 –15.

GhanaWeb.com. (2016). Ghana’s cab drivers fight Uber. GhanaWeb. [online]. Available from:

https://www.ghanaweb.com/GhanaHomePage/business/Ghana-s-cab-drivers-fight-Uber-

467331 [Accessed July 17, 2019].

Gherardi, S. (2000). Practice-based theorizing on learning and knowing in

organizations. Organization, 7(2), pp.211 –223.

Geels, F.W. (2002). Technological transitions as evolutionary reconfiguration processes: a

multi-level perspective and a case-study. Research Policy, 31(8–9), pp.1257–1274. [online].

Available from: http://linkinghub.elsevier.com/retrieve/pii/S0048733302000628.

Geels, F.W. (2005). Processes and patterns in transitions and system innovations: refining the

co-evolutionary multi-level perspective. Technological Forecasting and Social Change, 72(6

SPEC. ISS.), pp.681–696.

Geels, F. and Schot, J. (2007). Typology of sociotechnical transition pathways. Research

Policy, 36(3), pp.399-417. (online) Available at: https://doi.org/10.1016/j.respol.2007.01.003

(Accessed 17 Mar 2018)

Geels, F. (2011). The multi-level perspective on sustainability transitions: responses to seven

criticisms,' Environmental Innovation and Societal Transitions. Elsevier B.V., 1(1), pp. 24–40.

(Online) Available at: doi: 10.1016/j.eist.2011.02.002. (Accessed 16 Jan. 2018).

307

Georgano, N., and Munro B., (2008). The London Taxi. Great Britain: Shire Publications.

Georg, C. and Rose, M. (2016). Licensing metered taxis does more harm than good – South

Africa should stop it. The Conversation. [online]. Available from:

https://theconversation.com/licensing-metered-taxis-does-more-harm-than-good-south-africa-

should-stop-it-64563 [Accessed May 4, 2022].

Gillespie, T. (2014). The relevance of algorithms. In T. Gillespie, P. Boczkowski & K. Foot,

ed. Media Technologies: Essays on Communication, Materiality, and Society. Cambridge:

MIT Press, pp. 167 – 193.

Glasner, J. (2021). These countries have the most startup investment for their size. Crunchbase

News. [online]. Available from: https://news.crunchbase.com/startups/countries-most-startup-

investment/ [Accessed May 23, 2022].

Goods, C., Veen, A. and Barratt, T. (2019). Is your gig any good? analysing job quality in the

Australian platform-based food-delivery sector. Journal of Industrial Relations, 61(4), pp.502–

527.

Gorman, G. et al. (2005). Qualitative research for the information professional: A practical

handbook. Facet Publishing.

Gössling, S. (2018). ICT and transport behavior: a conceptual review. International Journal of

Sustainable Transportation, 12(3), pp.153–164.

Gledhill, J. (2007). Neoliberalism. In D. Nugent & J. Vincent, ed. A Companion to the

Anthropology of Politics. Chichester: Wiley, pp. 332-348.

Gregor, S. and Lee-Archer, B., (2016). The digital nudge in social security

administration. International Social Security Review, 69(3-4), pp.63-83.

308

Graham, M., Hjorth, I. and Lehdonvirta, V. (2017). Digital labour and development: impacts

of global digital labour platforms and the gig economy on worker livelihoods. Transfer, 23(2),

pp.135–162. [online]. Available from: https://doi.org/10.1177/1024258916687250.

Gray, J., Gerlitz, C. and Bounegru, L. (2018). Data infrastructure literacy. Big Data and

Society, 5(2), pp.1–13.

Grohmann, R. et al. (2022). Platform scams: Brazilian workers’ experiences of dishonest and

uncertain algorithmic management. New Media & Society, 24(7), pp.1611 –1631.

Guba, E., & Lincoln, Y. (1994). Competing paradigms in qualitative research. In N. K. Denzin

& Y.

S. Lincoln (Eds.), Handbook of qualitative research (pp. 105-117). Thousand Oaks, CA: Sage

Guest, G., Namey, E. and Mitchell, M. (2013). Participant observation. in: collecting

Qualitative Data: A Field Manual for Applied Research, 55 City Road, London: SAGE

Publications, Ltd pp. 75-112. Available at: <http://www.doi.org/10.4135/9781506374680>

[Accessed February 16, 2020]

Gupta, D. (2001). Everyday resistance or routine repression? exaggeration as stratagem in

agrarian conflict. Journal of Peasant Studies, 29(1), pp.89–108.

Gupta, A. and Wang, H. (2016). The reason Silicon Valley beat out Boston for vc dominance.

Harvard Business Review. [online]. Available from: https://hbr.org/2016/11/the-reason-

silicon-valley-beat-out-boston-for-vc-

dominance#:~:text=Silicon%20Valley%20benefited%20from%20a,noncompete%20covenant

s%20in%20employment%20contracts [Accessed April 27, 2022].

Gustin, S. (2012). Lyft: ride sharing startup Zimride hits the gas pedal in San Francisco |

TIME.com. [online] Available at: http://business.time.com/2012/09/04/need-a-lyft-ride-

sharing-startup-zimride-hits-the-gas-pedal/ [Accessed 22 Mar. 2018].

Gutmann, M. (1993). Rituals of resistance: a critique of the theory of everyday forms of

resistance. Latin American Perspectives, 20(2), pp.74 - 92. [online]. Available from:

https://www.jstor.org/stable/2633594 [Accessed August 20, 2020].

309

Guardian. (2021). The ban on Okada in Lagos State. The Guardian Nigeria News - Nigeria

and World News. [online]. Available from: https://guardian.ng/opinion/the-ban-on-okada-in-

lagos-state/ [Accessed March 15, 2022].

Guven, M. and Karlen, R. (2020). Supporting Africa’s urban informal sector: coordinated

policies with social protection at the core. World Bank Blogs. [online]. Available from:

https://blogs.worldbank.org/africacan/supporting-africas-urban-informal-sector-coordinated-

policies-social-protection-core [Accessed August 30, 2020].

Gwilliam, K. M. (2005). Regulation of taxi markets in developing countries: issues and options,

transport. TRN-3 (Online) Available at:

https://siteresources.worldbank.org/INTTRANSPORT/214578-

1099488338138/20522187/TRN-3_Taxi_Reg.pdf. [Accessed February 7, 2018].

Haggerty, K. D. and Ericson, R. V. (2000). The surveillant assemblage, British Journal of

Sociology, 51(4), pp. 605–622. doi: 10.1080/00071310020015280.

Hall, V., Kendrick, C, and Nosko, C. (2015). The effects of uber's surge pricing: a case

study, working paper. Available at: https://eng.uber.com/research/the-effects-of-ubers-surge-

pricing-a-case-study/ [Accessed September 17, 2019]

Hanrahan, B. V., Ma, N.F. and Yuan, C.W. (2017). The roots of bias on uber. ECSCW 2017 -

Proceedings of the 15th European Conference on Computer Supported Cooperative Work,

(March).

Hao, K. (2017). Map: All the places where Uber is partially or fully banned. Quartz. [online].

Available from: https://qz.com/1084981/map-all-the-places-where-uber-is-partially-or-fully-

banned/ [Accessed March 27, 2022].

Hartmans, A. and Leskin, P. (2019). The history of how Uber went from the most feared startup

in the world to its massive IPO. Business Insider. [online]. Available from:

https://www.businessinsider.com/ubers-history?r=US&IR=T [Accessed May 27, 2022].

310

Hardy, C., Phillips, N. and Clegg, S. (2001). Reflexivity in organization and management

theory: a study of the production of the research `Subject'. Human Relations, 54(5), pp.531 –

560.

Harris, D. (2016). On company Time: American Modernism in the Magazines. New York:

Columbia University Press.

Harris, S. and Krueger, A. (2015). A Proposal for Modernizing Labor Laws for Twenty-First-

Century Work: The “Independent Worker”. Brookings. [online]. Available from:

https://www.hamiltonproject.org/assets/files/modernizing_labor_laws_for_twenty_first_cent

ury_work_krueger_harris.pdf [Accessed May 5, 2018].

Harlow, S. (2012). Social media and social movements: Facebook and an online Guatemalan

justice movement that moved offline. New Media and Society, 14(2), pp.225–243.

Harvey, D. (2005). A brief history of neoliberalism. 1st ed. New York: Oxford University

Press.

Head, E. (2009). The ethics and implications of paying participants in qualitative research.

International Journal of Social Research Methodology, 12(4), pp.335–344.

Healy, J., Nicholson, D. and Pekarek, A. (2017). Should we take the gig economy seriously?

Labour & Industry: a journal of the social and economic relations of work, 27(3), pp.232–248.

[online]. Available from: https://doi.org/10.1080/10301763.2017.1377048.

Heeks, R. (2017). Digital economy and digital labour terminology: making sense of the “gig

economy”. “online labour”, “crowd work”, “microwork”, “platform labour”, etc.

Manchester Centre for Development Informatics. Working Paper 70. Available at:

https://www.gdi.manchester.ac.uk/research/publications/di/di-wp70/

Heery, E. and Abbott, B. (2000). Trade unions and the insecure workforce. In E. Heery, B.

Abbott & J. Salmon, ed. The Insecure Workforce. London: Routledge.

311

Heintz, J. and Pollin, R. (2003). Informalization, economic growth and the challenge of

creating viable labor standards in developing countries. SSRN Electronic Journal.

Helderman, R. (2014). Uber pressures regulators by mobilizing riders and hiring vast lobbying

network. The Washington Post. [online]. Available from:

https://www.washingtonpost.com/politics/uber-pressures-regulators-by-mobilizing-riders-

and-hiring-vast-lobbying-network/2014/12/13/3f4395c6-7f2a-11e4-9f38-

95a187e4c1f7_story.html [Accessed May 17, 2022].

Hennell, K., Limmer, M. and Piacentini, M. (2019). Ethical dilemmas using social media in

Qualitative Social Research: A Case Study of Online Participant Observation. Sociological

Research Online, 25(3), pp.473–489.

Heilman, M. (2012). Gender stereotypes and workplace bias. Research in Organizational

Behavior, 32, pp.113-135.

Hein, J., Evans, J. and Jones, P. (2008). Mobile methodologies: theory, technology and

practice. Geography Compass, 2(5), pp.1266 –1285.

Hern, A. (2016). Uber employees spied on ex-partners, politicians and Beyoncé'. The

Guardian. [online]. Available from:

https://www.theguardian.com/technology/2016/dec/13/uber-employees-spying-ex-partners-

politicians-beyonce [Accessed August 28, 2020].

Hill, W. (2008). Global business today. McGraw-Hill Irwin.

Hine, C., (2008). Virtual ethnography: modes, varieties, affordances. The SAGE handbook of

online research methods, pp.257 –270.

Hook, L., (2017). Uber: The uncomfortable view from the driving seat. [online] Ft.com.

Available at: <https://www.ft.com/content/c9a8b592-a81d-11e7-ab55-27219df83c97>

[Accessed July 18, 2021].

312

Hoge, P. (2014). Uber doubles reach to 200 cities in four months. San Francisco Business

Times. [online]. Available from:

https://www.bizjournals.com/sanfrancisco/blog/techflash/2014/08/uber-doubles-reach-200-

cities.html [Accessed February 11, 2018].

Hollander, J., A. and Einwohner R., L. (2004). Conceptualizing resistance. Sociological

Forum, 19(4), pp. 533 –554.

Holland, R. (1999). Reflexivity. Human Relations, 52(4), pp.463 –484.

Hochschild, A. (1983). The Managed Heart: Commercialisation of Human Feeling. Berkeley:

University of California Press.

Houeland, C. (2018). What is Uber up to in Africa? Developing Economics. [online]. Available

from: https://developingeconomics.org/2018/05/13/what-is-uber-up-to-in-africa/ [Accessed

November 1, 2019].

Howell, C. (2020). Rethinking the role of the state in employment relations for a neoliberal era.

ILR Review, 74(3), pp.739-772.

Ighie. (2019). RURA clamps down on drivers not using taximeters. Igihe. [online]. Available

from: https://en.igihe.com/news/rura-clamps-down-on-drivers-not-using-taximeters

[Accessed June 23, 2022].

ILO. (2016). Facing the growing unemployment challenges in Africa. Ilo.org. [online].

Available from: https://www.ilo.org/africa/media-centre/pr/WCMS_444474/lang--

en/index.htm [Accessed May 23, 2022].

Ingle, M. (2009). An historical overview of problems associated with the formalization of the

South African minibus taxi industry. New Contree, (57), pp.71–87.

International Labour Organisation (ILO). (2018). Women and Men in the Informal Economy:

a Statistical Picture. Geneva: International Labour Office. [online]. Available from:

313

https://www.ilo.org/wcmsp5/groups/public/---dgreports/---

dcomm/documents/publication/wcms_626831.pdf [Accessed April 4, 2020].

Irvine, A., Drew, P. and Sainsbury, R. (2012). Am I not answering your questions properly?

Clarification, adequacy and responsiveness in semi-structured telephone and face-to-face

interviews. Qualitative Research, 13(1), pp.87–106.

Isaac, M. (2014). Uber’s System for Screening Drivers Draws Scrutiny. Nytimes.com. [online].

Available from: https://www.nytimes.com/2014/12/10/technology/ubers-system-for-

screening-drivers-comes-under-scrutiny.html [Accessed June 15, 2019].

Iwayemi, Z. (2022). African unicorn: the most valuable startups by Africans you should know.

Nairametrics. [online]. Available from: https://nairametrics.com/2022/02/15/african-unicorn-

the-most-valuable-startups-by-africans-you-should-know/ [Accessed April 1, 2022].

Jamil, R. and Noiseux, Y. (2018). Shake that moneymaker: insights from Montreal’s uber

drivers. Interventions économiques, (60), pp.1–30.

Jarrahi, M.H. et al. (2021). Algorithmic management in a work context. Big Data and Society,

8(2). [online]. Available from: https://doi.org/10.1177/20539517211020332.

Jarrahi, M.H. et al. (2019). Platform management, boundary resources for gig work, and worker

autonomy. Computer Supported Cooperative Work: CSCW: An International Journal, 29(1–

2), pp.153–189.

Jarrahi, M. and Sutherland, W., (2018). Algorithmic management and algorithmic

competencies: understanding and appropriating algorithms in gig work. Information in

Contemporary Society, pp.578 –589.

Jaynes, V. (2019). The social life of screenshots: the power of visibility in teen friendship

groups. New Media & Society, 22(8), pp.1378 –1393.

Jensen, O.B. (2009). Flows of meaning, cultures of movements - urban mobility as meaningful

everyday life practice. Mobilities, 4(1), pp.139–158.

314

Jean-François, E. and Veselinovic, M. (2016). For Uber, Lagos could be bigger than London.

[online] CNNMoney. Available at: http://money.cnn.com/2016/05/23/smallbusiness/uber-

nigeria-lagos/index.html [Accessed May 14, 2018].

Jhaver, S., Karpfen, Y. and Antin, J. (2018). Algorithmic anxiety and coping strategies of

airbnb hosts. Conference on Human Factors in Computing Systems - Proceedings, 2018-April.

Johansson, A. and Vinthagen, S. (2016). Dimensions of everyday resistance: an analytical

framework, Critical Sociology, 42(3), pp. 417–435.

Jordan, J., 2003. Sabotage or performed compliance: rhetorics of resistance in Temp Worker

Discourse. Quarterly Journal of Speech, 89(1), pp.19 –40.

Kalleberg, A. (2000). Nonstandard Employment Relations: Part-time, Temporary and Contract

Work. Annual Review of Sociology, 26(1), pp.341 – 365.

Kalleberg, A. and Wachter, T. (2017). The US. labor market during and after the great

recession: continuities and transformations. RSF: The Russell Sage Foundation Journal of the

Social Sciences, 3(3), p.1.

Katz, C. (2004). Growing up global: economic restructuring and children’s everyday lives.

Minneapolis: University of Minnesota Press.

Katz, L.F. and Krueger, A.B. (2016). The rise and nature of alternative work arrangements in

the United States, 1995 - 2015. NBER Working Paper Series [online]. Available from:

https://scholar.harvard.edu/files/lkatz/files/katz_krueger_cws_resubmit_clean.pdf.

Katerina, R. (2014). The impact of globalization on the business. Economic Analysis, 47(3-4),

pp.83 –89.

Katzenbach, C. and Ulbricht, L. (2019). Algorithmic governance. Internet Policy Review, 8(4),

pp.0–18.

315

Kaye-Essien, C.W. (2020). Uberization as neoliberal governmentality: a Global South

perspective. Journal of Asian and African Studies, 55(5), pp.716–732.

Kazeem, Y. (2017a). Startups raised more money in Africa than we thought during 2016.

Quartz. [online]. Available from: https://qz.com/africa/921982/african-startups-raised-over-

366-million-in-2016-a-new-report-says/ [Accessed July 27, 2022].Kazeem, Y. (2017b).

Africa’s most valuable startup ecosystem is also the least lucrative for software engineers.

Quartz. [online]. Available from: https://qz.com/africa/943613/lagos-is-africas-most-valuable-

startup-ecosystem/ [Accessed May 14, 2022].

Kellogg, K., Valentine, M. and Christin, A. (2020). Algorithms at work: the new contested

terrain of control. Academy of Management Annals, 14(1), pp.366 –410.

Kene-Okafor, T. (2021). African payments company raises $170m, now valued at over $1b.

Techcrunch.com. [online]. Available from: https://techcrunch.com/2021/03/09/african-

payments-company-flutterwave-raises-170m-now-valued-at-over-1b/ [Accessed June 25,

2022].

Keskinen, P. (2020). worker empowerment in the era of sharing economy platforms in Global

South. In International Conference Proceeding Series. Association for Computing Machinery,

pp. 30 –34. [online]. Available from: http://doi.org/10.1145/3384772.3385147 [Accessed May

27, 2022].

Keskinen, P. and Theophilus, H. (2020). Worker empowerment in the era of sharing economy

platforms in Global South. ACM International Conference Proceeding Series, 2, pp.30–34.

Khafagy, A. (2021). NYC yellow taxi medallion crisis, explained. Documented. [online].

Available from: https://documentedny.com/2021/11/23/taxi-cab-medallion-explained/

[Accessed May 2, 2022].

Klitkou, A., Bolwig, S., Hansen, T. and Wessberg, N. (2015). The role of lock-in mechanisms

in transition processes: The case of energy for road transport. Environmental Innovation and

Societal Transitions, 16, pp.22–37. (online) Available at: 10.1016/j.eist.2015.07.005

(Accessed 19 Mar 2018)

316

Kiger, M.E. and Varpio, L. (2020). Thematic analysis of qualitative data: AMEE Guide No.

131. Medical Teacher, 42(8), pp.846–854. [online]. Available from:

https://doi.org/10.1080/0142159X.2020.1755030.

Kincaid, J. (2008). Taxi Magic: hail a cab from your iPhone at the push of a button. [online]

TechCrunch. Available at: https://techcrunch.com/2008/12/16/taxi-magic-hail-a-cab-from-

your-iphone-at-the-push-of-a-button/ [Accessed March 26, 2018].

Kinder, E., Jarrahi, M.H. and Sutherland, W. (2019). Gig platforms, tensions, alliances, and

ecosystems: an actor-network perspective. Proceedings of the ACM on Human-Computer

Interaction.

Kinsley, S. (2014). The matter of ‘virtual’ geographies. Progress in Human Geography, 38(3),

pp.364-384.

Kitchin, R. (1998). Towards geographies of cyberspace. Progress in Human Geography, 22(3),

pp.385-406.

Kitchin, R. and Dodge, M. (2011). Code/Space. Cambridge, MA: The MIT Press.

Kodransky, M. and Lewenstein, G. (2014). Connecting low-income people to opportunity with

shared mobility, institute for transportation and development policy, pp.1–41 (Online)

Available at: https://www.itdp.org/wp-content/uploads/2014/10/Shared-Mobility_Full-

Report.pdf [Accessed March 18, 2018]

Kolodny, L. (2010). Ubercab ordered to cease and desist. [Online] TechCrunch. Available at:

https://techcrunch.com/2010/10/24/ubercab-ordered-to-cease-and-desist/ [Accessed March

14, 2018].

Kenney, M., and Zysman, J. (2016). The rise of the platform economy. Issues in Science and

Technology 32, No. 3 (Spring)

Knee, J.A. (2018). Why some platforms are better than others. MIT Sloan Management Review,

59(2), pp.18–20.

317

Kufuor, K. (2019). Uber in Ghana: markets and institutions in the emergence of ride-sharing

taxis. Lancaster University of Ghana Law Journal, 5, pp.1-37. [online]. Available

from:https://repository.uel.ac.uk/download/a0e08701204d88196de3ab3b3a881927e477c4443

47e02737051f617cb78c827/505076/7493.pdf [Accessed March 5, 2020].

Kuris, G. (2014). Remaking a neglected megacity: a civic transformation in Lagos State, 1999

– 2012. Innovation for Successful Societies, pp.1999–2012.

Kwan, M. and Knigge, L. (2006). Doing qualitative research Using GIS: an oxymoronic

endeavor? Environment and Planning A: Economy and Space, 38(11), pp.1999 –2002.

Kwet, M. (2019). Digital colonialism: US empire and the new imperialism in the GS. Race and

Class, 60(4), pp.3–26.

Lafferty, G. (2010). In the wake of neo-liberalism: deregulation, unionism, and labour

rights. Review of International Political Economy, 17(3), pp.589 –608.

Lagos State Government (LSG). (n.d). About Lagos. Lagos State Government Website.

[online]. Available from: https://lagosstate. gov.ng/about-lagos/ [Accessed March 28, 2019].

Lam, D. et al. (2006). The San Francisco Taxicab Industry: an equity analysis. Richard &

Rhoda Goldman School of Public Policy University of California, Berkeley. [Online]

Available at: https://archives.sfmta.com/cms/rtaxi/documents/GoldmanStudyFINAL2006.pdf

[Accessed March 8, 2018]

Lakemann, T., & Lay, J. (2019). Digital Platforms in Africa: the "Uberisation" of Informal

Work. (GIGA Focus Afrika, 7). Hamburg: GIGA German Institute of Global and Area Studies

- Leibniz-Institut für Globale und Regionale Studien, Institut für Afrika-Studien. https://nbn-

resolving.org/urn:nbn:de:0168-ssoar-65910-4

Larner, W. (2000). Neo-liberalism: Policy, Ideology, Governmentality. Studies in Political

Economy, 63(1), pp.5–25.

318

Lawrence, F. (2016). Uber is treating its drivers as sweated labour. The Guardian. [online].

Available from: https://www.theguardian.com/technology/2016/dec/09/uber-drivers-report-

sweated-labour-minimum-wage [Accessed February 10, 2019].

Lawuyi, O. (1988). The world of the Yoruba taxi driver: an interpretive approach to vehicle

slogans. Africa, 58(1), pp.1 –13.

Lebaron, G. and Ayers, A. (2013). The rise of a new slavery? Understanding African unfree

labour through neoliberalism. Third World Quarterly, 34(5), pp.873 –892.

Lechuga, V.M. (2012). Exploring culture from a distance: The utility of telephone interviews

in qualitative research. International Journal of Qualitative Studies in Education, 25(3),

pp.251–268.

Lee, D. and Strauss, D., (2021). Uber agrees to classify UK drivers as workers entitled to

benefits. [online] Ft.com. Available at: <https://www.ft.com/content/8d39472b-12a1-4754-

bb11-b9149dab2872> [Accessed September 1, 2021].

Lee, N., Resnick, P. and Barton, G. (2019). Algorithmic bias detection and mitigation: best

practices and policies to reduce consumer harms. Brookings. [online]. Available from:

https://www.brookings.edu/research/algorithmic-bias-detection-and-mitigation-best-

practices-and-policies-to-reduce-consumer-harms/ [Accessed May 23, 2020].

Lee, M.K. et al. (2015). Working with Machines: The impact of algorithmic and data-driven

management on human workers. Proceedings of the ACM CHI’15 Conference on Human

Factors in Computing Systems, 1, pp.1603–1612. [online]. Available from:

http://dx.doi.org/10.1145/2702123.2702548.

Lehdonvirta, V. et al. (2015). Online labour markets and the persistence of personal networks:

evidence from workers in Southeast Asia. In American Sociological Association Annual

Meeting 2015. Chicago.

319

Lehdonvirta, V. (2016). Algorithms that divide and unite: delocalization, identity, and

collective action in microwork. In J. Flecker, ed. Space, place, and global digital work.

London: Palgrave-Macmillan, pp. 53 – 80.

Leszczynski, A. (2015). Spatial media/tion. Progress in Human Geography, 39(6), pp.729 –

751.

Levosky, A. and Greenberg, A. (2001). Organized Dynamic Ride Sharing: The Potential

Environment Benefits and the Opportunity for Advancing the Concept, Transportation

Research Board Annual meeting, pp. 1–15. (Online) Available at:

https://pdfs.semanticscholar.org/d7aa/330e2dea4437c6201bb8ecc6943caacc29ab.pdf

[Accessed March 18, 2018]

Lewis, P. (2014). The dysfunctional state of Nigeria. Chapter 3.

http://www.cgdev.org/doc/shortofthegoal/chap3.pdf. [Accessed July 19, 2020]

Li, Z., Hong, Y. and Zhang, Z. (2017). An empirical analysis of on-demand ride-sharing and

traffic congestion. In: Hawaii international conferences on system sciences. [online] Hawaii:

Scholar space. Available at: https://scholarspace.manoa.hawaii.edu/handle/10125/41152

[Accessed 31 May 2018].

Li, T. (2007). Governmentality. Anthropologica, 49(2), pp.275 – 281.

Lilja, M. and Vinthagen, S. (2018) ‘Dispersed resistance: unpacking the spectrum and

properties of glaring and everyday resistance’, Journal of Political Power. Routledge, 11(2),

pp. 211–229.

Liu, C. (2015). Worldwide internet and mobile users. [online] New York: emarketer. Available

at: https://insights.ap.org/uploads/images/eMarketer_Estimates_2015.pdf [Accessed May 6,

2018].

Lloyd, A. (2014). Following the red thread of information in information literacy research:

Recovering local knowledge through interview to the double. Library and Information Science

320

Research, 36(2), pp.99–105. [online]. Available from:

http://dx.doi.org/10.1016/j.lisr.2013.10.006.

Lobel, O. (2017). The Uber-Waymo Lawsuit: It Should Be Easy to Poach Talent, But Not

IP. Harvard Business Review. [online]. Available from: https://hbr.org/2017/06/the-uber-

waymo-lawsuit-it-should-be-easy-to-poach-talent-but-not-ip [Accessed June 21, 2022].

Lloyd-Evans, S. (2008). Geographies of the contemporary informal sector in the GS: gender,

employment relationships and social protection. Geography Compass, 2(6), pp.1885–1906.

Longhurst, R., (2003). Semi-structured interviews and focus groups. Key methods in

geography, 3(2), pp.143-156.

Lomas, N. (2017). Let's talk about Uber, air quality and urban air quality. Techcrunch.com.

[online]. Available from: https://techcrunch.com/2015/08/26/uber-london-impact/ [Accessed

May 23, 2022].

Lomas, N. (2015). UK high court rules in Uber's favour on taximeter dispute. Techcrunch.com.

[online]. Available from: https://techcrunch.com/2015/10/16/u-k-high-court-rules-in-ubers-

favor-on-taximeter-dispute/ [Accessed May 27, 2022].

Lomas, N. (2019). Uber is paying $3.1BN to pick up Middle East Rival Careem.Techcrunch.

[online]. Available from: https://techcrunch.com/2019/03/26/uber-is-paying-3-1bn-to-pick-

up-middle-east-rival-careem/ [Accessed May 30, 2019].

Lunden, I. (2022). Bolt raises $709M at an $8.4B valuation to expand its transportation and

food delivery super app. Techcrunch.com. [online]. Available from:

https://techcrunch.com/2022/01/10/bolt-raises-709m-at-an-8-4b-valuation-to-expand-its-

transportation-and-food-delivery-super-app/ [Accessed May 15, 2022].

Luo, Y. (2021). New OLI advantages in digital globalization. International Business Review,

30(2), p.101797.

321

Lyon, D. (2006). Title: 9/11, Synopticon, and Scopophilia: Watching and Being Watched. In

K. Haggerty & R. Ericson, ed. The New Politics of Surveillance and Visibility. Toronto:

University of Toronto Press, pp. 35 –54.

Mack, L. (2010). The philosophical underpinnings of educational research. Available at:

https://www.semanticscholar.org/paper/The-Philosophical-Underpinnings-of-Educational-

Mack/a98a3f11879f9d2a91f087c0b9191239add287c4 [Accessed May 12, 2018]

Madge, C. (2007). Developing a geographers’ agenda for online research ethics. Progress in

Human Geography, 31(5), pp.654–674.

Malecki, E. (2017). Real people, virtual places, and the spaces in between. Socio-Economic

Planning Sciences, 58, pp.3 –12.

Manfredi, V. (2013). The etymology of ogan ( ogã ) proves the Gùn-gbè origin of Candomblé

Jeje *. , (October), pp.13–15.

Mann, S. (2004). “Sousveillance” inverse surveillance in multimedia imaging. In 12th ACM

International Conference on Multimedia. New York, pp. 620 -627. [online]. Available from:

https://www.ics.uci.edu/~djp3/classes/2012_09_INF241/papers/sousveillance.pdf [Accessed

June 15, 2020].

Mäntymäki, M., Baiyere, A. and Islam, A.K.M.N. (2019). Digital platforms and the changing

nature of physical work: Insights from ride-hailing. International Journal of Information

Management, 49(April), pp.452–460.

Martini, D. (2017). International Regulatory Entrepreneurship: Uber’s Battle with Regulators

in France.// J. Domenic Martini, International Regulatory Entrepreneurship: Uber’s Battle with

Regulators in France, 19 SAN DIEGO INT’L L.J. 127, 148 (2017).

Mason, S., (2016). High score, low pay: why the gig economy loves gamification. [online] the

Guardian. Available at: <https://www.theguardian.com/business/2018/nov/20/high-score-low-

pay-gamification-lyft-uber-drivers-ride-hailing-gig-economy> [Accessed July 11, 2021].

322

Mastercard. (2019). The Global Gig Economy: Capitalizing on a ~$500B Opportunity.

[online]. Available from: https://newsroom.mastercard.com/wp-content/uploads/2019/05/Gig-

Economy-White-Paper-May-2019.pdf [Accessed July 5, 2020].

Mathiesen, T. (1997). The viewer society: Michel Foucault's 'Panopticon' revisited. Theoretical

criminology, 1(2), pp.215 - 234.

Mateescu, A. and Nguyen, A. (2019). Explainer: Algorithmic management in the workplace.

Data & Society Institute, 2015-April (February), pp.1–15. [online]. Available from:

https://doi.org/10.1145/2702123.2702548.

Mátùlúko, M. (2016). Afro is an indigenous online taxi booking service built with Africans in

mind. Techpoint Africa. [online]. Available from: https://techpoint.africa/2016/02/19/afro-

online-taxi-booking-review/ [Accessed December 3, 2018].

Mazmanian, M., Orlikowski, W. J. and Yates, J. A. (2013). 'The autonomy paradox: The

implications of mobile email devices for knowledge professionals', Organization Science,

24(5), pp. 1337–1357. doi: 10.1287/orsc.1120.0806.

McNicholas, C. and Poydock, M., (2019). How California's AB5 protects workers from

misclassification. [online] Economic Policy Institute. Available at:

https://www.epi.org/publication/how-californias-ab5-protects-workers-from-

misclassification/ [Accessed August 31, 2021].

Mcfarland, M. (2014). Uber's remarkable growth could end the era of poorly paid drivers. The

Washington Post. [online]. Available

from:https://www.washingtonpost.com/news/innovations/wp/2014/05/27/ubers-remarkable-

growth-could-end-the-era-of-poorly-paid-cab-drivers/ [Accessed April 17, 2022].

McLoughlin, I., Badham, R. and Palmer, G. (2005). Cultures of ambiguity. Work, Employment

and Society, 19(1), pp.67 –89.

Meagher, K. (1995). Crisis, informalization and the urban informal sector in sub-Saharan

Africa. Development and Change, 26(2), pp.259 –284.

323

Meagher, K. (2018). Rewiring the social contract digital taxis and economic inclusion in

Nigeria draft paper prepared for the unrisd conference overcoming inequalities in a fractured

world: between elite power and social mobilization., (November).

Meijerink, J. and Keegan, A., (2019). Conceptualising human resource management in the gig

economy. Journal of Managerial Psychology, 34(4), pp.214 –232.

Merriam, S. (2009). Qualitative research. San Francisco: Jossey-Bass.

Mobereola, D. (2006). Strengthening urban transport institutions: a case study of Lagos State.

Lagos: World Bank. [online]. Available from:

https://documents1.worldbank.org/curated/en/758551468288333387/pdf/418330WP0UNI0A

1s0case0dp0501PUBLIC1.pdf [Accessed January 28, 2018].

Mobereola, D. (2009). Lagos Bus Rapid Transit Africa’s first BRT scheme. SSATP Discussion

Paper No. 9. Available from: https://www.ssatp.org/sites/ssatp/files/publications/SSATP-

DiscussionPapers/DP09-Lagos-BRT.pdf [Accessed April 16, 2022].

Möhlmann, M. and Zalmanson, L. (2017). Hands on the wheel: navigating algorithmic

management and Uber drivers' autonomy. Proceedings of the International Conference on

Information Systems (ICIS 2017), December 10 –13, Seoul, South Korea.

Moore, H. C. (1902). Omnibuses and Cabs: Their Origin and History. London: Chapman &

Hall, Ltd. Available at:

https://archive.org/stream/omnibusescabsthe00moor/omnibusescabsthe00moor_djvu.txt

[Accessed January 2, 2018].

Motor Sport. (1961). Meet Michael Gotla, the Mini-cab Man., p.647.

Mourdoukoutas, E. (2017). Africa’s app-based taxis battle Uber over market share. Africa

Renewal. [online]. Available from: https://www.un.org/africarenewal/magazine/august-

november-2017/africa%E2%80%99s-app-based-taxis-battle-uber-over-market-share

[Accessed April 28, 2019].

324

Mundt, M., Ross, K. and Burnett, C.M. (2018). Scaling social movements through social

media: The Case of Black Lives Matter. Social Media and Society, 4(4).

Muntaner, C. (2018). Digital platforms, gig economy, precarious employment, and the

invisible hand of social class. International Journal of Health Services, 48(4), pp.597–600.

Mutekwe, P., 2019. Resistance and repression in Zimbabwe: a case study of Zimplats mine

workers. Review of African Political Economy, 46(160), pp.246 –260.

Mutongi, K. (2006). Thugs or Entrepreneurs? perceptions of matatu operators in Nairobi, 1970

to the present. Africa: Journal of the International African Institute, 76(4), pp.549–568.

National Bureau of Statistics (NBS) (2019). 2019 poverty and inequality in Nigeria:

executive summary. [online]. Available from: https://nigerianstat.gov.ng/elibrary/read/1092 ,

[Accessed August 18, 2020].

Nairaland. (2009a). Fashola commissions 1,200 Cabs. . Nairaland.com. [online]. Available

from: https://www.nairaland.com/277261/fashola-commissions-1200-cabs-pictures [Accessed

April 15, 2019].

Nairaland. (2009). New Fashola Cabs. Nairaland.com. [online]. Available from:

https://www.nairaland.com/283807/new-fashola-cabs [Accessed April 28, 2018].

Nairaland.com. (2017). Uber Drivers in Lagos Using Fake GPS. App to Inflate Fares – Report

- Autos - Nigeria. [online] Available from: <https://www.nairaland.com/4176587/uber-drivers-

lagos-using-fake> [Accessed September 6, 2019].

National Bureau of Statistics (NBS) (2018). 2019 Road Transport Data (Q4 2018). [online].

Available from: https://nigerianstat.gov.ng/elibrary/read/1092 , [Accessed August 18, 2020].

National Bureau of Statistics (NBS) (2016). Nigerian Gross Domestic Product Report Q2 2016.

[online] Available from: https://nigerianstat.gov.ng/elibrary/read/1092 , [Accessed May 16,

2018].

325

Nieborg, D.B. and Poell, T. (2018). The platformization of cultural production: theorizing the

contingent cultural commodity. New Media and Society, 20(11), pp.4275–4292.

Neilson, D. (2020). Bringing in the ‘neoliberal model of development’. Capital and Class,

44(1), pp.85–108. [online]. Available from: https://doi.org/10.1177/0309816819852746.

Niemietz, K. and Zuluaga, D. (2016). Turning statutory regulation into private regulation for

the UK's taxi industry. Institute of Economic Affairs.

Newlands, G. (2020) Algorithmic Surveillance in the Gig Economy: The Organization of Work

through Lefebvrian Conceived Space, Organization Studies, 42(5), pp. 719–737. doi:

10.1177/0170840620937900.

Noble, S. U. (2018). Algorithms of Oppression: How Search Engines Reinforce Racism. New

York: NYU Press.

Norris, C. and Armstrong, G. (1999). CCTV and the Rise of Mass Surveillance Society. Crime

Unlimited? Questions for the 21st Century, 10, pp.76 –98.

Novick, G. (2008). Is there a bias against telephone interviews in qualitative research?

Research in Nursing and Health, 31(4), pp.391–398.

Nowell, L.S. et al. (2017). Thematic analysis: striving to meet the trustworthiness criteria.

International Journal of Qualitative Methods, 16(1), pp.1–13.

Nurvala, J. (2015). ‘Uberisation’ is the future of the digitalised labour market. European View,

14(2), pp.231 –239.

Nwokoma, N. (2021). A third of Nigerians are unemployed: here's why. The Conversation.

[online]. Available from: https://theconversation.com/a-third-of-nigerians-are-unemployed-

heres-why-159262 [Accessed June 28, 2021].

326

Nwokpoku, E. et al. (2018). Nigerian labor laws: issues and challenges. World Applied

Sciences Journal, 36(1), pp.47 –54.

Nyumba, T. et al. (2018). The use of focus group discussion methodology: insights from two

decades of application in conservation. Methods in Ecology and Evolution, 9(1), pp.20 –32.

Nytwa.org, (n.d). New York Taxi Workers Alliance. New York Taxi Workers Alliance.

[online]. Available from: https://www.nytwa.org/ [Accessed June 21, 2022].

O'Brien, S. et al. (2018). CNN investigation: 103 Uber drivers accused of sexual assault or

abuse. CNNMoney. [online]. Available from:

https://money.cnn.com/2018/04/30/technology/uber-driver-sexual-assault/index.html

[Accessed June 28, 2019].

Obeng-Odoom, F. (2012). Neoliberalism and the urban economy in ghana: urban employment,

inequality, and poverty. Growth and Change, 43(1), pp.85 –109.

Ofstad, M. (2017). The rise of ridesharing platforms: an uber-assessment of bits and atoms.

MSC. University of Oslo.

Ogunfuwa, I. (2017). 90% of technologies used in Nigeria are imported – DG,

NOTAP. Punchng.com. [online]. Available from: https://punchng.com/90-of-technologies-

used-in-nigeria-are-imported-dg-notap/ [Accessed August 28, 2019].

Ojowolo, S. and Wahab, B. (2017). Municipal solid waste and flooding in Lagos metropolis,

Nigeria: Deconstructing the evil nexus. Journal of Geography and Regional Planning, 10(7),

pp.174 –185.

Ogunbodede, E. (2008). urban road transportation in nigeria from 1960 to 2006: problems,

prospects and challenges. Ethiopian Journal of Environmental Studies and Management, 1(1).

327

Okere, A. (2019). NIMC needs N132bn to register 200 million Nigerians. Punchng.com.

[online]. Available from: https://punchng.com/nimc-needs-n132bn-to-register-200-million-

nigerians-aziz-dg/ [Accessed August 3, 2020].

Olajide, O., Agunbiade, M. and Bishi, H., (2018). The realities of Lagos urban development

vision on livelihoods of the urban poor. Journal of Urban Management, 7(1), pp.21 – 31

Olajide, O. and Lawanson, T. (2021). Urban paradox and the rise of the neoliberal city: Case

study of Lagos, Nigeria. Urban Studies. [online]. Available from:

https://doi.org/10.1177/00420980211014461.

Olawoyin, O. (2021). Lagos overtakes Nairobi as Africa’s leading startup hub. Premium Times

Nigeria. [online]. Available from: https://www.premiumtimesng.com/news/more-

news/469338-lagos-overtakes-nairobi-as-africas-leading-startup-hub.html [Accessed May 18,

2022].

Olorounbi, R. (2019). Nigeria to give all of its 200 million people identity

numbers. Bloomberg.com. [online]. Available from:

https://www.bloomberg.com/news/articles/2019-09-20/nigeria-to-give-all-of-its-200-million-

people-identity-numbers [Accessed February 19, 2020].

Olukoju, A., (2013). Infrastructure development and urban facilities in Lagos. Ibadan: Institut

franais de recherche en Afrique.

Olowu, D. (1986). A Decade of Local Government Reform in Nigeria: 1976-

1986. International Review of Administrative Sciences, 52(3), pp.287 - 299.

Omobowale, A. and Fayiga, O. (2017). Commercial Motor Drivers, Transport Unions and

Electoral Violence in Ibadan, Nigeria. Development and Society, 46(3), pp.591 – 614. [online].

Available from: http://www.jstor.org/stable/90017821. [Accessed June 28, 2019].

Onatere-Ubrurhe, J. (2016). Developing A Benchmarking Methodology for the Nigerian

Transport Sector. PhD. University of Wolverhampton.

328

Onyekakeyah, L. (2018). Why Lagos is choking with traffic gridlock. The Guardian. [online].

Available from: https://guardian.ng/opinion/why-lagos-is-choking-with-traffic-gridlock-2/

[Accessed May 28, 2019].

Osborne, H. (2016). Uber loses right to classify UK drivers as self-employed. The Guardian.

[online]. Available from: https://www.theguardian.com/technology/2016/oct/28/uber-uk-

tribunal-self-employed-status [Accessed September 4, 2019].

Orelus, P.W. and Chomsky, N. (2014). Neoliberalism: The Rich Over the Poor: Noam

Chomsky and Pierre Orelus in Dialogue. Source: Counterpoints, 458, pp.65–76.

Oshodi, L. et al. (2016). Urban mobility and transportation. [online] Lagos: Heinrein Böll

Stiftung. Available at:

https://ng.boell.org/sites/default/files/160206_urban_planning_processes_digital_new.pdf

[Accessed June 5, 2018].

Otunola, B. (2019). The BRT and the danfo: a case study of Lagos’ transport reforms from

1999-2019. IGC Cities that Work Case Study. International Growth Centre, London, UK

Ormston, R., et al. (2014). The foundations of qualitative research. Qualitative research

practice: A guide for social science students and researchers, 2(7), pp.52-55.

Oyaniru, A. (n.d). Rethinking public transportation in Lagos: lessons from

history. Universityoflagos.academia.edu. [online]. Available from:

https://universityoflagos.academia.edu/OyarinuAbimbola [Accessed August 10, 2019].

Oyewale, A., Adelowo, C. and Ekperiware, M. (2018). Patenting and Technology

Entrepreneurship in Nigeria: Issues, Challenges and Strategic Options. Journal of Economics,

Management and Trade, 21(2), pp.1 –14.

Page, M. (2020). Dubai Property: an oasis for Nigeria's corrupt political elites. , (March), pp.1–

30. [online]. Available from: https://carnegieendowment.org/2020/03/19/dubai-property-

oasis-for-nigeria-s-corrupt-political-elites-pub-81306. [Accessed October 12, 2020]

329

Pasquale, F. (2015). The black box society: the secret algorithms that control money and

information. Cambridge, MA: MIT Press.

Paul, K. (2019). California advances bill that would 'lead the world' on gig worker rights. The

Guardian. [online]. Available from:

https://www.theguardian.com/technology/2019/aug/29/california-lawmakers-consider-bill-

that-would-lead-the-world-on-gig-worker-rights [Accessed January 28, 2020].

Peck, J., Theodore, N. and Brenner, N. (2009). Neoliberal Urbanism: Models, Moments,

Mutations. SAIS Review of International Affairs, 29(1), pp.49-66.

Petzinger, J. (2018). Taxify’s 24-year-old CEO is smashing Uber’s monopoly in Europe and

Africa. Quartz. [online]. Available from: https://qz.com/1303297/markus-villig-ceo-of-taxify-

is-smashing-ubers-monopoly-in-europe-and-africa/ [Accessed July 28, 2019].

Perlberg, S. (2013). The 20 Countries where people get kidnapped the most. Business Insider.

[online]. Available from: https://www.businessinsider.com/top-20-countries-by-kidnapping-

2013-12?r=US&IR=T [Accessed August 30, 2018].

Pilling, D. (2018). Nigerian economy: Why Lagos works. Ft.com. [online]. Available from:

https://www.ft.com/content/ff0595e4-26de-11e8-b27e-cc62a39d57a0 [Accessed May 31,

2022].

Phelan, S. and Dawes, S. (2018). Liberalism and neoliberalism. Oxford Research Encyclopedia

of Communication. [online]. Available from:

https://oxfordre.com/communication/view/10.1093/acrefore/9780190228613.001.0001/acrefo

re-9780190228613-e-176#acrefore-9780190228613-e-176-div1-1 [Accessed June 29, 2018].

Polletta, F. (1999). Free spaces’ in collective action. Theory and Society, 28(1), pp.1–38.

Pollio, A. (2019). Forefronts of the Sharing Economy: Uber in Cape Town. International

Journal of Urban and Regional Research, 43(4), pp.760–775.

330

Potts, D. (2008). The urban informal sector in sub-Saharan Africa: From bad to good (and back

again?). Development Southern Africa, 25(2), pp.151–167.

Public Utilities Commission of the State of California. (2010). Notice to Cease and Desist.

[online]. Available from: http://consumers.cpuc.ca.gov/carrierinvestigations/.

Purcell, C. and Brook, P. (2020). at least I’m my own boss! explaining consent, coercion and

resistance in platform work. Work, Employment and Society, p.095001702095266.

Radu, R. and Psaila, S. (2017). 'Uberisation' demystified: examining legal and regulatory

responses worldwide. In 5th Conference of the Regulating for Decent Work Network. Geneva:

DiploFoundation. [online]. Available from:

https://www.researchgate.net/publication/317973057_’Uberisation’_Demystified_Examining

_Legal_and_Regulatory_Responses_Wordwide [Accessed May 28, 2020].

Ramachandran, V. et al. (2019). The New Economy of Africa: Opportunities for Nigeria’s

Emerging Technology Sector. Washington: Center for Global Development.

Rani, U. and Dhir, R. (2020). Platform Work and the COVID-19 Pandemic. The Indian Journal

of Labour Economics, 63(S1), pp.163-171.

Rekhviashvili, L. and Sgibnev, W. (2019). 'Theorising informality and social embeddedness

for the study of informal transport. lessons from the marshrutka mobility phenomenon', Journal

of Transport Geography. Elsevier, 88(January), pp. 0–1. doi: 10.1016/j.jtrangeo.2019.01.006.

Raven, R., Schot, J. and Berkhout, F. (2012). Space and scale in socio-technical transitions.

Environmental Innovation and Societal Transitions, 4, pp.63 –78.

Rayle, L. et al. (2016). Just a better taxi? a survey-based comparison of taxis, transit, and

ridesourcing services in San Francisco. Transport Policy, 45, pp.168 –178.

331

Reuters. (2018). Egypt court orders suspension of Uber, Careem services in victory for taxis:

sources. Reuters. [online]. Available from: https://www.reuters.com/article/us-egypt-uber-

idUSKBN1GW1O2 [Accessed June 14, 2019].

Rip, A., and Kemp, R. (1998). Technological change. In S. Rayner, and E. Malone (Eds.),

Human choice and climate change, 2, pp. 327 – 399

Rizzo, M. (2011). ‘Life is War’: Informal transport workers and neoliberalism in Tanzania

1998-2009. Development and Change, 42(5), pp.1179 –1206.

Rodrik, D. (2017). The fatal flaw of neoliberalism: it's bad economics. The Guardian. [online].

Available from: https://www.theguardian.com/news/2017/nov/14/the-fatal-flaw-of-

neoliberalism-its-bad-economics [Accessed May 21, 2022].

Rochet, J. and Tirole, J. (2006). Two-sided markets: a progress report. The RAND Journal of

Economics, 37(3), pp.645 –667.

Rogers, B. (2015). Employment Rights in the Platform Economy: Getting Back to

Basics. SSRN Electronic Journal, (Research Paper No. 2015-33), pp.480 – 519.

Rosenblat, A., Levy, K., Barocas, S. and Hwang, T., (2017). Discriminating Tastes: Uber's

Customer Ratings as Vehicles for Workplace Discrimination. Policy & Internet, 9(3), pp.256

– 279.

Rosenblat, A. and Stark, L. (2016) 'Algorithmic labor and information asymmetries: A case

study of Uber's drivers', International Journal of Communication, 10, pp. 3758–3784. doi:

10.2139/ssrn.2686227.

Rosenbloom, B. (2007). The wholesaler's role in the marketing channel: disintermediation vs.

reintermediation. The International Review of Retail, Distribution and Consumer Research,

17(4), pp.327-339.

332

Rotmans, J., Kemp, R., Van Asselt, M., (2001). More evolution than revolution: transition

management in public policy. Foresight 3 (1), 15–31. (Online) Available at:

Doi:10.1108/JHOM-09-2016-0165 (Accessed 8 Feb. 2018).

Rosette, A., Leonardelli, G. and Phillips, K. (2008). The white standard: racial bias in leader

categorization. Journal of Applied Psychology, 93(4), pp.758-777.

Rheingold, H. (1993), The virtual community: homesteading on the electronic frontier,

addison-wesley, Reading, MA

Rupp, D., Vodanovich, S. and Crede, M. (2006). Age bias in the workplace: the impact of

ageism and causal attributions. Journal of Applied Social Psychology, 36(6), pp.1337-1364.

Russell, J. (2017). China’s Didi invests in Taxify, an Uber rival operating in Europe and

Africa. Techcrunch.com. [online]. Available from: https://techcrunch.com/2017/08/01/chinas-

didi-invests-in-taxify/ [Accessed May 28, 2019].

Rysman, M. (2009). The Economics of Two-sided Markets. Journal of economic perspectives,

23(3), pp.125 - 143.

Saal, P. (2018). Uber says it will have ‘zero’ tolerance for drivers who turned violent during

protest. Times. [online]. Available from: https://www.timeslive.co.za/news/south-africa/2018-

06-26-uber-says-it-will-have-zero-tolerance-for-drivers-who-turned-violent-during-protest/

[Accessed June 29, 2019].

Sade-Beck, L. (2004). Internet Ethnography: Online and Offline. International Journal of

Qualitative Methods, 3(2), pp.45–51. [online]. Available from:

https://doi.org/10.1177/160940690400300204

Sandberg, J. and Pinnington, A.H. (2009). Professional competence as ways of being: an

existential ontological perspective. Journal of Management Studies, 46(7), pp.1138–1170.

Satariano, A. (2019). This Estonian Start-Up Has Become a Thorn in Uber’s

Side. Nytimes.com. [online]. Available from:

333

https://www.nytimes.com/2019/04/23/technology/bolt-taxify-uber-lyft.html [Accessed June

25, 2019].

Savolainen, R., (2008). Everyday information practices: a social phenomenological

perspective. Scarecrow Press.

Scheiber, N., (2017). How uber uses psychological tricks to push its drivers’ buttons. [online]

Nytimes.com. Available at: <https://ceo.usc.edu/wp-

content/uploads/2019/02/2_How_Uber_Uses_Psychological_Tricks_to_Push_Its_Drivers_w

_Perf_Mgmt.pdf> [Accessed December 9, 2019].

Schildt, H. (2017). Big data and organisational design–the brave new world of algorithmic

management and computer augmented transparency, Innovation: Management, Policy, and

Practice. Routledge, 19(1), pp. 23–30. doi: 10.1080/14479338.2016.1252043.

Schillinger, H.R. (2005). Trade unions in Africa: weak but feared. Occasional Papers,

International Development Cooperation: Friedrich Ebert Stiftung, (March). [online].

Available from: http://library.fes.de/pdf-files/iez/02822.pdf.

Schmidt, A. (2017). ‘Digital Labor markets in the platform economy: mapping the political

challenges of crowd work and gig work”, Available at: https://library.fes.de/pdf-

files/wiso/13164.pdf [Accessed August 2, 2020]

Scotland, J. (2012). Exploring the philosophical underpinnings of research: Relating ontology

and epistemology to the methodology and methods of the scientific, interpretive, and critical

research paradigms. English Language Teaching, 5(9), pp.9–16.

Schor, J.B. et al. (2020). Dependence and precarity in the platform economy. Theory and

Society, 49(5–6), pp.833–861.

Schor, J.B. (2014). Debating the Sharing Economy. Great Transition Initiative, Available

at: http://www.greattransition.org/publication/debating-the-sharing-economy [Accessed

August 6, 2019]

334

Scott, J. (1990). Domination and the Arts of Resistance. Yale

University Press.

Scott, J. (1989). Everyday forms of resistance, Copenhagen Papers, No. 4, pp. 33–62.

Scott, J. (1985). Weapons of the weak. London: Yale University Press.

Sink, J. and Biles, R. (2016). Born in the bandwidth: digital native as pretext for age

discrimination. ABA Journal of Labor & Employment Law, 31, p.521. [online]. Available from:

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2887774 [Accessed December 3, 2019].

Sim, J. and Waterfield, J. (2019). Focus group methodology: some ethical challenges. Quality

& Quantity, 53(6), pp.3003 – 3022.

Shaheen, S. and Chan, N. (2016). ‘Mobility and the Sharing Economy-Potential to Facilitate

the First and Last Mile Public Transit Connections’, Built Environment, 42(4), pp. 573–588.

[Online] Available at doi: 10.2148/benv.42.4.573 [Accessed March 20, 2018].

Shapiro, A. (2018). Between autonomy and control: Strategies of arbitrage in the “on-demand”

economy. New Media and Society, 20(8), pp.2954–2971.

Shalhoub-Kevorkian, N. (2012). E-resistance and technological in/security in everyday life:

The Palestinian case. British Journal of Criminology, 52(1), pp.55–72.

Sheller, M. and Urry, J. (2006). ‘The new mobilities paradigm’, Environment and Planning,

38(2), pp. 207–226. [online] Available at: doi: 10.1068/a37268. [Accessed May 29, 2018]

Skinner, C. and Watson, V. (2017). The informal economy in cities of the GS. In G. Bhan, S.

Srinivas & V. Watson, ed. The Routledge Companion to Planning in the GS. London:

Routledge, pp. 140-151.

Skok, W. and Tissut, M. (2003). Managing change: the London taxi cabs case study. Strategic

Change, 12(2), pp.95-108.

335

Smith, J. and Tran, K. (2017). Smartphone adoption on the upswing in Nigeria. [online]

Business Insider. Available at: http://uk.businessinsider.com/smartphone-adoption-on-the-

upswing-in-nigeria-2017-4 [Accessed 6 May 2018].

Snider, L. (2002). Theft of time: disciplining through science and law. Osgoode

Hall Law Journal 40(1): 90 – 112.

Snider, L. (2000). The Sociology of Corporate Crime: An Obituary. Theoretical Criminology,

4(2), pp.169-206.

Sporton, D. (2013). ‘They control my life’: the role of local recruitment agencies in east

European migration to the UK. Population, Space and Place, 19(5), pp.443–458.

Stillwell, D. et al. (2016). National Travel Survey: England 2016. Available at:

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data

/file/633077/national-travel-survey-2016.pdf [Accessed April 7, 2018]

Stehlin, J., Hodson, M. and McMeekin, A. (2020). Platform mobilities and the production of

urban space: Toward a typology of platformization trajectories. Environment and Planning A,

52(7), pp.1250–1268.

Stevens, A. and Lavin, D., (2007). Stealing Time: The Temporal Regulation of Labor in

Neoliberal and Post-Fordist Work Regime. Democratic Comminiqué, 2(21), pp.41-56.

Stewart, A. and Stanford, J. (2017). Regulating work in the gig economy: What are the

options?. The Economic and Labour Relations Review, 28(3), pp.420-437.

Storrie, D. (2018). Non-standard forms of employment: Recent trends and future prospects.

Dublin: European Union. [online]. Available from:

https://www.eurofound.europa.eu/publications/customised-report/2017/non-standard-forms-

of-employment-recent-trends-and-future-prospects [Accessed January 4, 2020].

336

Strauss, A. and Corbin, J. (1990). Basics of qualitative research. Sage publications.

Sweeney, S. (2019). Uber, Lyft Drivers Manipulate Fares at Reagan National Causing

Artificial Price Surges. [online] WJLA. Available at: <https://wjla.com/news/local/uber-and-

lyft-drivers-fares-at-reagan-national> [Accessed 13 June 2020].

Sun, P. (2019). Your order, their labor: An exploration of algorithms and laboring on food

delivery platforms in China. Chinese Journal of Communication, 12(3), pp.308–323. [online].

Available from: https://doi.org/10.1080/17544750.2019.1583676.

Sundararajan, A. (2017). The sharing economy. Cambridge (MA): The MIT Press.

Sutherland, W. et al. (2020). Work Precarity and Gig Literacies in Online Freelancing. Work,

Employment and Society, 34(3), pp.457–475.

Tauss, A. (2012). Contextualizing the current crisis: post-fordism, neoliberal restructuring, and

financialization. Colombia Internacional, (76), pp.51-79.

Taylor, M. et al. (2017). Good work: The Taylor review of modern working practices. London:

Department for Business, Energy & Industrial Strategy.

Taylor, F. (1919). The Principles of Scientific Management. New York: Harper & Brothers.

Tellis, W. (1997). Introduction to Case Study. [online] 3(2). Available at:

https://nsuworks.nova.edu/tqr/vol3/iss2/4 [Accessed May 27, 2018].

Teo, T., Kimes, S. and Yong, Z. (2019). Case—The Rise and Fall of Taxi Club Management

in New York. INFORMS Transactions on Education, 20(1), pp.28-32.

TFL. (2021). Guidelines for CCTV systems in licensed London taxis and private hire vehicle.

Content.tfl.gov.uk. [online]. Available from: https://content.tfl.gov.uk/guidelines-for-cctv-in-

taxis-and-phvs.pdf [Accessed April 27, 2022].

337

Transport for London (TFL) and Uber London Limited, Licensed Taxi Drivers Association,

Licensed private hire association (2015) EWHC 2918.

Toner, J. (1992). Regulation in the Taxi Industry. Working Paper. Institute of Transport

Studies, University of Leeds, Leeds, UK.

Thelen, K. (2018). Regulating Uber: The Politics of the Platform Economy in Europe and the

United States. Perspectives on Politics, 16(4), pp.938-953.

Toner, J. (1996). English experience of deregulation of the taxi industry. Transport Reviews,

16(1), pp.79-94.

Topham, G., Hellier, D. and Gani, A. (2015). Uber wins high court case over taxi app. the

Guardian. [online]. Available from:

https://www.theguardian.com/technology/2015/oct/16/uber-wins-high-court-case-taxi-app-tfl

[Accessed April 15, 2022].

Timm, S. (2018). We've invested R2.3m in ride-hailing app - former security guard, prison

warden. Ventureburn. [online]. Available from: https://ventureburn.com/2019/07/taxi-live-

africa-ride-hailing/ [Accessed October 29, 2020].

Tsotsis, A. (2011). Shervin Pishevar Announces $20 Million Menlo Ventures “Talent Fund”.

[online] TechCrunch. Available at: https://techcrunch.com/2011/09/13/shervin-pishevar-

announces-20-menlo-ventures-talent-fund/ [Accessed May12, 2018].

Tassinari, A. and Maccarrone, V. (2020). Riders on the Storm: Workplace Solidarity among

Gig Economy Couriers in Italy and the UK. Work, Employment and Society, 34(1), pp.35–54.

Uber. (2016). The question of congestion. Uber.com. [online]. Available from:

https://www.uber.com/en-gb/blog/london/the-question-of-congestion/ [Accessed May 27,

2022].

338

Uber Africa Newsroom. (2016). Uber signs SOU with Ghanaian Ministry of Transport. Uber-

africa.africa-newsroom.com. [online]. Available from: https://uber-africa.africa-

newsroom.com/press/uber-ghana-uber-signs-sou-with-ghanaian-ministry-of-transport

[Accessed May 8, 2018].

Uber.com. (2017). Uber and Germaine Auto. Uber Blog. [online]. Available from:

https://www.uber.com/en-NG/blog/uberxgermaine/ [Accessed January 11, 2018].

Uber Africa Newsroom. (2016). Uber and PWC help close the transport-industry tax gap in

Nigeria. Uber-africa.africa-newsroom.com. [online]. Available from: https://uber-

africa.africa-newsroom.com/press/uber-and-pwc-help-close-the-transportindustry-tax-gap-in-

nigeria [Accessed May 16, 2019].

Urry, J. (2004) The System of Automobility, Theory, Culture & Society, 21(5), pp. 25–39.

(online) Available at: doi: 10.1177/0263276404046059. (Accessed 18 May 2018)

US Bureau of Labour Statistics. (2012). The Recession of 2007 - 2009.

Vanguard. (2018). Uber monthly pass. Vanguard. [online]. Available from:

https://www.vanguardngr.com/2018/08/uber-monthly-passenger-base-in-nigeria-hits-267000/

[Accessed December 5, 2019].

Vallas, S.P. (2019). Platform Capitalism: What’s at Stake for Workers? New Labor Forum,

28(1), pp.48–59.

Veen, A., Barratt, T. and Goods, C. (2020). Platform-Capital’s ‘App-etite’ for Control: A

Labour Process Analysis of Food-Delivery Work in Australia. Work, Employment and Society,

34(3), pp.388–406.

Vinthagen, S. and Johansson, A. (2013) ’Everyday resistance”: Exploration of a concept and

its theories’, Resistance Studies Magazine, (1), pp. 1–46. Available at: http://resistance-

journal.org/wp-content/uploads/2016/04/Vinthagen-Johansson-2013-Everyday-resistance-

Concept-Theory.pdf [Accessed July 16, 2020]

339

Vilakazi, T. and Chisoro-Dube, S. (2017). Two very different responses to Uber - Kenya and

South Africa. The Conversation. [online]. Available from: https://theconversation.com/two-

very-different-responses-to-uber-kenya-and-south-africa-74262 [Accessed July 10, 2022].

Vygotsky, L. (1978). Mind in society. Cambridge, MA: Harvard University Press

Wacquant, L. (2012). Three steps to a historical anthropology of actually existing

neoliberalism. Social Anthropology, 20(1), pp.66–79.

Wallsten, S. (2015). The competitive effects of the sharing economy: How is Uber Changing

Taxis? New York: Technology Policy Institute. Available:

https://www.ftc.gov/system/files/documents/public_comments/2015/06/01912-96334.pdf

[Accessed June, 2018]

Walsh, D. (2017). Dilemma for Uber and rival: Egypt’s demand for data on riders. Nytimes.

[online]. Available from: https://www.nytimes.com/2017/06/10/world/middleeast/egypt-uber-

sisi-surveillance-repression-careem.html [Accessed August 30, 2019].

Waters, F. and Woodcock, J. (2017). Far from seamless: a workers' inquiry at Deliveroo -

viewpoint Magazine. Viewpoint Magazine. [online]. Available from:

https://viewpointmag.com/2017/09/20/far-seamless-workers-inquiry-deliveroo/ [Accessed

January 5, 2018].

Watson, K. (2014). Easy Taxi: A company born out of frustration. BBC News. [online].

Available from: https://www.bbc.co.uk/news/business-27991612 [Accessed August 28, 2019].

Waring, P. (2016). Taylorism transformed: Scientific management theory since 1945. UNC

Press Books.

Wazid, M. et al. (2013). Hacktivism trends, digital forensic tools and challenges: A survey.

2013 IEEE Conference on Information and Communication Technologies, ICT 2013, (Ict),

pp.138–144.

340

Weiz, R. (2001). Women and Their Hair: Seeking Power through Resistance and

Accommodation. Gender and Society, 15(5), pp.667-686. [online]. Available from:

https://www.jstor.org/stable/3081969 [Accessed November 27, 2021].

Wilkinson, S. (1998). Focus group methodology: A review. International Journal of Social

Research Methodology, 1(3), pp.181–203.

Williamson, A.E. and Burns, N. (2014). The safety of researchers and participants in primary

care qualitative research. British Journal of General Practice, 64(621), pp.198–200.

Willson, M., (2016). Algorithms (and the) everyday. Information, Communication & Society,

20(1), pp.137 –150.

Willis, R. (2019). Observations online: finding the ethical boundaries of Facebook research.

Research Ethics, 15(1), pp.1–17.

WIPO (2021). Global Innovation Index 2021: Tracking Innovation through the COVID-19

Crisis. Geneva: World Intellectual Property Organization

Wood, J. et al. (2019). 'Good gig, bad gig: autonomy and algorithmic control in the global gig

economy', Work, Employment and Society, 33(1), pp. 56–75. doi:

10.1177/0950017018785616.

Wood, J., Lehdonvirta, V. and Graham, M. (2018). Workers of the Internet unite? online

freelancer organisation among remote gig economy workers in six Asian and African countries.

New Technology, Work and Employment, 33(2), pp.95–112.

Woodcock, J. (2020). The Algorithmic Panopticon at Deliveroo: measurement, precarity, and

the illusion of control. Ephemera: theory & politics in organizations, 20(3), pp.67–95.

Woodcock, J. and Graham, M. (2020). The gig economy. Cambridge: Polity Press.

Woodcock, J. and Johnson, M. (2018). Gamification: what it is, and how to fight it,

Sociological Review, 66(3), pp. 542–558. doi: 10.1177/0038026117728620.

341

Woodcock, J. (2021). The Limits of Algorithmic Management. South Atlantic Quarterly,

120(4), pp.703 –713.

World Bank (2020). World Development Indicators. GDP growth (annual %), [online].

Available from: https://databank.worldbank.org/source/world-development-indicators

[Accessed June 20, 2021]

World Bank. (2019a). The Changing Nature of Work. Washington DC: World Bank. [online].

Available from: http://doi:10.1596/978-1-4648-1328-3 [Accessed July 4, 2020].

World Bank (2019 b). World Development Indicators. Unemployment, total (% of labour

force) (national estimate). [online]. Available from:

https://databank.worldbank.org/source/world-development-indicators [Accessed May 16,

2020]

World Bank (2019b). World Development Indicators. Unemployment, youth total (% of labour

force 15 - 24) (national estimate), [online]. Available from:

https://databank.worldbank.org/source/world-development-indicators [Accessed May 16,

2020]

World Bank (2018). World Development Indicators. Inflation, Consumer Prices (annual %),

[online]. Available from: https://databank.worldbank.org/source/world-development-

indicators [Accessed September 16, 2020]

Wright, C. and Brown, W. (2013). The effectiveness of socially sustainable sourcing

mechanisms: Assessing the prospects of a new form of joint regulation. Industrial Relations

Journal, 44(1), pp.20-37.

www.africareport.com. (2012). Metro Taxi, Nigeria. [online]. Available from:

https://www.youtube.com/watch?v=EC6VDrvsqwA&t=84s&ab_channel=infoafricareport

[Accessed June 10, 2019].

342

Xiao, A.H. (2019). “Oyinbo, Wole!”: Urban Rhythms and Mobile Encounters in the Lagos

Transport Systems. Urban Forum, 30(2), pp.133–151.

Yin, R., (1984). Case Study Research: Design and Methods. Beverly Hills, Calif: Sage.

Yin, R. (2003). Case study research: Design and methods (3rd ed.). Thousand Oaks,

CA: Sage

Young, R.A. and Collin, A. (2004). Introduction: Constructivism and social constructionism

in the career field. Journal of Vocational Behavior, 64(3), pp.373–388.

Young, M. and Farber, S. (2021). Ride-hailing platforms are shaping the future of mobility,

but for whom? In A. Zwick & Z. Spicer, ed. The Platform Economy and the City: Urban Peril

and Promise in the New Digital Economy. McGill-Queens University Press, pp. 1-30.

Zaccheaus, B. (2017). LAGOS AND MOTOR VEHICLE ADMINISTRATION. Lagos State

Government Blog. [online]. Available from: https://lagosstate.gov.ng/blog/2017/07/05/lagos-

and-motor-vehicle-administration/ [Accessed October 3, 2020].

Zembylas, M. (2019). The affective dimension of everyday resistance: implications for critical

pedagogy in engaging with neoliberalism’s educational impact. Critical Studies in Education,

62(2), pp.211-226.

Zuboff, S. (2015). Big other: Surveillance capitalism and the prospects of an Information

civilization. Journal of Information Technology, 30(1), pp.75-89.

Zukin, S. (2020). Seeing like a city: how tech became urban. Theory and Society, 49(5–6),

pp.941–964.

Zuylen-Wood, S. (2015). The Struggles of New York City’s Taxi King. Bloomberg.com.

[online]. Available from: https://www.bloomberg.com/features/2015-taxi-medallion-king/

[Accessed April 16, 2022].

343

Zwick, A. (2018). Welcome to the gig economy: neoliberal industrial relations and the case of

Uber. GeoJournal, 83(4), pp.679–691.

344

8. Appendices

8.1. Appendix A: Ethical Approval for Research

I applied for a data management plan online (DPM) before applying for an ethical clearance

in June 2018. This was used for the first phase of fieldwork. A subsequent DMP application

was submitted in July 2019.

8.1.1. Appendix A1: Data Management Plan Application online

See the pages below.

345

346

347

348

349

350

351

352

353

354

355

356

8.1.2. Appendix A2: Ethical Approval letter

These DMPs led to my ethical clearance approvals for both applications, as seen below.

357

358

359

360

8.2. Appendix B: Fieldwork Data collection Approach and Instruments

8.2.1. Appendix B1: Participant Information Sheets

Please note that the Participant information sheets below incorporate sections for both SSIs,

FGDs, and participant observation measures to avoid duplicates.

Working Title: Ride-hailing apps and the impacts on Urban Mobility in the GS: A

Socio-technical approach to Taxi Regimes in Lagos

Participant Information Sheet (P.I.S.)

This P.I.S. should be read in conjunction with The University privacy notice

You are being invited to take part in a research study [as part of a student project – Ride-

hailing apps and the impacts on Urban Mobility in the GS: A Socio-technical approach

to Taxi Regimes in Lagos]. Before you decide whether to take part, it is important for you

to understand why the research is being conducted and what it will involve. Please take time

to read the following information carefully and discuss it with others if you wish. Please ask

if there is anything that is not clear or if you would like more information. Take time to

decide whether or not you wish to take part. Thank you for taking the time to read this.

Who will conduct the research?

Name: Daniel Arubayi

School: University of Manchester

Address: The University of Manchester

Oxford Rd. Manchester, M13 9PL, U.K.

What is the purpose of the research?

To understand and examine how ride-hailing technologies such as Uber in Lagos impact or

benefit the taxi services and actors such as drivers, passengers, transport authorities and so

forth.

Why have I been chosen?

361

The participant has been chosen because he/she is a key consumer/provider of this service

and is perceived to have an understanding of how ride-hailing services, taxis and urban

mobility in general work in Lagos. The overall number of participants for this study would

be around 70 participants.

What would I be asked to do if I took part?

The researcher will be conducting an interview regarding the aforementioned course of study.

Participants who agree to be interviewed would be required to respond to interview questions

when asked. Interview/focus group questions would be centred around experiences, usability,

perception or understanding, control, labour management, recommendations, and likes and

dislikes of how ride-hailing services such as Uber operate. Please note that interview

sessions would be duly recorded for analytical purposes.

What will happen to my personal information?

To undertake the research project, I will need to collect the following data about you:

- Name (Not compulsory)

- Gender

- Age range (applicable to drivers and passengers only)

- Income range (applicable to drivers and passengers only)

- Mobile phone number or any preferable communication platform of your choice (not

compulsory)

For audio recordings, the following would be observed:

• An audio recorder would be used for recording during interview sessions or focus

group discussions. No sensitive information would be required. Only basic personal

information would be required, as mentioned above.

• A verbal consent would be recorded in the circumstance where the participant is

unable to read or understand a written consent.

• This would be an audio-only recording

• Audio recordings will be transcribed and analysed for the purpose of this research.

362

[Only] the research team will have access to this information.

We are collecting and storing this personal information in accordance with the General Data

Protection Regulation (GDPR) and Data Protection Act 2018, which legislate to protect your

personal information. The legal basis upon which we are using your personal information is

"public interest task" and "for research purposes" if sensitive information is collected. For

more information about the way we process your personal information and comply with data

protection law, please see our Privacy Notice for Research Participants.

The University of Manchester, as Data Controller for this project, takes responsibility for the

protection of the personal information that this study is collecting about you. In order to

comply with the legal obligations to protect your personal data, the University has safeguards

in place, such as policies and procedures. All researchers are appropriately trained, and your

data will be looked after in the following way:

The study team (the researcher and supervisors) at the University of Manchester will have

access to your personal identifiable information, that is, data that could identify you, but they

will anonymise it as soon as it is transcribed. However, your consent form, contact details, etc

will be retained for at least three years on the University of Manchester data storage system.

You have a number of rights under data protection law regarding your personal information.

For example, you can request a copy of the information we hold about you, such as audio

recordings. This is known as a Subject Access Request. If you would like to know more

about your different rights, please consult our privacy notice for research and if you wish to

or write to The Information Governance Office, Christie Building, University of Manchester,

Oxford Road, M13 9PL. at the University, and we will guide you through the process of

exercising your rights.

You also have a right to complain to the Information Commissioner's Office, Tel 0303 123

1113

Will my participation in the study be confidential?

Your participation in the study will be kept confidential and only visible to the study team of

this project.

363

The following would be observed for audio recording:

- The researcher would be responsible for transcribing the information from the

recordings

- Any personal information would be removed from the final transcript

- The recordings would be backed up on the University of Manchester data storage

cloud and Dropbox business data storage.

- Data online or transcribed would be kept for the period of study (3 years and a

maximum of 4 years).

- Where a participant mentions a piece of personal information in an interview, such as

the name of a person or company, it would be anonymised or muted using voice

masking software.

- The research team (researcher and supervisors) would have access to the research

information

What happens if I do not want to take part or if I change my mind?

It is up to you to decide whether or not to take part. If you do decide to take part, you will be

given this information sheet to keep and be asked to sign a consent form or respond to a

verbal consent. If you decide to take part, you are still free to withdraw at any time without

giving a reason and without detriment to yourself. However, it will not be possible to remove

your data from the project once it has been anonymised and forms part of the dataset, as we

will not be able to identify your specific data. This does not affect your data protection rights.

Will my data be used for future research?

When you agree to take part in a research study, the information about your health and care

may be provided to researchers running other research studies in this organisation. The future

research should not be incompatible with this research project and will concern digital

technology platforms and mobility. These organisations may be universities, N.H.S.

organisations or companies involved in health and care research in this country or abroad.

Your information will only be used by organisations and researchers to conduct research in

accordance with the U.K. Policy Framework for Health and Social Care Research.

This information will not identify you and will not be combined with other information in a

way that could identify you. The information will only be used for the purpose of health and

364

care research and cannot be used to contact you regarding any other matter or to affect your

care. It will not be used to make decisions about future services available to you.

Will I be paid for participating in the research?

Transportation costs, packed lunch and drinks would be given to participants who attend

focus group discussions and offer their time for interviews.

What is the duration of the research?

Interview sessions would vary across participants as follows:

- Drivers and passengers (at least 2 hours interview per person or less).

- Other participants such as LAMATA representatives, BRT, ride-hailing firms, a start-

up for online payment (at least 1-hour 30 minutes interview per person or less)

- Consent forms would be given to participants 24 hours before commencing any

interview.

- Alternatively, if there is any language barrier, such as a difficulty in reading consent

forms, it would be duly explained verbally at least 24 hours before commencing.

Focus Group Discussions would vary across participants as follows:

- For platform drivers (FGD ranging between 45 minutes to 1 hour 30 minutes per

session).

- For conventional taxi drivers (FGD ranging between 45 minutes to 1 hour 45 minutes

per session).

- For passengers (FGD ranging between 45 minutes to 1 hour 45 minutes per session)

- Consent forms would be given to participants 24 hours before commencing any

focus group discussion.

- Alternatively, if there is any language barrier, such as a difficulty in reading consent

forms, it would be duly explained verbally at least 24 hours before commencing.

365

Where will the research be conducted?

- For drivers and passengers, interviews/focus group discussions would be conducted at

taxi ranks, within the vehicle or any public place mutually agreed upon.

- For ride-hailing drivers' interviews/focus group discussions would be conducted at

public places such as eateries, within the vehicle, screen and certification centres.

- For passengers, interviews would be conducted in homes, offices, or any public place

mutually agreed upon.

- For transport authorities, ride-hailing firms and start-ups, interviews would be

conducted in their offices after an appointment is booked.

- Interviews would be conducted in any public place as mutually agreed between

researcher and participants.

Will the outcomes of the research be published?

Interested participants would be informed of the findings as well as any publication which

comes out of the analysis.

Who has reviewed the research project?

This project has been reviewed by the University of Manchester Research Committee –

School of Environment, Education, and Development division.

What if I want to make a complaint?

If the participant has any minor or formal complaints, read the information below.

Minor complaints

If you have a minor complaint, then you need to contact the researcher(s) in the first

instance. For further complaints, kindly send an email to: [email protected];

James.Z.Evans@manchester.ac.uk or call +44 (0)161 275 0969

366

Formal Complaints

If you wish to make a formal complaint or if you are not satisfied with the response you have

gained from the researchers in the first instance, then please contact

The Research Governance and Integrity Manager, Research Office, Christie Building,

University of Manchester, Oxford Road, Manchester, M13 9PL, by emailing:

research.c[email protected] or by telephoning 0161 275 2674.

What Do I Do Now?

If you have any queries about the study or if you are interested in taking part, then please

contact the researcher(s): daniel.arubayi@manchester.ac.uk or call +44 (0)161 275 0969

This Project Has Been Approved by the University of Manchester's Research Ethics

Committee [ERM reference number]

8.2.2. Appendix B2: Participant Consent Form

This research had two separate consent form versions based on separate fieldwork phases. To

avoid duplications, both versions for both SSI and FGD have been incorporated into one

document.

Consent Form

If you are happy to participate, please complete and sign the consent form below:

S/N

Activities

Initials

I confirm that I have read the attached information sheet (Version 1, Date

05/10/2018) or (Version 2, Date 10/07/2019) for the above study and have

had the opportunity to consider the information and ask questions and had

these answered satisfactorily.

I understand that my participation in the study is voluntary and that I am free

to withdraw at any time without giving a reason and without detriment to

myself. I understand that it will not be possible to remove my data from the

project once it has been anonymised and forms part of the data set.

367

I agree to take part on this basis

I agree to the interviews/focus group sessions being audio recorded.

I agree that any data collected may be published in anonymous form in

academic books, reports, or journals

I agree that the researchers/researchers at other institutions may contact me

in future about other research projects.

I agree that the researchers may retain my contact details in order to provide

me with a summary of the findings for this study.

I understand that there may be instances where during the course of the

interview/focus group, information is revealed, which means that the

researchers will be obliged to break confidentiality, and this has been

explained in more detail in the information sheet.

I agree to take part in this study.

Data Protection

The personal information we collect and use to conduct this research will be processed

in accordance with data protection law as explained in the Participant Information

Sheet and the Privacy Notice for Research Participants.

________________________ ________________________

Name of Participant Signature Date

________________________ ________________________

Name of the person taking consent Signature Date

The original copy of the consent form would be retained by the research team, while one

copy would be printed for the participant.

368

8.2.3. Appendix B3: Detailed Ethical Considerations.

Typologies

Semi-structured interviews and Focus Group

discussions

Mobile participant Observations

Online Participant Observations

Informed

consent

In both SSIs and FGDs, consent was sought from

participants. For SSIs, a participant information

sheet describing my project and a consent form were

issued to participants both in-person and via media

platforms at least 24 hours before interviews. In the

case of ride-hailing platforms like Taxify and

Max.ng, an email and hardcopy consent forms and a

participant information sheet were sent to these

companies. However, only Max.ng agreed to an

interview.

For FGDs, a copy was sent out to platform drivers

and taxi drivers that showed interest. In addition,

consent was also read out and audio-recorded before

the discussion started because most participants did

not have the time to go through my participant

information sheet and consent forms.

Sim and Wakefield (2019) reiterate the lack of

information or inadequate disclosure as a weakness

These were widely informal observations

and conversations, as stated earlier.

Platform drivers were recruited specifically

for SSIs in scenarios where this merged

with a more structured interview. Out of 40

trips of observation, four drivers were

exclusively recruited for telephone-based

SSIs with consent forms and participant

interviews given to them before further

discussion. For the remaining 36 trips,

drivers were informed about my status as a

researcher.

Scholars such as Madge (2007), for instance,

argue for the mandatory requirement for

consent concerning public information in

online environments. However, with

progressive digital development and public

information on platforms, the difficulty of

following conventional consent patterns has

been increasingly blurred and difficult to

practice (Henell et al., 2020). According to

Willis (2019, pp.3), there are two scenarios

where a researcher might evade informed

consent, such as when the data is treated as

'documentary or textual information' and

when 'research involves observation of

human subjects in a public space. Willis

(2019) observes Facebook news feeds as both

textual but entwined as 'human subject

research' where news from online

369

for FGDs, which is a similar challenge to interviews

(Allmark et al., 2009).

The lack of information or inadequate disclosure as

a weakness in both interviews and FGDs of a

participant can influence their involvement level

(Allmark et al., 2009; Sim and Wakefield, 2019). In

this case, participants were informed of every detail,

such as the purpose of the study, duration, potential

benefits and risks, an overview of what will be

covered and other themes.

environments reflects offline experiences and

vice versa. Like Willis, I liken my Uber &

Taxify Platform observations where consent

cannot be collected from every person in a

public space.

In my study, I was able to submit a consent

form and information sheet to the

administration of the Uber & Taxify private

group before I was accepted. Following this,

I observed drivers passively and actively

between April 2019 to September 2019.

Consents were collected for screenshots used

in this study. Consents for saved posts were

not sought, mainly when not used.

Confidentiality

and Anonymity

Longhurst 2016, views these typologies as the most

important for both SSI and FGDs Participants are

entitled to know if their information would be

confidential, especially if the information given

becomes sensitive. In the analysis of the data, the

names of participants were anonymised using

pseudonyms – that replacing the real names of

participants with fake names. This, I believe, was

Considering I was also a participant during

the trips assigned to this study, the driver

information aspects on trip screenshots are

confined to the app and censored.

Screenshots that were used were

anonymised or censored using photo-

editing software.

The consent form given to both the platform

group administration, drivers recruited via

this process, and screenshots used by drivers

all included a column for confidentiality and

anonymity. The names and profile pictures

were all censored using photo-editing

software. Quotes from posts used in this

study were paraphrased, and the names were

370

substantial in preserving the identity of participants

in this study.

pseudonymised to protect drivers' identities

and prevent online search queries that lead

back to them.

Risk of Harm

The risk of harm is often an attribute of studies that

involve vulnerable people, bereavement research,

medical research, and general sensitive information,

which could potentially undermine the safety of

both the participant and interviewer in both SSIs and

FGDs (Allmark et al., 2009; Sim and Wakefield,

2019).

The amount of sensitive information collected in

this study was minimal in both SSIs and FGDs;

these include participants' names, information of

platform fraudulent behaviours and plans, plans for

a driver-led platform start-up. Steps of

confidentiality and anonymity provided through

consent forms were substantial in preserving the

participants from potential harm such as disruption

of existing social relationships with colleagues,

embarrassment or adverse employment (in this case,

independent contractual) consequences (Sim and

Wakefield, 2019). The pseudonyms given to

The trips involved in mobile observations

were safe and unharmful to drivers. I was

aware of any potential harm that could

affect drivers if I interviewed them during

trips. As mentioned, this could cause a

distraction and potentially an accident

during a trip. The informal conversation

with drivers was usually initiated by the

drivers at their convenience. There were no

conversations with drivers that were not

talkative

The risk of harm also applies to online

environments like the platform driver group

on Facebook used in this research. Hennell et

al. (2020) outline how sensitive information

such as publishing usernames from online

communities could potentially affect the

participant's reputation, family life, career

online and offline, and generally how people

perceive them. Again, additional measures

such as censoring usernames on screenshots,

profile pictures, and confidentiality assurance

via consent forms prevent participants' risk of

harm.

371

participants preserve their identity and ensure safety

from unforeseen circumstances.

Private vs

Public Data

The overt nature of social media content

creates ethical challenges in data collection

techniques, analysis and sharing of

information because of unclear boundaries

between public and private domains (Hennell

et al., 2020). This debate of whether

information posted in public domains has

occurred over the last decade remains

unclear. It has been argued that data that can

be accessed without any website registration

can be considered in the public domain

(Sudweeks and Rafaeli, 1996). However, this

also depends on the expectations and

characteristics embedded in social media

platforms. For example, private chat rooms

of groups with friends on social media

platforms would require different kinds of

accessibility and governance (Hennell et al.,

2020).

372

The information from Uber and Taxify

private groups were private, requiring an

application to join the community, which

could be accepted or denied. However, the

information, as well as the administration

across platforms, could be semi-public. This

is because the information is shared in both

publicly accessible groups and other

privately accessible groups. Also, within the

private community of over 17,000 members,

platform drivers are aware of the potential of

information to be public. This is why there

are adverts for more private WhatsApp

groups for various reasons, as discussed

previously.

However, the information collected was

private, hence the need to censor profile

photos and names on screenshots.

Incentives and

Payments

Participants, especially drivers, were reimbursed for

their time, transport, and general contribution in this

study. No cash payments were given to participants.

However, payments for transport costs, feeding or

Each trip taken was charged accordingly,

except for the four drivers exclusively

recruited for interviews that were also

further compensated.

Platform drivers recruited via the Facebook

platform for SSIs were compensated, as

discussed in the SSI column.

373

mobile data in the case of telephone-based

interviews were given to drivers.

For FGDs, snacks, food from roadside sellers, and

drinks were given to participants after each session.

The last FGD Meeting with the NUPEPD union

leaders consisted of snacks, drinks and three power

banks for their contribution.

374

8.3. Appendix C: Researcher's guide and Coded Participant Information

This section comprises the interview and focus group guides for the different participants in this study.

8.3.1. Appendix C1: Focus Group Discussion Guide for Ride-hailing Platform drivers

Qualitative Methods

Questions

FGD with Ride-hailing

Drivers

Descriptive and Demographic Questions

1. What platform do you drive for?

2. Why do you think it is flexible?

3. Do you have a family?

4. Are you educated or a graduate?

5. What's your age range 18 – 25; 26 – 30; 31 – 40; 40 and above?

Gadget specification

6. What kind of phones are used for Uber and Taxify?

7. Do Uber and Taxify provide you with phones and a monthly data allowance?

8. What networks (MTN., 9 mobile etc.) work better with the app?

9. Do your cars have trackers, stickers or dash cams?

10. What is the current car specification for drivers?

Registration

11. Talk me through the registration process for Uber and Taxify?

12. Do you have a copy of the Uber exam aptitude question, or is it online?

Safety, resistance, and Coping Mechanisms

375

13. When you have a problem with a passenger, how long does it take for Uber/Taxify to respond? What is the first thing

you do?

14. Does it reflect on the app as a pending message or follow-up message?

15. Have you ever used the Taxify SOS button? Does it even work?

16. What are some of the strategies as a driver do you use to protect yourself before, during and after a trip?

17. Between Uber and Taxify, which platform investigates conflicts between drivers and passengers more?

18. Have you used the app before? Why?

19. How did drivers react to the price slash on Uber?

20. How did you discover about the lockito app?

21. What cash transfer or bank app do you usually use?

22. Has the Uber/Taxify platform enabled offline trips for you?

23. Is Uber/Taxify aware of the use of the Lockito app? How are they combating it?

24. What are some other ways of bye-passing the app?

25. Why do drivers resist the app? Do they feel restricted?

Driver motivation.

26. What motivates you to drive for Uber/Taxify?

27. Are you fulfilled driving for Uber/Taxify?

Price competition and regulations

28. Do price regulations and commissions motivate you to choose between platforms?

29. Do promos/incentives motivate you to work longer or not?

30. Between Uber and Taxify, which platform rewards drivers more?

31. Need to know about the insurance and regulations/restrictions from Taxify and Uber

Livelihood and income

32. When did you start driving for Uber/Taxify?

33. Is it your car, or are you working to own the car?

34. Do you drive for Uber/Taxify part-time or full-time?

35. Part-time – if you have another job, why drive for Uber/Taxify?

36. Fulltime – Why do you drive for them full-time?

376

37. If you had another job or a better-paying job, would you drive for them?

38. The money made from this job, is it enough to cater for your needs and your family?

39. Compared to 2/3 years ago, are you making more or less now? Why is it like that?

Algorithm Management and Ratings

34. On what basis do you rate a riders high (4 or 5 stars) or low (2 or 1 star)

35. How do you react to surge? Do you target surge areas or work during surge times?

36. Do riders ratings motivate you to drive better?

37. Do promos entice you to work longer and perhaps make more income?

38. Do the ratings of a rider before a trip influence your decision to decline or accept a trip? (i.e., such as low rating and high

ratings)

F.G.D. with Ride-

hailing Union

Platform Emergence

1. How many ride-hailing platforms are there in Lagos?

2. Does this union plan to integrate and evolve traditional taxi unions? If yes, how?

3. Does this union incorporate other innovative platforms such as bike-hailing platforms in Lagos?

4. Does the state ministry of transport support the emergence of ride-hailing unions?

5. How does the reality of the city Lagos affect drivers' everyday labour?

Challenges & Resistances

6. What are the challenges that led to the formation of your platform union?

7. How are unions able to effectively bring drivers together against platforms?

8. Do platforms secretly disrupt union plans considering revelations from a few drivers?

9. How can unions tell if platforms are disrupting their plans, and what are you doing to avoid or resist such attempts?

10. How are the challenges of ride-hailing platforms different from traditional taxi/transport labour in Lagos?

11. How are platform unions different from traditional taxi/transport unions in Lagos today?

Platform Competition

12. How has the competition between Uber and Bolt impacted your labour?

13. Do the vehicle owners add to the challenges of ride-hailing drivers? If yes, how?

377

14. Why do drivers patronise more international platforms compared to indigenous platforms?

378

8.3.2. Appendix C2: Semistructured Interview Guide for Ride-hailing Platform Drivers

Interview guide

Semi-Structured

Interview guide with

Rid-hailing drivers

Descriptive and demographic questions

1. What platform do you drive for?

2. Why do you think it is flexible?

3. Do you have a family?

4. Are you educated or a graduate?

5. What's your age range 18 – 25; 26 – 30; 31 – 40; 40 and above?

Platform registration and driver preferences

6. Which platform do you prefer and why?

7. Before signing up for Uber, what are the kinds of questions do they ask in their assessment test?

8. When did you start driving for Uber or Taxify?

9. Why did you start driving for Uber or Taxify?

Labour Nature, Income and Working relationships

10. Did you have a previous job before Uber?

11. How did you feel when you made your first income on the platform?

12. Did you read the terms and conditions of the contract before signing up for it?

13. Is there any difference with what is happening currently in reality?

14. What plan are you under, your personal car, hire-purchase, or weekly car hire?

15. How do you get information about a person that wants to lease his/her car on hire-purchase?

16. How much do you remunerate to the owner of the car weekly, and how much do you keep for yourself?

17. On average, how much goes into maintenance, buying fuel and other miscellaneous activities?

18. Generally, how is your relationship with the owner of the car?

19. How many car managers have you had?

379

20. Are platform companies aware of this?

21. Do car owners treat you fairly?

22. What was your previous work experience like?

23. Do you understand your right as a worker, such as employee benefits, even though platforms do not recognise you as

an employee?

24. Sign-up contacts as fleet management

25. Why did you quit working/driving for Uber?

26. Was Uber a transition or side job for you?

27. How has working on this platform helped with your decision to choose your current job?

28. What is a typical day for you like?

29. Does this work affect your work-life balance? Does it give you time for your family (if you have one) and other activities?

30. Health is wealth. Does it give you time to rest?

Challenges in working for platform ride-hailing companies and coping strategies

31. What are your challenges on these platforms? (After challenges have been outlined, you'd probe further on specific

challenges)

32. How do you cope with these challenges?

33. What are some of the challenges or difficulties that hinder you from working in a day?

34. How do you deal with traffic congestion?

35. How did the fare reduction of 40% by Uber in 2017 impact how you work?

36. Do vehicle managers, riders add to your challenges on the job?

37. Are platform companies helping you deal with challenges or not?

38. How do you deal with conflicts with riders?

39. Do you think the platform companies are fair to you and your work?

40. Why do you think they are not fair to you? (No driver protection, no employee benefit or pension after they quit

41. What do you think can be done better, and what do you suggest?

42. In terms of safety and security, where are the red-hot zones in Lagos, and how do you avoid them? Around what times?

Driver coordination and unionisation (Resistance/Coping strategies)

380

43. How do you coordinate yourselves with other drivers? (Are you part of any groups on social media such as Whatsapp

groups, Telegram, Driver forums, and so forth). Why?

44. Are you aware of any E-hailing union in Lagos? If yes, which one? If no, why not?

45. Why are you part of the E-hailing union? Any preferences?

46. I understand that there have been recent strikes from the union, are you aware?

47. If yes, was this strike just via the app, or it also meant drivers refused to work completely for platforms? In other words,

did drivers take offline trips on these days?

48. How did you respond to the May 8

global E-hailing strike called #appsoff?

Algorithmic assignments and interaction between apps and drivers.

49. What do you think about the in-app bonuses, and how do you respond to them?

50. What do you think about surge fares, and how do you respond to them?

51. How do you feel and understand in-app messages that nudge you to keep driving? (give an example)

52. On average, how many hours do you work daily and weekly? (Take screenshot)

53. How many minutes does it take for you to cancel or accept a trip?

54. How do the following signals make you feel about yourself and your job?

A. Low or high activity score

B. Number of hours worked per day or per week

C. Acceptance rates and cancellation rates

D. High or low ratings from a rider (do you consider the ratings of passengers before picking them up?). If no, why

not?

E. Distance travelled

F. Percentage of cash or card trips.

55. Do you prefer cash or card trips? (Why do you prefer any of these options)

56. Do you prefer short or long trips? (Why do you prefer any of these options)

57. Which areas are you certain of receiving trip requests and why those key areas, and around what times?

58. How do you game the system for requests?

59. What coping strategies do you have to optimise your time on the app in order to make more income? (Lockito; Battery

saving mode; offline trips; phone 6).

60. Have you ever been blocked? Why?

61. Does it record how many times you have been blocked on the app?

381

Commission

Artefacts: Screenshots of app messages; employment contracts.

SSI guide with

traditional taxi Drivers

Description and demographic questions

1. Age range (18-25; 26 – 30; 31 – 40; 40 and above)

2. Educational qualifications

3. Do you have a family?

Registration and taxi industry requirements

4. How long have you been driving taxis?

5. What are the requirements for working for your taxi company?

History, taxi Labour and unionisation

6. Are there employee benefits? (What are they?)

7. What Union or association are you part of?

8. What is the difference between NURTW, RTEAN, LSTDCOAN and other unions?

9. What do you think about platform work and its drivers?

10. Do you pay monthly dues?

11. Do you have monthly meetings?

12. What is a typical workday like for you?

13. In comparison to 5 – 7 years ago, on average, how much do you make a month?

14. Why do you think your income on the job has been reduced?

15. On average, how many hours do you work daily? (What time to what time

16. What kind of discussions do you have in union/association meetings?

17. Who sets the prices before any trip is taken?

18. How do you know which areas to park and wait for rides?

19. As a taxi driver of this park, can you park in another car park or in another public place such as a hotel without getting

queried?

20. In your park, how are trips assigned to each worker? (turn-by-turn basis or phone call trips?)

382

21. Before Uber, what were the strategies for attracting more jobs or trips daily?

22. What are the strategies for coping with the scarcity of trips now? Or How do you cope with the scarcity of trips now?

Challenges and resistance/coping strategies of drivers

23. What have been the main challenges of taxi work for you since you started it?

24. Has your patronage from passengers also been reduced? (Why do you think so?)

25. On average, how many passengers did you get daily, and how many do you get now?

26. In terms of safety and security, where would you say are the red-hot zones in Lagos, and how do you avoid them?

27. Have you heard about platform companies such as Uber and Taxify?

28. Do you think they are responsible for the changes you have been experiencing in your work? If yes, why do you think so?

29. How have you been coping since Uber came into the picture?

30. What strategies have helped you access passengers and get more trips?

31. Which of these strategies are most effective?

32. Have you considered ever joining platform companies such as Uber? (if yes, why, if no, why not?)

33. Have you at least considered using the technology they use through apps in accessing passengers?

34. Did you hear about Afro-Taxi back in 2012/2013?

Perception of technology in the taxi business, awareness, and usage

35. If apps are introduced into your work now, will you be willing to learn how to use them, or you would still prefer the

traditional way of accessing passengers?

Regulation of taxis and platform companies.

36. Based on the current regulations and policies, do you think the government has failed the taxi industry?

37. What do you think can be done to rectify this?

S.S.I. guide for

Transport ministry

History of taxis and early Platforms

383

representatives &

union representatives

1. Can you outline a brief history of traditional taxis, including early innovative platforms like Afrocabs?

2. What is your perception of the rampant rise of ride-hailing platforms?

3. What is the idea behind the modernisation of the taxi system? Why did Lagos state decide to modernise it?

4. What counts as traditional taxi companies?

5. Are red cabs part of traditional or early taxi companies?

6. What were some of the failures of traditional taxis? How and why did traditional taxis fail?

Emergence of Platforms

7. With the emergence of platform companies, is it difficult for this administration to regulate them?

8. When platform companies register their platforms, do they do this at the state ministry of transport?

9. Did traditional taxi drivers use taximeters in Lagos?

10. Considering the emergence of ride-hailing platforms, how are the interests of traditional taxis being protected? Are they

even?

11. If state-owned platforms are commissioned, how do they intend to make profits?

12. How would state-owned platforms improve on the disadvantages of ride-hailing platforms?

13. Does the transport ministry intend to regulate vehicles against third-party ride-hailing platform owners? If yes or no,

why?

Driver challenges perceptions, data management and unionisation

14. What is your perception of safety & security between ride-hailing platforms and traditional taxis?

15. Does the state practice effective data management, and Is there a database for taxi registrations?

16. With the platform technology, do you think in the next couple of years, Lagos state is going to platformise taxis? If so, do

you think it would favour people?

17. Does the ministry work with traditional taxi unions as well as platform unions? If yes, how?

18. Is there a plan to unite platform unions & traditional taxi unions?

19. Are traditional taxis in the formal or informal sector?

20. What is the difference between NURTW and RTEAN?

21. Why is the government fighting taxis and tagging them as illegal?

22. The daily levies collected by these unions are they legal?

23. Do they pay taxes?

384

S.S.I. guide for

International Platform

companies (Uber &

Taxify)

Platform Emergence and Challenges

1. What prompted your company (Uber, Taxify) to move into the Nigerian market?

2. Why did you choose Lagos and Abuja first?

3. What series of funding motivated a move to Africa and Nigeria in particular?

4. What were some of the peculiar difficulties that you encountered in launching in Lagos?

5. Particularly, did you experience any regulatory difficulty and are you experiencing any at the moment?

6. How were you able to convince drivers to be a part of your platform?

7. Did the Nigerian market influence some of your policies, such as cash payments?

8. In what ways is your company impacting the taxi industry and overall mobility in Lagos?

9. Are you aware of any E-hailing union of drivers?

10. Have you considered working with the E-hailing union of drivers? If yes, why? If no, why not?

S.S.I. guide for Local

platform companies

(Oga-taxi)

Platform Emergence and Challenges

1. What inspired you to start your company?

2. Where do you draw your idea of E-hailing from?

3. Why did you choose to start in Lagos?

4. Are you self-funded, backed by any global series of funding, or being funded locally?

5. What were some of the peculiar difficulties that you encountered in launching in Lagos? (follow-up questions)

6. Do some of these difficulties still exist now?

7. Particularly, did you experience any regulatory difficulty and are you experiencing any at the moment?

8. How were you able to convince drivers to be a part of your platform?

9. In what ways is your company impacting the taxi industry and overall mobility in Lagos?

10. Are you aware of any E-hailing union of drivers?

11. Have you considered working with the E-hailing union of drivers?

12. If yes, why? If no, why not?

13. Are international ride-hailing companies affecting your company's profit gains? Or do you think it is healthy competition

at the moment?

385

8.3.3. Appendix C3: Coded Thematic Data and Pseudonymised Participant Information.

This section presents the thematic coded data and an overview of participants who participated in this research. Also, it includes an overview

of participant observations, including online observations from the Facebook platforms and mobile observations of ride-hailing platform drivers.

The data collection took place between September 2018 to January 2019 and July to September 2019. It is important to note that the second phase

of data collection was more effective than the first phase because of my improved understanding of the subject of ride-hailing workers and

platformised labour. Also, the first phase of fieldwork was disrupted by the campaign for the 2019 election due to violence and lack of safety in

Lagos. While SSI interviews were 46, the total number of participants, including FGDs, equates to 70 people. It is worth noting that 10 participants

overlapped across SSI and FGDs, reducing the total from 70 to 60 people.

C1: Demographic and contextual participant information in this study.

Methods

Pseudonyms of

Participants

Gender

Age

Group

Marital

Status

Level of

Education

Income Range Weekly

(before

remittances/expenses)

Income Range Weekly (after

remittances/expenses)

SSI with Ride-

hailing drivers

Charles

31 – 40

Single

BSC

N60,000 – N70,000

N20,000 – N30,000

Koffi

31 – 40

Married

HND

N60,000 – N70,000

N20,000 – N30,000

Efe

26 – 30

Single

B.Eng.

N50,000 – N60,000

N10,000 – N20,000

Henry

26 – 30

Single

BSc

N50,000 – N60,000

N10,000 – N20,000

Jude

31 – 40

Single

BSc

N50,000 – N60,000

N20,000 – N30,000

Jonathan

31 – 40

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

Thomas

26 – 30

Married

SSCE

N50,000 – N60,000

N10,000 – N20,000

George

31 – 40

Single

HND

N60,000 – N70,000

N20,000 – N30,000

386

Akpos

31 – 40

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

Jacob

31 – 40

Married

MSc

N60,000 – N70,000

N20,000 – N30,000

Obus

26 – 30

Single

MSc

N60,000 – N70,000

N20,000 – N30,000

Eguono

31 – 40

Married

BSc

N60,000 – 70,000

N20,000 – N30,000

Samuel

26 – 30

Single

MSc

N60,000 – 70,000

N/A

Temi

40 &

above

Married

HND

N60,000 – N70,000

N20,000 – N30,000

Junior

31 – 40

Single

BSc

N50,000 – N60,000

N10,000 – N20,000

Okoro

31 – 40

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

Uche

31 – 40

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

Akin

40 &

above

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

Osahon

31 – 40

Single

BSc

N50,000 – N60,000

N10,000 – N20,000

Akpan

31 – 40

Married

HND

N50,000 – N60,000

>N10,000

Dipo

31 – 40

Married

B.eng

N50,000 – N60,000

N10,000 – N20,000

Abel

40 &

above

Married

BSc

N60,000 – N70,000

N20,000 – N30,000

Abiodun

31 – 40

Single

BSc

Not disclosed

Jolomi

40 and

above

Married

BSc

Not disclosed

N.B: Drives for

both platforms &

traditional taxis

Olawumi

40 &

above

Married

BSc

N50,000 – N60,000

N10,000 – N20,000

387

SSI with Local

Taxi drivers

Kehinde

60 &

above

Married

HND

N/A

N5, 000 – N10,000

Joseph

40 &

above

Married

SSCE

N/A

N5,000 – N10,000

Deji

40 &

above

Married

SSCE

N/A

N5,000 – N10,000

Adebanjo

40 &

above

Married

HND

N/A

N5,000 – N10,000

Ahmed

40 &

above

Married

HND

N/A

N5,000 – N10,000

Adedayo

40 &

above

Married

BSc

N/A

N5,000 – N10,000

Jeff

40 &

above

Married

BSc

N/A

N5,000 – N10,000

Tom

31 – 40

Married

SSCE

N/A

N5,000 – N10,000

Isaac

40 &

above

Married

SSCE

N/A

N5,000 – N10, 000

Taiwo

60 &

above

Married

SSCE

N/A

N5,000 – N10,000

Ewoma

26-30

Single

BSc.

N/A

Teniola

26 –30

Single

MSc

N/A

388

SSI with Riders

Chidera

18 – 25

Single

BSc

N/A

Akunna

31 – 40

Single

MSc

N/A

Edeki

26 – 30

Single

MSc

N/A

SSI Transport

ministry

representatives,

lecturer, and

Venture

Capitalist

Mustapha

41 &

above

Married

BSc

N/A

Lateef

60 &

above

Married

MSc

N/A

Funmilayo

41 &

above

Married

PhD.

N/A

Abiola

31 – 40

Married

MSc

N/A

SSI for platform

bike-hailing

company Max.ng

Benjamin

31 – 40

Single

MSc

N/A

Kene

26 – 30

Single

N50,000 – N60,000

N10,000 – N20,000

Noah

26 – 30

Single

Not disclosed

Alex

31 –40

Single

BSc

N60,000 – N70,000

N20,000 – N30,000

389

FGD of 6 ride-

hailing drivers

October 26

2018

Charles

31 – 40

Single

BSc

N60,000 – N70,000

N20,000 – N30,000

Efe

31 – 40

Single

BSc

N50,000 – N60,000

N20,000 – N30,000

Felix

31 – 40

Married

BSc

N60,000 – N70,000

N20,000 – N30,000

FGD of 5 Ride-

hailing drivers,

November 11

2018

Charles

31 – 40

Single

BSc

N60,000 – N70,000

N20,000 – N30,000

Jude

31 – 40

Single

BSc

N50,000 – N60,000

N20,000 – N30,000

Paul

40 &

above

Married

BSc

Not disclosed

Okoro

40 &

above

Single

N50,000 – N60,000

N20,000 – N30,000

Akin

40 &

above

Married

BSc

N50,000 – N60,000

N20,000 – N30,000

390

FGD of 5 of

Ride-hailing

Drivers

November,

2018

Samson

40 &

above

Married

BSc

Not disclosed

Dayo

31 – 40

Single

N50,000 – N60,000

N20,000 – N30,000

Timothy

31 – 40

Single

N50,000 – N60,000

N20,000 – N30,000

Emmanuel

31 – 40

Single

Not

disclosed

Not disclosed

Charles

31 – 40

Single

BSc

N60,000 – N70,000

N20,000 – N30,000

Dipo

31 – 40

Married

B.eng

N50,000 – N60,000

N10,000 – N20,000

Onome

31 – 40

Married

Not

disclosed

Not disclosed

Salifu

31 – 40

Married

Not

disclosed

Not disclosed

391

FGD of 8

platform union

members, August

27, 2019,

Thomas

26 – 30

Married

SSCE

N50,000 – N60,000

N20,000 – N30,000

Ejiro

41 &

above

Not

disclosed

Not

disclosed

Not disclosed

Shade

31 – 40

Married

BSc

N50,000 – N60,000

N20,000 – N30,000

Ibukun

41 &

above

Married

Not

disclosed

Not disclosed

Oreva

31 – 40

Single

Not disclosed

Source: Author's fieldwork, 2018; 2019

Note: While the Naira is constant, the exchange rate varies with time. Here, using Oanda rates, £1 = 531. The drivers mentioned in this article

work for both Uber and Bolt. Some drivers, such as Charles, Koffi, and Jude, have a higher income range overall because they both own their

vehicles and are very strategic in gaming trips. Expenses from interviews include fuel, food, mobile phone data and unforeseen vehicle maintenance

such as a bad tyre or faulty part of the vehicle. The maintenance did not include insurance, mobile calling units, and car wash expenses. Also,

compared to 2021, the earning rate was better than now, although it had started to dissipate from 2017, as highlighted in the discussion.

392

8.3.4. Appendix C4: Relevant Results from a Failed Survey Experiment

Considering the opacity of ride-hailing platforms in Lagos, I also sort to triangulate

information from drivers by posting on their social media groups, as mentioned in the

reflexivity and positionality section in chapter three. While I do not count these 11 driver

responses as part of my sample, it is critical to highlight that the income ranges and working

hours mentioned above show close similarities. However, it is essential to note that not all

drivers, such as Charles, Koffi and Jude, who own their vehicles, make high-income ranges.

Some drivers who own their vehicle but make less often choose platforms as side gigs.

Figure C4.1: Responses for vehicle ownerships

Source: Author’s fieldwork, September 2019

Figure C4.1 shows the percentages of people who own their vehicles. Most people above

claimed to own their vehicle (54.5%), while 27.3% and 18.2% claim to be involved in a hire

purchase agreement and rentals, respectively.

393

Figure C4.2: Weekly Income ranges (Before expenses)

Source: Author’s fieldwork, September 2019

Figure C4.2 shows the income ranges of drivers before weekly costs and remittances to

vehicle owners. The majority of the driver earnings range from less than N50,000 to N60,000

weekly, i.e., 36.1% and 27.3%, respectively.

Figure C4.3: Weekly Income Ranges (After expenses)

Source: Author’s fieldwork, September 2019

394

Figure C4.3 shows the income ranges of drivers after weekly costs and remittances have been

subtracted from their earnings. Out of 11 driver responses, 54.5% claimed to earn between

N10,000 to N20,000, and 27.3% earned between N20,000 to 30,000 weekly.

Figure C4.4: Average Weekly hours per week.

Source: Author’s fieldwork, September 2019

Finally, on average, most drivers work between 50 to 70 hours per week at 27.3% weekly.

Another group of drivers work between 80 to 90 hours every week.

395

Notes

According to the court case TFL v Uber and Others (2015), the calculation of fares is determined by

two servers; Server 1 stores the long-term data for Uber, while Server 2 is responsible for the

calculation of fares and determining the fare structure, which applies in different contexts, in this case,

London. https://www.judiciary.uk/wp-content/uploads/2015/10/tfl_-v_uber-final_approved-2.pdf

[Accessed October 13, 2022]

According to the Cambridge dictionary (n.d), a taxi is “a car with a driver who you pay to take you

somewhere” (Cambridge dictionary, n.d). This definition is influenced by the historical accounts of the

hackney carriage in England and the taxicab in the United States. In England and the US, taxis mean

Hackney carriage, street taxi, taxi which can be hailed from the street. While PHVs, minicab service,

taxicab, dispatch vehicle , for-hire vehicle, black car are only available via pre-booking. In the GS, it

could mean, auto-rickshaws, Tuk-tuk, metered taxi; shared taxi, Kabu-kabu, small public service, and

so forth which are all accessible via street hail.

Tokunbo is derived from the Yoruba name ‘Adetokunbo which means a crown; Ade – returns; Bo –

from overseas; (Ti) Okun – is given to a child born away from home. (Ezeoha, et al. 2019; p. 188).

From field observation and interviews with taxi drivers between September 2018 to January 2019.

This was observed at the Eko Hotel and Suites Motor Park, in Victoria Island Lagos, November 2018.

There is a conflicting report on this, because both the director of transport policy and head of transport

operations stated that it was a fee of N5 million a year with at least 100 vehicles. But in 2015 when the

advent of e-hailing companies was well present, it appears that the law was revised because it states a

minimum of ’50 vehicles per franchise, and N100,0000 (£215) per vehicle in (Vanguard, 2015).

However, the head of taxi operations stated that the scheme was reviewed under the Lagos state public

transport policy.

Translation: There was a time metro came, they were using taximeter in their cars. Yellow taxis do

not have taximeter. We do not use all those things. We used it many years ago, but now we have

cancelled it. About 10 to 20 years ago. Some passengers do not like that package because if there is a

go-slow or traffic for instance, it never stops reading. It is usually a problem for passengers.

396

Nairaland is a reputable forum where people discuss different topics as well as market their services.

Evidence from this space shows that corporate cabs are still functional. However, experience from

living in Lagos, interviews and observations from the field showed that the company is non-existent or

struggling to make a comeback.

One chance’ is a term in Lagos used to describe private vehicles, taxis or buses who connive to

manipulate passengers in order to rob them of their belongings such as money, jewelleries and so forth.

In some cases, these criminals use charms or other fetish means to manipulate victims. In the case of

expatriates or well-to-do Nigerians, these drivers would kidnap them and seek ransom from the

companies or families responsible for them.

This was the day Uber arrived in Lagos, Nigeria. https://www.uber.com/en-NG/blog/lagos-your-

secret-ubers-have-landed/ [Accessed June 19, 2019]

In an interview with Michael Akindele, the former manager of Uber disclosed some of the

company’s successes and challenges in launching in the Nigeria market, as well as plans to create more

employment the unemployed, underemployed, and entrepreneurs. JohnRhoda. (2016). The CEO Watch

- Ebi Atawodi of Uber Nigeria. [online]. Available from:

https://www.youtube.com/watch?v=L_fbhzUSyw4&list=PL_atoEL_-bGqDkFpu6ln-

hT1iSL9eD5aB&index=9&ab_channel=JohnRhoda [Accessed July 10, 2019].

Requirements for being an Uber driver. Note that these requirements change with time, according

based on platform updates and any state regulatory requirements

https://www.uber.com/ng/en/drive/requirements/ [Accessed June 19, ,2019]

Directions on how to become a Bolt driver. https://blog.bolt.eu/enng/how-to-become-a-bolt-driver-

in-nigeria/ [Accessed 28 March 2019]

This was the Oga-taxi app on Google Playstore before it went defunct. Note that, The Oga Taxi app

page https://play.google.com/store/apps/details?id=com.oga.driver&hl=en&gl=US [Accessed July 18,

2018]. However, the only available link about the platform now is its Facebook page:

https://www.facebook.com/ogataxi [Accessed 17 May 2022]

This is an archive interview with TV360 Nigeria. Here, Loko Udoko explains their business model;

strategy for emerging in the market; plan for localisation of the platform to suite the Lagos space;

difficulties in access to funding and other challenges and successes experienced in since Oga-taxi launch

. TV360 Nigeria (2019). Meet the CEO of Nigeria's first indigenous taxi service. [online]. Available

397

from: https://www.youtube.com/watch?v=VOFf9UA1HGg&t=256s&ab_channel=TV360Nigeria

[Accessed December 16, 2019]

The only indication of their existence were the two drivers who stated they had it as a supplementary

platform but did not receive rides on it. The last post on their Facebook page was in January 2019,

another indication of their invisibility in the sector.

These were platforms I mapped out from a list of registered platforms with the state ministry of

transport at Alausa Lagos. The head of taxi operations at the time, disclosed some of these, while I

found others online. However, there others like Lolotaxi, Clickabs which were not on the list, but drivers

mentioned in interviews. It was difficult to determine which were now defunct because drivers did were

not aware of most of them.

This was observed on driver forums as well as during interactions with drivers.

This was observed in August 2019, when Charles who was formerly a platform driver, established

his own fleets and managed vehicles for other vehicle owners. He highlighted using installed trackers

to monitor drivers’ movements, to reduce drivers who cheat vehicle owners on the job, by using vehicles

for private affairs.

The equivalence of N68,285 is (£147)

Activity score measures the total number of trips divided by number of requests.

https://support.taxify.eu/hc/en-us/articles/115002946174-I-have-issues-with-activity-score [Accessed

18 October 2019]

Uber.com. How Star Ratings work. https://www.uber.com/us/en/drive/basics/how-ratings-work/

[Accessed July 2020]

Bolt.eu. General Terms for Drivers https://bolt.eu/en/legal/terms-for-drivers/. [Accessed 19 April

2020]

Note that this frequently changes in response to competition and price wars between competitors. At

the moment according to the Uber estimator, base fare is N200; minimum fare is N400; N11 per minute

and N60.01 per km. It is important to note that these increase or decrease across time.

https://www.uber.com/en-ng/blog/uberselect-your-step-above-the-everyday/ [Accessed May 16,

2018]

398

These amounts change with inflation. At the time of this study these were the estimates for popular

used vehicles for platforms (Toyota Corolla and Toyota Camry). Tokunbo cars were discussed in the

previous chapter.

For example, the author has registered for a NIN since 2014, and in 2021 has not received it despite

the incessant efforts to do so.

Judging by the comment section of these two scenarios. Drivers were more confident in Uber

compensating the driver in comparison to Bolt. Although some colleagues advised the driver to change

his phone, a majority of them expressed that it had happened to them; it is Bolt’s problem, not the

device.

Assembly bill No.5, September 2019. This is an act to amend Section 3351 of, and to add Section

2750.3 to, the Labor Code, and to amend Sections 606.5 and 621 of the Unemployment Insurance Code,

relating to employment, and making an appropriation, therefore.

Few hours to the protest in September 2019, comrade George reached out to ask for comrade Dipo’s

contact number in order to join forces in the protest. However, previously in our interview, George who

was part of the PEDPA association spoke against the leadership of Dipo as the union president of

NUPEDP, questioning their objectives. It was interesting to see them coming together as one voice.

According to information on the website, its features include simulation of static locations and

enabling smartphones to follow fake itineraries while having control over the altitude, G.P.S. signal

accuracy and speed. (Lockito-app.com). Geofence refers to a virtual boundary set up around a

geographical location which is triggered when an app or other software uses GPS, RFID, Wi-Fi or

cellular data to trigger a pre-programmed action via a mobile device or RFID tag (White, 2017).

I visited the Bolt office a few times in search of these so-called Uber-boys based on Charles

description, but I could not locate them.

Uber boys, all they do is research and finding out new practices to gain. They do not go to work.

They can use one ATM card to initiate multiple trip requests to places like EPE among 5 to 6 of them.

Each of them would collect about N20,000 (£34) each. I know a guy who died as a result of some of

these practices.

In January (2019) I attended the Workshop on Promoting Road Safety in a Smart City organised by

the Lagos State Ministry of Transportation in Collaboration with the Lagos Chamber of Commerce and

399

Industry. At this event, the former commissioner for transportation Mr Ladi Lawanson, highlighted his

lack of trust in Uber and the importance of regulating the platform. The commissioner went ahead to

state that the service was unsafe and fraudulent because of series of reports about women and drivers

who have been attacked. He also discussed his personal experience: saying that he used Uber briefly,

but he noticed that amounts were deducted unknowingly from his account until he had to shut it down.