Content uploaded by Andrea Pollio
Author content
All content in this area was uploaded by Andrea Pollio on Aug 24, 2023
Content may be subject to copyright.
INTERNATIONAL JOURNAL OF URBAN AND REGIONAL RESEARCH
DOI:10.1111/1468-2427.13200
1
— ALGORITHMIC SUTURING: Platforms,
Motorcycles and the ‘Last Mile’ in Urban Africa
, ’
Abstract
The ‘last mile’ is not only a powerful metaphor of contemporary life, but also the
tangible site of a challenge, whether for governments wanting to reach their citizens or
companies wanting to reach their customers. In urban Africa this challenge is compounded
by the fragmented material condition of cities. As a result, a growing number of tech
companies have been compelled by the possibility of creating digital platforms that
address the unique logistical configurations of African cities, often enrolling informal
systems such as motorcycle taxis to address spatial and economic fragmentation. Through
the perspective of three Nairobi- based startups that incorporate motorcycle taxis into
their last- mile platforms, this article illustrates how processes of ‘algorithmic suturing’
knit together the loose ends of splintered urban networks thanks to platform business
models that visualize the last mile as a site of optimization. In parallel with common
understandings of suturing within African infrastructure debates which foreground
makeshift practices of the urban poor, this article argues that algorithmic suturing is a
speculative endeavour through which urban fractures are made legible as sites of value.
By stitching together city fragments, these platforms envision large data- driven urban
economies which interface with informal mobility networks and the shifting urban
demographic of the lower- middle class.
Introduction
The ‘last mile’ is a powerful metaphor of contemporary life, shifting meanings
across contexts and among dierent actors. Supply chain managers, for example, define
the last mile as the ultimate leg of goods reaching their final customers, whether a
transportation hub, a warehouse or an individual person (Hayes,2021). For development
technocrats in Africa the last mile refers to the distance between households and
centres where public services such as healthcare are available. For telecommunication
companies the last mile allows individual users to connect to the backbone of bigger
communication networks. For AI scholars the last mile captures the ‘ghost work’ that
humans need to perform when algorithms fall short (Gray and Suri,2019).
Despite these various meanings, there is a thread running through these
definitions. Consider, for example, the schematic representation in Figure1, which
accompanies the Wikipedia entry for ‘last mile’ (as of July 2022).1 In the visualization,
the last mile is represented by the loose ends of a networked tree. These loose ends are
not just the ultimate conduits of networked systems; they are also the terminals where
infrastructural grids scatter and break, where they frictionally interface with each other,
and where urban splintering (Graham and Marvin,2001) and suturing (De Boeck and
1 Using a few examples (namely, tree trunks vs. root hairs, rivers vs. drip irrigation, interstate highways vs. back roads,
and intercontinental cables vs. user internet access), the Wikipedia caption illustrates the capillary and peripheral
nature of the last mile in a networked system.
© .
.
This article would not have been possible without the collaboration of our wider team: Rike Sitas, Prince Guma,
Alexis Gatoni and Alicia Fortuin. A draft of this work was presented at the Beyond Splintering Urbanism workshop
in Autun, where we benefited from Steve Graham and Sophie Schramm’s close readings and the generous
support of Simon Marvin, Jon Silver, Alan Wiig, Jonathan Rutherford and Colin McFarlane. Andrea Pollio’s work
has received funding from the European Union’s Horizon 2020 research and innovation programme under the
Marie Sklodowska- Curie grant agreement No 886772. Liza Rose Cirolia and Jack Ong’iro Odeo received funding
from the Volvo Education and Research Foundation (VREF) under the Mobility and Accessibility in African Cities
programme (MAC).
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
POLLIO, CIROLIA AND ODEO 2
Baloji,2016) become a matter of laborious coordination between dierent technical
systems. Because of these frictions, reaching the last mile is the costliest and most
challenging part of many businesses.
At a time of emerging forms of what Moritz Altenried(2019) has termed
‘logistical urbanism’, global platform companies are making these ‘problems’ of last- mile
coordination their core business model: Uber with e- hailing; Alibaba with drop-
shipping; Glovo with Q- commerce;2 as well as Amazon (and many others) with their
mechanical Turks.3 In urban Africa
—
the setting of this article
—
these last- mile
challenges are compounded by the fragmented material condition of cities. Low- density
sprawl, under- maintained road networks, lack of addressability and fractured service
2 Q- commerce or ’quick commerce’ (sometimes referred to as ’e- grocery’) is the on- demand outsourcing of grocery
shopping to platform workers.
3 Amazon Mechanical Turk is a platform for hiring remotely located ‘crowdworkers’ to perform discrete on- demand
tasks. Although this is not necessarily an urban phenomenon, there is a distinct urban bias to the geographies of
on- demand work.
A visual representation of the last mile
—
the thinner lines or loose ends
—
in a
hierarchical network (source: Wikipedia, from user Dycedard; no changes were made to
the image: https://upload.wikim edia.org/wikip edia/en/3/32/The_last_mile_hiera rchy.svg)
High capacity, long distance conduits
Examples:
Examples:
Tree trunks
Root hairs
Drip irrigation
Capillaries
Appliance cords
Back roads
User Internet access
Rivers
Arteries and veins
Power grid
Interstate highways
Inter
continental fiber
Widely share
d
costs
Locally shar
ed
costs
Lower capacity, short distance conduits
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3ALGORITHMIC SUTURING
delivery systems express both obstacles and opportunities for the optimization of last-
mile problems through digital platforms.
Building on this insight, the article focuses on startups that employ digital
platforms to develop last- mile economies by incorporating motorcycle taxis into their
algorithms. As in other Southern cities, in large African metropoles such vehicles are
perfectly formatted for the kind of fast- paced and short- distance movements necessary
to overcome spatial and economic fragmentation. They already constitute a lifeline
for the mobility of people and goods where larger vehicles fail to provide sucient
capillarity. And so motorcycles have compelled both global and domestic tech companies
to find ways of subsuming riders into platforms that seek to address the last- mile
problem with dedicated digital solutions.
While most scholarly work on this phenomenon has rightly focused on the
issue of ‘platform labour’ (Van Doorn,2017), highlighting the forms of precarity
and exploitation that are beholden to the digitization of mobility in urban Africa
(see Pollio,2019; 2021; Anwar and Graham,2020; Doherty,2020; Iazzolino,2021;
Odendaal,2021; Anwar et al.,2022; Arubayi,2021), our focus on startups deploys a
dierent entry point. Specifically, the article charts how these fledgling companies make
the ‘work’ of last- mile coordination legible, by visualizing the opportunities through
which motorcycle taxi networks can be mobilized to address economic and spatial
fractures in African cities. As detailed in what follows, we use the notion of ‘seeing like
a last- mile business model’ (inspired by James Scott’s1998 book Seeing Like a State)
and insights from media studies to show how startups themselves understand and make
legible the last mile, as well as how we, as researchers, can methodologically encounter
the last mile, not (only) as a metaphor, but as an algorithmic practice.
In doing so, we describe ‘algorithmic suturing’ as a key operation at the nexus of
digital technologies and motorcycle taxis. Algorithmic suturing is the knitting together
of the loose ends of splintered urban networks and informal economic activities through
platform business models that visualize the last mile as a site of optimization and value
creation. This conceptual contribution emerges from ongoing research on the
platformization of two- wheel logistics in Nairobi, Kenya (Sitas et al.,2022). Much as in
many African cities, Nairobi’s motorcycle- taxis
—
called boda boda
—
are used to carry
people, goods and parcels. They are one of the fastest ways to move through the city and
form an integral part of everyday life. They generally fall into what is called
‘paratransit’4
—
a blurry category that straddles public and private modes of movement
and includes minibuses (usually 14- to 16- seater vans), tuk tuks (three- wheelers) and
motorcycles. As a prosthesis to splintered network economies, boda boda operate where
other mobility and logistics options are too costly, too cumbersome, or simply not
flexible enough to address the last- mile problem. Increasingly, they do so through data-
driven platforms, as will be shown in our three case studies.
Drawing from this work on Nairobi, the article makes two arguments. First, we
suggest that the last mile renders the loose ends of networks legible as a problem of
coordination, as a potential business opportunity, and as a key site of optimization
—
this is, after all, the most expensive part of global supply chains. Seeing like a last mile,
which we intend here more narrowly as a last- mile business model, thus reveals some
of the hidden scripts that algorithmically suture the city, and some of the behind-
the- scenes processes of coordination through which new urban networks are given
eect and substance thanks to last- mile platforms. This last- mile view also creates new
‘leads’
—
things made visible that otherwise might not be. We highlight some of these in
the conclusion
—
for example, the importance of the Chinese internet industry in urban
Africa
—
in the hope that they will inspire scholars to take these emerging algorithmic
economies seriously.
4 See Pirie(2014).
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 4
Second, we argue that algorithmic suturing challenges some of the conventional
notions within Southern infrastructure scholarship concerning how and by whom
material and network improvisation takes place.5 While acts of suturing are often read
as a practice of urban survival at the margins of the city, last- mile business models
extend this story. Platforms stitch together city ‘fragments’ (McFarlane,2018),
developing algorithmic interfaces between precarious workers and the growing
demographic of the lower- middle class in urban Africa. This lower- middle class
demographic is not a discrete category or group, but rather a market both envisioned
and, in part, produced by platformed service delivery. Aspirational processes of ‘middle-
classing’ have been addressed in the financial anthropology of African capitalism
(James,2021), yet their role in shaping urban change (Mercer,2014; 2020) is also evident
in the future- making practices of last- mile algorithms.6
Overall, our contribution to urban scholarship stems from a desire to produce
a meso- level analysis of urban platforms, without dissolving their specificity into a
frontierist critique of global technocapital’s search for new profit opportunities in
African cities (Ouma,2017), nor seeking to offer a fine- grained investigation of the
infrastructural labour upon which the former depend. While cognizant that these too
are crucial analytical entry points, we suggest that a shift in perspective
—
see Qadri
and D’Ignazio(2022) for a different but complementary approach to this one
—
or
better, taking for a moment the perspective of last- mile startups, sheds light on some
of the emerging urban geographies of platformization in Africa and their implications.
Before moving on to these last- mile ‘views’, however, the next three sections provide,
respectively, a description of algorithmic suturing, an overview of digitization in the
boda boda sector in Kenya, and a detailed account of our methodological approach.
Algorithmic suturing
The concept of algorithmic suturing draws on the insight that urban life in Africa
is often made possible by practices of infrastructural repair, mending and patchwork.
Specifically, Filip De Boeck and Sammy Baloji use the metaphor of the ‘suture’ to
describe the junctures and seams through which urbanites
—
often against all odds
—
find ways to ‘fill the gaps, overcome the hiatus, design realignments and thereby redefine
the zero … into a possibility, a something else, a surplus’ (De Boeck and Baloji,2016:
16). More broadly, it was AbdouMaliq Simone’s fitting yet sometimes misunderstood
notion of ‘people as infrastructure’ that drew attention to the machinic acts of suturing
through which life is reworked and made possible in contexts of extreme disposability
and precarity (Simone,2021).
With ‘algorithmic suturing’ we combine this attention to the ‘mathematics’ of
collective urban life (Simone,2021: 1343) with an expansive definition of algorithms
(Gillespie,2014). More than just lines of code (Seaver,2017), algorithms have a genealogy
that predates digital technologies (Striphas,2015; Daston,2022). An algorithm is in fact
‘an abstract diagram that emerges from the repetition of a process, an organization of
time, space, labor, and operations’, as well as ‘the division of this process into finite steps
in order to perform and control it eciently’ (Pasquinelli,2019: 6). Most importantly,
Pasquinelli further writes, ‘an algorithm is an economic process, as it must employ the
least amount of resources in terms of space, time, and energy, adapting to the limits of
the situation’ (ibid.). Moreover, as Lilly Irani explains, confining algorithms to a purified
notion of software ‘automation’, devoid of the multiple forms of labour intrinsic to it,
risks reinforcing the narrative adopted by platform companies, which essentially makes
5 On a related subject, see Cirolia et al.(2021), Günel(2021), Lemanski(2021), Cirolia and Pollio(forthcoming).
6 In this article we are not concerned with (nor would we have the data for) defining the sociological contours of this
demographic category (the lower- middle class) but we treat it as an emic category that our informants used to
describe an under- served market and therefore an opportunity for processes of algorithmic optimization.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5ALGORITHMIC SUTURING
some forms of work invisible to the benefit of others (Irani,2015). Put differently,
algorithmic suturing also captures the limits of algorithms, in that they are themselves
always in need of spatial and other fixes (Pollio,2021).
The concept of ‘algorithmic suturing’ thus dovetails with two important strands
of scholarship. First, as we have seen, work dedicated to understanding how, in African
cities, hybrid, alternative and incrementally improvised configurations suture disrupted
urban fabrics both physically and metaphorically (Silver,2014; De Boeck and Baloji,2016;
Baptista,2019). These scholars have also challenged the binary between networked and
post- networked systems (Akallah and Hård,2020; Cirolia et al.,2021), charting the
strategies of ‘technological bricolage’ that exist on the ground (Lemanski,2021) and
stretching the ontological scope of what constitutes infrastructure (e.g. Larkin,2004;
Kimari,2021; Simone,2021). Second, algorithmic suturing speaks to a growing body of
literature that specifically maps the patchworked nature of digital technologies in urban
Africa, oering a perspective on the incomplete, piecemeal and adaptive remaking of
platforms, and challenging simplistic visions of domination and leapfrogging (Guma
and Mwaura,2021; Guma,2022; Guma and Wiig,2022; Odendaal,2021). This work
fits into a wider body of scholarship on digital transformation and entrepreneurship
in the continent (Friederici,2018; Pollio,2020; 2022a; Guma and Monstadt,2021;
Odendaal,2023).
Motorcycles, algorithms and the materiality of African cities
Our focus on motorcycles as a last- mile device is not incidental. Their
contribution to the movement of people and goods in Africa is vital to the everyday pulse
of cities (Agbiboa,2020).7 Where private car ownership is a relatively new addition to
the ascent of the middle class, reliance on alternative modes of movement has long been
a necessity. In many African cities, motorcycles are indeed the most eective vehicle for
short, fast trips. For a city like Nairobi, conservative estimates by the Boda Boda Safety
Association put the number of motorcycle riders at fifty thousand (Omulo,2021)
—
a total
that doesn’t include unregistered operators.8 Owing to this ubiquity in Kenya’s capital
city (as well as in smaller urban centres) a mixture of internationally imported and
locally assembled motorcycles enables a distributed mobility service that complements
larger vehicles such as minibuses (locally called matatu9 and generally used by passengers
for longer trips) and delivery vans (used for the movement of larger quantities of goods).
The modal choice of the motorcycle, particularly for last- mile transportation, is
a response to the material conditions of the African city.10 In the context of large,
sprawling urban footprints, getting passengers and goods to the peripheral outskirts or
to the crevices of high- density suburbs requires a fuel- ecient, agile vehicle. Such
agility of movement is likewise necessary to bypass the thick trac which clogs highways
and intersections, and to overcome incomplete and under- maintained road networks
(for example, navigating rutted surfaces unsuited for cars). Nairobi is a particularly
fitting vantage point from which to observe this: a city splintered by early colonial plans
(Ese and Ese,2020) and contemporary large- scale bypasses (Guma et al.,2023), and
dotted with leafy, middle- class cul- de- sacs, busy commercial malls, hyper- dense
7 The same is true in other large Southern cities too. See, for example, the work of Claudio Sopranzetti in relation to
Bangkok (2013), Sam Nowak(2021) and Qadri and D’Ignazio(2022) on Jakarta, and Kevin O’Neill for Guatemala
City (2022).
8 Due to the absence of an official database, the exact number is unknown, but it is likely to be much higher. The
National Crime Research Centre (Opondo and Kiprop,2018), for example, reported that 1,393,390 motorcycles
had been registered with the National Transport and Safety Authority as of February 2018. The Kenya National
Bureau of Statistics (KNBS,2022) also reports that the number of motorcycles registered in 2021 rose to 285,203
from 246,705 the previous year. How many of these motorcycles are operating as boda boda in each city/region
remains undocumented, however.
9 For a history ‘from below’ of this unique, embattled transit network, see Kenda Mutongi’s Matatu (2017). An
interesting visualization project of the matatu network is captured in Klopp et al.(2017).
10 Boda boda, as Joyce Nyairo(2023) has brilliantly explained, are also part of the making of citiness itself.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 6
suburbs, master- planned estates, sweeping informal settlements, warehousing precincts,
and ‘plotted’ (Karaman et al.,2020; Maina and Cirolia,forthcoming) peri- urban fringes.
In this context, motorcycles are a vital and rapid undercurrent. They stitch together
fragmented parts of the city, hybridizing service delivery (Jaglin,2014) and overcoming
the brittle and enduring legacy of colonial planning, postcolonial projects of modern city
building, and forced informality (Ese and Ese,2020). Alongside passenger transit,
motorcycles have for decades also been used for last- mile delivery, as they are not only
fast but also aordable to maintain and operate
—
all of which contributes to reducing
the supply- chain costs for logistics companies.
Despite this important role of motorcycles in urban economies, most African
governments have taken punitive measures against the sector (see Goodfellow,2015).
While two- wheelers were initially conceived as a solution to austerity measures
introduced for structural adjustment (Rizzo,2002), today they are seen as an unwieldy
urban industry in need of some form of regulation. However, as the riders are mostly
young men, they also constitute an important political clientele. Central and local
governments have therefore pursued ambivalent and contradictory practices. The most
draconian of these policies have been implemented to address issues of passenger
safety
—
measures which artificially segregate using motorcycles to move people around
from using them to move and deliver goods. Conversely, eorts to modernize urban
mobility generally include banning motorcycles from particular parts of cities (usually
city centres), and supporting investment in large- scale transport projects such as Bus
Rapid Transit (BRT) systems
—
both of which fail to overcome the incredible need and
demand for distributed, capillary delivery and pillion services. The Kenyan government,
in particular, has long turned a blind eye to the boda boda sector
—
allowing the
emergence of competing and self- regulating voluntary associations
—
but it is now
involved in a regulatory attempt that, if successful, will radically transform the industry.11
In this context, the tech community has been increasingly compelled by
motorcycle mobilities over the last few years. This also aligns of course with a much
larger trend, whereby digital innovation seeks out ‘untapped’ niches in city systems,
looking for sites where platform solutions may be able to solve what appear to be urban
problems. Some of the digital innovation sits in the e- hailing space (allowing end- users
to connect to riders through mobile phone- based applications such as Uber boda or
Safeboda) (see Doherty,2020). However, the invisible majority are focused on improving
the business models of companies that utilize motorcycles for last- mile, express and on-
demand deliveries (Cirolia et al.,2023).
In Nairobi, a city seeking to implement ambitious smart city plans (Guma and
Monstadt,2021) and long dubbed Africa’s Silicon Savannah (Graham and Mann,2013),
a booming startup ecosystem has emerged in the last decade as a melting pot of local
entrepreneurs, diaspora returnees with experience in the global tech industry, foreign
companies, and international venture capital (Rosenberg and Brent,2020). All these
players have thus contributed to an explosion of platform experiments addressing
not just motorcycle mobilities but many facets of the Kenyan economy, from payment
technologies to agriculture (Mann and Iazzolino,2021). While some of these platforms
are international, accessing venture capital to finance the trial and interaction of their
oerings, several experiments are in fact driven by home- grown, bootstrapped startups,
many of which only last a few months after being launched and then wilt away or
11 Recent legislative amendments to the National Transport and Safety Authority Act seek to give the latter agency
power to establish systems and procedures for the registration, licensing and operations of these categories of
vehicles, just as it does with other motor vehicles. Additionally, county governments are empowered to enact
regulations for boda boda operations within their jurisdictions. In Nairobi, boda boda operators will in future be
registered by the county government according to the cooperative society to which they belong and their zone of
operation; motorcycles will be fitted with digital plates containing information about the riders.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7ALGORITHMIC SUTURING
pivot their business models to find more profitable last- mile market niches (Cirolia et
al.,2023).
As we will see, the motorcycle is just one, albeit crucial, component of the
platform arrangements through which last- mile startups ‘see’ the city. It is a vehicle (in
all senses) for us to chart the practices of algorithmic suturing that seek to fill the gaps
of a fragmented city, both spatially and economically.
A note on method
Our approach in this article is usefully summarized by the metaphor of
‘seeing like a last- mile business model’, which explicitly borrows from Scott’s critique
of ‘authoritarian high modernism’(1998). The original phrase, ‘seeing like a state’,
described the multiple acts of ‘bringing into view’ which administrative designs need
in order to simplify, and therefore govern, complexity. While Scott’s insights have
been variously challenged
—
both in their empirical validity and their applicability to
corporate capitalism (e.g. in Ferguson’s Seeing Like an Oil Company[2005])
—
the notion
that the making legible of things (for example, through measurement protocols, cadastral
maps and census taxonomies) is a crucial register of power has generated a wealth of
contributions in various disciplines
—
from Dourish’s Seeing Like an Interface(2007) to
Seaver’s Seeing Like an Infrastructure(2021), to name just two.
In a more science and technology studies (STS)- inspired interpretation of ‘seeing
like’ as a performative act, whereby legibility works both on what sees and on what is
seen, John Law(2009) instead argues that realities are enacted through the methods
that are purportedly meant to simply bring them into view. This insight lies at the centre
of our methodological approach to understanding how last- mile ‘views’ render network
fragmentation legible as a problem of platform coordination, whilst enabling us as
researchers to follow additional leads. We would also argue that this intuition informs
much of the critical scholarship on digital platforms
—
showcasing how the data
processes through which algorithms operate are never just descriptive, but follow
multiple performative logics12 (as described by Fourcade and Healy in Seeing like a
Market[2017]).
We are equally inspired by a body of scholarship in media and cultural studies
which has shown the importance of network imagination, arguing that networks are
material and metaphorical infrastructures of sensibility that mediate and narrow our
experience of the world (Terranova,2004; Galloway and Thacker,2013; Munster,2013).
According to Patrick Jagoda, for example, networks are realities that exist at the edge
of sensibility. They are at once something that we see and a way of seeing, inextricably
both material and metaphorical (Jagoda,2016). Infrastructures, therefore, cannot be
thought of separately from the ways in which they are visualized. And as we have seen
in Figure1, the last mile is a key component of network visualization, pointing to its
peripheral terminals. Visualization, in this context, is not just a matter of ‘pictures’, but
a broader process of rendering visible (Halpern,2015).
Building on these different and at times disconnected contributions, for us
‘seeing like a last- mile business model’ is both a descriptive device that captures the
ways in which motorcycle taxis become inscribed in the algorithmic suturing of the
loose ends of urban networks, and a methodological orientation. It is a methodological
orientation because seeing like a last mile means charting the business models of these
platforms: the algorithmic scripts through which the ends of networks are made visible
as sites of coordination and optimization. As Liliana Doganova has shown in her work
(e.g. Doganova and Eyquem- Renault,2009), business models are indeed particular
kinds of market abstractions: ways of seeing that participate in the making of economic
12 See, for example, Wendy Chun’s Discriminating Data (2021).
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 8
realities. In other words, last- mile business models are one of the ‘legibility’ scripts
through which fractured urban networks are coordinated into digital platforms. When
algorithmically visualized, fragments and sutures become more than just metaphors
of urban life; they become discrete processes of spatial and economic optimization
(Pasquinelli,2019).
Cognizant that the perspective of platform startups may be contested or
reinforced by how the riders themselves ‘see’ the algorithmic economies in which they
are imbricated, as Qadri and D’Ignazio write in Seeing like a Driver(2022), our practical
strategy to see like a last mile involved a scoping of all Nairobi- based digital platforms
that incorporate boda boda in their operations and a mapping of both their value chains
and supply chains to understand the dierent roles boda boda play in each business
model. This work included trial- and- error practices of ‘playing’ with the apps, trying
out websites, making orders ourselves, speaking to customer services, and analysing
online information available for each platform, as well as accessing private archives
that document investments in digital companies. In the cases we now move on to
discuss, we were also able to conduct traditional ‘corporate interviews’ with founders
and other platform managers (Schoenberger,1991). To do full justice to this empirical
richness, the three cases below are presented as detailed vignettes that showcase how
each business model sees fragmented urban economies as sites of algorithmic suturing,
viewed through the eyes of those whom Jane Guyer(2016) calls ‘platform attendants’
—
that is, the creators and managers of last- mile platforms.
Seeing like a last mile in Nairobi
Global e- hailing companies such as Uber and Bolt, Africa’s largest e- commerce
platform Jumia, and Delivery Hero- owned e- grocer Glovo are undoubtedly the most
visible operators in Kenya’s last- mile urban economies. They are not, however, the
sole players. As our research shows (Sitas et al.,2022), not only do dozens of smaller
platforms exist on the market, seeking to create alternatives to these oligopolies, but
several digital startups are also active in the value chain of these bigger platform firms.
Through their last- mile business models, many of these fledgling companies onboard
motorcycle riders into the functioning of their algorithms, with the aim of using
informal boda boda networks as a spatial suture that reduces logistical costs and creates
additional avenues of optimization.
In this section of the article, we chart the business models of three such
startups
—
three smaller platforms that are using boda boda to address a number of
last- mile problems
—
ranging from the exploitative nature of precarious work, to the
addressability of lower middle- income residents, to the reconfiguration of hub- and-
spoke logistics in Kenya. The first company, Cleanify, is a gig- work platform for domestic
workers. The other two companies, Shoppist and Dasher, are an e- commerce platform
and a last- mile logistics provider, respectively. For each of these cases, we map how
their business models frame boda boda as a prosthetic link to issues of algorithmic
optimization, and foreground what these ways of seeing make legible about urban
economies in Nairobi. We therefore ask the reader to take the perspective of these
businesses and to ‘see’ the splintered urban realities as coordination problems that can
be bridged by motorcycle taxis.
— Networks of waiting among last- mile workers
Soft- spoken and eloquent, Mike13 is the co- founder of Cleanify, a four- year- old
startup oering on- demand laundry and household cleaning services. With a degree in
statistics and a passion for crunching data, Mike does not have the boisterous panache of
13 All names of people and companies in this article are pseudonyms that we used to protect the confidentiality of
our interviewees.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
9ALGORITHMIC SUTURING
the prototypical startup founder; by his own admission, he had to learn how to successfully
promote his company by imitating the seemingly limitless confidence of others. To do so,
when he and his co- founder started the company, they joined one of the co- working
spaces that Nairobi is famous for. Rubbing shoulders with other ‘techies’ taught them how
to perform at pitching events and competitions. Four years later, and despite the COVID- 19
slow- down, Cleanify is running at a profit, provides employment for more than 200
women, and recently raised capital to switch from a web- to a mobile- based application.
Cleanify’s business model looks very simple on paper: it is a gig- work platform
through which households can hire the services of a cleaner. By charging a fee on
each transaction, Cleanify replicates the business model of large platform companies,
matching freelance labour with local demand. The work of the algorithm also seems
rather simple: by geolocating both homes and cleaners, the platform connects the supply
and demand of household cleaning services. However, as Mike explains, things are more
complicated than this, both in the demand and in the supply of these services. As most
middle- class Nairobians already have regular (often live- in) domestic help, the demand
for gig work comes from the lower- middle class, especially from households that cannot
aord the services of a regular or live- in domestic help and which (more importantly)
lack cleaning appliances such as washing machines. In response to this demand, families
rely on the informal work of women sometimes referred to as mamafua or mamasafi
(Kiswahili for cleaning mother/lady). Mamafua are known to be found in specific
waiting areas, usually at the interstice between wealthier and poorer neighbourhoods.
They rely on the odd jobs that neighbourhood families have for them, usually hand-
washing clothes and, occasionally, heavy- duty housework.
Cleanify recruited and trained a number of these women, taught them how to
use their smartphones to receive and locate a ‘call’, assigned each of them a unique
identification (ID) code, and linked them to an algorithm through which the logistics of
the last- mile problem could be optimized. It is at this point that motorcycle taxis come
into play. Much like these women, boda riders wait in specific areas of the city. Connecting
these two networks of waiting spaces, Cleanify’s algorithm determines whether a job
will require the support of a nearby motorcycle taxi and, because mamafua cannot
normally aord this commuting service, adds a half- dollar fee (50 Kenyan Shillings) to
the end customer’s bill. In this way, the platform relies on the boda riders’ unparalleled
knowledge of the city, addressing a logistical problem inherent to freelance housework.
Yet Mike is adamant that this last- mile problem was not just a profit opportunity.
He explains:
You know, platform gig companies are very exploitative because their business
model functions on very small margins. Our idea was to start from the opposite
side of the business model and ask how these small margins could actually help
the livelihoods of these … precarious workers while addressing a need of an
under- served market.
As he further notes, it is in the marginal price difference between informal
work and gig work that Cleanify’s last- mile innovation lies. End customers are willing
to pay slightly more than they would normally because the quality of the service is
better, faster, more consistent and more reliable
—
not least, thanks to the boda riders.
It is in this willingness to pay (marginally) more that Mike and his business partner
saw the possibility of injecting a different kind of rationality. Contrary to platform
business models that seek to drive the cost of labour down, Cleanify operates to render
fluctuating informal gains and working conditions more stable. The unique ID, the
iterative training (which allows for returning customers), the plug- in of motorcycle
taxis, and the algorithmic distribution of calls now provide every mamafua with a
monthly income that is above the national minimum wage.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 10
The last- mile algorithm, however, is not just a piece of software. It entails a
careful process through which Cleanify enrols various infrastructures, including human
networks that make the algorithmic suturing possible. In Mike’s words:
Each area now has a group leader, and the group leader is the one who
coordinates all this. So when an SMS goes to a lady and she needs a rider, she
goes to the group leader, who is usually stationed in the same waiting area that
we have selected as a base, and she arranges a ride. Usually our bases are close
to boda boda waiting areas. And the company reimburses the riders at the end
of the day. It’s a dynamic web network that we are trying to make work through
the platform. The first step was to envision the interfaces between these smaller
networks and from there imagine how some key players could function as
mediators between the platform and the complex dynamics of each area.
As Cleanify grows, so does the number of networks with which it seeks to
interface. Informal financial infrastructures are the latest addition. When Mike and his
partner saw that mamafua were starting table- banking groups with their savings, they
prototyped a financial scheme that allows cleaners to buy shares in a savings account
owned by the company. In this way, each saving club member can borrow or withdraw
what they contributed (as with a normal table- banking protocol), while also benefiting
from a basic interest rate that cash- based saving does not yield. These are all very small
improvements, Mike admits, but they became visible by looking at the margins of the
many fractured networks and treating them as data problems of algorithmic distribution:
by seeing Nairobi’s urban fragments as a challenge for last- mile coordination.
What we thus begin to see, through Cleanify’s last- mile business model, are
the complex operations through which disjunctions in urban economies
—
such as
domestic housework
—
become legible as problems that require algorithmic suturing.
In these operations, boda boda are one of the infrastructural optimizations through
which platforms seek to interface with various precarious networks in the city: from
cleaning services to pillion commuting and even informal banking. Rather than simply
disciplining labour, or replacing more secure forms of employment with gig work,
Cleanify’s algorithm is designed to optimize marginal gains at the intersection of two
demographics
—
the lower- middle class and informal workers
—
that not only constitute
the bulk of Nairobi’s dwellers, but are both extremely sensitive to changes in the price
of piece work, whether to earn a living wage or to pay for essential services. The next
two case studies will speak more directly to the question of aordability for the lower-
middle class, and also to the unique last- mile problems that arise through creating
e- commerce solutions for a shifting and elusive demographic group (Darbon,2019)
that defies sociological definitions but very clearly shapes how digital platform startups
visualize their markets.
Before moving on to the next company, it is worth noting one last important
thing about Cleanify. Like any startup, Mike and his co- founder have ambitious plans for
the future of the business. These plans include a distributed network of Cleanify- owned
ghost laundromats which will increase the number of heavy- duty cleaning options
available to its customers. It goes without saying that boda boda will provide the logistics
between these laundromats and households. In other words, seeing through a last- mile
business model also sheds light on the speculative urban networks that are not there yet,
but are waiting in the wings of future algorithmic iterations.
— Last- mile addressability in e- commerce
With a decade of experience in marketing for e- commerce companies, Li landed
in Nairobi from southern China early in 2019 to join the marketing department of a
Chinese retailer selling fast- moving consumer goods (FMCGs). He now works as the
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
11ALGORITHMIC SUTURING
head of marketing for Shoppist, one of the up- and- coming e- commerce platforms in
East Africa, operating in Kenya and Uganda. For many young managers of international
Chinese companies, Li explains, moving to Africa is an ambitious career move, usually a
trampoline to somewhere else in the world. In his case, however, Shoppist represents a
rather unique journey, because the company was actually founded in Kenya in the mid-
2010s by a Chinese expatriate, a former Huawei executive.
The inspiration for Shoppist was obviously the Chinese company Alibaba and the
rags- to- riches story of its founder Jack Ma, who made a fortune by creating the largest
e- commerce company on the globe (by gross merchandise value). Alibaba’s success, Li
argues, lay in its capacity to experiment with and diversify its business models, from
drop- shipping to consumer- to- consumer options. For Shoppist’s founder, Alibaba and
other Chinese giants like Jingdong (JD.com) had something valuable to teach Africa’s
fledgling e- commerce sector: reaching the customers who were usually left out from
online shopping platforms was not just about marketing prowess, but also about last-
mile logistics, both financially and geographically.
However, as Li explains, ‘e- commerce in China is very unique.… In Kenya, we
[could] not just copy Alibaba or Amazon. And so that’s why it was important to have
local intelligence’. Local intelligence, as he later elaborates, meant more than hiring
a team of locals for the operations team. It entailed relying on their understanding of
the East African market, on their knowledge of specific urban economies of logistical
distribution (in this case, the boda boda network), as well as on an iterative, cumulative
process of data- powered learning.
From the very beginning, these experimental processes translated into a series
of small adaptations and tweaks to more traditional e- commerce models. The first of
these changes was the integration of M- Pesa,14 Kenya’s mobile currency system. This was
no small feat, even though M- Pesa has an application programming interface (API) for
embedding the mobile currency as an online payment option. The reason other online
retailers of foreign goods only allow purchases by credit card, Li speculates, is because
of the currency risk. But with M- Pesa, Shoppist became one of the first online retailers
where Kenyans could buy aordable phones made in China in the same way they used
to do in a physical shop: with a one- o, fully paid, cash- like transaction.
The other change in the business model stemmed from the intuition that
Shoppist did not need to function completely as either a drop- shipping company (i.e.
with no warehouse) or a stockist (i.e. with large- scale warehousing facilities). Shoppist’s
hybrid model is based on a central warehouse located close to Nairobi’s international
airport (in an area that is becoming more and more densely packed with this kind of
logistical facility) and pop- up warehouses in Mombasa and Kisumu that it can use
during peak- order periods (such as Christmas time). A three- tiered product taxonomy
thus ensues: things that are already in the warehouse (where they never stay for more
than a month); things that are sold by other stockists in Kenya (and are in third- party
warehouses); and things ordered directly from China (which therefore only stay briefly
in the Nairobi warehouse before being dispatched).
For this carefully orchestrated supply model, Li argues, coordinating the
last mile is as important as understanding the suppliers. In fact, understanding the
supplier side depends on last- mile data and coordination. For goods to move quickly
to their end customers, two motorcycle- enabled solutions are crafted to resolve two
addressability issues. The first is the fact that Nairobi addresses may or may not be
accurately geolocated, or even have a specific house number
—
boda riders know the
city well enough to fill this gap. The second is the total lack of an address, because
some Shoppist customers live in parts of the city that for one reason or another are
14 This local mobile money platform was originally developed by telecom operator Safaricom. See Guma and
Mwaura(2021) for an infrastructural perspective on last- mile mobile telephony configurations in Kenya.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 12
unaddressable. In this case, the last- mile solution consists of little delivery hubs called
Shopposts. Deliveries to these hubs are fulfilled daily by boda riders and they have the
double function of providing a delivery address for people without one, as well as for
those customers who cannot receive deliveries at home during working hours. As Li
explains, Shoppist customers are mostly young people who do not belong to the upper-
income brackets of Kenyan society; consequently, costlier delivery options on Amazon or
other e- commerce platforms are simply prohibitive. The Shoppost option is essential for
this type of customer because they are less addressable than middle- class households.
While Shoppist employs a small number of boda riders directly, the majority
of them are subcontracted through business- to- business (B2B) last- mile logistical
platforms. The reason for this choice is that specialized two- wheel delivery companies
have developed more ecient distribution algorithms; as we will see in the next case
study, they are better at optimizing delivery routes across urban centres, and their
riders have much better performance indicators than could be achieved with in- house
operations. At the same time, keeping a few riders on the payroll means that Shoppist
also has a degree of flexibility and can adapt to unexpected market fluctuations,
particularly in Nairobi. Such a system, Li concludes, did not happen just by chance, but
evolved organically using the last mile itself as a data source, constantly monitoring and
experimenting with feedback loops to improve the final leg of logistical coordination,
while ensuring that small tweaks to the dispatch system would not aect the aordability
of Shoppist’s e- commerce oering.
These are small but crucial aspects of Shoppist’s business model which speak
to the ways in which platform algorithms seek to mend disjointed urban fabrics. First,
Shoppist visualizes the goal of selling aordable products to its target customers (the
urban lower- middle class and youths) as a problem of addressability. Second, creating
an e- commerce market from scratch requires incremental and iterative data- driven
experiments. This is very much in line with what has been observed about large platform
companies, which experiment with the city as a testbed for new products and services
(Mattern,2016). What the example of Shoppist also shows us, however, is how these
experiments are framed as potential solutions to logistical problems that are unique to
a city like Nairobi and to which boda boda networks present a gap- filling opportunity
—
whether by knowing intimately the complicated urban fabric of a booming postcolonial
city or by buering against the fluctuations in demand for last- mile services. Once again,
as with Cleanify, the lower- middle class represents a key driver of these new economies
of algorithmic suturing. Yet as with Cleanify’s business model too, we catch sight of how
these new companies speculate on new urban spaces (such as ghost warehouses and
distributed delivery hubs) as the future interfaces through which algorithmic economies
will latch onto the city.
— A last- mile crouching tiger
To find one of the specialized last- mile companies that support e- commerce
platforms like Shoppist, one doesn’t need to travel very far from the headquarters of
the latter. At the southern end of Mombasa road, Nairobi’s main trac artery, stands
a growing cluster of depots and other logistics operations, attracted by the proximity
of Kenya’s main international airport, the terminus of the newly built standard-
gauge railway, and access to the country’s busiest highway. At the crossroads of these
networked systems, Dasher occupies two large warehouses painted in bright colours,
one being the distribution centre and the other a sorting facility.
Dasher is one of an increasing number of small last- mile companies that oer
logistics services through a network of boda riders across Kenya. Unlike bigger platforms
like Uber and Bolt, Dasher only operates in the B2B market, oering a plug- and- play
delivery service to e- commerce businesses. This is explained to us by Jenny, a minute,
animated young woman who is head of sales for the company, which was started by
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
13ALGORITHMIC SUTURING
a small team of Chinese expatriates in late 2020. Kenya was not their first market,
however. Dasher is the subsidiary of a conglomerate already active in half a dozen
African countries, as well as, more recently, in the Middle East. These subsidiaries all
report to headquarters in Shanghai but, as Jenny elaborates, they only do so once a
year, and Dasher is an independent business unit operating as a startup. The investment
came from one of China’s leading express delivery companies, a logistics firm that
made its fortune on the heels of e- commerce giants like Alibaba and JD.com and which
now handles a staggering volume of more than ten million parcels a day. Its expansion
into the African market replicates the model that the company used in China, with a
growing cluster of subsidiary startups that gradually formed a comprehensive network
of logistical services, from line- haul first- mile to express last- mile solutions.
In Kenya, Dasher is spearheaded by Kevin, a tall young man who cut his teeth as
a logistics manager in southern China during the heyday of the e- commerce boom, ten
years prior to relocating to Nairobi. Sitting with Jenny in a small oce separated from
the sorting facility by a glazed wall he maps out Dasher’s network for us. At the centre
of the network stands the warehouse complex in which they welcomed us. Parcels
coming in from overseas, for example, are scanned in the sorting facility. This is a
manual operation but it is augmented by technological equipment imported from China
and by in- house software that Dasher built itself to automate its logistics management.
In the future, Kevin plans to import a fully automated sorting belt, but so far the volume
of parcels does not warrant a large- scale machine. From the sorting facility, items move
across to the distribution centre. Here, parcels are assigned to variously sized trucks and
vans which head out to smaller distribution hubs across the city and in most of Kenya’s
other counties.15 These are either small shop fronts with a loading bay or, in some cases,
other businesses acting as franchisees. From these hubs, boda riders bridge the last- mile
gap, delivering parcels to end customers on behalf of Dasher’s B2B clients. These hubs
are also drop- o points for domestic deliveries: once an item is scanned in at one of
these hubs, it travels to the main distribution centre in Nairobi from where it is line-
hauled to the destination hub and eventually delivered by local boda riders. This
centralized system responds to the need to optimize the interface between the middle
and the last mile, Jenny explains, pointing to the second warehouse building:
Everything goes through this DC [distribution centre]. So, for example, between
Meru and Mombasa trucks would be half full. But because they all go through
this DC, we make sure that all the routes are optimized … In logistics, the most
important thing is the network; the more a network is capillary, the bigger your
abilities.
The ultimate capillaries of this network are indeed the boda riders, she goes on
to explain, showing us the algorithmic system that renders each rider and each parcel
a discrete data point thanks to the barcodes that are scanned at key points of a delivery.
Unlike other large on- demand labour platforms oering parcel services, Uber
included, Dasher does not rely on a casual workforce. The riders are all formally
employed, although their wages are piece- based. Only during peak periods such as
Black Friday are temporary riders added to the base team, which, after just one year of
operation, already comprises some 200 members (ranging from 75 in Nairobi itself to
just a couple in smaller cities). While casual labour would be a cheaper option, Dasher
decided to invest in training couriers so as to baseline their customer service quality and
make sure their riders are indeed data collectors as well as couriers. Although neither
Jenny nor Kevin mention it explicitly, and only refer to a ‘politeness’ problem, it is
15 Since the constitutional reform of 2010, Kenya has been divided into 47 counties which vary greatly in both size
and population.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 14
clear that the decision to onboard riders as employees is a way to address the negative
perceptions which surround the boda boda sector in Kenya.
In many respects, Dasher is not unique in its business model or in the way that
it formalizes boda riders as direct employees. Research into platforms has shown time
and again that on- demand work can make labour less rather than more precarious,
but not necessarily dierent or better (Schor et al.,2020). Nonetheless, by examining
the ways in which Dasher visualizes its last- mile network we catch a glimpse of how
business models envision the future of urban mobility infrastructures. Both Kevin and
Jenny explain that Dasher’s aim is to become Africa’s ‘number one last- mile service
provider’ dedicated to e- commerce. This is something of a pipedream, as they admit,
since e- commerce in Africa is still only a very small market. But they are playing the
‘long game’, says Kevin, arguing that it only takes two ingredients for e- commerce to
boom: a functional, diused online payment system and eective last- mile logistics:
If anyone can solve these two things, online payment systems and last- mile
logistics, e- commerce will be pumping … No country can escape e- commerce.
I don’t know who will be the last winner. In America it’s Amazon, in China it’s
Alibaba, and in different countries they have their own. I don’t know who will be
the winner in Africa. But whoever it is, they’ll need us.
So why did we come here? At this phase, we are building the night work. We’ve
connected every county, trained our team, taught them how to be ready. We are
waiting like a crouching tiger.
In playing this long game16 while African e- commerce develops, Kevin is already
working on the distributed infrastructure that a boom in the demand for last- mile
delivery will mean for a city like Nairobi. He has ordered twenty smart cabinets from
China. These cabinets, which will be distributed across the city to shopping malls and
large corporate oces, will function as the last- mile endpoints to Dasher’s delivery
services. Enabled with Internet- of- Things (IOT) technology and fully automated, the
cabinets sport the same bright colours as the company’s brand and are equipped with
small lock- boxes. Every day, boda riders will fill and take from these cabinets across the
city
—
addressing the same issue of addressability that Shoppist is trying to solve with its
Shoppost model.
The prototype of one of these smart cabinets, soon to become a fixture of urban
life in Nairobi, sits just outside the oce where Jenny and Kevin are talking to us. In
many respects, Dasher makes some of the same issues that we have already seen in
the two previous case studies legible; once again, as problems of last- mile algorithm
coordination. Motorcycles, in Dasher’s business model, contribute to a unique hub-
and- spoke logistics model in which the gap- filling capacities of the boda boda sector are
visualized as both capillaries of a larger distribution network and the data- gathering
terminals of this network. In other words, algorithmic suturing is always more than
just software- based ‘tricks’ (Pasquinelli,2019: 6): it entails riders, motorcycles, cabinets,
warehouses, and so much besides. What is also worth noting
—
which Dasher illustrates
even more than the other two cases
—
is the speculative nature of these platforms; not
only does Dasher currently operate at a loss, waiting for an e- commerce boom that may
or may not materialize, but it does so by speculating on a spatial network of distribution
that is currently in the making but not there yet. As we move now to the conclusion
16 In a way, this confirms and extends C.K. Lee’s(2017) argument that corporate China in Africa is often involved in
practices of profit optimization rather than profit maximization. While Lee’s point is specifically about state capital
as opposed to other varieties of capital, our research suggests that Chinese tech startups in Kenya also operate
across diverse temporalities of profitability.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
15ALGORITHMIC SUTURING
of this article, we will return to the importance of these future- making algorithmic
practices for urban research.
Conclusion
What can be seen about a city when one starts with the business models of last-
mile startups? How do they see, make sense of and make visible the ends of splintered
urban infrastructures and economies? Of course, these are complex questions that can
be cut and sliced in many ways (Qadri and D’Ignazio,2022). Owing to the ubiquity of the
motorcycle in the provision of last- mile logistics in African cities, we have intentionally
selected startups that tap into Nairobi’s (in)famous boda boda sector as our entry point
into this way of seeing and making legible. Boda boda, as we have shown, operate as
a prosthetic tie that helps suture the fragmented urban infrastructures of Nairobi.
Increasingly, they do so through digital platforms, and through processes that we have
labelled using the concept of ‘algorithmic suturing’: the more- than- spatial patching of
urban fragments through practices of data- driven optimization.
Our contribution with this article is therefore both empirical and conceptual. On
an immediate level, our case studies show how last- mile algorithmic business models
seek to stitch together splintered infrastructures by making them legible as sites of
coordination. More specifically, what last- mile business models visualize
—
and therefore
seek to address
—
are the gaps in the urban economies that they can tap into. In this
sense, our research speaks to a body of Southern infrastructure and Southern urbanism
scholarship which has shown how in Southern cities networked systems (where they
exist) are gap- filled and sutured by practices of material improvisation and the marginal
technological survival of the poor. Yet the last- mile examples we have narrated in this
article point to something different: stitching together the loose ends of splintered
urban infrastructure is about more than material improvisation, or individual responses,
or marginal practices of survivability. Algorithmic suturing is, in fact, a deliberate,
elaborate set of experimental business models through which urban fractures become
legible as problems of optimization and, therefore, as sources of value. In other words,
acts of suturing are not just performed by informal economies through make- do thrift,
but by deliberate algorithms that seek to integrate the latter into (perhaps) profitable
data- driven platforms.
Moreover, while there is a tendency (particularly among scholars who have
sought to reconcile top- down readings of infrastructure and platforms with everyday
lived experiences) to assume that suturing happens at the individual level, or at small
collective scales, the last mile is about big numbers: it is about producing algorithmic
stitches that do not address single splinters of urbanity but enrol whole interfaces
between inconsistent economies, using the ends of these networks
—
the last mile
—
as a
constant data source to recursively coordinate better and optimize more. Put dierently,
the ‘infrastructural heterogeneity’ often attributed to Southern cities (Lawhon et
al.,2018) is also the calculated outcome of highly scalable business models. Uber and
the like know this very well, and to leave this insight with them represents a missed
opportunity for urban research on digital infrastructure, both in the African city and
beyond.
Seeing like a last mile is also a methodological intervention. Like any method, as
John Law reminds us(2004), our approach makes some things legible while others fall
into the background. We have mentioned, for example, how the exploitation of platform
labour (Van Doorn,2017) is not the focus of this article. At the same time, our perspective
inductively calls attention to important undercurrents sweeping across African cities.
The cases we have outlined hint at some of these issues, which we cannot address in
full here. However, it is worth highlighting a few of the issues which merit a much
richer urban inquiry. The last mile sees, for instance, the possibilities that lie within the
needs and aspirations of the lower rungs of Africa’s urban middle class, as well as the
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 16
inescapably tight margins of this demographic as drivers of urban change. The last mile
also reveals the impressive strides made by Chinese tech and e- commerce companies,
so often overlooked by scholars concerned with China’s large- scale and state- driven
investments (Pollio,2022b). Finally, these business models also foreground the quiet
emergence of new spaces of urban logistics such as dark kitchens, dark laundromats
and ghost warehouses (Shapiro,2022), or even more mundane urban ‘objects’ like
smart cabinets. Last- mile startups envision and speculate on these spaces and objects
before they are materialized, when they are still being considered, planned and tested.
Algorithmic suturing, we have shown, is speculatively experimental.
On a more pragmatic level, the last mile sheds light on the diverse forms of value
promised by these platformed urban economies and the various shades of invisible
work and knowledge that sustain them (Irani,2015). There is a need, we believe, to
bring the perspective represented in this article into scholarly research that centres
questions of labour, recognizing that labour dynamics are also and fundamentally
shaped by contextual practices of algorithmic suturing. Without a fuller appreciation of
this nexus, we are already witnessing the emergence of regulations (and the purposeful
lack thereof ) which negatively aect those whom they are meant to protect. As engaged
researchers, we will seek to make these insights speak to regulators and industrial
players through the work that we do beyond academic publishing. In the meantime, our
hope is that scholars of cities and infrastructure will take up the research challenges that
the last mile makes visible, some of which we have outlined in this article.
Andrea Pollio, Department of Urban and Regional Research and Planning (DIST),
Polytechnic of Turin and African Centre for Cities, University of Cape Town,
Rondebosch, 7701, South Africa, andrea.pollio@polito.it
Liza Rose Cirolia, African Centre for Cities, University of Cape Town, Rondebosch,
7701, South Africa, liza.cirolia@gmail.com
Jack Ong’iro Odeo, Department of Human Geography, Stockholm University, SE-
106 91 Stockholm, Sweden, jack.odeo@gmail.com
References
Agbiboa, D.E. (2020) How informal transport
systems drive African cities. Current History 119.817,
175– 81.
Akallah, J.A. and M. Hård (2020) Under the historian’s radar:
local water supply practices in Nairobi, 1940– 1980.
Water Alternatives 13.3, 886– 901.
Altenried, M. (2019) On the last mile: logistical urbanism
and the transformation of labour. Work Organisation,
Labour & Globalisation 13.1, 114– 29.
Anwar, M.A. and M. Graham (2020) Hidden transcripts of
the gig economy: labour agency and the new art of
resistance among African gig workers. Environment and
Planning A: Economy and Space 52.7, 1269– 91.
Anwar, M.A., J.O.I. Odeo and E. Otieno (2022) ‘There is no
future in it’: pandemic and ride hailing hustle in Africa.
International Labour Review 162.1, 23– 44.
Arubayi, D. (2021) Documenting the everyday hidden
resistance of ride- hailing platform drivers to
algorithmic management in Lagos, Nigeria. South
Atlantic Quarterly 120.4, 823– 38.
Baptista, I. (2019) Electricity services always in the
making:informality and the work of infrastructure
maintenance and repair in an African city. Urban
Studies 56.3, 510– 25.
Chun, W.H.K. (2021) Discriminating data: correlation,
neighborhoods, and the new politics of recognition.
MIT Press, Cambridge, MA.
Cirolia, L.R. and A. Pollio (forthcoming) Spectrums of
infrastructural hybridity: insights from urban Africa for
a propositional research agenda. In O. Coutard and D.
Florentin (eds.), Handbook of infrastructure and cities,
Edward Elgar, Cheltenham.
Cirolia, L.R., T. Hailu, J. King, N.F. da Cruz and J. Beall (2021)
Infrastructure governance in the post- networked
city: state- led, high- tech sanitation in Addis Ababa’s
condominium housing. Environment and Planning C:
Politics and Space 39.7, 1606– 24.
Cirolia, L.R., R. Sitas, A. Pollio, A.G. Sebarenzi and
P.K. Guma (2023) Silicon savannahs and motorcycle taxis: a
Southern perspective on the frontiers of platform urbanism.
Environment and Planning A: Economy and Space. doi:
10.1177/03085 18X23 1170193
Darbon, D. (2019) The political role of the African middle
class: the over- politicization of an elusive category. In
The Oxford Research Encyclopedia of Politics [WWW
document]. URL https://oxfor dre.com/polit ics/displ
ay/10.1093/acref ore/97801 90228 637.001.0001/acref
ore- 97801 90228 637- e- 739
Daston, L. (2022) Rules: a short history of what we live by.
Princeton University Press, Princeton, NJ.
De Boeck, F. and S. Baloji (2016) Suturing the city: living
together in Congo’s urban worlds. Autograph, London.
Doganova, L. and M. Eyquem- Renault (2009) What do
business models do? Innovation devices in technology
entrepreneurship. Research Policy 38.10, 1559– 70.
Doherty, J. (2020) Motorcycle taxis, personhood, and the
moral landscape of mobility. Geoforum. doi: https://
doi.org/10.1016/j.geofo rum.2020.04.003
Dourish, P. (2007). Seeing like an interface. In OZCHI ’07,
Proceedings of the 19th Australasian conference
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
17ALGORITHMIC SUTURING
on computer- human interaction: entertaining user
interfaces, November 2007 [WWW document]. URL
https://doi.org/10.1145/13248 92.1324894
Ese, A. and K. Ese (2020) The city makers of Nairobi: an
African urban history. Routledge, London.
Ferguson, J. (2005) Seeing like an oil company: space,
security, and global capital in neoliberal Africa.
American Anthropologist 107.3, 377– 82.
Fourcade, M. and K. Healy (2017) Seeing like a market.
Socio- Economic Review 15.1, 9– 29.
Friederici, N. (2018) Hope and hype in Africa’s digital
economy: the rise of innovation hubs. In M. Graham
(ed.), Digital economies at global margins, MIT Press,
Cambridge, MA.
Galloway, A.R. and E. Thacker (2013) The exploit: a theory of
networks. University of Minnesota Press, Minneapolis,
MN.
Gillespie, T. (2014) The relevance of algorithms. In T.
Gillespie, P.J. Boczkowski and K.A. Foot (eds.), Media
technologies: essays on communication, materiality,
and society, MIT Press, Cambridge, MA.
Goodfellow, T. (2015) Taming the ‘rogue’ sector: studying
state effectiveness in Africa through informal transport
politics. Comparative Politics 47.2, 127– 47.
Graham, M. and L. Mann (2013) Imagining a silicon
Savannah? Technological and conceptual connectivity
in Kenya’s BPO and software development sectors.
Electronic Journal of Information Systems in
Developing Countries 56.1, 1– 19.
Graham, S. and S. Marvin (2001) Splintering urbanism:
networked infrastructures, technological mobilities
and the urban condition. Routledge, London and New
York, NY.
Gray, M.L. and S. Suri (2019) Ghost work: how to stop
Silicon Valley from building a new global underclass.
Houghton Mifflin Harcourt, Boston, MA.
Guma, P.K. (2022) The temporal incompleteness of
infrastructure and the urban. Journal of Urban
Technology. doi: https://doi.org/10.1080/10630
732.2021.2004068.
Guma, P.K. and J. Monstadt (2021) Smart city making?
The spread of ICT- driven plans and infrastructures in
Nairobi. Urban Geography 42.3, 360– 81.
Guma, P.K. and M. Mwaura (2021) Infrastructural
configurations of mobile telephony in urban Africa:
vignettes from Buru Buru, Nairobi. Journal of Eastern
African Studies 15.4, 527– 45.
Guma, P.K. and A. Wiig (2022) Smartness beyond the
network: water ATMs and disruptions from below in
Mathare Valley, Nairobi. Journal of Urban Technology,
1– 21.
Guma, P.K., J.A. Akallah and J.O. Odeo (2023) Plug- in
urbanism: city building and the parodic guise of new
infrastructure in Africa. Urban Studies. doi: https://doi.
org/10.1177/00420 98023 1158013
Günel, G. (2021) Leapfrogging to solar. South Atlantic
Quarterly 120.1, 163– 75.
Guyer, J.I. (2016) Legacies, logics, logistics. University of
Chicago Press, Chicago, IL.
Halpern, O. (2015) Beautiful data. Duke University Press,
Durham, NC.
Hayes, A. (2021) Last mile: what it means in reaching
customers. Investopedia [WWW document]. URL
https://www.inves toped ia.com/terms/ l/lastm ile.asp
(accessed 24 January 2022).
Iazzolino, G. (2021) ‘Going Karura’: colliding subjectivities
and labour struggle in Nairobi’s gig economy.
Environment and Planning A: Economy and Space. doi:
0308518X211031916
Irani, L. (2015) Difference and dependence among digital
workers: the case of Amazon Mechanical Turk. South
Atlantic Quarterly 114.1, 225– 34.
Jaglin, S. (2014) Rethinking urban heterogeneity. In
S. Oldfield and S. Parnell (eds.), The Routledge
handbook on cities of the global South, Routledge,
London and New York, NY.
Jagoda, P. (2016) Network aesthetics. University of Chicago
Press, Chicago, IL.
James, D. (2021) Life and debt: a view from the South.
Economy and Society 50.1, 36– 56.
Karaman, O., L. Sawyer, C. Schmid and K.P. Wong (2020) Plot
by plot: plotting urbanism as an ordinary process of
urbanisation. Antipode 52.4, 1122– 51.
Kimari, W. (2021) The story of a pump: life, death and
afterlives within an urban planning of ‘divide and rule’
in Nairobi, Kenya. Urban Geography 42.2, 141– 60.
Klopp, J., D. Orwa, P. Waiganjo, W.S. Williams and A. White
(2017) Informal 2.0: seeing and improving urban
informal practices through digital technologies. The
digital matatus case in Nairobi. Field Actions Science
Reports 16, 39– 43.
KNBS (2022) Economic survey 2022, Kenya National Bureau
of Statistics [WWW document]. URL https://www.knbs.
or.ke/downl oad/econo mic- surve y- 2022/
Larkin, B. (2004) Degraded images, distorted sounds:
Nigerian video and the infrastructure of piracy. Public
Culture 16.2, 289– 314.
Law, J. (2004) After method: mess in social science research.
Routledge, London and New York, NY.
Law, J. (2009) Seeing like a survey. Cultural Sociology 3.2,
239– 56.
Lawhon, M., D. Nilsson, J. Silver, H. Ernstson and S. Lwasa
(2018) Thinking through heterogeneous infrastructure
configurations. Urban Studies 55.4, 720– 32.
Lee, C.K. (2017) The specter of global China. In The Specter
of Global China, University of Chicago Press, Chicago,
IL.
Lemanski, C. (2021) Broadening the landscape of
post- network cities: a call to research the off- grid
infrastructure transitions of the non- poor. Landscape
Research. doi: https://doi.org/10.1080/01426
397.2021.1972952
Maina, M. and L.R. Cirolia (forthcoming) Ring roads, revived
plans, and plotted practice: the multiple makings of
Nairobi’s urban periphery. Habitat International.
Mann, L. and G. Iazzolino (2021) From development state
to corporate leviathan: historicizing the infrastructural
performativity of digital platforms within Kenyan
agriculture. Development and Change 52.4, 829– 54.
Mattern, S. (2016) Instrumental city: the view from Hudson
Yards, circa 2019. Places [WWW document]. URL
https://place sjour nal.org/artic le/instr ument al- city- new-
york- hudso n- yards/ ?cn- reloa ded=1
McFarlane, C. (2018) Fragment urbanism: politics at the
margins of the city. Environment and Planning D:
Society and Space 36.6, 1007– 25.
Mercer, C. (2014) Middle class construction: domestic
architecture, aesthetics and anxieties in Tanzania. The
Journal of Modern African Studies 52.2, 227– 50.
Mercer, C. (2020) Boundary work: becoming middle class in
suburban Dar es Salaam. International Journal of Urban
and Regional Research 44.3, 521– 36.
Munster, A. (2013) An aesthesia of networks: conjunctive
experience in art and technology. MIT Press,
Cambridge, MA.
Mutongi, K. (2017) Matatu: a history of popular
transportation in Nairobi. University of Chicago Press,
Chicago, IL.
Nowak, S. (2021) The social lives of network effects:
speculation and risk in Jakarta’s platform economy.
Environment and Planning A: Economy and Space. doi:
0308518X211056953.
Nyairo J. (2023) The boda boda (r)age: economies of
affection in the motorbike taxis of Kenya. English
Studies in Africa 66.1, 109– 23
Odendaal, N. (2021) Platform urbanism and hybrid places
in African cities. In A. Aurigi and N. Odendaal (eds.),
Shaping smart for better cities, Academic Press.
Odendaal, N. (2023) Disrupted urbanism: situated smart
initiatives in African cities. Bristol University Press,
Bristol.
Omulo, C. (2021) How Nairobi CBD became one big boda
boda bazaar. Nation 16 November [WWW document].
URL https://nation.afric a/kenya/ count ies/nairo bi/
how- nairo bi- cbd- becam e- one- big- boda- boda- bazaa
r- 3620524.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
POLLIO, CIROLIA AND ODEO 18
O’Neill, K.L. (2022) Eat or be eaten: motorcycle taxis in
Guatemala City. Urban Geography 43.4, 483– 500.
Opondo, V.O. and G. Kiprop (2018) Boda boda motorcycle
transport and security challenges in Kenya. National
Crime Research Centre, Nairobi. URL https://www.
crime resea rch.go.ke/wp- conte nt/uploa ds/2020/11/
Repor t- on- Boda- boda- Motor cycle - Trans port- and- Secur
ity- Chale nges- in- Kenya - 2018.pdf
Ouma, S. (2017) The difference that ‘capitalism’ makes: on
the merits and limits of critical political economy in
African Studies. Review of African Political Economy
44.153, 499– 509.
Pasquinelli, M. (2019) Three thousand years of algorithmic
rituals: the emergence of AI from the computation
of space. e- flux Journal 101 [WWW document]. URL
https://www.e- flux.com/journ al/101/27322 1/three
- thous and- years - of- algor ithmi c- ritua ls- the- emerg ence-
of- ai- from- the- compu tatio n- of- space/
Pirie, G. (2014) Transport pressures in urban Africa:
practices, policies, perspectives. In S. Parnell and
E. Pieterse (eds.), Africa’s urban revolution, Zed Books,
London.
Pollio, A. (2019) Forefronts of the sharing economy: Uber
in Cape Town. International Journal of Urban and
Regional Research 43.4, 760– 75.
Pollio, A. (2020) Making the Silicon Cape of Africa: tales,
theories and the narration of startup urbanism. Urban
Studies 57.13, 2715– 32.
Pollio, A. (2021) Uber, airports, and labour at the
infrastructural interfaces of platform urbanism.
Geoforum 118 (January), 47– 55.
Pollio, A. (2022a) Acceleration, development and
technocapitalism at the Silicon Cape of Africa.
Economy and Society 51.1, 46– 70.
Pollio, A. (2022b) The digital silk road as planetary
intelligence: a story of China in Africa. In R. Singh, R.L.
Guzmán and P. Davison (eds.), Parables of AI in/from
the majority world, Data & Society Research Institute,
New York, NY.
Qadri, R. and C. D’Ignazio (2022) Seeing like a driver: how
workers repair, resist, and reinforce the platform’s
algorithmic visions. Big Data & Society 9.2. doi:
20539517221133780.
Rizzo, M. (2002) Being taken for a ride: privatisation of
the Dar es Salaam transport system 1983– 1998. The
Journal of Modern African Studies 40.1, 133– 57.
Rosenberg, L. and A. Brent (2020) Infrastructure disruption
in ‘Silicon Savannah’: exploring the idea of the creative
class and their relation to quality of place in Nairobi,
Kenya. International Journal of Urban and Regional
Research 44.5, 809– 20.
Schoenberger, E. (1991) The corporate interview as
a research method in economic geography: the
professional geographer 43.2, 180– 89.
Schor, J.B., W. Attwood- Charles, M. Cansoy, I. Ladegaard
and R. Wengronowitz (2020) Dependence and
precarity in the platform economy. Theory and Society
49.5, 833– 61.
Scott, J.C. (1998) Seeing like a state: how certain schemes
to improve the human condition have failed. Yale
University Press, New Haven, CT.
Seaver, N. (2017) Algorithms as culture: some tactics for
the ethnography of algorithmic systems. Big Data &
Society 4.2. doi: 2053951717738104.
Seaver, N. (2021) Seeing like an infrastructure: avidity and
difference in algorithmic recommendation. Cultural
Studies 35.4/5, 771– 91.
Shapiro, A. (2022) Platform urbanism in a pandemic:
dark stores, ghost kitchens, and the logistical-
urban frontier. Journal of Consumer Culture. doi:
14695405211069983.
Silver, J. (2014) Incremental infrastructures: material
improvisation and social collaboration across post-
colonial Accra. Urban Geography 35.6, 788– 804.
Simone, A. (2021) Ritornello: ‘people as infrastructure’.
Urban Geography 42.9, 1341– 48.
Sitas, R., L.R. Cirolia, A. Pollio, A.G. Sebarenzi, P. Guma and
A. Rajashekar (2022) Platform politics and silicon
savannahs: the rise of on- demand logistics and mobility
in Nairobi and Kigali. African Centre for Cities, Cape
Town.
Sopranzetti, C. (2013) The owners of the map: motorcycle
taxi drivers, mobility, and politics in Bangkok. Harvard
University Press, Cambridge, MA.
Striphas, T. (2015) Algorithmic culture. European Journal of
Cultural Studies 18.4/5, 395– 412.
Terranova, T. (2004) Network culture: politics for the
information age. Pluto Press, London.
Van Doorn, N. (2017) Platform labor: on the gendered
and racialized exploitation of low- income service
work in the ‘on- demand’ economy. Information,
Communication & Society 20.6, 898– 914.
14682427, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2427.13200 by Cochrane Romania, Wiley Online Library on [23/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License