3 Economic Geographies of Digital Work in Africa

Workers perform a diverse range of activities both in BPOs and on platforms, and one of our objectives in this book is therefore to understand the digital economy as a continuum with a great deal of heterogeneity. In this chapter we will provide a snapshot of how and where these diverse work activities get done in Africa, showing that African workers remain very much a part of contemporary digital capitalism. They perform a wide range of digital work activities from diverse locations—from an office block in the centre of a lively metropolis, to a makeshift room in a town recovering from civil war, as well as a multitude of bedrooms, cafes, and libraries across the continent. In summary, this chapter will provide a visual and descriptive outline of the various types of digital activities being performed in newer spaces that are connecting to the global information economy. In doing so, we ask what types of work get done in Africa, and what that means for value creation and capture.

As part of our study, we examined the outsourced services landscape in five case study countries by collecting information on the presence of BPO firms (both foreign and domestic), their country of origin, locations of their customers and clients, source of work, presence of intermediaries, kinds of operations such as captive or outsourced, types of activities in the value chains (e.g. low value-added or high value-added), and the distribution of firms in each country. We focused specifically on BPOs—that is, on entry-level activities in the outsourcing sector’s value chains (Fernandez-Stark et al., 2011), including both voice-based (telephone calls) and non-voice services (e.g. emails and webchat) involving customer support, data-capture, outbound sales, document conversion, transcription, and digitization.

The rapid diffusion and penetration of ICTs in Africa has led to a nascent BPO industry in recent years (Anwar and Graham, 2019; Benner 2006; Benner and Rossi, 2016; Kleibert and Mann, 2020). These BPO operations, such as call and contact centres, provide African economies with an important entry point into digital outsourcing production networks. With the uptake in digital production generating a ‘tradability revolution’ in services—now considered a key component for economic development (UNCTAD, 2004)—BPOs have become a key policy initiative in a number of African countries, including Kenya, South Africa, Nigeria, Uganda, Ghana, and Egypt (see Ndemo, n.d.). However, the outsourcing industry is buyer driven (Fernandez-Stark et al., 2011), characterized by asymmetrical power relations, and highly cost sensitive which often benefit lead firms (Peck, 2017) to the detriment of smaller firms and local assets. It therefore places some limits on regions and their institutions’ attempts to plug into these networks and gain from these dynamics (Gereffi and Lee, 2016). Indeed, we found a varied landscape of outsourcing in our case study countries (Table 3.1).

The literature on value chains and production networks has been useful in making sense of the various production nodes in the global economy (Coe and Yeung, 2015; Gereffi, 1999, 2005; Henderson et al. 2002). One of the important concepts in the GPN literature has been that of ‘strategic coupling’, defined as ‘a mutually dependent and constitutive process involving particular ties, shared interests, and cooperation between two or more groups of economic actors who otherwise might not act in tandem to achieve a common strategic objective’ (Yeung, 2016: 54). But coupling is also ‘time–space contingent’ and geographically and institutionally dependent (Yeung, 2016: 57). It can change over time, and in different geographical contexts can even lead to decoupling of an entire region, defined by disinvestment, the exit of foreign firms, and loss of access to foreign markets (see Horner, 2014; MacKinnon, 2012, 2013).

Broadly speaking, strategic coupling is evident in South Africa, with the South African government actively promoting the industry through a variety of incentives. Foreign lead firms have established captive operations (i.e. run and managed by the parent firm) as well as outsourced their services to local vendors in South Africa. Some of the key lead firms are Teleperformance, WNS, Webhelp, and Aegis, all of whom have set up offshore delivery centres in the country. Webhelp has a workforce of 4,000 workers across six different offices in South Africa. A host of regional shared service operations of global corporations like Amazon, Microsoft, British Petroleum, Shell, and Lufthansa have also emerged.¹ Despite the increasing presence of international lead firms in South Africa, a majority of the BPO operations are still serving the domestic market. This is not surprising given that most call centres globally (86 per cent) serve the local rather than international market (Holman et al., 2007). However, the BPO industry in Kenya, Nigeria, Ghana, and Uganda can be characterized by structural coupling, defined by unequal power relations in networks with lead firms attracted to a region for their labour surpluses, and often associated by branch plant syndrome (Mackinnon, 2013).² These destinations are primarily seen by lead firms as low-cost destinations for low value-added work, and the countries’ dependence on third-party vendors or intermediaries make them prone to decoupling.³ BPO operations tend to be footloose, meaning that production can relocate with relative ease. In fact, we found examples of contraction in the outsourcing industry in some locations in Kenya, Ghana, and Nigeria, with BPO operations closing down.

The existing literature has already identified the potential and limits of technological connectivities in influencing BPO operations in Africa (Graham et al., 2015; Mann and Graham, 2016). One of the central arguments in this literature is that while digital infrastructure overcomes some constraints for African enterprises (namely technical skills), there are various other forms of control and imbalanced power relationships embedded in these global production networks which need to be overcome (Foster et al., 2018). In the following we discuss African countries’ attempts to integrate into digital outsourcing production networks and the uneven landscape that results.

Opportunities to develop BPO operations have improved for only a handful of African countries, including South Africa, Egypt, and Mauritius. This is surprising, since in our interviews with executives of BPO firms, government officials, and private sector associations, almost everyone pointed to similar value propositions of their respective countries, including a skilled workforce with English language skills, low labour costs, favourable time zones, and good internet connectivity. Ironically, Ghana and Kenya have both ranked higher than South Africa for the last two years in the offshore locations rankings of various management consultancy firms. Yet, their offshore segment has not grown, while South Africa is slowly becoming a destination of choice for offshore work; albeit still not as popular as India and the Philippines (Anwar and Graham, 2019). Various economic, socio-political, and reputational factors were noted by several industry experts as key for the BPO industry to develop further in South Africa.

As the second biggest economy on the continent, a member of the BRICS bloc, and the only African member of the G20, South Africa acts as a gateway to regional African markets. It is also relatively politically stable, and its infrastructure and education levels are considered by industry members to be optimal for offshore work to take place. While cost remains the main factor in locating BPO operations—in particular labour cost, which accounts for 60 per cent of operational expenditure (Interview, pers. comm., BPO Executive Johannesburg, 2016)⁴—there are combinations of other factors that can shape the offshoring landscape. According to an industry expert from the US, the buyer community in the US is looking to minimize their business risk and reduce their dependence on India and the Philippines, and South Africa offers to them low cost and a skilled workforce (Interview, pers. comm., Johannesburg, 2016). For BPO firms in South Africa the declining value of the South African Rand in the post-2008 period means the country has become reasonably competitive compared to locations such as the Philippines and India for entry-level work. According to one business executive from an international firm operating a call centre in South Africa:

We bill a large portion in [GB] Pounds, and so the worse the Rand is, more feasible and stronger our financial performances are because we have revenue in Pounds and we have a cost base (e.g. wages and workspace) in Rand, which is a dream.

Interview, pers. comm., Johannesburg, 2016

South Africa also has a stable base of graduate entrants to the labour market. An estimated 150,000 people graduate every year from South African universities into the labour market, and most of them are considered to be suited for entry-level BPO work (BPESA, 2015a). One of the officials of the industry body, Business Process Enabling South Africa (BPESA), told us that the availability and scalability of a skilled workforce at competitive prices remains key to South Africa’s success (Interview, pers. comm., Johannesburg, 2016). The fact that South Africa has one of the highest youth unemployment rates in the world means that the government’s industrial policy action plan is geared to supporting sectors which can contribute to employment generation, hence its prioritization of the ‘business process services’ (BPS) sector; a point confirmed to us by a Department of Trade and Industry (DTI) official (Interview, pers. comm., Johannesburg, 2016).⁵

The South African government’s policy initiatives, such as generous cash incentives especially for outsourced services (in place since 2014) provide a further impetus to attracting foreign lead firms in the hopes of creating jobs for South Africans. According to the DTI official we interviewed, these incentives are provided with a view to narrowing the cost gap between South Africa and other destinations to secure contracts for offshore markets (Interview, pers. comm., Johannesburg, 2016). The incentive programme was further revised in 2018, to include cash grants for firms creating at least fifty ‘offshore jobs’ within the first three years. For entry-level work (e.g. inbound customer services), a firm gets paid US$9,375 per job created over five years. There is also a 20 per cent one-off ‘bonus incentive’ for creating more than 500 jobs and maintaining these for five years.⁶ By comparison, Egypt provides subsidies for vocational training, telecommunications costs, and rents (Reuters, 2010b), and has an export-based cash rebate programme for IT firms (ITIDA, 2018). No similar incentives were found in any of the other four case study countries, although alternative initiatives exist.

Spatial zoning policies seem to be the key planning instrument used by governments in Kenya and Ghana to kick-start the nascent IT and BPO sector—however, with relatively little success in attracting investments and creating jobs. Examples include dedicated free trade zones or parks such as Konza City (Kenya), Tema ICT Park (Ghana), and the BPO Incubation Centre (Uganda). Such zoning policies have precedents outside Africa, for example in India where the government has allowed 100 per cent foreign direct investment into the IT-sector driven special economic zones and tax exemptions on export and import (Anwar, 2014). The Kenyan state has experimented with exemption from corporate income tax and stamp duty, employment-based grants, and duty-free import of materials in Konza City. The Kenyan government also subsidized bandwidth costs for BPO companies in 2007–08 by redirecting funds from an e-government project of the World Bank (cited from Mann and Graham, 2016). The Ghanaian government, meanwhile, has set up the Accra Digital Centre on the site of an old public works warehouse in Accra, which, according to its managing director, provides cheap office space and ancillary services to attract BPO companies to the site (Interview, pers. comm., Accra, May 2017).⁷ Unfortunately, the Tema ICT Park in Ghana, which is expected to host IT and BPO firms, had only one tenant (a local private sector firm doing customer services work) when we visited their offices in 2017 (Figure 3.1).⁸ Similarly, the BPO Incubation Centre (Uganda) had only one big regional firm, which was doing sales and customer retention campaigns for a cable television company and their customers in East Africa.

Markets do not exist in a vaccum, they need the protective covering of state institutions (Polanyi, 2001). Despite the push for privatization and deregulation on the continent (under the stuctural adjustment programmes) by the World Bank and IMF, the need for markets to be supported and developed in close collaboration with state was clearly evident in our work on the outsourcing sector. Several BPO executives and private sector associations conceded that the state needs to play an active role for the local BPO industry to develop and flourish. For them state can help with the brand marketing abroad, especially in buyer locations such as the US and United Kingdom (UK). This helps create awareness and also improve the image of the country in order to attract offshore work. However, The level of institutional capacity of the state and its relevant institutions varied across our case-study countries.

The South African government, through the DTI, led the country’s marketing effort as far back as 2002–03 in partnership with the private sector and management consultancies to formulate a sector development strategy for the BPO industry, which has made it easier to sell the country’s location abroad, attract investment, and create jobs (Interview, pers. comm., BPESA official, 2016). The South African Government has a close relationship with BPESA and is actively involved in supporting the sector through marketing. As one of the DTI officials explained to us:

Another part of our strategy [for development of the outsourcing sector] is investor promotion, what we call broad based marketing. So, creating awareness in an offshore market is about the fact that South Africa is the place to go to do BPO. We have a South African value proposition that gets offered to a prospective investor that wants to come in. We also provide all the investor facilitation and logistic services, work permits, etc. It makes it easier to get into South Africa. And then obviously there is the talent development path which is where the skills development strategy falls into. And then another part of the government’s role is to mobilise industry. The way we do that is by working through with the industry body [BPESA]. Although we do not fund the industry body, we have a working relationship with the industry body in order to support them in terms of marketing abroad.

Interview, pers. comm., Johannesburg, September 2016

An important aspect of the South African state’s role in the outsourcing sector is the increasing number of business trips and participation in buyer forums, which have created awareness among international buyers. In collaboration with BPESA, the South African government has been leading the way in this regard. In 2019 alone, BPESA delegates along with DTI officials went on several marketing and investment promotion trips to the UK, Australia, and the US—three major offshore markets for South Africa (Anwar and Graham, 2019). The DTI’s new Global Business Services (GBS) incentives were launched in London in 2018.⁹ At the launch, the BPESA’s CEO was quoted as saying ‘[The] government’s continued support for this sector is applauded and hopefully sends a clear message to investors in the UK and elsewhere that South Africa is very serious about your business and will continue to put great effort and resources into attracting and retaining your investment in South Africa’ (DTI, 2018).

The South African BPO industry is now the biggest in our sample in terms of number of firms and jobs created, with only Egypt as its closest rival (see Table 3.2). Of course, South Africa is an economic heavyweight on the continent with its well-developed infrastructure and better education levels, which many smaller economies such as Uganda simply cannot match. Even though both Kenya and Rwanda are considered to have potential for offshore services, their IT and BPO industry lags behind South Africa’s.

International bodies such as the Netherlands Trust Fund (NTF) have funded projects in Kenya and Uganda with the aim of enhancing the export competitiveness of the IT and IT-enabled service sector (ITC, 2015). The NTF in partnership with the International Trade Centre has selected a few small and medium enterprises to increase their capacity and create and validate an export plan, thus improving the capacity of trade support institutions and expanding business linkages and partnerships in target export markets. The CEO of a local Ugandan firm told us that ‘the NTF project provided Ugandan companies skills development training, evaluated their marketing materials online and offline, and offered advice on rebranding their websites and business cards’ (Interview, pers. comm., Kampala, 2017). The NTF project helped bring together a group of companies to form a private sector association, the Alliance for Trade and Information Technology and Services (ATIS), to undertake advocacy work and marketing of Uganda as a destination for offshore work (Interview, pers. comm., ATIS Official Kampala, 2017). However, it has failed to bring in more offshore work. They also noted lead firms’ prejudices about Africa as a location for BPO work. The CEO of the local Ugandan firm also told us that international clients often ask, ‘Uganda is known for tourism, it is not known for software work, so give us a reason why we should actually come to you.’

Many BPO executives told us a number of governance challenges associated with their participation in global production networks (Gereffit et al., 2005), which brand marketing alone cannot overcome. For example, lead firms exercise strict control in the networks for quality (e.g. through business standards certification from the likes of International Organization for Standardization (ISO)) which can prevent new local firms from winning international contracts. These are mostly voluntary, and often costly. Buyers also have specific requirements for certain types of accreditations, including for data security, project management, and quality control (Anwar and Graham, 2019). Lynnette Morris, Chair of the South African Technical Review Committee for South African Business Standards, revealed that some US firms like IBM and Microsoft are already using their own business standards. Some of them cost at least one million ZAR (almost US$75,000) for South African firms to accredit themselves, in order to gain the reputation and trust required for securing international contracts. This point was also made to us by the CEO of the Kenyan Information Technology and Outsourcing Services (KITOS), a trade association for outsourcing and IT firms in Kenya (Interview, pers. comm., November 2016). As a result, some smaller domestic firms struggle to operate profitably in the face of competition from international firms, expensive certification costs for business standards, and their lack of capacity and market reputation to secure contracts. The CEO of one Ugandan firm told us that one of their potential clients in the US once asked them if they have Microsoft-certified developers and software engineers. He went on to explain the kind of conditions clients place on African firms before awarding any contract:

[T]here are so many conditions you have to meet before you can get work from international buyers. Clients ask do you have professional certifications and do you have Microsoft certified developers or engineers. I believe if we overcome this, then we should be able to compete internationally. But at times there are many conditions for us. One of them is your certifications. You must be having the industrial certifications. Then the way you deliver your work. Do you have your project management system, like quality control? Then what is your delivery approach? Do you have servers where you do your development? How are your security controls? The clients would say Why should I come to Africa? The US itself has firms and organisations and also, we know where to outsource, we will go to India.

Interview, pers. comm., Kampala, November 2017

Nigeria faces these reputational challenges more evidently. As the most populous country in Africa, Nigeria has the potential for a BPO market due to its growing population which can act both as a consumer market and also as the supplier of the labour force for the BPO industry. But its image among the business community abroad is a major hinderance to attracting offshore work. Nigeria is known for online scamming and there are also security and terrorism concerns in the northern parts of the country. The CEO of a local BPO firm told us:

In Nigeria offshore work is a gold mine for us in the outsourcing space but that has not taken off. There are snippets of it occurring but not fully . . . There are lots of reasons why. Obviously, the political landscape. In the early days, it used to be technology and connectivity but we have since moved on from that—you know that the technology is available, connectivity is available now. For example, a conference bridge service (through IP PBX)¹⁰ that allows international calls to be routed directly to us from the US and the UK. But I think the biggest challenge is the confidence in the country. Well think of Nigeria as an address. As a service provider or product provider in the UK or in Europe you are thinking of offshoring to a destination somewhere across the world where when you turn on the news you are seeing Boko Haram. And you are thinking, am I making the right decision here. If you just listen to the news people are wary of coming to Nigeria, let alone outsourcing their business to Nigeria. So that for me I think is a big challenge, there is an image challenge.

Interview, pers. comm., Lagos, March 2017

Similarly, one of executives of the Association of Outsourcing Professionals in Nigeria (AOPN) told us that the image and perception of Nigeria among international buyers is of a corruption-ridden country, which ‘is working seriously against us and the government is not helping the matters’ (Interview, pers. comm., Lagos, March 2017). The National Information Technology Development Agency (NITDA), the Nigerian government agency tasked with developing the local IT industry, struggles with its leadership and a lack of a clear policy framework to create a political–economic space in which the outsourcing industry can germinate.

The NITDA officials we interviewed maintained that the government is actively trying to improve the image of the country and letting the business community know that ‘we are not corrupt people. We are not lazy people. We are very disciplined people and we are hardworking people’ (Interview, pers. comm., Abuja, 2017).

According to one of the NITDA’s officials, there are three main challenges that Nigeria’s BPO industry faces: lack of a clear policy framework, poor infrastructure (electricity, transport, internet), and the country’s image. When we conducted our fieldwork in early 2017 the government was holding discussions about a new outsourcing bill that would include several incentives, including tax breaks. The NITDA officials told us that in order to enhance the perception of Nigeria the state needs to set up a dedicated office for outsourcing, step up its marketing campaigns abroad among the business and buyer communities, and demonstrate the suitability of the business environment in the foreign media (Interview, pers. comm., Abuja, 2017). It took the COVID-19 pandemic in 2020 for the NITDA to unveil a new National Outsourcing Strategy for Nigeria 2020–25, which contains some fairly bold claims—such as one million direct and indirect jobs and 500 new outsourcing entrepreneurs employing an average of fifty staff to be created within three years of implementing the strategy (NITDA, 2020). Nigeria already had a National Outsourcing Policy, dating from 2007, which contained less ambitious targets of 10,000 jobs and ‘500 new outsourcing entrepreneurs employing an average of twenty staff within the next three years of implementing the outsourcing programme’ (NITDA, 2007). While the targets are bigger in the new policy, the mechanisms to achieve these aims have remained virtually the same. For example, the ‘special Outsourcing Development Fund’, which was to be established under the 2007 policy, features again in the 2020 policy, suggesting it has not been set up yet during the last thirteen years. We observed similar issues in Kenya.

The Kenyan case is somewhat paradoxical in the sense that it has a slightly better image than Nigeria among international buyers, is noted for its growing IT and tech-hub scene, and has been dubbed the ‘Silicon Savannah’ in the press (Financial Times, 2016). However, its BPO industry did not take off in the way that industry experts, policymakers and development organizations predicted almost a decade ago. Kleibert and Mann (2020) have noted that the Kenyan government did not put in place a proper strategy to develop the BPO sector, and in fact, preferential treatment for certain firms has meant that many inexperienced domestic firms struggled to compete and were crowded out of the market. The CEO of KITOS told us that:

there was a lot of push with the studies by McKinsey that highlighted Kenya’s potential for international BPO work and I think that was what was sold . . . Unfortunately, IT and BPO has not been in the forefront of the government’s priority. Maybe tourism or other things have been very forefront for Kenya. The branding from the government I do not think is very clear on the BPOs and the sector I think is getting smaller. Interview, pers. comm., Nairobi 2016¹¹

Recently, the ICT Authority Kenya, the government body responsible for Kenya’s wider ICT sector, has been very active in promoting various digital economy activities. It has drafted a Strategic Plan 2019–23 ‘to take advantage of the “Fourth Industrial Revolution”’ (Kenya ICT Authority, 2019: 3), although the plan has little to say about the outsourcing industry and its growth. The BPO industry in Kenya (and also Uganda) is now reliant on a handful of firms and new intermediaries (who are foreign entities) for the bulk of its offshoring work, which affects local value creation and capture within the country. The increasing reliance on foreign firms for offshore work and a focus on low value-added activities (e.g. customer support services and data entry) make decoupling very likely.

By the time we did our fieldwork in 2016–17, many foreign firms had already left Kenya, Ghana, and Nigeria. According to one of the managers of a local firm in Kenya, the government’s marketing campaign in 2006 (before the fibre-optic undersea cables landed in Mombasa in 2009), did help bring some work to Kenya, but the government failed to develop the necessary infrastructure (e.g. adequate office spaces, electricity supply) and a skilled labour force (Interview, pers. comm., Nairobi, 2016). As a result, companies which came to Kenya for offshore work left because they could get the work done with a skilled workforce more cheaply elsewhere, for example in the Philippines. We met the owner of a newly established local firm doing entry-level digitization work for the domestic banking and financial sector in Ghana. He told us that previously he used to work for a US firm called Affiliated Computer Services (ACS), which began operations in 2000 in Ghana and left in 2013. Similarly, another US-based firm, Teletec, left Ghana in 2014. Because outsourcing value chains are cost driven, buyers often shift the risks to vendors (i.e. the company providing or selling the service), including infrastructure and sunk costs, making it easier for the buyer firms to relocate to other places if the cost goes up or service quality declines, leaving local suppliers struggling to stay in the market.

The local firms in most of our case study countries lacked managerial experience, technological know-how, personal connections, and some were facing financial and operational difficulties after the 2008 financial crisis, despite a recent reduction in telecommunications costs and improvements in internet connectivity. For example, Kencall, the winner of the Best Non-European Call Centre at the Call Centre Focus Conference in 2008, was facing liquidation due to unpaid debts. We visited their call centre just off the Mombasa Road on the southern outskirts of Nairobi only to find it closed.¹² The CEO of a small IT services firm which also runs call centre operations in Nairobi told us that finance has been their main challenge to expanding their operations, to undertake complex high value-added work, and attract offshore work (Interview, pers. comm., Nairobi, 2016). At the time we did our interviews the company was primarily doing transcription work subcontracted by other bigger local and regional buyers. In South Africa, there were also indications of domestic firms shutting down operations due to foreign competition and governance challenges (e.g. maintaining business standards and data security in the value chain, see Anwar and Graham, 2019). The development implications of this uneven landscape of outsourcing are briefly explored next.

Data on the BPO industry on the continent remains sketchy at best. Most countries in Africa do not have good-quality official data, and estimates derived from reports by development organizations and management consultancies remain scattered and outdated, which makes comparable analysis even more difficult. We therefore attempted to collect estimates on the size of the wider IT or IT-enabled services (ITES) industry in our case study countries, since some official figures on the ITES industry also include the BPO segment (Table 3.2). The data come from a range of sources, including interviews with private-sector associations and secondary sources (reports and publications), and physical trips to firms’ offices. In 2014, the IT sector made up around 2.7 per cent of the gross domestic product of South Africa, slightly higher than the agriculture sector but less than the tourism sector (Statistics South Africa, 2017). Figures for Nigeria and Uganda remain unavailable. We further collected information on the number of firms doing BPO work in our case studies, along with Egypt and Mauritius, to understand the size of the market on a larger regional scale. In Kenya, we were told by an official at KITOS that they had a total of sixty members, of whom ten were doing entry-level BPO work, primarily data entry, transcription, inbound customer support, and sales, while they had no record in their database of any firm doing high-value work such as knowledge process outsourcing. In Uganda, a newly formed private-sector association, ATIS, has twenty-four registered firms (Interview, pers. comm., ATIS official, 2017). Comparable figures for Ghana and Nigeria are more difficult to ascertain. During our fieldwork, we found only twelve companies doing entry-level BPO work in Accra.¹³ In Nigeria, we interviewed officials from the Association of Outsourcing Professionals Nigeria (AOPN) and the Nigeria Association of Information Technology Enabled Outsourcing Companies (NAITEOC). While they connected us to a few local industry players, neither organization had estimates of the BPO industry. We triangulated web-based research and field visits to the known firms’ offices in Lagos and Abuja in 2017 to build a database of just fourteen BPO firms.¹⁴

There is no denying the fact that the improvement in internet connectivity in Africa since 2009 has increasingly made some of its economies competitive in wider global outsourcing value chains (Anwar and Graham, 2019; Kleibert and Mann, 2020; Mann and Graham, 2016). Other elements have also contributed to the development of the BPO industry in several parts of Africa. One in particular has been the trend in the Indian BPO industry of moving away from entry-level work (generally classified as low value-added) into high-value and skilled professional services including complex IT-enabled services. India has some of the world’s best-known and most valuable IT and software development companies in the world.¹⁵ Many Indian IT companies owe their success to outsourced business process functions for clients or corporations in high-income countries like the US and UK (Peck, 2017; also Parthasarathy, 2004; Parthasarathy and Aoyama, 2006). More recently, the national industry body, the National Association of Software and Services Companies (NASSCOM), has started using the term Business Process Management (BPM) in its reports and various other outputs as they reorient their value proposition from low-cost locations to domain expertise and high value-added offerings such as data science and artificial intelligence capabilities (see NASSCOM, 2018).¹⁶

This shift has opened up the space for some of the low-value work to be relocated and re-outsourced to other locations which are cost effective. This point was emphasized by a Ugandan CEO, who told us that ‘the Indian BPO industry is looking to sell themselves for higher-end professional services such as the web support services and cloud computing. The cost of doing bottom of the value chain work such as inbound customer service, transcription and digitization is increasing in India which means many African countries can fill the gap as our costs are fairly low’ (Interview, pers. comm., Kampala, November 2017). That said, without accurate data it is hard to ascertain how much BPO work is being redirected from India to Africa. It is clear, however, that low value-added activities such as inbound customer services, digitization, transcription, and content generation have emerged in our sample—the bulk of these services being primarily targeted at domestic markets. That said, in South Africa, some firms were doing various forms of high value-added work (see Anwar and Graham, 2019: 213–214).

The growth of domestic market-oriented BPOs we see in our sample is also reflective of services demand coming from mobile operators such as MTN, Vodacom, Safaricom, and Airtel, who have a growing subscriber base. For example, in South Africa, telecom operators are the biggest segment buying entry-level services, both from foreign firms and domestic firms, such as customer and technical support (BPESA, 2018). Spurred on by improvements in digital infrastructure technologies, most notably broadband fibre-optic cables, the potential for low value-added BPO work to be done on the continent remains. However, digital economy activities can also lead to new forms of intermediation with implications for economic development (through value capture) (Murphy et al., 2014).

The contraction of the local BPO industry in Kenya and Uganda has created an economic space for re-intermediation by new types of players, which we term ‘third party coupling’, through US-based intermediaries. These intermediaries are not lead firms as in direct buyers of services. Instead, they source work from Silicon Valley corporations and then subcontract that work to local firms in Africa. In essence, they are becoming the ‘bridge’ between Silicon Valley and the ‘Silicon Savannah’.

From the Kenyan perspective, Mann and Graham (2016) noted that trust was a major issue facing local firms in sourcing offshore work directly, as well as the power of incumbents. Indeed, many local industry players became reliant on new foreign intermediaries. One of these US-based intermediaries started with a 70-seater centre doing entry-level data annotation and image tagging work (also referred as machine learning and artificial intelligence work) for Silicon Valley corporations. Their initial business model was based on subcontracting services delivery to local firms in Kenya. They now have offices in both Kenya and Uganda. In 2016, when we visited their offices, they had expanded their inhouse delivery centres, and do limited subcontracting to local firms. At the time of our fieldwork, this intermediary had capacity for 600 employees in a 200-seater office space in Nairobi, with three eight-hour shifts a day (Interview, pers. comm., business manager, Nairobi, 2016).

Since these intermediaries have the significant advantage of a base in the US, they control a sizeable portion of the entry-level work coming into the East African region. Overall, these intermediaries provide much-needed work for local firms to stay afloat in the market. One of the managers of a local Kenyan firm told us that the industry has not grown and the demand for work has slowed down, however, they have survived because of their dependence on this international intermediary, which provides them with a majority of their business contracts (generally short-term) doing transcription work (Interview, pers. comm., Nairobi, 2016). That said, some local firms had taken a direct approach to bringing offshore work into their respective countries, given that intermediaries can significantly affect the value capture and upgrading opportunities for regional economies.

The director of the BPO segment of one of the big regional IT firms, with an extensive presence in East Africa, told us that international intermediaries were important for them to break entry barriers into these value chains, though this intermediation affected the value that they were able to capture. In 2014, they set up offices in California and approached clients directly, bypassing the intermediary, and now have long-term contracts with two of the biggest Fortune 500 companies, doing data capture and data processing.¹⁷ The director also told us these intermediaries were effectively charging a huge mark-up to take a large cut of the profit, so when they approached Silicon Valley clients directly with their pricing for similar work, the clients were surprised and decided to buy those services directly from them (Interview, pers. comm., Nairobi, 2016). Similarly, in Nigeria, two local firms told us that they had succeeded in securing small contracts from US-based clients. According to the chief operations officer of one of these firms, the buyer was unhappy with the insurance sales service they were getting from a company in India, so they outsourced a portion of that work to Nigeria, despite slightly higher cost per agent in the country (Interview, pers. comm., Abuja, 2017). In essence, while enhancing digital footprints (e.g. through online marketing and improving websites) can integrate firms into global networks, physical presence in buyer locations is still a significant factor for firms to be able to win contracts for offshore work.

There are strong indications of regionally oriented value chains emerging in outsourcing in Africa (Kleibert and Mann, 2020). In the BPO sector, where foreign-owned firms have shown less enthusiasm for engaging in the local or regional markets in our sample, Indian-diasporic businesses are filling the gap. The Indian diaspora has long been present in Africa and is received with a mixture of friendliness and hostility in local socio-political circles. Some observers have called Indian entrepreneurs in East Africa ‘settled strangers’ (Oonk, 2013), while in West Africa, the colloquial term used is ‘briefcase businessmen’ (according to an Indian immigrant we interviewed in Nigeria). Some of the well-known BPO firms run by Indian diasporic businesses in Africa are ISON, Technobrain, Acreaty, and Simbatech.¹⁸ Another big IT firm, Tech Mahindra, which is a subsidiary of the Indian conglomerate Mahindra Group, has offices in thirteen African countries and serves primarily local markets. According to the regional head at Tech Mahindra, their company deals mostly in call centre operations for telecom services providers such as MTN and Airtel (two of the major networks in Africa) (Interview, pers. comm., Lagos, March 2017). Similarly, a high-ranking ISON executive told us that if telecom operators in sub-Saharan Africa outsourced customer services, 90 per cent went to ISON (Interview, pers. comm., Lagos, March 2017).¹⁹

These regionally oriented players and networks have the potential for wider industry-level upgrading into complex value chain activities, for example data management and legal process outsourcing. The shifting of local firms’ priorities towards domestic and regional value chains is also reflective of a trend towards regionally oriented value chains as a pathway to economic development (Morris et al., 2016). The expansion of regional value chains can generate South–South trade, which is considered more important for high-value capture among African firms than North–South value chains and networks (see Franssen, 2019). However, it should be pointed out here that in the value chain literature, there is a recognition that the risk in value chains shifts towards lower-end suppliers (see Selwyn, 2018), as has been the case for supermarket value chains in Southern Africa, where supermarkets’ demand for private standards and requirements affects upgrading among the lower-end suppliers (Nair et al., 2018). While the potential for local firms to upgrade into complex value chains and high value-added activities remains to be seen in Africa, the employment generation potential through digital outsourcing is evident.

In the context of high levels of working poverty across the continent, employment within the newly emerging outsourcing industry is seen as a critical economic development on the continent. Grand claims are often made by governments about job creation through the digital economy, with BPO work predicted to be a provider of formal-sector jobs to the African population (Ndemo, n.d.). Similarly, the World Bank’s 2016 World Development Report noted the potential of digital technologies in creating jobs in low- and middle-income countries, both through internet-enabled offshoring (e.g. call centre work) and online work (World Bank, 2016). The Rockefeller Foundation’s Digital Jobs Africa initiative, a US$100 million project, was designed ‘to create new jobs’ and employment opportunities in the IT sector, including BPO and online outsourcing jobs.

There is some evidence that jobs are emerging on the continent. South Africa’s BPO industry, for example, supports around 236,000 jobs, yet this represents only 0.6 per cent of South Africa’s total labour force. However, data on employment in the outsourcing industry is sketchy for our other case countries. In Kenya, a country which was once regarded as best placed to tap into outsourcing value chains, data on job creation are often inconsistent and speculative. The Kenyan government expected Konza City (a US$14 billion project conceived in 2008) to create 200,000 high value IT-enabled services and entry-level BPO jobs in the zone (Ncube and Ondiege, 2013). Yet, expectations from projects like Konza City to kick-start digital outsourcing services in Kenya remains far-fetched (see Saraswati, 2014). In fact, at the time of writing in 2021, Konza City is still under construction.

Remote work has gone mainstream and the genie is not going back into the bottle.

CEO of Upwork, quoted in Consumer News and Business Channel (CNBC, 2020)

We have already considered the emergence of the gig economy in Chapter 2. Such has been the uptake of gig economy platforms around the world that various terms such as ‘gigification’ (Veen et al., 2019) and ‘platformization’ (Casilli and Posada, 2019) are now commonly used to underscore their importance for the world economy. There is no doubting the potential of gig economy platforms to provide new services (Srnicek, 2016). Platforms are implicated in moving jobs from one part of the world to another (Graham and Anwar, 2019; Kässi and Lehdonvirta, 2018), commodifying work (Anwar and Graham, 2020a; Rani and Furrer, 2020; Wood et al., 2019), and expanding contingent employment relations around the world (Morgan and Nelligan, 2018; Panteli et al., 2020; Todolí-Signes, 2017). Put simply, gig economy platforms have unleashed tremendous change in work and employment relations around the world (Anwar and Graham, 2020a; Berg et al., 2018; Prassl, 2018). With the onset of the COVID-19 pandemic, remote work has gained even more traction, as the quote from the CEO of Upwork above illustrates (CNBC, 2020). Latest estimates by Kässi et al., (2021) suggest that there are 165 million registered worker profiles globally on 162 digital labour platforms for remote work.

In the context of Africa, Insight2impact in their database²⁰ list ninety-one freelancing platforms that connect buyers and sellers of remote work in Africa. African workers use a variety of major platforms, including Upwork, Fiverr, Freelancer.com, PeopleperHour, and Figure Eight (previously CrowdFlower). Other platforms are less well-known but nevertheless still widely used by African workers for niche work such as writing, including Uvocorp, iWriter, and Gotranscript. Indeed, workers usually register profiles on multiple platforms. To examine the geographies of remote work, we studied Upwork, the world’s largest platform in terms of registered workers, and the most commonly used by African workers.

To understand the economic geographies of remote gig work in Africa, we used Upwork’s filter menu (Figure 3.2).²¹ This allows users to select freelancers based on their locations, hourly rate, skills set, number of hours worked, and income earned. While the filter menu provides enough quantitative information to get some sense of the size of remote work and the relative market shares of countries, the filter mechanism is nevertheless a ‘black box’; that is, Upwork doesn’t publish details on how its back-end data feeds into the information returned by the search interface. On their website, Upwork claims to have 12 million workers with 2,900 skills.²² However, on 22 May 2019, when the screenshot was taken (Figure 3.2), Upwork had 2.3 million searchable worker profiles. Upwork also cancels certain workers’ accounts and deletes their profiles, which essentially means workers can easily be excluded from these platforms (Figure 3.3).

Another difficulty when trying to estimate the number of workers on Upwork is that interactions between different filters considerably affect the aggregated numbers reported through the search form, requiring cautious interpretation. For example, workers who have worked at least one hour will almost certainly have earned at least one dollar, but there are also workers listed who have earned at least one dollar, but worked zero hours. Those workers might have been hired on piece-meal contracts. Additionally, the numbers are aggregated into large groups. For instance, it is only possible to select freelancers who have earned at least one dollar, 100 dollars, 1,000 dollars, or 10,000 dollars. However, despite these caveats, the filter menu provides a means to assess the size and composition of the labour supply on Upwork in Africa, and worldwide.

To assess the relevance of different online labour platforms in African countries we also collected data on aggregate monthly clicks from SimilarWeb.com, a web-traffic monitoring website, between May 2018 and November 2018. The tool provides web analytics, including insights into the number of visitors to a web page, and the country from where these visits took place (using visitor IP addresses). The SimilarWeb-API allowed us to automate the collection of aggregated visitor statistics for our five case study countries.

Upwork is one of the leaders in the remote gig economy, with roughly one third of the overall market share in Africa, as measured by the monthly visits to the website. Figure 3.4 shows the relative market shares of the five most important digital work platforms as measured by the aggregated monthly clicks in the six months between May 2018 and November 2018 (data source: SimilarWeb). The African online labour market is dominated by the global market leaders Upwork and Fiverr, with Upwork being the largest or second largest platform in all African countries. Overall, the market share of Upwork is 36 per cent (similar to Fiverr’s market share), but it is even higher in a number of countries such as Egypt and Kenya, two of the largest online labour suppliers from Africa. Freelancer.com also has a substantial market share in some countries, while Peopleperhour and Guru have only a tiny share of the African market.

To get a comprehensive perspective on the overall supply and demand trends of remote work worldwide, we collected data from Upwork.com by setting up a web-scraping tool based in Python to access Upwork’s main search interface. To do so, we used the ‘Selenium’ package, a web-based automation tool that simulates a human user opening a web browser and clicking through web pages. This scraping tool allowed us to automatically apply different filters to Upwork’s search interface, such as ‘country’, minimum number of ‘hours billed’, and ‘amount earned’. Each filter results in different numbers of freelancers being returned. We therefore performed several hundred search requests to obtain detailed figures on supply, demand, and hourly rates for all the African countries we were interested in. Given the need for relatively long sleeping intervals in the scraping tool between requests (in order to avoid CAPTCHA controls), we needed several weeks to complete the data collection. During that time, overall numbers of supply and demand on the platform appeared largely constant. We noted slight fluctuations on a daily basis, but these did not seem to be systematic in any direction. To estimate the demand for online labour from our five case study countries we counted the number of open projects listed on Upwork. Both the global supply of workers on Upwork and the demand for their labour are highly concentrated (Figure 3.5). Much of the labour demand comes from high-income countries, primarily the US, UK, Germany, Canada, and Australia. With 33,000 projects listed on Upwork, out of 70,000 projects listed at the time of data collection, the US accounts for around half the global demand for online labour.

The labour supply is also concentrated in a few countries, and low- and middle-income countries are some of the largest suppliers of this digital labour. The three largest labour suppliers—namely, the US, India, and the Philippines—were home to more than 1.13 million registered workers, or close to half of the total registered workers on Upwork, at the time of data collection (Table 3.3). Indeed, these three countries represent the forefront of global outsourcing value chains: the US is the leader in terms of demand for outsourced services, and India and the Philippines together account for well over half of the global market share in terms of global services delivery (Beerepoot et al., 2017; Peck, 2017). These three countries therefore enjoy a network effect and reputational advantage in online labour markets. With just over 120,000 registered freelancers on Upwork, the African continent as a whole has a global market share of just 5.2 per cent, which is less than the Philippines (Table 3.3).

Regionally, there is some unevenness in the labour supply on Upwork. While countries on the Asian continent and most Latin American countries are major labour suppliers, almost all African countries have a relatively low labour supply, with most being in the lowest quartile of global distribution in terms of labour supply on Upwork. Two of the key factors influencing labour supply from the continent are the relatively low penetration of ICT tools, and costly broadband services. In fact, ITU (2019) has noted that internet penetration in Africa, despite a recent surge in mobile phone adoption, is still the lowest in the world, with affordability and lack of digital skills among the population representing key barriers to uptake. Although there are some significant labour suppliers on Upwork such as Egypt, Kenya, South Africa, Nigeria, and Morocco, the other African countries either have small workforces far below the global median, or are absent altogether (Figure 3.6).

Despite these large differences in labour supply, most African countries show a similar composition of remote labour in terms of job categories. Figure 3.7 shows the composition of online jobs within each African region and for the market leaders (i.e. the US, India, and the UK) on Upwork. In contrast to India, most African countries provide fewer workers in ‘IT’ than in ‘Admin support’. In the largest labour-supplying countries, most jobs are conducted in the ‘writing and creative’ category. Such jobs include translation and blog writing. The second largest category is ‘Admin support’, including data entry, virtual assistance, and transcription. The other main categories are ‘IT’ including software development or analytics, and ‘Legal and consulting’ including tasks like accounting, paralegal jobs, and business plan writing.

If we expand our analysis and include more platforms, the regional and global outlook stays roughly the same. For example, the Online Labour Index (OLI) (Kässi and Lehdonvirta, 2018) measures the supply and demand of online labour across countries and occupations by tracking the number of projects and tasks across a number of platforms in real time. Though limited to only English language platforms, it tracks 70 per cent of the market by traffic. In June 2020, the US was the leader in terms of the employer market share on the major platforms around the world, followed by India and the UK (Figure 3.8). Worker distribution is also concentrated, with Asia contributing more than 65 per cent of the global labour force (Figure 3.9), and just three countries, namely India, Bangladesh, and Pakistan, contributing more than 50 per cent of the world’s remote labour force (Figure 3.10).

All this said, the COVID-19 pandemic will undoubtedly affect the world of work in the near future. The pandemic and the lockdown imposed by countries around the world has forced employers and workers to shift towards home-based working,²³ and a few large enterprises around the world are already advocating remote working to be the standard way of organizing work in the future (UN News, 2020).²⁴ Furthermore, the pandemic has affected businesses and threatened job losses around the world, with the ILO estimating that nearly half of the world’s workforce is at risk (ILO, 2020a). In the US alone, an estimated 42.6 million workers had applied for unemployment benefits by June 2020 (Lambert, 2020). Many of these workers could look for online work to earn a living. This could mean changes to some of these geographies of remote work.

The imbalances in the supply of remote digital labour we have discussed above translate into an even more skewed global distribution of income and wages. Figure 3.11 compares the market shares of the different world regions in two groups: (a) amount earned, and (b) hourly rate. Africa, Latin America, and the Middle-East together account for less than 15 per cent of the overall market share; and account for even less in the high-income bracket (a) and high hourly rate groups (b). Africa’s market share is low overall, and it is basically absent in the high income groups (a) and high hourly rates (b).

From the above observations, we can safely say that there is excess supply of online labour globally (Graham and Anwar, 2019). This can significantly affect the bargaining power of workers on platforms (Anwar and Graham, 2020a; Graham et al., 2017b). Second, most workers who earned more than US$1,000 have come from India, Pakistan, the Philippines, the Ukraine, and Russia. In particular, India and the Philippines have the largest market shares in the high-income bracket (that is, US$1,000+ and US$10,000+) (Figure 3.11a). That said, while workers from these countries perform the majority of online tasks, they earn low wages per hour. The high hourly rate groups (US$30+ and US$60+ per hour) are dominated by workers from high-income countries, such as the US (Figure 3.11b). The issue of bias among employers towards workers of the same nationality is also well known in the gig economy (Ghani et al., 2014 and World Bank, 2016). Additionally, people from low- and middle-income regions—in particular Africa—are even more marginalized in the remote gig economy. Low overall market shares are coupled with low incomes and low wages.

The gig economy is also structured along gender lines, with a varied distribution of men and women globally. The Online Labour Observatory estimates that female workers make up about 39% of the labour force in the remote gig economy (Stephany et al., 2021). A 2018 study by the ILO covering 3,500 gig workers in seventy-five countries found that a third of those were women (Berg et al., 2018). More men are also likely to participate in the gig economy than women in the US (Farrell et al., 2018) and Australia (Churchill and Craig, 2019). In Europe more men take part in the gig economy than women, with the sole exception of Italy (Huws et al., 2017).However, a study done by Research ICT Africa across seven countries found that more women than men participated in remote work from Ghana, Kenya, Nigeria, and Tanzania, despite a significant gender divide in terms of internet access (Research ICT Africa, 2017).

We also scraped data of South African remote workers [3500 profiles or 25% workers] by total income earned on Upwork in November 2018 (Figure 3.12). We applied facial-recognition software on the workers’ profile images to get estimates of users’ gender. Details on the face-recognition algorithm can be found in Dehghan et al. (2017). The authors report high accuracy of their algorithm through having trained it on more than four million images from more than 40,000 individuals. The algorithm also reports confidence scores which help the user to assess the accuracy of the estimate. If the image is, for instance, of low resolution, the confidence score might be lower. The results do come with some uncertainty. To assess the uncertainty involved, the software reports probabilistic estimates of gender of the individuals on the profile images. In our case, the algorithm identifies the users’ gender with high average confidence: the algorithm identified users as ‘male’ with a mean confidence of 94.8%, ‘female’ users with a confidence of 94%; for users identified as ‘White’ the mean confidence is 98.4%, for ‘Black’ users it is 91.1%, and it is 83.9% for ‘Asian’ users. Overall, the software provides outstanding results in image recognition. While this is not the ultimate method to identify someone’s gender, it gives us some indication of gig economy’s socio-economic distribution. Overall, we found slightly more male users (54 per cent) than female ones. But there are more women in the subset of users who successfully conducted projects on the platform: 54 per cent within the US$1+ and US$100+ groups, 52 per cent in the US$1,000+ group, and 56 per cent in the US$10,000+ group.

To gain insights into the racial composition of the gig economy workforce in Africa, we analysed South African workers’ profiles in the high-income bracket on Upwork manually to understand who succeeded in earning a higher income. At the time of analysis in June 2019, ninety-one of the ninety-nine workers from South Africa, who had earned more than US$10,000 and worked over 1,000 hours, were white South Africans (person with lighter skin tone). Access to technology is lowest among the low-income groups or poor (Research ICT Africa, 2018), who often happen to be Black South Africans (persons with darker skin tones) (Statistics South Africa, 2014). Hence, they hardly succeed in earning high-income on platforms.

Contrastingly, White South Africans have better access to education and better material conditions (resulting from the institutional and political manipulations of the colonial and the apartheid regimes) than Black South Africans, who So, our categorization of remote workers in South Africa based on their racial categories is to highlight how race as an idea or a social construct and its political–economic application has inter-generational impacts in society and on inequality (also see Box 3.1). The point here is that the current socio-economic imbalances in South Africa are amplified in the new gig economy as well.

This brings us to an important question. How do these digital work activities affect African workers’ lives and livelihoods? In Chapter 4, we will discuss job quality issues in the remote work and call centre work that we have discussed above.

This study was part of the bigger GeoNet project funded by the European Research Council 2013, ERC-2013-StG335716—GeoNet (see Graham et al., 2017a).