The race to digitize commerce in sub-Saharan Africa
How Jumia and Facebook are competing, even though they’re playing different games
Tiffany* sells makeup, and is evaluating a new display shelf where her product would be at eye level with passing shoppers. That visibility could lead to new customers, but she’s not convinced it’s worth the trouble, because almost all of her customers currently find her on Instagram. And actually, that shelf isn’t even in her store — it’s in the front of a retail locker space where she rents a container in the back to hold her inventory. Tiffany keeps some of her most popular products on hand here so that customers can easily drop by and see the merchandise, but otherwise she runs a completely virtual, just-in-time business. And paying for a physical footprint, even if it’s only a few square feet of shelf space, might not give her the return she needs on scarce capital.
Tiffany is among millions of entrepreneurs in sub-Saharan Africa taking advantage of popular digital tools to transform their business. Formal marketplaces such as Jumia or Takealot have successfully established a traditional model of e-commerce, while new platforms from Paystack, Flutterwave, and others provide the underlying tools for independent sellers. Despite these formal offerings, the vast majority of digitally connected small businesses are using Facebook, WhatsApp, and other consumer platforms to fulfill an increasing number of commercial use cases. This “social commerce” is often heralded as a leveler of the playing field, as even the smallest businesses can typically use social platforms for a low cost.
To help make sense of this mix of formal and informal, of purpose-built and appropriated, I developed a framework for comparing the range of digital solutions available to today’s entrepreneurs. For those with a commercial interest in this sector, the framework offers a lens for understanding how different solutions solve distinct parts of the value chain, and what this means for the likely addressable market of each solution. For policymakers and those interested in inclusive economic growth, the analysis highlights the implications of fundamentally different operational structures between formal ecommerce and social platforms, and why we should worry about how previously informal economic activity becomes formalized.
From analog to digital
Small businesses in sub-Saharan Africa looking to sell online have multiple options to consider; which I’ve simplified here into four categories. These are by no means mutually exclusive, but for clarity I have separated them in the analysis.
The most accessible and likely approach for most small businesses is to use those consumer products that are already familiar and ubiquitous, especially Facebook, WhatsApp, YouTube, and Instagram. Entrepreneurs can create a page on Facebook and/or Instagram, post photos or videos of products, communicate with prospective customers, respond to customer reviews, and use built-in advertising tools to boost her page or specific product posts.
Businesses may also sell on formal marketplaces such as Jumia, Kilimall, or Konga. These online marketplaces aggregate supply and demand into a single platform, providing sellers with immediate access to a large audience — for a price. Like Amazon, they offer sales support with payment processing, fulfillment, and delivery, as well as back office business tools such as inventory management and analytics; typically they charge a commission on sales.
More digitally savvy businesses might go a step further and publish their own standalone website using one of a growing number of e-commerce tools, such as Paystack Commerce or Flutterwave Store, two new solutions from Nigerian fintechs that combine payment processing and back office functions with a website builder into a single platform (essentially, Shopify for Africa).
And another option for FMCG (fast-moving consumer goods) retailers are B2B platforms that connect businesses with suppliers and financing, such as Trade Depot in Nigeria or Sokowatch in East Africa, both of which aggregate demand from retailers to provide discounted wholesale pricing and deliver consumer products to the retailer.
The retail value chain
To understand how these digital solutions are reshaping retail in sub-Saharan Africa, it’s helpful to examine the business functions that each solves for. Retail businesses like Tiffany’s have a relatively straightforward value chain: retailers buy goods from upstream suppliers (inbound logistics), manage inventory and resources (operations), engage with customers (marketing), and, when they’re lucky, make sales and deliver product (sales/outbound logistics). Looking at each category of the value chain in turn, we can summarize the benefits different digital solutions provide to offline businesses looking to digitize.
Figure 1. Where different digitization solutions impact the retail value chain
The B2B marketplaces or supply chain solutions that service this part of the value chain are primarily trying to secure better pricing for retailers by aggregating demand. For example, Sokowatch in East Africa promises not only lower pricing but also financing and complete last-mile delivery to its retailers. These solutions are currently limited to FMCG, where high sales volume and commodity products make it easier to aggregate orders.
All of the e-commerce marketplaces provide warehousing services to eliminate the need for physical storage facilities. This is especially convenient if the retailer also pays the platform to handle fulfillment and delivery, but these services eat into margins, and many smaller businesses prefer to handle warehousing and fulfillment themselves. Businesses that sell on a marketplace are often also selling directly to customers via physical shops or social media, so the platform’s inventory tracking systems will only cover a portion of their business, reducing its utility.
This is where digital options far outperform anything offline. A small retailer with no online presence can only market to those prospects who pass in physical proximity of the shop. Although advertising costs are low, the maximum size of the funnel is limited. But once the business starts marketing online, they can easily increase the reach of their advertising by many orders of magnitude. This is especially true of social media, where the smallest entrepreneur can put up a Facebook page and start bootstrapping eyeballs via their personal networks before transitioning to paid ad spend to “boost” their page or posts.
Marketplace platforms also offer substantial reach given the installed base of customers who use the site, and sellers pay to promote their products higher in the search or browsing results. One could argue that advertising dollars spent on an e-commerce platform should outperform those on social media, as customers on the former are more likely to be closer to a purchase decision, i.e. deeper in the sales funnel. But marketplace platforms can be brutally competitive for many product categories, and the structure of the market doesn’t allow for much retailer branding or other differentiation. You live and die by the transaction, with little opportunity to build a brand and a following — which is exactly what social media excel at.
Because standalone domains exist on the open web, they rely on driving traffic from other sources — search and social — to the domain, and therefore don’t benefit from any installed base of buyers.
Sales/ Outbound logistics
One of the bigger service offerings from both the marketplace platforms and the independent domain platforms, like Paystack Commerce, is the ability to process different kinds of payments. For some product categories and businesses this is important, but the majority of e-commerce payments in sub-Saharan Africa are still COD (cash-on-delivery). And in markets where mobile money is popular, buyers can make non-cash payments (pre- or post-delivery) directly into mobile money wallets. So the payment processing function is most valuable only for customers who are willing to pay upfront and want to pay with a card or bank transfer.
Last-mile delivery to the customer is infamously challenging in most markets, and one of the key services that marketplace platforms offer to their sellers. But typically the delivery price is passed on to the buyer, not built into the product price. Many businesses work with local couriers or contract with new logistics platforms like Sendy in East Africa; many retailers let the customer choose a preferred logistics provider. Because most small retailers have less expensive options for delivery, paying the marketplace to handle delivery only makes sense if the retailer has sufficient volume such that the overhead of managing independent deliveries outweighs the cost savings.
Follow me for top-line growth
One helpful way to think about how each solution solves a different business function is to categorize it as impacting either (a) top-line growth or (b) bottom-line profitability. The latter category includes solutions that streamline procurement; process payments; digitize inventory management; or provide warehousing, fulfillment, and delivery services. Efficiencies in any of these can lower a retailer’s operational costs, but these will be incremental gains, especially for the smallest retailers.
On the other hand, digital advertising channels enable entrepreneurs to grow their business by reaching many more customers than they would otherwise. Social media platforms in particular allow businesses to substantially and cost-effectively increase their potential customer base, thereby scaling top-line growth and revenue.
In the language of unit economics, this is about very large reductions in CAC (customer acquisition costs) vs. incremental improvements to COGS (cost of goods sold). Obviously both are crucial metrics to a business, but have different implications at different stages of business maturity. For any business, reducing COGS and thereby improving margins, by say 10% can easily be the difference between break-even and profitability. But for the smallest businesses, where sales numbers are low, improving margins by 10% may not actually translate into anything meaningful in terms of absolute value. On the other hand, if a small business can double or triple sales at a lower CAC, that revenue opens up all kinds of opportunities for staffing up or expanding the business in different ways.
Using the framing of the value chain analysis above, here’s a simple 2x2 illustrating which types of businesses are most likely to benefit from each type of digital solution. In addition to the maturity or size dimension, which as described above is a primary factor in determining the relative value of increasing reach and top-line sales vs. operational efficiencies, it plots average sales price on the other axis to serve as a proxy for the types of products and income level of the customers buying them.
Figure 2. The sweet spot for each type of solution
Social commerce occupies the lower left because it’s essentially free to use (minus paid advertising) and provides the best opportunity for growing the customer base with a lower CAC, the two most important factors for the smallest businesses. Social commerce likely has a ceiling in terms of average sale price, partly because of trust (i.e., consumers tend to place more trust in retailers on formal marketplace platforms, which may have buyer protection programs), but also as a function of how prospective buyers are most likely to search for and purchase products: Higher-income consumers are more likely to use a laptop and web-based search engines to find products, making them more likely to shop from marketplaces and standalone website domains. Lower-income consumers are more likely to only have a mobile device and spend more time within the confines of social media feeds, where social commerce reigns.
Ecommerce platforms like Jumia and Konga are used as sales channels by all kinds of businesses, but are most helpful to those that have grown to have multiple employees and higher sales volumes, where the operational and logistics services the platforms offer will have a greater impact on efficiency and overall profitability. Because of the additional formality and high transaction fees — Jumia sellers pay a sales commission of 10%-20% depending on the product category — vendors are less likely to focus on really low-value items, keeping the average sales prices higher than what you see with social media businesses.
Standalone domains supported by e-commerce platforms like Paystack Commerce are most appropriate for the largest businesses with the highest average sales price. Such firms already have established customer bases and are looking to control more of their brand without giving up margin to a marketplace. Because they exist on the open web, they have to drive traffic from search and social, which typically means a higher CAC and thus requires higher average sales prices.
Finally, B2B sourcing platforms have a hard ceiling on average sale price given that (at least in current incarnations) they are limited to FMCG and other high-volume, commodity products. The economics of bulk ordering and delivery costs mean that smaller shops with lower turnover are less likely to enroll given the expected cost benefits.
The future will be formalized
A natural question to ask is to what extent online commercial activity will migrate away from the social commerce model and toward the more formal paradigms of e-commerce marketplaces and standalone domains that are dominant in the West. I would argue that while the business logic that has the vast majority of African businesses choosing informal activity on social platforms won’t change anytime soon, those social platforms will increasingly push that economic activity into more structured and legible forms.
For the most part, social commerce businesses are appropriating tools designed for personal communication to use for business purposes; as such, the platforms don’t have good ways of knowing about these transactions being coordinated through their ecosystem. WhatsApp Pay and the new Facebook Shops are obvious steps toward improving the platforms’ visibility into commercial activity, and much more will come. The goal isn’t necessarily to compete head-to-head with e-commerce platforms with a fully integrated offering, but to capture enough of the activity to develop a completely new purchase behavior dataset that can be monetized via their ad engines. Imagine advertisers being able to target their ads on Facebook to those users who have “purchased a health and wellness product for $10-$20 within the last 30 days from a seller with 5,000–25,000 followers.”
The result will be an asymmetric, top-down vs. bottom-up competition between the formal marketplaces and social media: Formal marketplaces will try to broaden their value proposition to reach lower to the smallest-scale businesses and spin the network effects flywheel, while social media will attempt to formalize its offerings, not to compete directly as a solution provider, but to develop one of the most comprehensive and valuable datasets on consumer economic activity.
The competition is asymmetric in that the two platform types have fundamentally different business models. Social media platforms, and for all intents and purposes this means Facebook, are ad-based; they make activity free in order to accumulate eyeball-hours. The marketplaces rely on the transactional revenue of the commissions they charge sellers on each sale. Put another way, the marketplaces have to make the unit economics work for each transaction on the platform, which Jumia has estimated to require something like a 10%-20% margin, a significant cost that is prohibitive to many smaller retailers. Facebook, on the other hand, can afford to offer (a more limited set of) services for free because the resulting activity will substantially increase the value of its ad inventory. That data and resulting value may take a while to accumulate, but as I’ve argued previously, Facebook can use its tremendous cash reserves from its high-ARPU user base in the West to cross-subsidize unprofitable initiatives in the Global South for as long as it needs. In any war of attrition, Facebook doesn’t lose.
This disconnect between Facebook’s revenue model and its consumer-facing services presents challenges not only for regulation (e.g., antitrust policy based on consumer welfare in the form of pricing), but also assessments of its net benefit to the resource-constrained populations who rely on it. As with Free Basics, its subsidized internet access initiative, Facebook’s commerce offerings provide very real value to millions of individuals with few truly affordable alternatives. Yet the long-term implications of allowing it to continue to envelop and become the de facto provider of adjacent sectors of the digital economy are immense.
In the context of social commerce, that risk may be under-appreciated, given that most of the commercial activity moving onto the platform was informal and unreported to begin with. The sheer size of the informal sector — an estimated 86% of employment in Africa — makes it a key focus for governments grappling with economic growth policies. And digitized network technology is key in state efforts to make legible the invisible, as large-scale identification and currency initiatives have shown. Yet as more informal trade moves onto and is structured by social media, we may face a future in which the most complete and accurate source of informal economic activity is held not in Abuja, Nairobi, or Pretoria, but in Menlo Park. The future of trade may indeed be formalized, but by a platform logic with a distinctly private agenda.
The unique market conditions of sub-Saharan Africa mean the dominant model of digital commerce that emerges won’t look like digital commerce in Southeast Asia, China, or the West. Savvy and ambitious entrepreneurs like Tiffany will continue to cobble together the services they need to grow their businesses digitally, but widespread adoption will require solutions that are accessible and affordable to a broader base of sellers. For the vast majority of these micro businesses and solo entrepreneurs, the driver of value is the ability to cost-effectively increase their reach to a wider market of customers. And if this is indeed the race, it is Facebook’s to lose.
Thanks for input and suggestions from Sammy Muraya (SM Kollectionz), Winnie Makena (EGD Luggage Center), Sabina Naamwinbong (Glammy Accessories), Bode Oyelami (Hekmore Systems), Mavis Dzedu (Hotfro), Annabel Schiff, Grace Natabaalo, Stephen Deng, and Jonathan Donner.
* A fictional persona. At Caribou Digital we often create composite personas to anonymize interviewees. In a prior version of this article, I used a Biblical name for the persona which I now understand could be perceived as offensive. I very much regret this oversight, and have updated the article with a new name.
 A note on scope: Many other models of digitally-enabled commerce exist, especially in China and Southeast Asia, where platforms like Pinduoduo and Go-Jek have successfully proven different operational models. While there are arguments for why group purchasing, influencer-based revenue models, and other “social” practices may gain traction in sub-Saharan Africa, these are currently not widespread and thus not covered here. For simplicity, I also exclude classifieds marketplaces (e.g., Jiji/OLX), which are lower-cost for the retailer but don’t offer the same range of services as formal e-commerce marketplaces.
 Jumia reports 65% to 95% of orders are COD; social commerce orders are even more likely to be cash https://newsroom.mastercard.com/mea/press-releases/delivering-cashless-e-commerce-in-africa/.
 This is especially true for solo entrepreneurs, where the nature of the hustle makes it difficult to accurately account for time/labor costs. Yes, opportunity costs are real, but for the smallest shops an inefficient process that takes 50% longer but is free will win out over paid services that automate or streamline the task. The cost calculations change once the business is large enough for paid labor.
 Facebook already buys purchase behavior data (including offline) from external data aggregators in order to inform its user profiling, but this would be first-party data of transactions on its platform and concretely tied to the individual’s account, so infinitely more valuable.
 For reference, Amazon says it averages 8%-15%; https://sell.amazon.com/pricing.html#referral-fees.