Startup Stories

Yellow: Powering Africa’s future

The problem Yellow is solving

A key part of the venture investment decision is figuring out whether a problem is big enough to warrant a new company solving it. One of the reasons why investing in, or building businesses in Africa is unique is that the problems are obviously large to anyone looking, but it’s the execution that is so damn hard. It’s for this reason that the number of operators tackling problems in most African countries is minuscule relative to the rest of the world. This is why Yellow’s story is so great. When they started, 90% of Malawians had no access to electricity – the problem was obvious. But at the same time, in Malawi there is next to no infrastructure regarding logistics, retail, digital penetration (no electricity, no smart phone) or other things you would expect to have at your disposal. For this reason, almost no one was solving this problem. Yellow is now electrifying the country faster than the government itself, creating untold benefit to their customers and building a large business at the same time. What’s not to love? Let’s get into it.

To understand the challenge that Mike Heyink faced when he started Yellow, you need to understand what infrastructure tools they had at their disposal in Malawi when he first started out. My questions about what the logistics network was like that they work with was met by cackles of laughter from both Mike and Ross (the CEO and CFO). “Logistics is organized ad hoc by individual operators of retrofitted minibus taxis from South Africa. They’re very efficient at getting things around, but don’t expect RTT trucks. The roads and delivery volumes just can’t support that”.  There is no online marketplace, there are no shopping malls. The retail that there is dominated by localized spaza shops. Given the lack of banking infrastructure, there is no credit. The government’s budget is filled majority by donations (the timing of which leads to the grid being extended piecemeal when such donations come in). Not exactly the dream business environment.

Malawi logistics – getting it done their way

A solution to the need for electricity came from a technology called Paygo, a financing mechanism developed in Kenya and made in China. The version of this that Yellow sells includes a solar panel, battery, light and radio (you can see it here). Electricity is generated from the solar panel, stored in the battery and powers a light, a phone charger and a radio. They come in various forms, but that’s the basic idea. As product market fit goes, they couldn’t be better, given their major competition is candle light and torches. The system provides 20x more output for the same price and the difference this can make in people’s lives is clear and self-evident (and non-linear). Importantly, Paygo technology results in the box being automatically disabled each month unless a mobile money payment is made making financing feasible. If people stop paying, their box switches off. At 20x better product, customers will want to pay. Customers don’t have significant disposable income, making these a financed product – 24 months payment and they keep it thereafter. Given payment for the boxes happens on delivery from China, providing these boxes needs a lot of funding for working capital.

Mike’s journey

Mike, previously at Meteir private equity in South Africa, came across several of these working capital funding opportunities for other operators in the country, which although Meteir wasn’t interested in funding, Mike was interested in exploring. After conducting extensive research and finding local partners in Malawi, Mike set off with 400 systems for a 3 week trip which he thought would easily sell out. In the end, the single sale was made in the taxi on the way back to the airport. His single customer didn’t pay. Business is easy on Excel, difficult in reality. But Mike didn’t give up when many would have, and the solution to selling these systems was actually straight forward.

While financial literacy in Malawi is low, their ability to assess relative value of goods is high. When Mike dropped the price by 40% and made them comparable to candle light, the sales flew. Economists would call this is extreme price elasticity. Malawians would call it common sense. So Mike was off to the races and now faced the next challenge – building for scale.

What’s Yellow really up to

Given Yellow’s first product is electrifying households, it would be easy to see themselves as some kind of private utility, starting with solar panel-based power systems to individual households, before shifting into larger solar systems to places covered by the grid – this is the vision for a lot of Yellow’s competitors. Yellow sees life a bit differently, focusing on the retail aspect of what they’re doing. They’re building a version of Alibaba/Amazon for countries whose retail penetration consists mostly of spaza shops. For Amazon, the route to customer is direct, through the internet, where they focus on the marketplace they’ve built. Buyers go straight to Amazon’s website, where they expect a vast range at good prices and trust Amazon to get it to their door within a day.

If you think about that journey vs the infrastructure Yellow has to play with, you get a proper sense of the challenge they’re overcoming. Their customers largely have no access to the internet, ruling out that distribution channel, there is no credit system in place, meaning ‘rapid checkout’ that Amazon is famous for, can’t be done and finally, fulfilment can’t be outsourced to big logistics players, as there aren’t any.

Yellow has spoken about being inspired by Silicon Valley’s (“SV”) methods in the solutions they created. This inspiration is far looser than most startups looking into SV for inspiration, where in reality they’re just copying and adapting to their home market. With Yellow, the challenges were truly unique. The inspiration was to stare down a hard problem and solve it scalable by going digital as much as possible. That’s where the help ends though. There is no copying anyone as they’re coming to the first set of answers in the market. I emphasise this to try and illustrate what a leap they were making out there and why there wasn’t ferocious competition – most thought it wasn’t possible.

The first problem is getting to customers. This has largely been solved before in Africa through agent networks, whereby local sellers go door to door, or organise gatherings in the community to explain the product being sold. Once a few sales start happening in a community, word of mouth kicks it off. This isn’t new and has been used before. What is new is the management of this agent network. Traditionally, they would be managed through a hierarchy of WhatsApp groups (WhatsApp penetration in Malawi is high and operated on feature phones). Agents are a part of a regional group, whose manager is part of the national group etc. The difficulty is tracking both the coverage of agents, the uniformity of the sales process and the relative performance between them. While Yellow started with WhatsApp (and still uses it for communication), they didn’t see it as scaling well as a tool, with too much manual processing needing to take place in the background. Their solution was far rounder and all encompassing, wrapping up both the scalability of their agents, along with the credit check, rapid onboarding and monitoring of their supply.

Agent community meetings, small…
…to large






The Yellow machine

They built a centralized database where each agent, sale and customer was added, with the ability to add ever more characteristics to agents and customers in time. On top of this, they built a front end mobile application, which in tandem with Upya (their CRM app) is where their agents input all sales-related data in real time, and in turn receive information from Yellow on commission targets, tasks and other important information. On top of this they built three things which in tandem make up rump of Yellow’s machine. (i) An incentive system to keep agents aligned with Yellow, (ii) A credit algorithm to rapidly approve sales (this is still a work in progress) and (iii) a data analytics platform on the backend for Yellow management to monitor everything, no matter how big it got, fully centralized, fully remotely.

The incentive system was built partially to overcome the potential issue of agents selling their own goods on the side, but mainly to drive sales agents to work as hard as possible on selling Yellow products, knowing that major rewards were available should they hit certain milestones. Indirectly, this also gives Yellow the ability to create a benchmark for an agent and how their sales should ramp up over time. Any poor performance will get picked up automatically, when Yellow can intervene. This incentive system is seemingly working so well that there is potential in it as a separate product down the line.

The second problem was rapidly approving credit applications. Given the lack of savings by their buyers, these products were always sold on credit. Their first product had the design feature of only being accessed when a code is input, which code is only generated via SMS when payment (via mobile money) is made each month. As such, management of defaulting payers is quite easy. Not all products are going to have the ability to switch them off for non-payment (especially consumables). A credit algorithm is needed. A regular credit check encompasses modelling previous credit history, earning stability, history in the banking system and for some lending, your asset register. None of this is possible in rural Malawi for people not earning a fixed salary. Yellow have created their own model taking in various inputs including payment history each household that they come across. This model isn’t production ready, but holds huge potential. For now, they are starting to re-lever the Paygo-enabled box that they first sold once it has been paid off. The further benefit (and partial reason its non-linear) of using Paygo vs candles is that customers get an asset vs a consumable product and can use it as collateral to buy other things.

While the credit model is still being refined, Yellow relies on their incentive model to negatively incentive agents from selling the product to buyers who are bad credit risk, which they often know through community referrals etc. A defaulting debtor results in a negative strike on the agent. In this way, the agents are like regional banks, with Yellow being the central bank. Yellow controls the amount of credit by ‘fining’ agents who are providing reckless lending. It’s a good interim model, which they can use as a bridge to perfecting their actual credit model.

Ross and a few agents

Having both an incentive system and credit system that work is necessary for Yellow’s success, but the issue with both of them is ability to scale. This is where the data analytics comes in and is really the key to Yellow getting large in a sustainable way. The obvious major difference between the markets Yellow are targeting and the developed world is the lack of Yellow’s customers being online, either through smart phone or regular internet access.

Before the internet, personal retail in the way Amazon does it was impossible. No retailer could know when consumers wanted to order or what they wanted, and thus the model of brick and mortar retail, along with catalogue sales thrived. The internet enabled a raft of technologies which culminated in the ability for buyers to generate secure profiles online, where they could browse, pay for and have delivered certain goods to a specified address.

Same problem, solved differently

In Malawi, most people don’t have smart phones or access to the internet, making this solution impossible. Incidentally, the consumer spending power isn’t large enough to incentivize investment in shopping malls, so retail has consisted of spaza shops. Agent networks try to overcome this problem by having individuals go door to door and sell directly to customers. Most players track this activity through layered networks of WhatsApp, with data inputs taking place manually on the backend. Manual inputs never scale very well, being expensive and cumbersome – this business model either caps at a single product or customer pool.

Yellow is trying to get as close to a fully digital retail experience without the customer having access to a smart phone. They do this by leveraging as much technology as they can through the agents smartphone that they walk around with. Every agent is onboarded digitally, with the major requirement being a smart phone. The smart phone is key for the location data and ability to run the Yellow mobile app. Location data is used to map each potential customer in an area as well as an overlay of the agents working in the area (they obviously don’t want agents duplicating efforts), while the app is key for the inputs to the Yellow database. Agents visit potential customers or address gatherings in public areas to do their sales. Each gathering, or individual house visit is input immediately into the application. When customers buy something, the transaction is logged through the app, along with the location, key assets of the borrower, what they were using before, and any other information that yellow would find useful. In the developed world, this profile creation is done by buyers themselves. In Yellow’s world, this is done by agents. Once the information is collected, the process is much the same. Yellow have a database with all customers and agents, what they’ve ordered, when and where. They visualise this on a dashboard in order to track all activity against certain metrics. All incentive activity is pushed through the application as well, managed in a centralized way with the only requirement being agents with smartphones.

The major downside with this model is that customers can only buy things when the agents happen to be at their door. This is just the reality of where markets like Malawi are at. The beauty of the model though, is that as the Malawian consumer comes online (either through Yellow’s work or others), they could sign into the Yellow application themselves and immediately begin transacting. So long as Yellow can match the entries on their database, their model will become increasingly valuable as the Malawian consumer becomes richer. Yellow is not taking a pure bet on Malawi/Uganda’s economies getting big. Their first products are sold at healthy margins (rightly so given the benefit they give the end customer), meaning their business is successful just on that basis. Looking forward though, assuming a level of economic growth in the country, Yellow’s value will really be unlocked down the line, and this value could be massive. At the same time, the productivity growth that is unlocked by getting consumers access to electricity has been evidenced around the world. Yellow is making their own ‘upside case’ more likely by pushing Malawi forward.

Yellow’s positioning

The infrastructure that Yellow has built also makes their framing of competition quite different. While it may seem like they are competing with other sellers of solar units, what they’re really competing for is the retail market place in these countries. In order to win this market place, they need a lot of things to happen. For them, they need to get data on as many customers as possible to build the demand side of the marketplace, as well as manage supply and logistics of goods that they order. In time, they’ll need to add more products from external suppliers to make the market place more valuable for consumers and in so doing benefit from the network effects that successful retail market places do. Externally, Yellow needs consumers in the markets where they operate to get richer in order to afford to get online and buy other goods. This in turn requires significant economic growth. For countries so behind the infrastructure curve, there will be tremendous economic gains through productivity growth from things as simple as getting consumers access to electricity (along with healthcare, roads etc). Yellow is also directly pushing this forward, but so are various of their ‘competitors’. So long as Yellow’s network is deep enough, they’re more than happy for others to supply goods that benefit the economic consumer. As the spending power of their consumers grows by an extra dollar per day, so Yellow’s market gets bigger and bigger. In a weird way, they’re cheering on all the investment that’s going into their markets, whether from themselves or anyone else and even if investment is in selling the same products that they’re selling right now.

A business making a real difference. Electricity enabling education.
The endless opportunity

So what does Yellow’s business look like in 10 years time? Having successfully made a mark in Malawi, they launched a similar offering using the same playbook in Uganda. Being a bigger country with c. 43m population and more developed infrastructure (40% electrified vs 10% in Malawi), there is more competition on the solar product. While this isn’t Yellow’s endgame, they do need the footprint in order to build the marketplace. Successfully building this out is a key challenge for them as they grow across the region. But assuming they solve this problem, as with the myriad problems they have solved to date, the dream is to be the retail marketplace across Africa, pushing the continent straight past brick and mortar retail right into the a digital marketplace with unlimited choice. Developing a retail landscape in the digital era will be even more interesting. With no established brick and mortar players, Yellow could well end up doing both – starting online and going offline for fulfilment, click and collect, showroom etc (Amazon is planning on opening thousands stores next year – brick and mortar is a key part of the future retail mix).

While currently agents hold the entire product catalogue in their smart phones, Yellow would hope to make this part of their model redundant. As their customers come online, they’ll match their profiles with Yellow’s data and order directly. Yellow’s role will be to attract suppliers that their customers are demanding, enable the logistics and have their agent network arrange the last mile logistics for them (there isn’t sufficient infrastructure to outsource this at the moment). Connecting suppliers and customers is currently impossible at distance, and opening this up should prove incredibly lucrative to the team that gets it right. Pinduoduo in China showed the potential in this, connecting rural buyers direct to farmers in China, becoming a $100bn company at the same time.

Yellow’s assets at this point would be numerous. To start with, their centralized incentive model is likely to be inherently valuable itself. Much like Discovery in South Africa showed the value in pushing consumers to be healthier through incentives, enabling a large workforce to achieve a goal that’s aligned with your business will prove valuable in many instances. This piece of the business could be sold as a set of tools (aka a platform) in time should they choose. Then the credit model. If they find a way to bank the previously unbankable in these markets, that would also be incredibly valuable. There is no shortage of funds to lend to consumers with a proven track record of paying it back, but there are almost no funds available for consumers with no track record and no data points at all. By adding a layer of credit history to new customers, Yellow could unlock various finance products, both for themselves and others.

Then of course the marketplace itself. Retail in the united states was transformed by Amazon, but was already 100% penetrated. The scope of possibility of digital first retailers is better illustrated in China, where Alibaba and Tencent ended up supporting entire ecosystems off the back of the channel they opened. If Yellow is successful in creating a retail marketplace that works in rural Africa, they will likewise have an opportunity to benefit in huge ways. If they open up their tools to enable new businesses to operate on their infrastructure, they’ll get a first look at businesses starting up. Much like Tencent does, they’ll be able to invest in companies that look promising and promote them through their marketplace to increase the chance of success. Tencent has done this many times over and have turned into more of a capital allocation machine than purely operators of the Wechat ecosystem.

While this vision is a long way off, these are the possibilities that come with solving what look from the outside like intractable problems. Consultants have been talking about the ‘billion person untapped retail opportunity in Africa’ for years, but it’s companies like Yellow who are taking the risk, building the infrastructure and enabling the market that will likely be the ones to benefit. Historically unstable politics, almost no infrastructure and a low income consumer base are but some of the major challenges operators like Yellow face, but with over 40,000 customers to date, they must be doing a fair bit right.

It’s hard not to be inspired by companies like Yellow, taking bets where most think it’s too hard/risky. And no one should be rooting harder for them than the customers they are serving in markets like Malawi. They recently raised over $3m from investors in South Africa, which money they are using to provide customers with an asset that generates electricity. This in turn increases productivity and drives sustainable economic growth. This growth will increase spending power and wellbeing of citizens, while increasing the value of Yellow’s business. This is capitalism at its best – taking risk capital to improve consumers lives at a rate faster than that managed by government itself. I wish them all the best!


  1. Really interesting article James. I feel more informed now. Thanks for sharing their business model and your insights and explanation.

Leave a Reply

Your email address will not be published. Required fields are marked *