ImpactAlpha, July 11 – Mansoor Hamayun couldn’t have anticipated the data revolution when he launched Bboxx more than a decade ago.
Back then, the U.K.-based solar provider was designing and selling small-scale solar home systems for families in Rwanda that didn’t have access to the electricity grid—or to financial services and many other basic services. Today, the company provides pay-as-you-go water, electricity, gas and appliances, as well as loans to millions of low-income individuals in nearly a dozen frontier markets.
“In terms of predicting where we are today, we definitely didn’t do that,” Hamayun tells ImpactAlpha. “Maybe someone else did.”
Back in 2010, machine learning, big data and smart devices required highly specialized skills. “You had to find people with PhDs who were super-experts on those topics,” he recalls. “Those technologies were not very scalable.”
By now, those functions are largely productized, allowing companies in all sectors and sizes to leverage their data for business advantage. Feeding Bboxx’s trove of customer data through modern predictive algorithms has opened a raft of opportunities for a pioneer of affordable off-grid energy. The company sells irrigation systems to farmers. It sells liquid petroleum gas for cooking. It sells TVs. Bboxx has raised nearly $200 million in debt and equity, according to Crunchbase.
Importantly, its customer insight has enabled Bboxx to double down on poor customers while many of its peers have struggled to build scalable businesses for the base of the economic pyramid. The company has evolved as energy hardware and software have gotten better, cheaper and more accessible. Bboxx recently launched a new entry-level solar home product, called Flexx, and is partnering with the governments of Rwanda, Togo and the Democratic Republic of the Congo to deliver energy access to households that the “pure capitalistic model doesn’t reach,” Hamayun says.
Bboxx launched around the same time as d.light, M-Kopa, Barefoot Power, Zola Electric and off-grid solar companies, all of which shared a similar mission of ending global energy poverty. Some of those early peers have moved upmarket. Others have failed and folded.
ImpactAlpha spoke with Hamayun about Bboxx’s 12-year journey, and how “big data” has helped the company deepen, rather than diminish, trust and closeness with its customers.
ImpactAlpha: Everyone is in the data business these days, even though that word rarely came up in the early days of pay-as-you-go solar, financial inclusion and other emerging sectors serving low-income households. How has data changed your business from, say, a decade ago?
Hamayun: When we started back in 2010, even before pay-as-you-go was a word for the sector, the business was mostly about how to increase the affordability of our products and services for consumers and how to make their purchases bite-sized. We discovered that by making those purchases bite-sized [through pay-as-you-go financing], the volumes increased drastically, which resulted in a new problem: we had a portfolio of thousands and thousands of customers that was more complex than any Excel spreadsheet or paper and pen could manage.
Now we have millions of customers and enormous datasets on customers for whom very little information existed historically. And now we are able to use modern machine learning and data science techniques to establish a credit score per household; predict product behavior, like when a battery will fail, and maintenance needs; or—what’s most exciting for me—predict the next product and service a household needs, whether that’s a bigger TV or a refrigerator or a smartphone.
That’s where things are starting to move in our sector: using that data to predict things rather than just manage and control things. And our realization was that data on customers was actually what is necessary to manage distributed products and customers and to manage our operations.
ImpactAlpha: The natural by-product of mobile money, connected devices and call centers is a lot of data. How much did you anticipate the impact of “big data”?
Hamayun: In terms of predicting where we are today, we definitely didn’t do that. Maybe someone else did. But if you look back to 2010, machine learning, big data, IoT were still extremely specialist functions. You had to find people with PhDs who were super-experts on those topics. Those technologies were not very scalable.
Today a lot of it is productized—you get the output without really needing to understand the details of the algorithm behind it. So it’s becoming easier for companies to take advantage of machine learning without needing to reinvent it.
ImpactAlpha: Bboxx has expanded its range of products and services beyond just small-scale solar home systems. Now you sell irrigation systems to farmers and appliances to households; you provide loans to the financially-underserved. How has Bboxx navigated where its resources are best directed?
Hamayun: I often see that Bboxx’s real contribution at the household level is not solar power and lights and TVs, but giving people a pathway to economic self-development and empowerment. The opposite to poverty is not increasing dollars per day, it’s having access to products and services.
Our key skill is having last-mile relationships and managing them in a scalable way. It just happens that last-mile relationships often start with a solar home system. People then have different pathways to getting to more products and more services.
Increasingly, we see ourselves more or less as a sort of a “super platform” that is connecting a digital relationship with a household with a physical relationship that is initially enabled by some sort of basic utility, whether it’s pay-as-you-go water, gas or electricity. But that’s not the endpoint of that relationship.
ImpactAlpha: How come pay-as-you-go solar became the entry point for other inclusive products and services? Was it just because solar companies like Bboxx were most successful in building these last-mile networks, versus other types of product and service providers?
Hamayun: I think two things happened. One was that we sold a product that was super sticky—it passed the toothbrush test: people use it multiple times a day. Because people are buying something from you on a daily basis that means you have a very active relationship, and that is a very unique thing in the demographics that we serve. How many things do they buy daily? Very few.
If you can structure [those purchases] that in a digital way, where you can scale, then the more data you collect, the more you know, the more bespoke services you can provide. That was a realization for us: that the scale and intimacy of our customer relationships are not opposing, but they are actually collinear.
Then [it was about] our ability to cross-sell. Once we have an agent network, a shop, a warehouse, our ability to sell a loan or an additional product or service and deliver it in the cheapest way possible is probably better than anyone else.
From the customer perspective, everyday they make a payment to us is testament to their trust. So with that trusted relationship, if you suggest to them, “Hey, would you also like this?”, their hesitation level is lower than for a completely new provider.
ImpactAlpha: In the early days, the majority of your customers were low-income, many of whom lived below the international poverty line. Is that still predominantly who Bboxx serves today?
Hamayun: About 90% of our customers remain in that category. There are some exceptions. In the Democratic Republic of the Congo, even big cities have no or very limited electricity.
ImpactAlpha: Meaning you’re serving customers on a wider income spectrum in the DRC?
Hamayun: Right, versus like a place like Rwanda or Kenya or Togo, where our focus is very much rural.
We actually want to do more, because pay-as-you-go is an incredible method to reach a lot of people, but it doesn’t reach everyone. We just launched our Flexx brand last year, which is an even more entry-level system, to be able to reach segments that our normal range of systems could not.
At the same time, we’ve been working with governments and other multilateral institutions in Togo and Rwanda, as well as the DRC, to provide subsidies to our end customers. The pure capitalistic model doesn’t reach everyone; subsidy programs can help some of the most fragile people get access to energy.
ImpactAlpha: What have you seen from your customers over the years in terms of income growth or changes in buying power?
I was recently in Togo visiting a few of our clients. We have a project there with the Ministry of Agriculture to sell water pumps [to farmers]—we have a few thousand pumps now installed.
Having access to water is pretty basic, but until then, farmers were relying on the mercy of the weather to ensure a certain farm output. I don’t know exactly how much climate change and other factors have affected or predictability of weather in that region, but one thing is for sure, farmers’ ability to manage their business is a lot harder when they don’t have predictable access to water.
One interesting thing with the farmers I met: they were starting to grow chilis, which as I understand, require more water than other crops. But chilis command a higher price. So our customers are choosing to grow something new, that has a higher market value, because they have access to water.
It’s very hard to quantify exactly what is related to us versus other interventions, but I do think we have a fairly significant part to play in that economic empowerment.
The population that we serve is so economically stretched that if they did not see value in our products, they would not make their daily payment. They make that choice daily. About 20% to 30% of customers upgrade with us annually.
ImpactAlpha: Beyond data, what are you paying most attention to as a disruptor to the pay-as-you-go solar sector?
Hamayun: One key change is the massive growth of agent networks. No company sells directly to consumers, so there are thousands of people running small businesses as agents within each of these networks. The off-grid solar industry as a whole probably has one of the biggest agent networks, after telecoms.
Agents are like the Uber drivers of our space. You have to earn the trust of the agent before you earn the trust of the customer, because you sell to the customer through the agent. So the agents are really the real client, and there’s competition for the agent and [need to prove] value proposition to the agent.
I don’t think anyone thought about agents in that way a few years ago.
ImpactAlpha: Many of your early peers and competitors are still around (some aren’t). You’ve each moved in different directions. How do you feel about that evolution?
Hamayun: We all started a similar time, but started with different strengths. And I think everyone is playing to their strengths, whether it’s market access or hardware knowledge or something else.
The problem we’re trying to solve is so complex. One hundred countries do not have full access to electricity. There are different demographics within a market, and different business models that are needed to address them.
The pressure point in the space is profitability: how do we manage scale and profitability at the same time? There are different views on that. You can be a small company but be very big in one country. You can sell bigger systems at low volumes and have the same revenue as a company selling a lot of small systems.
What is happening is a lot more innovation in this space. Although we started from a similar point, innovation has never been greater.