Five years, one paying customer: building for the informal sector

Ted Boulou, as told to Hardware Things

In 2012, Ted Boulou moved to Dakar, Senegal to work on rural electrification projects across Francophone West Africa for the World Bank and IFC. Ted, a Cameroonian native, soon became interested in informal markets and had an idea to develop solar-powered devices to manage inventory, keep track of suppliers, as well as keep financial records. Speaking to How We Made It In Africa in 2015, Ted mentioned that the full concept was to become some kind of financial institution that could facilitate micro-loans for the same businesses in order to help them grow. The company he co-founded (and led as CEO) was christened Somtou.

In September last year, Ted joined WeLightAfrica in Madagascar as Chief Operating Officer, effectively shutting down the company after 6 years of operation. Just before that, we had a chance to chat to about what went down, their product development struggles, and what he would do differently today.

Hardware Things: How did Somtou start?

Ted Boulou: In 2015 it started as a side business while I worked at the World Bank. We had two bold ambitions at the same time: the market ambition to target the informal sector and the product ambition to make the devices locally. I have since learnt that one can’t have both, it took us five years to understand that. To do anything from scratch is a nightmare because you’re struggling on both the technical and market sides. We started with using 3D printing, Raspberry Pis, some CNC cutting to try and build the first prototype. While I enjoyed it as an engineer is was a mess as a business owner. The ecosystem in Dakar was not mature enough for what we wanted to do.

I understand that at some point you moved back to Cameroon, was the experience different?

It was still difficult. I mean, even companies in the West struggle with hardware and tend to develop more software now. Whether in Senegal or Cameroon, we could not find design for manufacturing experts. While the Raspberry Pi prototype worked, it was quite slow. When you want to work with small shops and the prototype is slow—because they do very quick transactions—you realize you need to build your own device made for purpose that would be fast. We tested the device with different types of businesses: shops, tailors—as they do slower transactions so our prototype could handle these. Then we ran into energy management issues.

What did you learn from all these tests? And in interacting with small shops vs tailors, what was different in the way they used the devices?

That people do not always say what the real painpoint with using your product is. In Dakar, people are too nice [with feedback]. They won’t tell you but when you look at their faces when interacting with the product, you see that you are far from where they want you to be in terms of the technology.

First, the speed of transactions. Small shops start receiving customers from the moment they open their door, if you lose one transaction your device is useless because you’ve messed up the balance sheet. With craftsmen or tailors, they have more time. They can record one transaction an hour and then take time to draw the design or other activities.

Second, the data collection is different. The key component of working with shops in the informal sector is to collect meaningful data like the brands in inventory, the Nestlé or milk. The amount of data craftsmen have is marginal in comparison, while interesting in it’s own right.

Overall, with every person we met they always wanted a device [like ours] that would make their lives easier. On this point, I have no doubt. But our solution was not ready.

Did you try to manufacture in China or just white label?

At the beginning, we wanted to do everything ourselves. People would tell us “stop doing that, go to China” and we would say no that we wanted to build it ourselves. And maybe that’s part of entrepreneurship, trying to be a hero and not listening to people. But now, I’ve just found the perfect device that would be the closest to what I would want and we would just buy it from China. We can put in a SIM card and just focus on the business part to start.

Your device must have gone through different revisions, what were you testing out?

You know, in order to learn what sellers wanted we tried different things. We must have put together anything that could be connected to the Raspberry Pi. We knew they needed scales to weigh goods, so we tried a measuring scale once. We knew there would be energy issues so we tried versions with solar panels. But towards the end, we just went with an incorporated barcode scanner.

How did the business side of things go?

We’ve tried a lot of things over the years and we tested the market. At our peak, we had about 8 people on staff and with the money we raised we usually hired ad-hoc staff so that we could have a long runway. In five years, we got only one paying customer—a neighbourhood shop that sells provisions.

One thing I learnt that showed the need for a product like Somtou is that the owners of the shops and the person who sells in the shops are usually different people. The owners are very interested in inventory and having a real-time view of their sales, so they know what’s happening in the shops. So while the device is in the shop, the owner can log into an app to see this information.

This begs the question, could Somtou just have been an app?

Hardware is more defensible than software, that was our initial thinking. Second, we thought of Somtou as a professional, work tool that would be dedicated to shop tasks. We did not want staff to have it on their personal phones and imagined a cheap enough device that worked very efficiently. Shop owners were also aligned in that vision.

Also, there were items in shops that did not have labels or barcodes that could be scanned. Like bread, for example. So QR codes would have to be created with stickers for these items on a shop-by-shop basis, this was simpler to do with one device per shop.

_{Sign up with your email address to receive the monthly newsletter.}