On a quiet evening in December 2010, while walking from Kenyatta University main campus towards Kahawa Wendani (students’ preferred off-campus residence), a thought rushed through my mind – I partially remembered a computational trick I read from a recreational mathematics book in High School. A problem, I couldn’t remember the title nor the author – Google couldn’t help either. I went over to my nerd friend Elvis to discuss whether we could build a system to find information that’s generally not easy to get. He was fascinated by the concept and we spent the rest of the night hammering out the idea. Next day on our way to campus, Elvis stopped a lad and asked him how they say ask in his mother-tongue, he responded ‘KWETHA’, and Kwetha was born.
Kwetha endeavored to transform the information search experience through an ask and find format. We figured my problem could be broken down to a series of questions which can lead to the getting the answer. Fresh from studying a course in Advance Artificial Intelligence, I was ready to start building the algorithms. Elvis emptied his bank account, rented an office space, bought furniture and hired a developer. The week after that, he contacted a lawyer friend and registered Doban Africa Ltd. In less than a week, we had an MVP product, a registered company and an office. There was a lot of excitement around what the product could become, and I pitched it to fellow students. They loved it and joined in as Doban Africa employees.
We had a unique proposition, instead of returning a web page as with normal search, we returned just an answer (Google instant was not available at the time, and ask.com was manual). The concept and product went viral, and within 2 months we were the 8th most visited website in Kenya.
At that point, we set to capitalize on the growing visibility and launched two new products, NaiRoutes for matatu directions and NaiTraffic for giving traffic information utilizing Swift Global and Access Kenya cameras. This significantly increased the number of employees, and I had to take the role of Operations Manager. In this period, we focused on building awesome products and not revenue. Given the increased number of employees, we had to find a revenue stream in order to pay them. We didn’t have an advertising model or potential advertisers to pitch to. So we built another product, pricelist.co.ke, a product listing website for comparing prices across different stores – vendors would pay a fee to list their inventory.
Marketing the new product also turned out to be an uphill battle. A lot of merchants feared losing customers because of higher prices, their inventory was in printed papers and they didn’t want to commit time to enter the records to our database. Dead End! Remember at this time we are running 3 products with a staff of 10 employees. The solution was to find another hot product.
We settled on D-trail, a parcel tracking Android application that would allow companies to manage delivery logistics easily and cheaply. The application utilized QR-codes to embed parcel information as well as location, enabling security, tracking and confirmation of delivery when they are done.
This had a mild success, two companies showed interest in piloting the project with a firm commitment for paid service if the pilot went well. At this point, we made a decision to kill all products that weren’t bringing in revenue in favour of D-trail. Doban Africa was now a Software-as-a-Service (SaaS) entity. The technical team's focus would be on maintaining the backend, and a marketing team would bring in clients. The setup was easy to scale and required minimal manpower. Time for a few retrenchments 😉.
All was well until one client dropped and the cost of replacing smartphone skyrocketed. Back to the drawing board.
Since cash was the problem, we opted to get into hackathon competitions to win price money that could sustain the company. We put our energies into one project that Elvis was working on part time – Zahanati, a mobile application that can detect malaria by deducing a narrative of symptoms. After working on the app for 48 hours with barely any sleep, we lost the competition to our good friend Nick Hargreaves. The mood was somber, we were sure we would win the competition, we even included a translation module that would switch response to Swahili if a user used more than 50 percent Swahili in the descriptive text. A consolation we were featured in Reuters, Al Jazeera, and CCTV.
The final straw was to convert Doban Africa to a data analysis company. We ran this for a while, encountering problems with big companies hesitant to pass their data to a start-up despite our track record. After two years in the company, many nights spent in the office, and crashing at a lot a friends’ houses, I opted to leave Doban Africa, having learnt many lessons.
So where are we now? Elvis is onto his next startup, MobiAds, a mobile marketing company whose flagship product is Royalty King, an Android application that displays an ad every time a user get a phone call. The users get points for each ad viewed, which they can redeem for airtime or data bundles. For me, I currently work for iHub Research as Data Lab manager where I’m in-charge of Data Science projects.