SafariShare · 8 March 2022 · 3 min read
The question of adding card payments comes up periodically at SafariShare. The argument is that cards signal a mature product and are familiar to certain user segments.
We keep deciding against prioritising it. Here is the reasoning.
Mobile money is not an alternative payment method in Uganda. For the majority of the population, it is the primary one. MTN Mobile Money and Airtel Money together cover tens of millions of registered users. People receive salaries, pay rent, settle bills, and pay for goods through mobile money every day.
Bank accounts are common in cities, but mobile money is ubiquitous. A passenger booking a trip from Gulu has almost certainly used MTN Mobile Money this week. They may or may not have a Visa card.
Designing a payment flow for the broadest possible group of Ugandan users means starting with mobile money. Not as a fallback. As the foundation.
The integration work took longer than expected. The payment APIs for MTN and Airtel work differently from each other. Confirmation flows are asynchronous: a user initiates a payment, a prompt appears on their phone, they confirm, and the platform handles the callback. When that callback arrives late, or the user dismisses the prompt and tries again a minute later, the booking system needs to handle every state gracefully.
Edge cases accumulated. A successful payment the platform had not yet registered. A pending booking the user believed was confirmed. Each scenario needed a defined outcome that did not create a bad experience for either side.
The version we have now handles these cases. Earlier versions did not.
Adding mPesa extended the platform to cross-border routes and East African travellers coming in from Kenya. For any route with an international dimension, having mPesa available matters. It also matters for trust — seeing your payment provider listed on an unfamiliar platform reduces the friction of trying it.
We have not added cash-on-arrival. If payment happens at the end of a trip, the platform cannot confirm whether a booking is real, cannot release a seat if the passenger does not show, and has no basis for managing disputes.
Mobile money upfront commits the passenger. It protects the driver. It gives the platform a record. That trade-off is correct.
Payment infrastructure shapes behaviour. When payment is familiar and reliable, completion rates go up. When it requires unfamiliar steps, people abandon the flow.
Building for Uganda means building for how Uganda's economy works. Mobile money first. That principle will hold for whatever we build next.