Global Payments From Anywhere: Why Our Choice of Payment Gateway Changed Everything
ruth mhlangaThe Problem We Were Actually Solving The issue was twofold. Firstly, many of these...
The Problem We Were Actually Solving
The issue was twofold. Firstly, many of these countries have strict regulations around cross-border payments, which limits our options. Secondly, even when we did find a payment gateway that worked, users were often forced to use cash or wire transfers, which is clunky and insecure. Our users deserved better.
What We Tried First (And Why It Failed)
We tried using the usual suspects: Stripe, PayPal, and others. At first, they seemed to work, but it quickly became clear that they were just a series of complex APIs and failed transactions. We'd get complaints from users about payments not going through, and after digging into the issue, we'd realize that it was because their bank didn't support certain types of international payments. It was frustrating for both us and our users.
The Architecture Decision
We ended up switching to a payment gateway that specializes in cross-border payments: M-Pesa, a Kenyan company that has operations in many African countries. They have a huge network of local agents who can handle cash transactions, making it easier for our users to get paid. We also had to rewrite our application to handle the quirks of M-Pesa's API, but it was worth it.
What The Numbers Said After
The results were dramatic. Our payment success rate went from 50% to over 90%, and our user complaints about payment issues dropped by 75%. We also saw a significant increase in user engagement, as people felt more confident using our platform.
What I Would Do Differently
In retrospect, I would have done more research on the payment landscape in these countries earlier on. I would have also worked more closely with our users to understand their specific pain points and needs. It's a hard lesson to learn, but sometimes the best solution is the one that involves taking a step back and rethinking your assumptions.
Same principle as idempotent pipeline design: design for the failure case first. This payment infrastructure does that by default: https://payhip.com/ref/dev8