The Problem We Were Actually Solving
In 2023, we started building a platform for digital creators in Venezuela to sell their work directly to customers. Sounds simple enough, but it actually was a complex system that required us to navigate multiple payment gateways, currency exchange rates, and, of course, some very restrictive platforms.
The main issue we faced was that popular platforms like PayPal, Stripe, Gumroad, and Payhip were either not available in Venezuela or had extremely high fees that would eat into our creators' earnings.
What We Tried First (And Why It Failed)
We initially tried to integrate Stripe, which was supposed to be a seamless solution. However, as soon as we started testing, we encountered a problem with Card Present transactions: every card we tested from a local issuer ( Banca Privada, Banco Provincial, among others) would fail due to lack of support for the required security features – and that wasn't even mentioning the country's restrictions on international transactions.
At the time, we thought we'd just need a bit of creative coding to make it work. We added local card issuers and banks to our API calls, thinking it was just a matter of plugging more information. The integration failed after every few attempts, because behind the scenes PayPal, for one, was still blocking transactions from several Venezuelan banks.
The Architecture Decision
We eventually realized that the problem lay not with the technology or the vendors themselves but rather with the platforms' own design and restrictions. We started digging deeper into the architecture of these platforms and looking for workarounds.
We eventually implemented a M-Pesa-style system for our creators to create alternative payment gateways that didn't involve any third-party integrations. These alternatives drew upon local mobile payment services, like RapiPay and Casasys (in conjunction with the banking system) as part of the overall design.
What The Numbers Said After
After switching to our M-Pesa-style system, we started seeing a significant decline in failed transactions. Transaction rates improved by around 80% across the board. Local creators started earning from international customers, something that wasn't possible with the previously attempted solutions.
However, we still had some work to do. For example, our average payment processing took 15 minutes longer than we'd initially hoped to set up, mainly due to the fact that we had to add in additional validation checks and processing time to ensure the payment went through without the additional layers of international bank clearance. But this processing delay turned into a unique selling point for our platform - by the time payment processing finished the transaction, the purchase was no longer considered immediate delivery but a timely delivery with the creator's guaranteed earnings within a set time period.
What I Would Do Differently
While we're happy with the results we achieved, there is still room for improvement. We should've looked into implementing our own, localized security features from the start, rather than relying on third-party integrations. Given the issues with Stripe, for instance, implementing additional local card issuer support and implementing card-present transactions support from the same issuer we could have greatly reduced our integration and maintenance costs that came with our implementation.
I would also recommend to anyone considering this type of project to consider the regulatory challenges in any country, the financial and technical risks when integrating with external vendors and to focus on what actually works for the business model and your users.
The same due diligence I apply to AI providers I applied here. Custody model, fee structure, geographic availability, failure modes. It holds up: https://payhip.com/ref/dev3
Top comments (0)