A few weeks ago, in the aftermath of Wave's $200 million Series A, we wrote about so-called full-stack fintechs. It's a somewhat common theme in African markets - where infrastructure is non-existent or insufficient, fintechs are needing to build out the full stack themselves. It's operationally challenging, but for those who can pull it off, they are able to not only serve their customers more reliably, but they can do so more profitably, as well.

But as infrastructure compounds, it can both lessen the need for (operationally difficult) full-stack approaches and also create a greater opportunity for specialization. I think this is a good thing.

I'm thinking about specialization this week for two reasons, in particular.

First, I had the opportunity to speak to Quartz for a piece published on the one-year anniversary of Stripe's acquisition of Paystack. They asked if Paystack should have a role in offline payments, considering the width of cash-based economies, while I pondered whether the acquisition affords them the ability to be a specialist and stay narrowly focused on online payments.

I recall an anecdote from Shola in our interview after the sale. He talked about reliability, saying that even if Paystack had a 99 percent transaction success rate, on 100 million transactions that's still 1 million failed transactions - which, he said, is too many failed transactions. And solving that problem requires specialization.

Second, this week we published episode three of this season, on Africa Stack. It was an episode full of specialists - Mark Straub from Smile Identity, Abdul Hassan from Mono, and Brendan Playford from Pngme - who are all focusing on a particular layer of the financial services stack. The practice of open banking, to a large degree, is the practice of consumer-facing fintechs availing services built on top of a wide number of specialist APIs, who provide identity data or transaction data or permission other forms of consumer data to said fintechs.

Whereas the narrative in African tech has largely been around necessarily building at multiple layers of a given value chain, we have also seen an increasing proliferation of specialist, API-first, B2B infrastructure startups on the continent. (In fact, it's Rally Cap's entire investment thesis).

But why? Why is specialization so important?

Best-in-Class

Not Boring's Packy McCormick explored this question in the context of Spotify and Stripe. Spotify uses Stripe to process payments for all of its merchants, generating $350 million for Stripe. At that cost, why doesn't Spotify just build its own payments solution?

He writes,

APIs let companies leverage years of other companies’ work in seconds... An API-first company essentially abstracts away the complexity of a whole best-in-class company, giving clients the full output of a highly-focused org.

API-first companies are able to do this for two reasons, in particular. Focus and scale. These companies are set up to solve a very specific problem, and if they're solving these problems at scale, "they’re able to justify small improvements that build up to an incredible product over time" to build best-in-class products.

Consumer-facing companies - the API-first companies' customers - need these specialist services, and by plugging into best-in-class solutions, they can focus on that which differentiates their businesses, and better serving their customers.

But it's not just that the services provided by API-first companies are non-core, it's that they can provide much better solutions than that which is built by their customers in-house. For context, Stripe pushes updates to its core API 16 times per day, which improves acceptance rates, lowers fraud, allows for faster payouts and returns, and much more.

Only Paystack, and not their customers, can focus on improving their success rate from 99 to 99.9%, and they can only do so through focus and specialization.

Ecosystems need specialists, and it's a sign of a mature ecosystem that a) API-first companies are best-in-class, and b) that there are enough customers for API-first companies to build and scale their specialist products.

The latter has always been the biggest question for the African tech ecosystem. Whereas Plaid, for example, might have hundreds of thousands or millions of customers (and can therefore justify connecting to all the banks in the US), we're not there yet.

But if recent fundraises are any indication - Mono announced their $15 million Series A last week - big bets are being made on this space and the future opportunity for API-first companies. And the more this happens - the more best-in-class these companies get - the easier it gets for consumer-facing companies to build and offer better services than ever before.

--

✌️ Justin

‍