Stripe announced Stripe Financial Connections today. The product is similar to Plaid's, providing an aggregated API layer to banks. Behind the scenes, Stripe and Plaid are connecting to private partner APIs provided by aggregators. The aggregators are companies like MX and Finicity that do the hard work of connecting to over 15,000 banks in the U.S. For banks without APIs, Stripe and Plaid screen scrape (I was wrong about the value of screen scraping).
While the FinTech stack is worth diving into in and of itself, there's more generalized lessons to learn from this.
- The M:N API Problem is real and valuable to solve (or at least patch). For the most valuable problems, data consumers are willing to deal with terrible data quality and leaky abstractions.
- Data gravity's pull is weakening. Banks have the valuable information (your spending habits and worth) but they simply can't hold onto it (even with regulation).
- Value chain economics have never been more important. Consumer FinTech is notorious as a game of finding the lowest CAC (customer acquisition cost) strategies in a sea of expensive-to-obtain-yet-valuable consumers.
Today, the layering is extensive. I suspect the natural equilibrium is for the intermediaries to be cut out. Companies will either specialize in generating consumer demand (banks, end-user applications) or developer demand (APIs and infrastructure). API aggregators naturally will 1) partner with leaky connectors directory and 2) move down the stack.
For more thoughts on aggregation in the digital age, Ben Thompson of Stratechery has written extensively about an idea he calls Aggregation Theory that he applies to the companies like Google, Amazon, and Meta.