FinTech Credibility

I worked at Affirm for 3½ years, before “Buy Now Pay Later” (BNPL) was coined as the instantly-approved, low-duration loan popular with online retailers today. Some would argue that the entire business model was only possible because of our previously low-interest rate environment and that rising interest rates have exposed the fragility of their profit margins. Unfortunately, this has meant that the publicly-traded stock has also been battered pretty hard for the last 18 months.

But! At least the company is still around, biding its time for more favorable macroeconomic tailwinds. Most of the BNPL startups that came onto the market in 2020–21 have either shut down completely, consolidated/sold themselves, or drastically reduce their footprint as capital costs became untenable and new regions too difficult to operate1. It turns out that building a sustainable lending business is harder than most people thought, and the low-interest rate environment helped mask the depth of underwriting and operational excellence necessary.

Anyway, I was reminded of this when I read that Apple Card had lost Goldman Sachs $1b so far, and the bank wants to pass off the partnership to someone else who can make something of it. The reasons for such a massive loss are the usual factors that plague fintech—and finance—companies: loose underwriting, overextending credit, and not controlling for fraud and default rates. Oh yeah, Goldman is also the banking partner for Apple’s BNPL product, which launched in March.

Fintech has a superpower: with the right timing, the right product can produce hockey stick growth in users and capital. The software powering this financial service tends to scale horizontally—eventually, faster than comparable non-technical, i.e., people-driven, services. Companies get tripped up by regulations and underwriting: with so much money flowing through these systems, a misstep in a couple of basis points amasses to a lot of money, an effect that can drive revenue as easily as losses.

It turns out that this is pretty hard!

I’m reminded of a similar gold rush that played out early in Square’s life as a startup. As we were hitting our stride and saturating small businesses with the iconic square credit card reader2, we saw a field of competitors crop up: some from expected companies, others seemingly out of nowhere. Even banks got in on the product category; tellingly, the executives launching these services knew they were commoditizing their merchant services offerings and felt like they had to respond, even if it meant cutting into their profit margins.

As we see a decade later, most of these card readers don’t exist anymore. Certainly, the timing of magstripe credit cards and the introduction of EMV and contactless payments in the US provided a narrow window when that type of simple hardware could facilitate payments, but pretty much all of the new entrants competed on price. Square had charged a flat 2.75% fee, and everyone else had to come up with something lower, sometimes with additional conditions. These companies knew they’d be losing a bunch of money trying to compete strictly on a percentage basis, but the goal was perhaps to stick around long enough for Square to run itself to the ground, or gain enough market share to start raising prices, à la the Uber monopoly strategy. For everyone other than maybe PayPal, the gambit did not work.

Back to Apple and their forays into fintech, I will give them credit for building out their services over several years—starting initially with Apple Pay, and now fleshing it out into a full financial services arm of the firm, while striking up partnerships to offload risk. It goes to show that mere technology is not sufficient, and even tech + money does not guarantee immediate success, as evidenced by efforts like the Amazon credit card reader or Facebook’s failed Libra cryptocurrency. To apply technology to finance successfully requires patience, persistence, and working through the details of the business to squeeze out a small advantage—in investment terminology, some alpha.

  1. Affirm itself shut down its Australian business earlier this year.

  2. The most popular Gen 2/3 readers plugged into phones’ audio jacks, required no additional power, and were amazing little hardware devices.

Share this article
Shareable URL
Prev Post

Mutual AI Escalation

Next Post

Iterative Recruiting

Read next