I have heard for years people criticizing the payments industry as a low margin commodity biz. What has changed? Or they just got it wrong?
replies(4):
“In the tech world, particularly after PayPal was sold to eBay, the dogma for many years was ‘Don't ever do payments, it's too hard and fraud will blow up and destroy you.’ And during the first internet wave, that was true. It was a challenge to get the type of data and information you needed to really deal with fraud at scale,” he says. “But fast forward to today and take a look at companies such as Stripe and Affirm. It seems that the fraud problems weren’t as bad as we thought and we’ve since developed systems, processes and data science tools that just didn’t exist back then. So if you’d applied first principles thinking five or 10 years ago, asking if those assumptions were still true and digging into whether it was still too hard, then maybe you would have come to a very different conclusion about starting a payments company and gotten ahead of the curve.”
http://firstround.com/review/future-founders-heres-how-to-sp...
Until we see Stripe's financials I would not assume anything about healthy margins. Square also had a lot of hype in the early days but turns out their processing margins were really low and most of their current valuation is from expanding into other products like business loans.
Stripe's valuation is similarly based on a hypey "grand vision" to offer other services that actually make money, not by magically making businesses pay more for commodity services.
If all you are doing is taking credit card numbers, taking money out, and handing it to merchants, then sure, there's relatively little value there, and the competition at that level can be fierce.
But what you find in companies like Uber or Amazon today is not just taking credit cards: They run marketplaces where signing on is easy. They will take very different payment forms in different countries: The world is not really all about credit cards in US dollars. You need to fight different kinds of credit card fraud, build a reporting infrastructure, integrate with some traditional accounting system, and connect it all to your logistics in some fashion. A large marketplace will have dozens, if not hundreds of people dedicated to payments, and that without caring of the little bit of taking credit cards. It's in those areas where payment companies don't have feature parity, and a big part of what you'd pick one over another, especially as a startup. And in differentiation there's money, not in taking all the physical retail sales for Walmart.
I suspect that eventually a few big players will really be feature complete with each other and margins will go down, but maybe there is a big winner there, who has an expensive to replicate product, or whose product relies on so much data that competition is impossible. If that were to happen, that kind of big winner's valuation would be far higher than what we see today from Stripe, Square, Paypal or Alipay.
Fraud is only as bad as we know. And it's very bad, payment providers just aren't going to go around telling you about it.
I worked for a payment provider and it's very very difficult to eat fraud because it wipes out low-margin profits quickly. It's a tough dance - there isn't a lot of wiggle room between being too harsh in rejecting payments and being too lenient in allowing payments. And criminals are very, very good at moving on to another attack vector once you figure out what they are doing.
At some point in the 1930s a paper was written arguing from first principles that a jet engine could not work to propel a plane. Axial flow compressors were to inefficient and radial-flow compressors generated too much drag.
What the author didn't know is that another already written paper had come up with significant improvements to axial-flow compressors by using airfoil-shaped blades, which combined with improved metallurgy was more than sufficient for a working turbojet.
So it was a perfectly sound argument for why something was impossible, but it was still wrong because the assumptions were changing.
The tech industry looks like this all the time; in many of pg's essays he mentions that a lot of successful ideas look like toys, and part of the reason for that is this core concept; as the underlying assumptions change something can morph from a toy to big business overnight.