Restaurants are the natural home of impulse purchases. Would you like a third bottle of that wine? Would you like to see the dessert menu? What the hell, why not. All you need to do is say the word, and it all just appears, fresh and delectable for your consumption, before you’ve so much as paid a penny. Eventually, of course, the bill comes — essentially, it’s an invoice listing everything that you’ve already consumed. Then you pay that invoice, and leave.
This is a very sensible way for restaurants to operate. They don’t all work that way: at fast-food joints, for instance, or coffee shops, you tend to pay for what you’re consuming before you consume it. At that point, if you want more, you have to pay again. Which is just one reason why you rarely see people doing that. But generally, the extra amount that people order before the bill arrives more than makes up for the fact that some tiny percentage of them might try to dine-and-dash. Paying is never very pleasant, and if you force people to do it before they’re get what they want, that’s going to reduce both the number of people who buy things and the number of things that they buy.
When we order food in a restaurant, we know that we’re going to be paying for it, literally, in the future — but thanks in part to hyperbolic discounting, even pushing the moment of truth back half an hour or so makes us more prone to running up a tab right now. Similarly, we spend more on credit cards than we do on debit cards: it’s always easier to spend money in the future than it is to spend money in the present.
Online merchants, especially ones who have a lot of mobile shoppers, face a similar problem to restaurants. They want to encourage impulse purchases, but it’s hard to make an impulse purchase when you’re laboriously typing in your name and address and credit card number. The ideal solution would be for would-be purchasers to be able to just press a button, and presto, the item is ordered: the buyer can worry about exactly how to pay for it tomorrow.
That’s the promise behind Klarna, in Europe, and Affirm, which launched today. You click a button, and the item is ordered and on its way to you; the merchant is actually paid by Klarna or Affirm, and not by the purchaser. It then becomes the job of the intermediary — Klarna, or Affirm — to invoice the buyer and chase down the payment, long after the actual purchase has been made.
The two companies work on slightly different models. Klarna uses credit: it’s essentially lending money to the purchaser, and charging interest. Affirm, by contrast, charges the merchant, rather than the customer. Merchants already pay an interchange fee so that they can accept credit cards as payment; paying a similar fee to Affirm will surely be worth it, if they can convert a greater percentage of shopping carts into actual purchases. Also, Affirm seems to be a lot more mobile-native than Klarna.
At heart, however, the two shops are selling much the same product: a way of making online shopping as painless as possible, with payment pushed off until tomorrow. It’s a pretty good idea. But what’s interesting to me is the way that Affirm founder Max Levchin is touting Affirm’s know-your-customer algorithms: the site will identify who you are using Facebook, pull in lots of other data including your Zip code and your mobile device ID, and use all of that information to predict how likely you are to pay the bill once you receive it.
Levchin’s a big fan of such predictive usage of data:
At PayPal, where I was the CTO, we succeeded because we gained deep understanding of the immense quantities of behavioral data that we captured in processing millions of transactions per day. We learned so much about our customers, that we could predict their intentions, and prevent vast majority of intentional fraud.
This is clever, but it can also be a weakness. The reason why so many fintech startups are aiming at PayPal is that people don’t like PayPal; and the one of the main reasons that people don’t like PayPal is precisely its sophisticated fraud-detection algorithms, which tend to throw up a lot of very annoying false positives.
Obviously, Affirm needs to know who you are, and where to find you, so that it can invoice you for the stuff that you’ve bought online. And if you do end up being rejected when you try paying with the Affirm button, then the worst-case scenario is that you’re just back to the status quo ante, forced to pay with a credit card or similar. Still, people don’t like being instantly profiled as untrustworthy; the problem, of course, is that it’s precisely the untrustworthy people with no intention of paying who are likely to be flocking to Affirm in an attempt to order free stuff.
Affirm is trying to make buying stuff on your phone as easy as two taps, without being sunk by massive coordinated fraud. If it works, merchants will surely love it — and won’t much care about the fees it charges. Also, if it works, it will probably end up being bought by Facebook. Which would only exacerbate the worries that people already have about multi-billion-dollar corporations monetizing their personal data. Buying stuff online might become a lot easier. But there’s bound to be a privacy cost, somewhere.