The coming world of ubiquitous price discrimination

It's going to be annoying and weird — but with upside!

If we raised the price of a Slow Boring subscription tomorrow, some of you would probably drop your subscriptions. But others would pay the higher fee. Conversely, if we dropped prices, some of our free subscribers would probably sign up. I did not settle on the $8-per-month/$80-per-year pricing through any elaborate quantitative analysis. I didn’t hire a team of McKinsey consultants to develop the strategy, and I have no idea whether the price is anything close to optimal. In effect, we’ve been charging less over time due to inflation, but our nominal revenue has risen faster than overall price inflation, so I’m happy.

In principle, though, I’d like to be much more successful. One way to do that would be to write dramatically better pieces.

Another would be a more advanced pricing system. If I could x-ray your brains and charge $80/year to those who would cancel if I charged one penny more but bump it up to $85 or $97 or $103 or precisely what each individual one of you would be willing to pay, I could make a lot more money. Similarly, I could use my brain x-ray to offer customized discounts to people who’d be willing to give the site a try for a lower fee. I could, in theory, generate a lot of extra revenue from the exact same product if I had my brain scanner.

Of course, I don’t have a brain scanner, and I can’t do that.

But “price discrimination,” where you sell the same product to different customers at different prices, is a longstanding feature of the business world. Instead of brain scans, sellers rely on broad demographic characteristics (senior discounts) or minor frictions (coupons) to make guesses about willingness to pay. Modern information technology has created a lot of new possibilities in this regard. On the internet, it’s feasible to have not just one or two or three different prices, but an arbitrary number of prices. It’s also easier to hide from any given customer the price others are seeing. And while modern machine learning can’t yet scan your brain, it’s certainly possible in principle for sellers to learn a lot about you and draw much stronger inferences than just “this guy is old.”

The increasing prevalence of price discrimination is changing consumer experiences, not always for the better. It’s also generating more policy buzz, with a school of thought emerging that any kind of algorithmic price setting is per se bad. Elizabeth Warren, who represents one of the states where happy hour discounts are illegal, is relentlessly opposed to the introduction of what she terms “surge pricing” into new retail arenas.

In the sense that leftists don’t want to concede an inch to the argument that high levels of government spending on Covid relief played a role in inflation, this heightened scrutiny of microeconomic pricing tactics makes sense. But since the relevant technologies are just going to keep getting better, it’s worth thinking in a more principled way about whether it’s really worth playing whac-a-mole with business model innovation.

Beyond arbitrary anchoring

When I began writing this, I was not at all sure what other people think about price discrimination. So, I did a poll of my Twitter followers to try to see where the conventional wisdom is, at least among Yglesias readers. And it turns out to be sharply split.