Britain Should Ignore America’s Personalised Pricing Panic

Suppose Tesco emailed you a £2‑off-deal for a coffee brand, a month after you stopped buying it. Would this worry you? Most shoppers, I suspect, would pocket the voucher happily without a second thought.

Describe that offer as “an algorithm using your purchase history to estimate your willingness to pay,” though, and some might get more nervous. Explain it instead as your friends paying more for coffee because of “surveillance pricing,” and many politicians would get a bee in their bonnet.

Algorithmic “personalized pricing” is producing something of a moral panic right now in the US. Last week the Federal Trade Commission proposed warning businesses that undisclosed use of personal data to set prices may fall under deceptive practices. The FTC’s chairman, Andrew Ferguson, said shoppers expect listed prices to be the same for everybody, not attuned to retailers’ estimates of what each will pay.

Really? Here in Britain, retailers, train operators, budget airlines, pub groups and ride-hailing apps discriminate using price between customers every day, just as companies do in the States. Think railcards for the young and the old, advance fares that punish last-minute bookers, Clubcard and Nectar offers tailored to shopping history and loyalty penalties in telecoms.

Flight prices, in particular, vary according to when you buy and how long your trip is, as companies predict whether you’re a leisure or business traveller, the latter being less price sensitive. Algorithms and AI didn’t invent price discrimination, in other words. Better data has just enabled more sophisticated, precise and cheaper profiling of individual customers in more industries.

It’s this shift from categories of people to individuals fueling distaste. New York already requires the magnificently ominous warning: “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.” Its legislature recently voted to ban surveillance pricing, while preserving specified discounts. Maryland has enacted a law barring large grocers and delivery services from using personalised data to raise food prices.

Britain has so far been refreshingly reluctant to leap to regulation, although there have been tremors in parliament. In 2016, Lord Whitty called it a “very dark art”. In 2018 Greg Clark, then business secretary, commissioned research into the prevalence and consumer effects of personalised pricing, warning companies against abusing customer data. A separate Competition and Markets Authority study that year found little evidence of overt personalised prices.

Its 2021 assessment highlighted the risks posed by opaque pricing, weak competition and the exploitation of inert customers, but admitted that personalised pricing could benefit customers too. The Bank of England says it is now widely adopted across all sectors, in the form of online accounts and loyalty cards.

So far then, we’ve been admirably in line with economic analysis. The popular perception is to see personalised pricing as akin to a monopolist charging everyone the maximum they are willing to pay, so hoovering up consumer surplus to boost company profits. But as Jidong Zhou, an economist at Yale, explains, that scenario assumes that there’s no meaningful competition and companies can perfectly price discriminate. If either assumption buckles, the economic impact of moving from uniform to personalised pricing becomes messier.

Where serving an extra customer costs little for a software platform, gym or online newspaper, personalised prices spread fixed costs across more customers, with discounts widening access. In a world where many firms have access to personalised data, Zhou’s published paper with Andrew Rhodes shows that companies would try to poach others’ customers, resulting in competitive personalised discounting. Information telling one business you are loyal — and ripe for exploitation — alerts another to the discounts needed to win you over.

Personalised pricing, then, is not inherently “good” or “bad.” Its precise impact in any context depends on an industry’s cost structure and competitive dynamics. As an example of this messy reality, Jean-Pierre Dubé and Sanjog Misra used randomised pricing experiments at ZipRecruiter and found that personalisation would raise profits by another 19 per cent over a uniform price and reduce total customer surplus. Yet over 60 per cent of customers would benefit. In many sectors, these price-sensitive customers may well be poorer.

What, then, would a prohibition on personalised pricing mean in practice? Not just a ban on companies squeezing you for more. Applied consistently, it also means no discount to someone with a product in their basket which they passed on, no electronic haggling and fewer ways for new companies to muscle in by undercutting big incumbent players on price.

And, if applied in the more populist way of banning personalised prices from being higher, businesses would often look to achieve the same economic outcome within the law. That might mean raising basic prices and offering more targeted discounts to price-sensitive customers. This would get them to the same place, albeit marketed very differently.

As this sort of pricing proliferates, will British equanimity survive the example of America’s panic? It’s easy to foresee how “personalised discount helps challenger poach customers” could be drowned out by future anecdotes about firms using “AI to charge grieving families more for flowers” here.

It’s also easy to imagine a Labour government promising to lower living costs trying to prove its vigilance on this misunderstood issue. Just last year, then the competition minister Justin Madders acknowledged the Commons business and trade select committee’s “sense that something is not always right” with personalised pricing, before adding: “I am not sure we have the evidence at the moment to support that assertion.”

Ministers should dwell on that distinction. A pricing practice feeling icky warrants investigation, not an assumption of consumer harm. Sometimes a company that knows more about you extracts more. Sometimes it offers a price to entice. And if this all raises profits? That’s precisely the signal for competitors to enter and compete.