Firms can now charge each shopper a different price, and regulators worry that the gains will flow from consumers to companies. But personalized pricing has two built-in limits: shoppers learn to hide how much they want things, and cheap purchases can be resold, writes Nursultan Mukhtaruly.
Editor’s note: This article is adapted from the author’s essay that won First Prize in Economics in the John Locke Institute‘s 2026 Global Essay Prize.
The price you see in an app may no longer be the price your neighbor sees. Late last year, an investigation by Consumer Reports and the Groundwork Collaborative asked 437 shoppers to add the same groceries to their Instacart carts at the same time. Nearly three-quarters of the items appeared at different prices. Instacart said these were randomized price tests, not prices tailored to the shopper based on their estimated willingness to pay, in what is called personalized pricing. The company said it ended the tests after the backlash. Ride-hailing apps already offer algorithmically targeted discounts to some riders and not others. Policymakers have noticed: Maryland now bars grocers from using personal data to raise prices, and in August the Federal Trade Commission proposed an enforcement policy on undisclosed personalized pricing.
It helps to be precise about what is new. Surge pricing during a rainstorm raises the fare for everyone in the area. Personalized pricing sets a fare for you, based on demographic and personal data, such as the zip code in which you live, the purchases you’ve made in the past, or your browsing history. There is already a decent amount of evidence to justify worrying about the impact of personalized pricing on consumers that also challenges ideas of fairness. However, there are also reasons to think that personalized pricing will not unfold the way its critics fear.
The fear: private central planning
When a firm charges a different price to every buyer, it gives up the market’s single, anonymous price. Instead, it starts doing privately what a central planner tries to do publicly: setting individual terms according to what it believes each person can bear. In the textbook extreme, the seller charges every customer exactly the most they would pay. Everyone who values the product more than it costs to make gets served, so nothing is wasted. But the seller keeps all of the gains from trade, and buyers are left no better off than if they had never bought at all.
Ramsi Woodcock, as summarized in ProMarket, argues that perfect price discrimination may amount to monopolization. However, the evidence that personalized pricing is extracting the welfare gains of trade from all consumers is more mixed. In a large field experiment at ZipRecruiter, Jean-Pierre Dubé and Sanjog Misra of the University of Chicago Booth School of Business found that personalized prices reduced total consumer gains, yet more than 60 percent of customers actually paid less. A single price shuts out everyone who values the product below that price. Personalization can offer those people a lower price and bring them into the market. So, if your preferred online marketplace is selling a pound of ground beef at $8 for a markup of 25% (the difference between the retail price and cost of production), under personalized pricing it may charge $10 to those it identifies as willing to pay more but only $7 to those unwilling to pay $8.
Yet, even when most shoppers pay less, the surplus—the difference between willingness to pay and what is actually paid—taken from consumers as a group goes to the firm. So, whether one fears personalized pricing will extract all consumer welfare and transfer it to firms or just some of it, many observers expect personalized pricing to be bad for consumers overall.
However, there are two reasons to think personalized pricing will not spread the way critics fear.
Reason one: shoppers learn to hide
The Dubé and Misra experiment captured a short window in which customers did not know that today’s purchase would shape tomorrow’s price. That will not last. Friedrich Hayek argued that central planners fail because the knowledge they need to optimally organize economic activity is scattered and cannot be fully collected. Modern data science gets around much of that problem, but it cannot overcome the flawed assumption that a consumer’s willingness to pay is fixed and waiting to be measured. People change their behavior, and it is plausible they will do so deliberately if they know companies are watching to charge them more.
Soviet planners discovered this decades ago. A factory that beat its quota was given a higher quota the next year, so managers learned to underperform on purpose. Economists call this the ratchet effect. Shoppers face the same incentive. If showing interest today raises your price tomorrow, the sensible response is to hide your interest: open the app less often, stop checking the same product again and again, clear cookies or switch devices, remove items from your cart, wait before buying, or simply buy less than you would like to. Research by Alessandro Bonatti and Gonzalo Cisternas shows how consumers who understand that their purchase history feeds a “consumer score” rationally cut back their purchases to protect their future prices.
The drawback of consumers obfuscating their data profiles is that for some 60% of consumers it will reintroduce the inefficiencies of purchases that do not happen when prices rise above their willingness to pay and that companies could have brought down under personalized pricing. And surely, even after adding noise to the data, firms will still be able to rely on some fixed demographic data. Those living in wealthy suburbs can still be charged more for the same good.
Reason two: cheap goods can be resold
Price discrimination has always depended on one condition: the customer who got the low price cannot resell to the customer who got the high one. Imagine a sneaker store that charges me, a student with little money, $50 for a pair and charges my professor $100 for the same pair. I can buy a second pair for $50 and sell it to my professor for $75. He saves $25, I make $25, and the store’s welfare gain from price discrimination evaporates. Wherever goods are physical or transferable—from sneakers to concert tickets on secondary resale markets—personalization can dissolve into arbitrage and gray markets.
Personalized pricing will be more protected where resale is difficult or impossible: a haircut, a ride, a consulting session, a gym membership, or a digital purchase tied to your identity. Airlines are a telling case. A ticket printed with the passenger’s name cannot be passed on, which is part of why air travel has long been the laboratory of price discrimination. Personalized pricing will matter most in the service and subscription businesses that increasingly dominate the digital economy, and that is why a lot of scholarship and regulatory activity is focusing on these businesses.
Where the worry is justified
There are limits to personalized pricing, and in the early days of its use we do not know yet how the economy’s participants—consumers, business partners, or regulators—will respond. Concerns are justified, especially where a few firms dominate a market and so competition does not force them to cut prices, where the product or service cannot be resold, and where customers are unaware they are being priced individually.
One development could make things worse: the same software that personalizes prices can also coordinate them. Researchers have shown that pricing algorithms can learn to keep prices high without any explicit agreement among competing firms to raise them, and courts are only beginning to sort out when that crosses the line into price-fixing. In 2024, the United States Department of Justice sued RealPage, the provider of a rental management software program, for collusion by coordinating competing landlords to share their non-public data to inflate rental prices and undermine competition. RealPage settled with the DOJ the following year.
This suggests where policy can help. Disclosure rules like those the FTC has proposed do more than inform shoppers. They speed up how consumers can respond and alter their behavior to undermine the consequences of personalized pricing, perhaps eventually learning how to minimize its harms while figuring out how to take advantage of its benefits. Protecting consumers’ right to resell what they buy will provide another powerful check on discriminatory pricing.
A firm that prices each of us individually is trying to solve a planner’s problem with a planner’s information. We should expect it to meet the planner’s enduring limit.
This article is adapted from the author’s essay that won First Prize in Economics in the John Locke Institute‘s 2026 Global Essay Prize.
Author’s Disclosure: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
