A Ban on Personalized Pricing Is Not Consumer Protection

Manitoba has introduced Canada's first ban on personalized pricing, and the federal New Democratic Party (NDP) has been pushing Ottawa to follow. Both measures promise to protect consumers from what they describe as an extractive practice. Whether they would actually do so is a different question, and the available evidence suggests they would not. A ban does not lower prices; it redistributes them. It eliminates discounts at the bottom of the distribution and raises the floor for price-sensitive shoppers, the very consumers these proposals claim to protect.

Manitoba's Bill 49 prohibits online retailers from using personal data to charge individual consumers more than the standard price for a good. The federal motion calls on Ottawa to do the same, in-store and online, labelling the practice “surveillance pricing.” Both proposals target the same core conduct: firms using data about an individual to charge that person more.

Ensuring consumers are treated fairly matters. But even the most prominent voices behind this push concede there is no Canadian smoking gun. The U.S. evidence they draw on, including the Federal Trade Commission’s reporting on pricing-as-a-service vendors, shows that the capability exists and that firms are selling it—not that anyone has documented widespread consumer harm. Manitoba is legislating against a capability, not a pattern.

Moving from “firms could do this” to “ban the category” skips a basic question: Does a ban actually make consumers better off?

The short answer is no. A ban on personalized pricing does not lower prices; it merely reshuffles them. Faced with a one-price-for-everyone rule, firms will set a price that works across their customer base, and that price will sit somewhere in the middle of the old distribution. The top comes down, and the bottom comes up. Price-sensitive consumers lose.

Manitoba's version of the rule does not change this dynamic. The bill only prohibits using personal data to raise prices, but not to lower them. In practice, however, firms are unlikely to run pricing systems that can only adjust in one direction. They are likely to either abandon personalized pricing altogether, producing the same result, or quietly reset the standard price upward and reframe everything below it as a discount. Because the bill bans price increases above a standard reference, firms can instead shift what counts as the reference.

Manitoba's bill and the federal NDP's motion are framed as responses to the affordability crunch many Canadian households are facing. But the consumers most likely to feel cost-of-living pressures, namely price-sensitive shoppers, households hunting for markdowns, and customers who respond to targeted coupons, are disproportionately the ones who benefit from the discounts these proposals would outlaw.

A field experiment on personalized pricing found that over 60 percent of consumers paid less under personalization. Those paying more were consumers with a higher willingness to pay; those paying less were consumers who would otherwise have been priced out. If the policy concern is protecting price-sensitive households, a ban moves in the wrong direction.

Both proposals also capture pricing practices that have little to do with profiling an individual's willingness to pay. Manitoba's definition covers any automated pricing that varies based on data about the consumer, including location data. The federal motion's "surveillance pricing" framing is broader still, sweeping in any use of personal data to charge different prices for the same product. That is a description of price discrimination, which is how loyalty programs, member pricing, and targeted coupons function. Delivery fees that differ by address, regional pricing, and inventory systems that route stock based on where customers live all rely on customer-specific data without inferring how much a particular shopper can afford to pay. These are longstanding features of market economies, not novel harms produced by Big Tech.

Dynamic markdowns on expiring perishables provide the clearest example. They respond to inventory and time, not to data about individual shoppers. A study of an online grocery retailer that implemented automatic markdowns on perishables found that the system reduced food waste in treated categories by 9.6 percent. Manitoba’s approach moves in the opposite direction, and the federal NDP's motion would do the same.

The conduct proponents may reasonably be concerned about is already illegal. Collusion among competitors is prohibited under the Competition Act, whether executed through a handshake or a shared pricing algorithm. Discrimination based on protected characteristics is prohibited under human rights law. Exclusionary conduct by dominant firms is prohibited under the Competition Act’s abuse-of-dominance provisions.

Earlier this year, the Competition Bureau completed its consultation on algorithmic pricing. Its posture is that existing law already addresses documented harms and that any further intervention should be proportionate and narrowly tailored. This is the agency with both the mandate and the evidence. If proponents want more aggressive intervention, their case should begin with the Bureau’s findings rather than sidestep them.

Canadians are right to be concerned about consumer fairness and responsible use of data-driven technologies. But banning personalized pricing is a blunt instrument that will not make Canada more affordable. While it allows policymakers to appear responsive to cost-of-living pressures, it does not actually address the specific harms proponents identify. A more effective approach would focus on transparency requirements, strong privacy protections, and clear prohibitions on unlawful discrimination, rather than eliminating beneficial price dispersion and harming the very consumers these proposals aim to protect.