Surveillance pricing has suddenly become a hot political topic in Canada, but most of what is being said about it is only half-right. Let me explain it as I would to a 10-year-old.
The cookie example
John and Jane both want to buy a cookie from Acme Cookie Co. online. John is browsing on a new MacBook from a nice postal code. Acme's algorithm clocks him as willing to pay more and shows him the cookie for $3.50. Jane is on a battered Android tablet from a low-income neighbourhood, and her phone battery is dying. The algorithm reads her as price-sensitive and shows her the cookie for $2.50. Same cookie. Same company. Same moment. No one is forced to buy.
Under Manitoba's new Bill 49, that is illegal. But now watch what happens next. Acme's lawyers read the Manitoba bill, and the next morning Acme lists the cookie at $3.50 for everyone. Then it quietly emails Jane a coupon for a dollar off. Jane pays $2.50. John pays $3.50. Exactly the same result as before, except now it is legal. Do you see the problem?
Practical and factual problems
Manitoba's bill and similar ones the NDP is pushing in Ottawa and Queen's Park all ban one direction of the same practice. Raising a price based on your data is out. Lowering a price based on your data is in. The bill explicitly preserves every loyalty program in the country. PC Optimum, Scene+, Air Miles, Canadian Tire Triangle: all untouched. Which means the moment the law passes, every retailer pivots from upward pricing to downward couponing and nothing changes for the shopper. The statute is obeyed. The NDP gets its press release. Jane still pays less and John still pays more, for exactly the same reasons as before.
That is the practical problem. The factual problem is worse. Our own Competition Bureau ran a consultation on algorithmic pricing last summer and published its "what we heard" report in January, drawing on 103 submissions. Its honest conclusion: separating surveillance pricing from ordinary dynamic pricing is hard. Britain's Competition and Markets Authority said the same thing back in 2018. It also found little evidence of true individualized pricing in the wild. The OECD survey of European studies lands in the same place.
Even the U.S. Federal Trade Commission's January 2025 staff perspective, the paper that kicked all this off, leaned on vendor-reported numbers and hypothetical examples, because trade-secret rules blocked it from naming a single respondent. Every regulator who has looked has come back saying the same thing: something is happening but it is not yet widespread and the evidence of systemic harm to consumers is not there.
Stakeholder perspectives
Two of the 26 stakeholder submissions to the Competition Bureau frame the choice. The Canadian Marketing Association cited research on Airbnb's pricing algorithms showing an 8.6 per cent revenue lift for hosts at prices 5.7 per cent lower for guests. Both sides gain. That is what price discovery looks like when it works. Option consommateurs, the Quebec consumer group, raised the opposite concern: algorithmic pricing makes it harder for shoppers to compare products and for regulators to catch collusion. Both groups are right about what they see. Neither argument gets you to a ban.
And we do not need one, because the tools to stop actual abuses already exist. The Bureau spent much of last year investigating RealPage and Yardi over algorithmic rent-setting and concluded it had not reached the Competition Act threshold. That is the system working.
