Everyday · Money & Consumers · Technology
What if the price on your screen was calculated for you?
The FTC has proposed an enforcement policy on personalized pricing — using personal data to estimate what an individual will pay. It is a proposal, not a ban, and the Commission says outright that Congress has not given it the power to prohibit the practice.
Americans are entirely used to prices moving. An airline seat costs more as the cabin fills. A hotel room costs more on a popular weekend. Petrol changes between Tuesday and Thursday. Concert tickets climb. A shop has a sale and then the sale ends.
None of that involves a seller deciding that you personally will probably pay more than the person standing next to you, for the same thing, at the same moment.
That is the distinction underneath a proceeding the Federal Trade Commission opened on 19 August 2026, and it is worth getting right, because almost every version of this story circulating at the moment has already lost it.
What the Commission means by the term
The FTC defines personalized pricing as the use of personal data to set prices according to the amount a company believes an individual consumer is willing to spend.
The data in question is the ordinary exhaust of using the internet: browsing history, purchasing behaviour, location, demographic information. What turns it into pricing is the inference drawn from it — an estimate of willingness to pay, or of whether this particular customer is likely to go and check another shop.
The practice also travels under the name surveillance pricing. The August statement is the Commission's most direct account so far of how it reads its existing powers against it — and reading the document rather than the coverage of it is the whole exercise here, because the two say noticeably different things.
The lifecycle state, before anything else
This is a proposal open for comment. It is not a rule, it is not a ban, and nothing about it is currently binding on any retailer.
After the comment period the Commission reviews what it receives and decides whether to finalise the statement, revise it or withdraw it. All three are live outcomes. A proposal is not a preview of a certainty.
And even a finalised enforcement policy statement is not a prohibition. It is the agency telling the market how it reads its existing powers — useful, consequential, and a different kind of instrument from a rule with the force of law.
The sentence in the document that most coverage leaves out
The proposed statement says, in its own summary, that Congress has not given the Commission the authority to prohibit personalized pricing.
That is the agency limiting itself in its own document, and it reframes the entire proceeding. The FTC is not proposing to outlaw individualised prices. It is saying that where the practice is used in a market whose customers do not expect it, and is not disclosed, existing consumer-protection law already reaches it.
The legal hook is section 5 of the FTC Act and its prohibition on unfair or deceptive acts or practices. The theory is a disclosure theory: a retailer who represents or implies that a price is the same for everyone, while in fact varying it by individual, risks deceiving the customer.
The Commission is not asking whether individualised pricing should exist. It is asking whether it can be done silently.
The Commission's own framing puts consumer expectation at the centre. Its chairman, Andrew Ferguson, said that when consumers see a listed price, they expect it to be the same price that everyone else sees
.
Why the same practice is unremarkable in one market and a problem in another
The proposed statement makes an observation more careful than the headlines around it: personalized pricing is a long-established norm in some markets and not in many others.
Nobody is scandalised that two people pay different prices for a car after negotiating, or that insurance premiums differ by driver, or that a tradesman quotes one figure for one job and another for the next. In those markets an individualised price is the expected shape of the transaction.
A shelf price is not that. A listed price on a website is not that. The expectation attached to a displayed retail price is that it is the price on offer at that place and time to whoever is looking — and the FTC's argument is that where a business quietly departs from that expectation using personal data, the departure is the thing that needs disclosing.
Which means the analysis is context-dependent by design. The same technique can be entirely ordinary in one setting and legally exposed in another, and the variable is what customers in that market reasonably expect.
Dynamic pricing and personalized pricing are not synonyms
This is the single most useful distinction on the page, and it is the one collapsed most often.
Picture an airline seat that goes from $300 to $360 because the aircraft is nearly full. Everybody looking at that seat sees $360. The price moved because the market moved.
Now picture the same seat priced at $300 for one shopper and $360 for another at the same instant, because a model predicts the second is less likely to comparison-shop. Nothing about supply changed. The price moved because of what the seller believes about the person.
Sophisticated systems can do both at once, and separating them from outside is genuinely hard. That difficulty is a reason for care in describing what you have observed — not a reason to treat the two as one thing.
Would two phones really show different prices?
This is the question that makes the subject interesting to ordinary shoppers, and it is where a documentary publication has to be most disciplined.
A regulatory proposal is evidence that an agency considers undisclosed individualised pricing important enough to address. It is not evidence that any particular retailer is showing you a different price today.
Turning the first into the second is how a proceeding becomes a conspiracy theory. The honest version is that this is an empirical question, and answering it requires controls:
- Can a price difference be reproduced for genuinely equivalent transactions?
- What data plausibly drives it — and is that an inference or an observation?
- Does the retailer disclose the practice anywhere?
- Can something ordinary explain it instead: geography, local inventory, a membership tier, a device-specific app promotion, a time-limited offer, an A/B test?
The last line is where most amateur experiments fail. Two phones showing two numbers is a screenshot, not a finding, until the ordinary explanations have been excluded.
“Just use private browsing” is not the answer
The FTC does observe that an informed consumer might take protective steps — a private browsing session, a VPN, or simply avoiding a retailer known to price this way. That observation is part of the agency's disclosure argument: a customer who knew could act, and a customer who does not know cannot.
It is not consumer advice that incognito mode gets you the lowest price, and it should not be repeated as if it were. A logged-in account, a loyalty number, a delivery address, a device fingerprint and a location signal all survive a private window. And most price variation you encounter has nothing to do with personalisation in the first place.
The behaviour that actually protects a household is duller and older than any of it: comparison. Check a second retailer. Check logged out as well as logged in. Check the app against the website. Look for whether member pricing explains the gap. On a purchase large enough to matter, record what you were shown and when.
That is not a defence against surveillance. It is the ordinary discipline that makes any pricing practice — personalised, dynamic or plain — less able to cost you money quietly.
What this page has not established
The federal proposal is not the only place this subject is moving. Individual states have legislated on data-driven pricing, and state privacy law plainly touches it — in California, the obvious question is how the California Consumer Privacy Act interacts with a price set from personal data.
None of that is established here. Everything above about the federal proposal rests on documents published by the Commission and read for this page. The state layer rests on commentary, and commentary is a lead rather than a source, so it is named as an open question instead of written up as a finding.
The practical consequence for a Californian shopper is worth stating even so, because it survives the uncertainty: a change in what retailers do may arrive from a state legislature, a state attorney general or a court well before the FTC finalises anything — or instead of it. Watching only the federal docket would be watching one strand of several.
Why this one is bigger than shopping
For two decades consumers have been told that data collection exists to personalise the experience. Mostly that has meant the advertisement, the recommendation, the order of the search results — things people have learned to tolerate, ignore or quietly resent.
Personalized pricing is what happens when the personalised thing is the price. Not what you are shown, but what you are charged.
That is a different proposition, and the difference is economic rather than technical. A household may accept a recommendation engine reading its history and object strongly to the same history setting its terms — and the reason is not confusion about the technology. It is that one costs money and the other does not.
What to watch, and what not to write down yet
Three things hold this record: the proposal on 19 August, a seven-day extension of the comment period, and the comment close on 25 September. After that the Commission decides.
This page will follow the proceeding rather than be quietly rewritten by it. If the Commission finalises the statement, that will be recorded as what it is — a finalised enforcement policy, not a prohibition. If it revises or withdraws it, that gets recorded too.
What will not happen is this page turning into “FTC bans surveillance pricing”, because the Commission has already said in its own proposal that it lacks the authority to do that. A headline that outruns the document is how a reader ends up believing they are protected by something that does not exist.
How to verify this yourself
- The proposed enforcement policy statement itself (19 August 2026) carries the definition, the long-established-norm distinction between markets, and the Commission's own statement about the limits of its authority.
- The FTC's 19 August press release records the 2–0 vote and the original 18 September deadline; the extension notice moves it to 25 September.
- Comments are filed electronically through the docket linked from the Commission’s own page for the proposed statement. The comment period is open to anyone, not only to businesses and their counsel.
- Everything asserted above about the federal proposal comes from those three documents. The state-law paragraph is deliberately framed as an open question, because no state instrument was retrieved for this page.
Where this page and the Commission's own record disagree, the Commission is right. A proposal under active comment is exactly the kind of record that changes.
Also on this desk
- Before you pay — the same verification habit, applied where the other party is a criminal rather than a retailer.
- What an airline owes you when a flight is cancelled — another market where the posted price and the actual entitlement are different objects.
- Working while collecting Social Security.