The seller on the screen may not have designed the system behind the offer. It may purchase software, advice or data from another business. This matters because the explanation a customer needs can be spread across several organizations.
One company may collect information. Another may combine it with other records. A third may recommend a price or choose an offer. The retailer may set the final constraints and display the result. This is an analytical description of possible roles, not a claim that every transaction passes through that exact chain. The investigation has to establish the actual chain in each case.
What the FTC documents establish
The FTC’s January 2025 interim staff perspective examined documents from six pricing intermediaries. Its coverage included tools for store-level pricing, customer groups, personalized offers and product ranking. The respondents collectively served at least 250 clients selling to consumers or other businesses. That number does not mean 250 retailers were proved to charge individualized prices for identical products. The report describes several kinds of input, including observed and inferred information, but does not establish that every input determined a final price. It makes neither a representative national prevalence estimate nor a finding of illegality. Some respondents disputed its descriptions. FTC staff perspective, pp. 1–3 and 7–9
Those limits do not make the documents unimportant. They make their contribution specific. They provide a reason to examine a real commercial infrastructure, including its objectives and division of responsibility. They do not license us to invent a victim, a transaction or an undisclosed practice by a named retailer.
The dissent by Commissioners Andrew Ferguson and Melissa Holyoak objected to releasing early impressions and hypothetical examples before a more complete account. It supported investigation while criticizing the presentation of preliminary evidence. That objection should be considered on its methodological terms. It does not demonstrate that the practices are absent, just as the majority’s concern does not determine their total cost. Commissioners’ dissent
A supplier’s claim is a claim
If a supplier says its system increases revenue, a buyer of that system should ask how the result was measured. Compared with what? Over which period? Was the comparison randomized? Were product availability, demand and other changes accounted for? Did the result persist after customers adapted?
There is an equally important consumer question: where did that extra revenue come from? It might arise from selling a better-matched product, preventing empty capacity or making a service easier to use. It might arise from extracting more money from existing buyers. These are different mechanisms, with different public implications. Revenue growth alone cannot identify them.
A serious evaluation should therefore report several outcomes. Revenue and profit matter for commercial viability. Final paid prices, access, product quality and customer retention matter too. For public policy, distribution matters: a gain spread across many people can coexist with concentrated losses. For privacy, the relevant cost can include a use of information that a person reasonably objects to even if the resulting price is lower.
This does not require us to believe that firms never care about customers. It requires us to stop treating a firm’s chosen performance measure as a complete account of everyone affected.
Where responsibility should sit
A retailer can outsource calculation. It should not be able to outsource its entire explanation.
That is my proposed standard. If a business chooses a pricing service, it should be able to describe the kinds of inputs the service uses, the actions those inputs can change and the constraints placed on them. It should have a way to investigate a credible complaint. The supplier, in turn, should provide records sufficient for that investigation without exposing another customer’s private information.
There are real trade-offs. Detailed records can increase storage and security burdens. Publishing exact models might expose commercial secrets or allow gaming. Small businesses may lack specialist staff. A workable policy should specify proportionate records and confidential regulatory access rather than demand that every shop publish its entire system.
But difficulty is not an answer to responsibility. When a company benefits from a system’s decisions, it is reasonable to ask it to maintain enough knowledge to answer for them. Otherwise the customer faces a peculiar arrangement: every organization in the chain can point to another, while the final bill remains perfectly enforceable against the buyer.
One practical test is to follow a disputed offer backward. Can the retailer identify the product and terms, the relevant time, the pricing rule version and the broad category of personalization? Can it say whether a discount was conditional on membership or another criterion? Can an authorized reviewer compare treatment across similarly situated customers? If these questions cannot be answered, the system is difficult to audit even when everyone involved insists it works well.
Current federal proposal, carefully stated
As checked on October 9, 2026, the FTC’s official record still describes its personalized-pricing enforcement policy statement as proposed. The comment deadline was extended to September 25. The proposal discusses undisclosed personalization under existing Section 5 standards and calls for meaningful disclosure of personalization, its basis and data types. It is not a blanket federal ban or an enacted new pricing rule. Its hypothetical examples are illustrations, not findings against named firms. The proposal also leaves open whether disclosure alone would cure every possible unfairness. This brief does not survey all state laws or decide the legality of a particular offer. Official proposal, deadline extension
The proposal brings us back to a question a supplier presentation may leave out: what does the customer understand at the moment of choice? A notice could reveal a practice while doing little to change the terms. That is a reason to evaluate the notice, the available alternatives and the underlying rule together.
My preferred direction is accountable use of information. A useful tool should survive clear questions about how it works and whom it serves. If the commercial case depends on buyers never understanding the bargain, that is a reason for closer scrutiny, not a measure of innovation we should celebrate.