The Federal Trade Commission released a proposed enforcement policy on August 19, 2026 that targets undisclosed personalized pricing. Public comments are due September 18, giving ecommerce operators a reason to examine how their pricing and promotion tools use customer data now.

This is analysis of a proposal, not a new final rule. The FTC has not banned personalized pricing, and the statement does not create an effective date for a disclosure requirement. It does explain how the agency may use existing Section 5 authority against practices it considers deceptive or unfair.

The proposal focuses on prices built from personal data

The FTC’s eight-page draft defines the concern as using consumer data and inferences to set individualized prices. Examples include browsing history, location, previous purchases and an estimate of how willing a particular person is to pay.

Where shoppers reasonably expect a generally available price, the agency says a business using personalized pricing should clearly disclose three things: that the price is personalized, the basis for the personalization and the types of data used. A vague “specially selected” label may not be enough.

The FTC also says it has not taken a position on whether some fully disclosed personalized-pricing practices could still be unfair. That unresolved question matters. Good disclosure may address the proposal’s central deception concern without guaranteeing that every pricing model is safe.

Dynamic pricing is not automatically personalized pricing

Online prices change for many reasons. A merchant may adjust a price for everyone because inventory is low, a promotion begins or competitors move. That is dynamic pricing, but it is not necessarily personalized around one shopper’s data.

The risk rises when two people see different prices for the same product at the same time because the system knows something about each person. The proposal specifically points to personal circumstances, behavior and inferred willingness to pay.

Segment-wide promotions require a closer look. A published student discount or loyalty price is different from secretly raising a price for a shopper whose browsing behavior suggests urgency. A coupon sent to lapsed customers may still use purchase history, but it also looks different from changing the base product price without telling the customer.

The draft does not draw a bright line around every discount. That is why the Ecommerce Innovation Alliance warned that the proposal could reach targeted discounts and loyalty pricing and urged brands to examine how offers are generated before the comment window closes.

Your software vendor does not own the customer-facing risk

Many stores do not describe their systems as personalized pricing. They use lifecycle marketing, conversion optimization, loyalty, experimentation or AI merchandising tools. The label on the dashboard matters less than the actual output.

Ask each vendor whether its software changes the price, creates a customer-specific discount or only changes product order and recommendations. Instacart’s AI cart-building system shows how customer history can shape product selection, but personalized recommendations are not automatically personalized prices. The distinction should be documented.

Map the inputs used for every customer-specific offer. Include account history, cookies, location, device data, referral source, loyalty status, abandoned carts and third-party audience attributes. Determine whether the customer consented to each use, especially when data collected for advertising or analytics is reused to set a price.

Then compare what two shoppers can actually see. Test logged-in and logged-out sessions, new and returning customers, locations and devices. Keep screenshots and system logs that connect the displayed price to the rule that produced it.

Disclosure must appear where the price decision happens

A privacy policy that says data may be used to “personalize your experience” may not explain that the actual price can change. The FTC proposal calls for clear and conspicuous disclosure of the pricing practice, its basis and the data types involved.

That suggests the notice should sit near the offer, before the shopper commits, rather than in a long terms page. Legal counsel should review the final language and placement. This article is operational analysis, not legal advice.

Pricing transparency also has a margin cost. If a personalized coupon becomes broader or easier to avoid, conversion and contribution can change. EcomCrew’s deal-profitability calculation for Prime Big Deal Days provides a useful template: calculate contribution per order and the extra volume required before expanding any discount.

Audit now, even though the policy is not final

Create an inventory of every system that can affect the price or discount shown to a shopper. Assign an owner, record the data inputs, identify the business rule and capture the current disclosure. Pause any practice that your team cannot explain consistently.

Preserve the ability to reconstruct an offer after a complaint. You should be able to show the price, discount, data category, model or rule version, disclosure and customer consent that applied at that time.

The next event to monitor is the FTC’s response to comments and whether it finalizes, revises or withdraws the statement. State laws may impose separate duties before then. For sellers, the practical goal is simple: know when two customers can receive different prices, why that happens and exactly what each customer is told.