If the Price Changes by Person, Your Marketing Team Needs Receipts.

A pricing system can call something a “personalized offer.” A customer may call it finding out the person beside them paid less for the same thing.

That gap between internal language and customer experience is where marketing gets dragged into a problem it did not design. The campaign promised value. The website collected the data. The pricing engine made the decision. Then the screenshot landed on social.

On August 19, the Federal Trade Commission proposed an enforcement policy statement on personalized pricing. The proposal deserves attention now, before legal sends a 14-page note and everybody discovers the launch diagram was drawn on a napkin.

If a different customer can see a different price, somebody should be able to explain the rule without saying “the model did it.”

What the FTC proposed

The FTC announced a proposed policy statement, not a final rule. The agency said it cannot prohibit personalized pricing in every circumstance. It also said that when people reasonably expect a price to be the same for everyone, a business using personal data to vary that price may need clear, conspicuous disclosure of the personalization, its basis, and the types of data involved.

The draft statement frames undisclosed data collection or use for personalized pricing as a potential unfair or deceptive practice under Section 5 of the FTC Act. Public comments are due September 18, 2026.

This is not legal advice, and the final position could change. It is a very good reason to find out whether your marketing stack is helping choose the price a person sees.

Start by naming the thing correctly

Three pricing practices often get tossed into one bucket. They should not be.

• Public price changes: everyone sees the same price change because of time, inventory, demand, or a published promotion.
• Eligibility-based offers: a discount follows a stated rule, such as membership, student status, contract tier, or a coupon.
• Personalized pricing: the price or promotion changes based on data about a person, household, device, behavior, or inferred willingness to pay.

The label on the project plan does not settle the question. A “loyalty offer” can still be customer-specific. A “dynamic price” can be public. Draw the actual decision rule.

Map every hand touching the price

Bring marketing, ecommerce, data, product, finance, and legal into one room. Then trace a price from base catalog to final screen.

List every input: inventory, time, location, referral source, device, browsing history, cart activity, loyalty profile, purchase history, predicted churn, customer segment, and third-party audience data. For each one, record where it came from, why it is used, how long it is kept, and who approved it.

The FTC’s earlier surveillance-pricing research examined systems capable of using details such as location, demographics, browsing patterns, shopping history, and customer behavior. Your vendor may call those inputs “signals.” Put their plain-English names in the diagram anyway.

If the data map is already a mess, run SigServe’s pixel and privacy audit first. A clever disclosure cannot rescue a team that does not know what it collects.

Write the disclosure before the campaign

Do not hide the pricing explanation in a privacy policy and expect it to carry the whole job. Draft the message where a customer will encounter the price. Make it specific enough to answer three questions:

1. Is this price or offer personalized?
2. What caused it to be personalized?
3. What types of personal data were used?

Then test the disclosure like campaign copy. Can a person see it before committing? Does it survive mobile? Does the message still make sense after a coupon, retargeting click, or loyalty login changes the screen? Is the explanation accurate in every channel where the price appears?

“Prices may vary” is impressively short. It is also useless if the real fact is that browsing and purchase history changed the offer.

Run a paired-customer test

Create controlled test profiles that differ by one input at a time. Visit the same product, at the same time, through the same channel. Record the displayed price, promotion, fees, disclosure, and checkout total.

Test logged-in and logged-out states. Test a fresh browser against a returning shopper. Test common geography and device changes. Check email, paid social, search ads, landing pages, account dashboards, and the final cart.

The goal is not to reverse-engineer every model weight. It is to catch outcomes the team cannot defend. Save screenshots and timestamps. If two profiles receive different prices, the launch owner should know which rule caused it and whether the customer received the right explanation.

Give promotions their own audit

Marketing teams often personalize the discount instead of the list price. Customers still experience the amount they are asked to pay.

Audit abandoned-cart offers, win-back emails, loyalty incentives, retargeting coupons, sales-assist discounts, and creator codes. Look for accidental discrimination, conflicting promises, expired creative, and segments based on data no one remembers approving.

Also check the public claim. “Our best price” becomes difficult to defend when the best price depends on a profile assembled behind the curtain. SigServe’s four-pass creative review gives the claim and proof a separate checkpoint before channel QA.

Build a pricing launch receipt

Before the system goes live, require one short record with:

• the business reason for varying the price or offer;
• the exact rule or model version;
• the personal and non-personal inputs used;
• the customer-facing disclosure and every place it appears;
• the paired-profile test results;
• the legal, pricing, data, and marketing approvers;
• the monitoring metric, complaint route, review date, and rollback owner.

Version the receipt when the data, rule, vendor, message, or channel changes. “Approved last quarter” does not cover the new retargeting signal someone connected on Tuesday.

What to ignore

Ignore the urge to rename personalized pricing until it sounds harmless. Ignore vendor assurances that never name the inputs. Ignore a disclosure that only legal can find. Ignore average conversion lift if complaints, cancellations, or screenshots show the trust cost arriving somewhere else.

And ignore the idea that this belongs entirely to compliance. Marketing owns claims, customer expectations, channel consistency, and the moment a strange price becomes a public story.

The 15-minute prelaunch check

1. Can we state why two customers might see different amounts?
2. Do we know every data type that can influence the price or offer?
3. Is the explanation visible before purchase and accurate on mobile?
4. Did controlled profiles produce the expected outcome?
5. Can one named owner stop the system when the outcome is wrong?

If the team cannot answer all five, keep the campaign in draft. A pricing model should not meet the customer before the people running it have been properly introduced.

About the Author

Brian Mundt

Director of Content Strategy

Brian helps organizations turn customer data, campaign promises, and complicated approval rooms into marketing people can explain without hiding behind the software.