Where Dynamic Pricing Does and Does Not Belong in Fashion

Not every changing price is the same practice.

A seasonal markdown applied to all shoppers is different from a price that changes with inventory. Both are different again from a system that estimates one individual’s willingness to pay from browsing history, location, device, or inferred characteristics.

Fashion retailers should stop using “dynamic pricing” as a single umbrella term. The decision, data, and disclosure determine whether a pricing system feels like ordinary merchandising or hidden surveillance.

Four distinct models

Calendar pricing

The price changes on a published or internally scheduled date: launch, promotion, end-of-season markdown, or clearance. Everyone who is eligible at that time sees the same offer.

Inventory-responsive pricing

The price changes according to aggregate stock position, sell-through, replenishment outlook, or product age. A retailer may protect the price of a scarce style or mark down excess units.

Demand-responsive pricing

The price changes with aggregate demand, channel conditions, region, or time. This is familiar in travel and ticketing but can feel unstable in fashion, where customers remember reference prices and may compare channels easily.

Personalized or surveillance pricing

The price or promotion changes because the system estimates what a particular person or audience will pay using personal, behavioral, or inferred data. The U.S. Federal Trade Commission used the term “surveillance pricing” in a 2025 staff study of pricing intermediaries. In August 2026, the FTC requested public comment on a proposed enforcement policy statement addressing undisclosed use of personal data for personalized pricing.

The August 2026 statement is a proposal under public comment, not a final blanket prohibition. The FTC also stated that it does not have authority to ban personalized pricing in all circumstances, while warning that undisclosed collection or use of personal data for pricing could implicate laws against unfair or deceptive practices.

Where changing prices can serve fashion customers

Transparent seasonal markdowns

Fashion has short selling windows, seasonal receipts, and products that may not replenish. A clear markdown schedule can move aging inventory and give shoppers an understandable tradeoff: buy early for selection or later for a possible lower price.

Inventory correction

Aggregate inventory signals can support earlier, smaller markdowns instead of a single destructive clearance. The rule should be item-based, documented, and consistent across shoppers in the same market and channel unless a legitimate disclosed condition applies.

Clearly defined member or channel benefits

A loyalty price, wholesale tier, employee benefit, or regional offer can be legitimate when eligibility is explicit and the comparison is not misleading. The ordinary price should remain real and available; a perpetual fake reference price is not transparency.

Operational incentives

Discounts for slower shipping, store pickup, preorder, bundle quantity, or final sale can reflect a genuine cost or inventory difference. The condition and consequence should be visible before the customer commits.

Where it does not belong

Hidden willingness-to-pay estimates

A fashion retailer should not quietly raise a price because a shopper appears affluent, repeatedly viewed the item, uses a particular device, came from a certain neighborhood, or seems unlikely to compare offers. Even when a particular implementation is not categorically prohibited, the trust cost is profound and the regulatory direction is toward more scrutiny and disclosure.

Manipulated scarcity

Inventory-responsive pricing becomes deceptive when paired with false countdowns, fabricated “only one left” messages, or a reference price that was never meaningfully offered. The pricing rule and the urgency claim are separate facts; both must be true.

Discrimination by proxy

Location, browsing behavior, purchase history, and inferred segments can correlate with protected or vulnerable groups. A model may reproduce unequal outcomes without receiving a field labeled with a protected characteristic. Review outcomes, not only inputs.

Price changes the frontline cannot explain

If store staff and customer service cannot reconstruct why a price changed, the system is not operationally ready. “The algorithm decided” is not a customer explanation or an audit trail.

A defensible pricing policy

1. Name the method. Separate calendar, inventory, aggregate demand, eligibility-based, and individualized rules.
2. Limit the data. Do not use personal or behavioral data merely because it is available.
3. Define the customer benefit. Faster stock correction, reduced waste, or a disclosed service tradeoff is clearer than extracting maximum willingness to pay.
4. Set update boundaries. Establish maximum frequency, minimum duration, market scope, and channel consistency rules.
5. Protect the reference price. Record when and where a comparison price was actually offered.
6. Disclose material conditions. Eligibility, timing, fees, and required actions should be difficult to miss.
7. Log every change. Preserve source inputs, rule version, old price, new price, timestamp, and approver.
8. Test outcomes. Look for unexplained differences across customer groups, devices, locations, and channels.
9. Provide recourse. Give staff a documented correction and escalation path.
10. Review the law at publication and launch. Rules differ by jurisdiction and continue to evolve.

The FTC’s business guidance on unfair or deceptive fees says dynamic pricing based on factors such as demand or inventory may be used so long as pricing information is not misleading. OECD work on algorithmic pricing also describes both potential efficiency benefits and competition risks. Neither point makes a particular implementation automatically acceptable.

The fashion-specific test

Before deploying a changing price, ask three questions:

1. Would we be comfortable explaining the exact rule on the product page?
2. Would two shoppers who compare screens understand why their prices differ?
3. Can we reconstruct the price and inputs six months later?

If the answer is no, the retailer does not have a pricing innovation. It has an accountability gap.

Sources and verification

• FTC: Surveillance Pricing Study Indicates Wide Range of Personal Data Used to Set Individualized Consumer Prices, January 17, 2025.
• FTC: Request for comment on proposed personalized-pricing enforcement policy statement, August 2026; proposal status verified August 31, 2026.
• FTC: Rule on Unfair or Deceptive Fees — Frequently Asked Questions, accessed August 31, 2026.
• OECD: Algorithmic Pricing and Competition in G7 Jurisdictions, 2025.