The Federal Trade Commission has moved its personalized pricing work out of the consultation phase and into the investigative one. In an interview published on October 5, 2026, FTC Chairman Andrew Ferguson said the agency already has active enforcement investigations into the practice and is preparing compulsory information orders under Section 6(b) of the FTC Act.
The detail that matters most for retailers is which industries he named. Ferguson pointed to delivery, rideshare and air travel as the sectors of greatest personal concern, according to a client alert published by law firm Wiley and a separate analysis in the National Law Review. General merchandise and grocery retail were not on that list.
The timing is pointed. One day earlier, the advocacy group Groundwork Collaborative published a report cataloguing Walmart pricing patents, and on October 5 and 6 Walmart answered with an open letter rejecting the characterization and surrendering one of the patents at the center of it.
In short
- The FTC is preparing Section 6(b) orders on personalized pricing, and Chairman Ferguson says enforcement investigations are already running.
- Delivery, rideshare and air travel are the named first targets, not supermarkets or general merchandise, which shifts near-term legal risk away from traditional retail.
- The comment record closed on September 25, 2026 with more than 3,700 filings on the August 19 proposed enforcement policy statement, docket FTC-2026-1057.
- Walmart will not renew patent US11687872B2, the cart-aware shelf label pricing patent, after Groundwork Collaborative published its patent review.
- State law is already binding: Maryland and Connecticut rules took effect on October 1, 2026, with New Jersey and Seattle arriving in 2027.
What the FTC actually said this week
Ferguson’s remarks came in a public interview rather than a formal Commission announcement, which is an important distinction for anyone trying to calendar a compliance deadline. No 6(b) orders have been published, no recipients have been named, and no Federal Register notice has set a response date.
What he confirmed is intent and sequencing. The FTC is drafting orders to compel information for the purpose of studying personalized pricing, and separately has enforcement investigations underway. Those are two different legal tracks that the agency can run in parallel.
Ferguson also made a jurisdictional point that explains part of his target list. The FTC does not hold consumer protection enforcement authority over airlines, which sits with the Department of Transportation, but it does retain authority to compel information from them for industry studies. That asymmetry lets the Commission map a sector it cannot directly police.
For retailers, the practical read is that the agency is building an evidentiary record before it writes final policy. A study is not a rule, and a 6(b) order is not a complaint.
The sequencing is still unusual. Agencies more often finish a guidance document before opening a study on the same subject, and running both at once leaves companies guessing which standard they will eventually be measured against.
How this differs from the 2024 study
The Commission has been here once before. In July 2024 the FTC voted unanimously to open a 6(b) study into personalized pricing, issuing orders to eight companies that supply the technology enabling individualized prices rather than to the retailers deploying it.
Staff released preliminary research summaries on January 17, 2025, in the final days of the previous administration. Two Republican commissioners, including Ferguson himself, dissented on the ground that the majority was publishing observations prematurely and without adequate analysis.
That dissent is now being read differently in light of this week’s comments. Objecting to how findings were released is not the same as objecting to the inquiry, and Ferguson is restarting it on his own terms with his own target sectors.
Why delivery, rideshare and air travel go first
The three named sectors share a structural feature that supermarkets largely lack: the price a customer sees is generated at the moment of request, by an algorithm, for that specific transaction. A rideshare fare, a delivery fee and an airfare are all quoted individually.
That makes the counterfactual tractable. An investigator can ask what two users searching the same route at the same second were quoted, and whether the difference tracks anything about the user rather than the trip. In a grocery aisle the posted shelf price is the same for everyone standing in front of it.
It also means the data trail already exists in the defendant’s own systems. These companies log every quote they generate, which is exactly the kind of record a 6(b) order can demand without the agency needing to run its own mystery shopping program.
Retail has absorbed most of the political attention on this issue, and most of the legislative output, as the spread of state surveillance pricing bans across grocery and food retail shows. Federal investigative attention is now pointing somewhere else.
The gap between political salience and legal exposure
Grocery prices are the most politically salient prices in the country, which is why Maryland wrote the first state ban around food retailers specifically. Legal exposure follows evidence, not salience.
Physical retailers have a defense that platforms do not: the posted price is public, visible to every shopper simultaneously, and in most states must be honored at the register. Personalization in that channel happens through coupons and loyalty offers, which are long established and which the FTC’s own proposal treats more leniently.
Platform pricing has no public posted price at all. Each quote is private to the recipient, which removes the single most effective check on discriminatory pricing and makes the practice far harder for consumers to detect unaided.
What a Section 6(b) order requires
Section 6(b) of the FTC Act lets the Commission compel companies to file special reports answering detailed questions, outside the context of any law enforcement action. Recipients are not accused of anything, which is a point companies frequently have to explain to their boards.
Practitioners advise treating a 6(b) order with the same seriousness as a subpoena or a civil investigative demand. A recipient must respond substantively or petition the Commission to limit or quash the order, and the production burden is routinely measured in months rather than weeks.
The questions in a pricing study would reach the architecture of the pricing stack. That typically means the input variables, the model features, the data vendors, the retention periods and the internal testing records showing what the company learned about how customers responded.
| Instrument | Legal basis | Accusation implied | Typical output |
|---|---|---|---|
| Section 6(b) order | FTC Act Section 6(b) | None | Public study or staff report |
| Civil investigative demand | FTC Act Section 20 | Possible violation under review | Closing letter, consent order or complaint |
| Enforcement policy statement | FTC Act Section 5 interpretation | None | Guidance on how staff will read the law |
| Section 5 complaint | FTC Act Section 5 | Yes | Litigation or negotiated consent order |
Why the study matters even without a rule
A 6(b) study produces a public report, and those reports have a long history of shaping the law without ever becoming one. They supply the factual predicate that later complaints cite, and they give state legislators drafting language something concrete to point at.
The 2024 study already demonstrated this effect. Its preliminary summaries circulated through state capitols during the 2025 and 2026 sessions and are visible in the definitional language several states adopted.
Walmart’s open letter and the patent it gave up
Walmart’s executive vice president of corporate affairs, Dan Bartlett, published an open letter addressed to Groundwork Collaborative president and chief executive Lindsay Owens. Retail Dive reported it on October 6, and The Shelby Report and the Washington Times covered it the same cycle.
The letter’s central line restates the formula Walmart chief executive John Furner used in September: “We price products, not people.” Bartlett wrote that Walmart does not and will not use an individual customer’s personal information, income, shopping history, urgency or willingness to pay to set an individualized price.
On electronic shelf labels, Bartlett argued the devices simply replace paper tags and display approved price changes more efficiently. The company has been rolling the hardware out across its US estate, and the economics of electronic shelf labels in grocery turn on labor savings from price changes rather than on any personalization capability.
The concession came on the patents. Addressing US11687872B2, the 2023 grant describing price adjustments via shelf labels based on what a shopper has already scanned, Walmart said the patent has issued but that it will not use it or renew it.
What the patent actually claims
The patent describes a system in which displayed prices change in response to the contents of a shopper’s cart as items are scanned. The example that has traveled furthest is a customer who scans tuna and is then shown a different price for mayonnaise.
Bartlett addressed that example directly, writing that putting tuna in a customer’s cart does not cause Walmart to charge that customer more for mayonnaise. The adjusted expiration date on the patent runs to April 2039, so the commitment not to renew forgoes more than a decade of protection.
Abandoning a patent is a meaningful signal because it is costly and public. It is also narrow: a pledge not to use one patent is not a pledge not to use the dozen others the report identifies.
What Walmart did not concede
The letter did not retire any other patent, and it did not dispute that the filings exist or that they describe what Groundwork says they describe. It disputed the inference that filing a patent indicates an intention to deploy it.
On Sparky, the company’s shopping assistant, Walmart said that higher average order values do not demonstrate personalized pricing and that information customers share with the assistant is not used to raise their individual prices. Bartlett closed by writing that the company welcomes scrutiny, but that criticism must be grounded in fact, not conjecture.
What the Groundwork report alleges
The report, titled “The Walmart Watchtower: Patents Speak Louder Than Pledges” and written by Lindsay Owens and Elizabeth Pancotti, was published on October 5, 2026. Its method is to read Walmart’s patent portfolio as a statement of capability and intent.
It catalogues filings covering price sensitivity profiling, discount targeting that routes larger discounts to value-conscious shoppers, coupon withholding from customers judged likely to buy at full price, and a price experimentation system the authors say tested daily swings of up to 10% across more than 20,000 items over five weeks.
Other filings in the catalogue reach further from pricing. They include systems for estimating the ages of children in a household from purchase patterns, inferring preferences from physiological signals, and profiling vehicles in parking lots.
The authors’ core argument is stated plainly: corporations do not spend the time and money, or hire highly specialized staff, to develop and patent technology only to file it away on a hard drive never to enter a store. That is an inference about intent, and it is the precise inference Walmart’s letter rejects.
| Patent | Year granted | What it describes | Walmart’s position |
|---|---|---|---|
| US11687872B2 | 2023 | Shelf label prices responding to scanned cart contents | Will not use or renew |
| US10896433B2 | 2021 | Prices set using customer profiles, purchase history and proximity | Not individually addressed |
| US10430817B2 | 2019 | Testing willingness to pay above historically absorbed levels | Not individually addressed |
| US12154159B2 | 2024 | Estimating children’s ages from purchase patterns | Not individually addressed |
The Sparky numbers
Two figures from the report have carried into the wider coverage. Groundwork says average order value rises by roughly 40% when customers engage with Sparky, and that about half of Walmart’s tens of millions of monthly app users interact with it.
Walmart does not dispute the engagement effect. It disputes the causal story attached to it, arguing that a higher basket reflects a better shopping experience and more items found, not a higher price charged on the same item.
Both readings are consistent with the published number, which is why the argument has not resolved. Distinguishing them requires the item-level price data that neither side has put on the table, and that a 6(b) order could compel.
This is the pattern across the whole dispute. Nearly every public claim about personalized pricing rests on capability evidence, such as patents, job postings and vendor contracts, rather than on observed prices paid by identifiable shoppers.
Capability evidence is persuasive in a press cycle and weak in a courtroom. The gap between those two standards explains why the FTC wants compulsory process before it commits to a legal theory.
How the policy statement would change disclosure duties
The proposed enforcement policy statement issued on August 19, 2026 is a disclosure instrument, not a prohibition. It says businesses may violate Section 5 of the FTC Act if they fail to explain clearly when a price is personalized, why it is personalized, and what data was used to personalize it.
The reasoning rests on consumer expectations. The Commission’s position is that shoppers expect prices to move with supply, demand and local market conditions, but do not expect a posted price to reflect an algorithm’s estimate of their individual willingness to pay.
Where that expectation of uniform pricing exists, the proposal would require affirmative disclosure. Where it does not, the duty is lighter, which is why loyalty discounts and advertised promotions sit more comfortably under the framework than silent individualized markups.
The comment period ran to September 25, 2026 after a seven-day extension, drawing more than 3,700 filings. When the FTC personalized pricing docket closed, industry commenters were already signalling vagueness and compelled-speech objections that point toward litigation over any final text.
The over-notification problem
Advertising and retail trade groups raised a practical objection that has not gone away. If the disclosure duty is drawn broadly enough to capture ordinary targeted promotions, every retailer ends up posting a personalization notice on nearly every price.
Notices that appear everywhere convey nothing, and the commenters argued this would degrade rather than improve consumer understanding. The counterargument from consumer groups is that the remedy is a narrower definition, not a weaker duty.
Ferguson’s sector list is informative here. Starting with industries where quotes are individually generated sidesteps the hardest definitional questions about loyalty pricing in physical retail.
The state and city patchwork retailers must already meet
While the federal framework remains proposed, state law has been in force for some time. Four states have enacted statutes with staggered effective dates, and two of them started this month.
Maryland’s Protection from Predatory Pricing Act and Connecticut’s disclosure requirement both took effect on October 1, 2026. Maryland’s is narrower, reaching food retailers and third-party delivery providers, while Connecticut’s covers retail sellers more generally.
New York’s Algorithmic Pricing Disclosure Act has been live since November 10, 2025, making it the longest-running of the group. New Jersey’s Fair Price Protection Act follows on August 1, 2027.
Cities have now joined. Seattle approved an ordinance covering grocers with 20 or more global locations that takes effect on September 1, 2027, and New York City is weighing a broader measure reaching most industries. The Seattle surveillance pricing vote established the template other councils are now copying.
| Jurisdiction | Measure | Effective date | Scope |
|---|---|---|---|
| New York State | Algorithmic Pricing Disclosure Act | November 10, 2025 | Disclosure when personal data sets a price |
| Maryland | Protection from Predatory Pricing Act | October 1, 2026 | Food retailers and delivery providers |
| Connecticut | Surveillance pricing disclosure law | October 1, 2026 | Retail sellers and delivery services |
| New Jersey | Fair Price Protection Act | August 1, 2027 | Broad consumer pricing |
| Seattle | Fair pricing ordinance | September 1, 2027 | Grocers with 20+ global locations |
| New York City | Pending surveillance pricing bill | Not set | Most industries, limited exceptions |
Why the patchwork favors a single national standard
Industry commenters have spent the year arguing that inconsistent state definitions make compliance impossible for any retailer operating nationally. A chain cannot run one pricing engine in Maryland and another in Virginia without rebuilding its stack around state borders.
That argument cuts in an unexpected direction. It gives large retailers a reason to prefer a federal disclosure standard, even a demanding one, over a growing set of divergent state prohibitions.
Roughly 40 bills were introduced across more than 24 states during 2026. If even a fraction pass in the 2027 sessions, the compliance case for federal preemption strengthens considerably.
What an enforcement case would have to prove
Ferguson’s reference to active enforcement investigations is the more consequential half of his remarks, and the harder one to act on. An investigation can become a Section 5 complaint, which carries remedies that a study does not.
Section 5 reaches both deceptive and unfair practices, and the two theories demand different proof. A deception case asks whether the company told consumers something misleading, including by omission, that a reasonable consumer would find material to a purchase.
That theory is the natural fit for undisclosed personalization, and it is the one the August policy statement is built around. If a company says nothing while quoting individualized prices, the omission itself becomes the violation, which removes any need to prove the pricing caused measurable harm.
An unfairness case is harder. Staff would need to show substantial consumer injury that shoppers could not reasonably avoid and that is not outweighed by countervailing benefits to consumers or competition.
Why disclosure is the Commission’s preferred route
Proving substantial injury from personalized pricing requires establishing the counterfactual price, which means showing what the consumer would have paid absent the personalization. In a market where every quote is individualized, that baseline has to be reconstructed from the company’s own model.
Deception sidesteps the problem entirely. The Commission does not have to decide whether personalized pricing is good or bad for consumers, only whether the company was candid about doing it.
This is why the proposal reads as a disclosure instrument rather than a ban, and why industry’s objections focus on vagueness and compelled speech rather than on the underlying theory. The disputed question is not whether the FTC can require disclosure, but how precisely it must define what triggers the duty.
The evidence a case would rest on
Internal experimentation records are the most dangerous category of document in this area. A/B tests that measure how much more a segment will pay establish both capability and knowledge in a single artifact.
Groundwork’s citation of a price experimentation system testing daily swings across more than 20,000 items illustrates the point, even though a patent filing is not evidence of deployment. Investigators reading that filing now know what question to ask, and a 6(b) order is the instrument for asking it.
Vendor contracts matter for a second reason. The 2024 study issued its orders to intermediaries supplying personalization technology rather than to retailers, which gave the Commission visibility into how the tools work across many clients at once without litigating against any of them.
What this means for retail pricing teams
The near-term federal risk for grocery and general merchandise has decreased, not increased. The sectors Ferguson named are elsewhere, and the policy statement remains unfinalized and contested.
The near-term state risk has increased and is already live. Two statutes took effect on October 1, and they bind conduct now regardless of what the FTC does next.
The medium-term risk is documentary. Walmart’s week demonstrates that a patent portfolio, which most retailers treat as a defensive asset managed by counsel, can be read publicly as a statement of intent and can force a chief executive into a position.
There is also a reputational asymmetry worth noting. Senators have already asked the FTC to examine how retail AI assistants surface products, as the request to probe Amazon and Walmart AI over domestic goods showed, and pricing allegations attach to the same systems.
| Sector | FTC enforcement authority | Named by Ferguson | Near-term exposure |
|---|---|---|---|
| Rideshare | Yes | Yes | High |
| Delivery platforms | Yes | Yes | High |
| Air travel | No, study authority only | Yes | Study exposure, no FTC enforcement |
| Grocery | Yes | No | State law, not federal |
| General merchandise | Yes | No | Reputational and state law |
Three questions worth answering internally
First, can the business produce, on demand, the list of variables its pricing engine consumes? A 6(b) response requires this, and firms that cannot assemble it quickly tend to discover the gap under deadline.
Second, does any patent in the portfolio describe conduct the company has publicly disavowed? That gap is what Groundwork exploited, and it is discoverable by anyone with a patent database subscription.
Third, does the loyalty program’s offer logic distinguish between rewarding a customer and charging a customer more? The FTC’s framework treats those differently, and so do the state statutes now in force.
What to watch next
The first checkpoint is publication of the 6(b) orders themselves. Until recipients are named and a Federal Register notice sets response dates, the sector list rests on an interview rather than a document.
The second is whether the Commission finalizes the August policy statement, modifies it in response to the 3,700 comments, or leaves it pending while the study runs. Leaving it pending would be the quieter path and would defer the compelled-speech fight that commenters have promised.
The third is the 2027 state legislative sessions, which open in January. New Jersey and Seattle rules take effect that year, and the bills that did not pass in 2026 return with the FTC’s own record to cite. Readers can track the Commission’s published materials on the proposal at the FTC’s announcement page.
The fourth is whether other retailers follow Walmart’s patent concession. One company abandoning one patent under public pressure sets a precedent that advocacy groups will test against the rest of the sector.
For now the position is stable but unsettled. Federal policy is proposed rather than final, federal investigations are confirmed but unnamed, and the only rules binding US retailers today are state ones that took effect this month.
Retailers reading this as a reprieve should note what changed and what did not. The FTC moved its attention to platform pricing, but it also moved from consultation to compulsory process, and the record it builds will outlast the current sector list.
Frequently asked questions
Is personalized pricing illegal in the United States?
Not as a general matter under federal law. The FTC’s August 2026 proposal would treat undisclosed personalized pricing as a potential violation of Section 5 of the FTC Act, but it remains a proposal. Several states, including Maryland and Connecticut as of October 1, 2026, now restrict or require disclosure of the practice.
What is a Section 6(b) order?
It is a compulsory order requiring a company to file a special report answering the FTC’s questions, issued under Section 6(b) of the FTC Act. It carries no accusation of wrongdoing and is used to study an industry or practice. Recipients must respond or petition the Commission to limit or quash it.
Which industries did Chairman Ferguson name?
Delivery, rideshare and air travel, described in an interview published October 5, 2026 as the industries of greatest concern to him personally. Grocery and general merchandise retail were not named. The FTC lacks consumer protection enforcement authority over airlines but retains authority to compel information from them for studies.
What did Walmart actually concede?
Walmart said it will not use or renew patent US11687872B2, the 2023 grant describing shelf label prices that respond to scanned cart contents. It did not retire other patents in its portfolio. Executive vice president Dan Bartlett restated that Walmart prices products, not people.
Do electronic shelf labels enable personalized pricing?
The hardware can change a displayed price quickly, but it displays one price to every shopper at the shelf. Walmart’s position is that the labels replace paper tags and display approved changes more efficiently. Critics argue the same infrastructure would be required for any future personalization, which is why several state bills address the devices directly.
How many comments did the FTC proposal receive?
More than 3,700, filed to docket FTC-2026-1057 before the record closed on September 25, 2026. The deadline had been extended by seven days from September 18. Commenters included advertising trade groups, technology associations, consumer organizations and retail trade bodies.
What is the difference between dynamic pricing and surveillance pricing?
Dynamic pricing adjusts a price for everyone based on supply, demand, inventory or time, and is long established and lawful. Surveillance or personalized pricing sets a different price for a specific individual based on that person’s data. State statutes generally target the second and exempt the first.
Does a loyalty discount count as personalized pricing?
Under the FTC’s proposed framework, offering a lower price through a disclosed loyalty program sits more comfortably than an undisclosed individualized markup. Several commenters asked the Commission to write an explicit carve-out for loyalty and discount programs. The final treatment is unresolved because the statement has not been finalized.
When will the FTC decide?
No date has been set. The Commission can finalize the policy statement, revise it, or leave it pending while the new 6(b) study proceeds. The study itself would typically take many months from the issuance of orders to any published report.