Walmart said this week that it would neither use nor renew a 2023 patent describing how the contents of a shopper’s cart could change the prices on its electronic shelf labels. The filing’s example: “if a customer has tuna fish, they may be offered a different price for mayonnaise.”
“The patent has issued, but we will not use it or renew it,” a Walmart spokesperson told Retail Dive. The patent was one of more than a dozen that Groundwork Collaborative, a progressive economic policy group in Washington, laid out on Monday in a report arguing that the $833 billion retailer is building the tools to profile shoppers and set what each one pays.
Walmart answered the same day with an open letter from Dan Bartlett, its executive vice president of corporate affairs, asking Groundwork’s president, Lindsay Owens, to “correct the record.” Common Dreams reported that the exchange came less than two weeks after Walmart’s chief executive, John Furner, promised customers in a Sept. 25 letter that the company would “price the product, not the person.”
Under the quarrel sits an older question, whether big retailers already charge different people different prices for the same thing, and the record answers it plainly. They have. Amazon did it with DVDs in 2000, Staples by ZIP code in 2012, Target through its app in store parking lots in 2019, and a half-dozen grocery chains selling through Instacart did it to hundreds of volunteer shoppers last year.
In nearly every documented case, the price moved with a random experiment or with where the shopper stood, not with a profile of the person. No one has shown Walmart charging one customer more than another. The patents describe what the company could build. Walmart says it uses none of them to set prices, and a Snopes fact check in September found no credible evidence that it does.
The law is moving faster than the evidence. Four states have now passed statutes on prices set with personal data, and Maryland’s took effect on Oct. 1. The Federal Trade Commission closed comments on Sept. 25, the day of Mr. Furner’s letter, on a proposed policy that would require companies to disclose personalized prices but that the agency’s chairman says cannot ban them. Most of the new state laws exempt loyalty programs and targeted discounts, though personal data can shape a discount as easily as a markup.
What the filings describe
The Groundwork report, written by Ms. Owens and Elizabeth Pancotti, the group’s vice president of policy, advocacy and research, gives each filing a nickname. The one it calls “Search, Then Surcharge” is U.S. patent 10,896,433, “Dynamic pricing systems and methods,” granted in January 2021 from an application that traces to 2014. It describes a system that “determines an identity of a user performing the search” and adjusts the listed price using that user’s loyalty, past returns, interests and distance from a store.
In one example, the patent lets electronics prices move 10 percent either way from the merchandiser’s price. Groundwork turned that range into its most quoted illustration: one customer paying $90 and another $110 for the same item. Those figures are the group’s arithmetic, not the patent’s text. The patent is active and, according to its public record, runs until 2035.
Other grants are older or narrower. A 2019 patent describes “probing” shoppers by marketing products priced above what they usually pay and recording how they respond. A related application lays out a five-week price test on more than 20,000 items, with daily swings of up to 10 percent, and says such tests “can harm the user experience.” That test did not use data on individual shoppers, the report says. A patent granted in 2024 estimates the number and ages of children in a household from its purchases and steers larger families toward bulk or cheaper goods.
The published applications reach further. One would compare a shopper’s heart rate or glucose against a sleeping baseline to predict “which flavor the customer is more likely to purchase.” Another would learn a person’s routine well enough to flag “that the person has not arrived at their usual 6PM-Wednesday dance class.” A third would photograph car suspensions in the parking lot to guess how much a shopper bought. Two describe withholding coupons from people likely to pay full price. At a 90 percent chance of purchase, one says, a discount “is not needed.”
Ms. Owens, the group’s president and chief executive, whose book Gouged came out Sept. 29, reads the filings as a statement of intent.
Corporations don’t spend years building high-tech pricing tools, hiring experts, and filing patents just to let them collect dust. They do it because they see a path to higher profit—one that runs right through everyday consumers’ budgets.
Mr. Bartlett’s letter rejects that inference. A patent describes an invention, he wrote, and that “does not mean that Walmart has deployed a capability, used it with customers or decided to do so in the future.” On the tuna example he was blunt: “Putting tuna in a customer’s cart does not cause Walmart to charge that customer more for mayonnaise.” He posted the letter on X with a note addressed to Ms. Owens by her handle.
Public claims deserve a public response. Especially when they’re inaccurate and misleading.@owenslindsay1 has repeatedly suggested Walmart uses or plans to use technology to charge people more based on their personal information.
What Walmart has promised
Mr. Furner, who succeeded Doug McMillon as chief executive in February, made three commitments in his letter: no prices based on who a customer is or the time of day, no use of what shoppers tell Sparky, the company’s A.I. shopping assistant, to raise prices or hide cheaper options, and data sharing that stays voluntary. “Using someone’s income, shopping history or moment of need to charge them more would violate the EDLP promise our business model is built on,” he wrote. “We won’t do it.”
Walmart executives had been fielding the question since spring. Kieran Shanahan, chief operating officer of Walmart U.S., told reporters at the company’s shareholder week in June that prices are set at headquarters and that a single store can receive 2,000 to 3,000 price changes, which the electronic labels spare workers from making by hand, Modern Retail reported. Asked whether Walmart would price by situation, he hedged once. “Never is a long time, so I’m loath to say never, but we don’t do that as a policy,” he said.
Sparky is where the two sides talk past each other. Dave Guggina, the chief executive of Walmart U.S., told investors in September that “when customers engage with Sparky, their average order value jumps 40%.” Ms. Owens cites the figure as evidence that the assistant pushes people to spend more. Mr. Bartlett wrote that bigger baskets do not show that anyone paid more for the same items, and that what customers share “is not used to raise their individual prices.” Both statements can be true. A larger order is a sales result, not a price.
Ms. Owens has asked Walmart to publish what data Sparky can see and how that information shapes what it recommends and buys. The company had not done so when the report came out. She made the case on X the week Mr. Furner’s letter appeared.
Walmart execs spent months bragging about how Sparky is boosting profits by getting customers to spend more.
Worse, the company has spent years developing and patenting technology that is specifically designed to vary prices based on consumers’ behavior and demand.
Groundwork also points to Walmart’s privacy notice, which it says the company updated in August to cover income, ethnicity, retinal imagery, license plates and a shopper’s precise location in a store. A Snopes fact check by Jordan Liles on Sept. 10 found that the Aug. 20 revision added only a section on voice interactions. Archived versions showed biometrics in the notice since 2020, household income since 2023, license plate readers since 2024, and device location and in-store cameras since at least 2012. The permissions are real, and older than the report suggests.
Twenty-five years of tests
In September 2000, Amazon customers comparing notes online found they had paid different prices for the same DVDs. The company said it had run a five-day test on 68 titles with random discounts of 20 to 40 percent, CNN reported. Jeff Bezos, the founder, apologized, said the discounts had nothing to do with customer demographics and called even a random test a mistake. Amazon refunded an average of $3.10 to each of 6,896 customers who had not received the lowest price, Computerworld reported.
In 2012, The Wall Street Journal tested Staples.com from every ZIP code in the country and found prices varying by location for about a third of more than 1,000 products, as Consumerist summarized. A Swingline stapler cost $14.29 in Scarsdale, N.Y., and $15.79 in Manhattan. The strongest predictor was distance to a rival OfficeMax or Office Depot, and the areas that got the discounts tended to have higher average incomes. The same investigation found Orbitz showing pricier hotels to people browsing on Mac computers.
In 2019, KARE 11 in Minneapolis found Target’s app raising prices as shoppers neared a store. A Samsung television listed at $499.99 rose to $599.99 when reporters pulled into the Minnetonka parking lot, and four of 10 randomly chosen items cost more once they were inside. Target said the app showed in-store prices in the store. Seven California district attorneys sued, and in March 2022 Target agreed to pay $5 million under a judgment that bars it from changing app prices based on a user’s location.
Kroger drew the next round. On Aug. 5, 2024, Senators Elizabeth Warren and Bob Casey wrote to the company that its electronic labels could let it “quickly update and display the customer’s maximum willingness to pay,” and asked about cameras it planned with Microsoft for shelf displays. Kroger said the label tests were meant “to lower prices more for customers where it matters most,” and later told Fast Company it had no plans for facial recognition in its stores.
Each of those episodes turned on chance or geography. None documented a price built from one person’s profile.
Instacart’s eggs
The case closest to the fear came from Groundwork itself. On Dec. 9, 2025, it published findings with Consumer Reports and More Perfect Union from a September experiment in which 437 volunteers in four cities put the same items from the same stores into Instacart carts. About three-quarters of the items showed up at more than one price, as many as five for a single product, CBS News reported. The average gap between highest and lowest was 13 percent, and the largest 23 percent. At one Safeway in Washington, a dozen eggs ranged from $3.99 to $4.79.
Whole baskets differed by about 7 percent, which the researchers estimated could cost a family of four $1,200 a year. The chains named included Albertsons, Costco, Kroger, Safeway, Sprouts Farmers Market and Target.
Instacart called the tests “limited, short-term and randomized” and said it did not use personal, demographic or user-level behavioral data to set prices. The F.T.C. said it was “disturbed by what we have read in the press,” and Reuters later reported an agency inquiry. Wegmans stopped using the software on Dec. 10. On Dec. 22, Instacart ended item price tests, writing that they had “missed the mark for some customers,” Consumer Reports reported. It kept tests of promotions and discounts.
New York’s attorney general, Letitia James, demanded documents from Instacart on Jan. 8, saying its disclosure could be “accessed by clicking fine print text” and was not clear and conspicuous. Ms. Owens announced the original findings in a thread on X.
Right now, Instacart is quietly running experiments on millions of us while we shop for groceries online. They are trying to figure out exactly how much they can get away with charging you for breakfast cereal, lunch meat, pasta, and everything in between. How do I know? 1/9
The episode proved that retailers will show different shoppers different prices at the same moment without telling them. It did not show that any of those prices were built from a shopper’s profile.
Shelf labels and the data
Walmart began testing electronic labels at a store in Grapevine, Tex., and in June 2024 announced plans to put them in 2,300 stores by 2026, Grocery Dive reported. It now expects them in every U.S. store by the end of this year, CNBC reported in March. A price change that once took up to a couple of days can now be made in minutes.
The largest test of what the labels do to prices is a working paper by Ioannis Stamatopoulos of the University of Texas at Austin, Robert Evan Sanders of the University of California, San Diego, and Robert Bray of Northwestern University. They examined more than 180 million product-level observations from 114 stores of one U.S. grocery chain in four states that installed the labels from 2022 to 2024. Before the labels, about one product in 20,000 showed a surge-like temporary increase on a given day, according to the university. Afterward, the rate did not change in any meaningful way, and discounts became slightly more common.
We find there is essentially no change in pricing behavior after these digital labels are installed.
The paper, posted on SSRN, has not been peer-reviewed and covers one chain. It also measures something narrower than the current fight. A shelf label shows one price to everyone in the aisle. Personal pricing in a store would have to run through a phone, a loyalty account or a cart scanner. That is the link the tuna patent describes: a Scan & Go cart wired to the label.
Weighing the argument
Ms. Owens’s strongest evidence that the tools exist comes from the government. In January 2025, the F.T.C. released preliminary findings from orders sent to eight pricing intermediaries, drawing on documents from six: Mastercard, Accenture, PROS, Bloomreach, Revionics and McKinsey. Staff found that details “like a person’s precise location or browser history can be frequently used to target individual consumers” with different prices, and that the firms served at least 250 clients, from grocery stores to apparel retailers.
The findings named no retailer and illustrated the point with a hypothetical: a shopper profiled as a new parent being shown pricier baby thermometers. Andrew Ferguson and Melissa Holyoak, the Republican commissioners, dissented. Days after he became chairman, Mr. Ferguson shut the public comment period, and the agency never published a final report.
Walmart’s best evidence is its business model. Ari Lightman, a professor at Carnegie Mellon University, told Retail Dive that the everyday-low-price reputation “doesn’t lend itself well to the idea of dynamic pricing.” The pledges are also public, and under the F.T.C.’s proposed policy, a retailer that suggests its prices are the same for everyone when they are not risks misleading customers. Their limit is wording. Both letters promise not to charge anyone more. Mr. Bartlett’s does not address coupons or personalized offers, and two of the filings Groundwork found describe withholding a discount, not raising a price.
The sharpest objection to the new bans comes from economists who see discounts at stake. Z. John Zhang, a marketing professor at the Wharton School and the only skeptic among five witnesses at an Aug. 4 Senate hearing, argued that personalized pricing can lower costs for price-sensitive customers and urged lawmakers to go after abuses of data and market power instead, according to a summary of the hearing. Brian Albrecht, chief economist of the International Center for Law and Economics, wrote in a column first published in the Financial Times that the campaign had “an odd target in its sights: discounts.”
Gov. Jared Polis of Colorado used the same reasoning when he vetoed his state’s ban on June 2, writing that it “would punish differentially lower prices, not just higher prices.” Groundwork’s answer is that bargain hunters hold prices down for everyone, and that personalization lets a store give discounts only to them while charging the rest more. The Staples case shows how both can be right at once. The discounts were real, and they went to places near competitors, which tended to be wealthier, while shoppers with fewer options paid more.
Phil Lempert, a grocery industry analyst, wrote in The Robin Report on Wednesday that “nobody has documented Walmart charging one shopper more than another,” gave the company credit for dropping the tuna patent and called that patent’s example “oddly specific for a defensive filing.” He also wrote that the Instacart tests were randomized, not tailored to income, ZIP code or shopping history. His conclusion turned on how long a pledge lasts.
I believe John Furner. I also think his promise will only be worth exactly as much as the next CEO decides.
What the law covers
New York moved first, with disclosure. Its Algorithmic Pricing Disclosure Act, in effect since Nov. 10, 2025, requires a notice reading “This price was set by an algorithm using your personal data.” The National Retail Federation sued, and Judge Jed S. Rakoff of Federal District Court in Manhattan dismissed the case on Oct. 8, 2025, finding the notice factual. The federation appealed to the Second Circuit. Penalties run to $1,000 per violation, and the attorney general had brought no formal enforcement action as of June, Arnold & Porter wrote.
Maryland’s Protection From Predatory Pricing Act, signed by Gov. Wes Moore on April 28 and in force since Oct. 1, is the first outright ban. It bars food retailers with stores of at least 15,000 square feet, and delivery services, from setting a personalized price from a shopper’s data. Only the attorney general can enforce it, after a 45-day chance to fix the problem, with fines of up to $10,000 per violation and $25,000 for repeats. Loyalty programs and targeted discounts are exempt, and Consumer Reports said the industry’s changes had left the law unlikely to protect Marylanders.
Connecticut’s law, signed June 4, reaches all retailers from July 1, 2027, and requires other businesses using price-setting software online to post “THIS PRICE WAS INCREASED BY A PRICE SETTING DEVICE USING YOUR PERSONAL DATA.” New Jersey’s, signed by Gov. Mikie Sherrill on July 23, covers groceries and household necessities and lets shoppers sue, the first such right in the country, according to Skadden. It also halts new electronic shelf label installations for a year starting Feb. 1, 2027.
New York’s Legislature passed a ban, the One Fair Price Act, in June. Gov. Kathy Hochul has until Dec. 31 to act on it, and business groups have urged her to change it. Colorado’s ban ended with Mr. Polis’s veto.
In Washington, the F.T.C. voted 2-0 on Aug. 19 to propose an enforcement policy statement under which collecting or using personal data for personalized pricing without disclosure could violate the law against unfair or deceptive practices. The agency extended the comment period by a week, to Sept. 25. Mr. Ferguson said the agency lacked the authority to ban personalized pricing “in all circumstances.” His case for the policy rested on what shoppers assume.
When consumers see a listed price, they expect it to be same price that everyone else sees
Congress has held one hearing. Senator Josh Hawley, Republican of Missouri, chaired the Aug. 4 session of a Senate Judiciary subcommittee at which Ms. Owens testified, and Senator Richard Blumenthal, Democrat of Connecticut, called for federal standards. Senator Ruben Gallego, Democrat of Arizona, introduced a federal One Fair Price Act on Dec. 9, 2025.

What comes next
The F.T.C. has set no date for a final policy statement. Ms. Hochul’s deadline on the New York ban is Dec. 31, and the retail federation’s appeal of the disclosure ruling was still pending at midyear. New Jersey’s label moratorium begins Feb. 1, 2027, and Connecticut’s ban on July 1, 2027. Walmart expects electronic labels in every U.S. store by the end of this year.
Walmart’s no-renewal pledge applies only to the tuna patent. Its adjusted expiration date is April 2039.

