Dynamic pricing in grocery: price the shelf, not the person

On 25 September, Walmart's CEO sent customers a public letter with a four-word pricing policy: "We price the product, not the person." No prices set from income, shopping history or a moment of need. Electronic shelf labels stay, but only to keep the shelf and the till in sync.

It reads like a PR move. It is also the clearest statement yet of where dynamic pricing in grocery is heading. The largest grocer in the US just drew the line that regulators are drawing anyway. Prices can move. They cannot move because of who is looking at them.

For a mid-market grocer, this is not a reason to stop modernizing pricing. It is a design brief. This memo covers what changed, where the legal line now sits, and what a CEO should ask the pricing team for before the next price change goes out.

What Walmart promised, and what it did not

The letter made three commitments. No variable prices based on shopper identity or time of day. No use of data shared with its AI shopping assistant to raise prices. Clear data policies with customer control.

Read it carefully and the scope is narrow. Walmart did not give up price changes. It did not give up electronic shelf labels. It did not give up promotions, rollbacks or member benefits. What it gave up is the idea that the price of a jar of coffee depends on the person holding it.

That distinction is the whole story. "Dynamic pricing" has become a single scare word in the press, but it covers two very different practices:

• Market-driven pricing. The shelf price moves because a cost changed, a competitor moved, stock is ageing, or demand shifted. Everyone who walks in that day sees the same price.
• Person-driven pricing. The price moves because a model estimated what this specific shopper is willing to pay. Two people in the same aisle can see different numbers.

The first is how retail has always worked, just faster. The second is what the public, and now the law, calls surveillance pricing. Walmart has chosen a side. Most grocers will have to as well.

The line regulators are drawing in grocery dynamic pricing

The pressure is not hypothetical. Across 2025, more than 100 price transparency bills were introduced in 33 states and Washington, D.C., many aimed squarely at grocery. In 2026 they started to pass.

• Maryland passed the first state law. Its Protection From Predatory Pricing Act took effect on 1 October 2026. It bars food retailers of 15,000 square feet or more, and third-party delivery services, from using personal data to set higher prices for individual shoppers. Fines reach $10,000 per violation and $25,000 for repeat offenders.
• Connecticut and New Jersey followed. New Jersey's law, signed on 23 July 2026, also pauses electronic shelf labels for a year while the state studies them. It takes effect on 1 August 2027.
• The FTC moved from study to policy. Its January 2025 study found six pricing and data intermediaries serving at least 250 clients, from grocers to apparel, with inputs that included precise location, browsing history and even mouse movements. In August 2026 it proposed an enforcement policy that treats undisclosed personalized pricing as a potentially deceptive practice.

European retailers have lived with a version of this for years. Since the EU's Omnibus Directive took effect in 2022, online sellers must tell shoppers when a price was personalized by automated decision-making, and every advertised price reduction must reference the lowest price of the previous 30 days. CEE markets are adding their own layers, as Croatia's anchor price rule showed this month.

Look at what the new laws leave alone. Maryland carves out promotions, temporary discounts, loyalty programs any shopper can join, and price differences driven by real costs such as delivery. New Jersey explicitly permits discounts, promotions and loyalty programs. None of them ban a grocer from changing a price because the market changed.

So the line is not "static versus dynamic." It is "what is the price based on, and can you prove it."

Dynamic pricing in grocery stores survives if it prices the shelf

Strip away the headlines and there is a large, safe space for store-level dynamic pricing. It is the space where every input describes the product, the store or the market. None of them describe the shopper.

• Cost. A supplier raises list price. The shelf follows, within margin rules.
• Competition. A nearby rival cuts the price on a key value item. You match, or decide not to.
• Freshness and stock. Day-old bread. Yogurt two days from its date. A markdown cadence clears it for everyone in the store.
• Demand patterns. Category elasticity, seasonality, weather, local events. Measured across all shoppers, never one.
• Price image. Keeping key value items sharp so the whole basket feels fair.

This is the work that actually moves margin for a grocer with 20,000 SKUs and a thin pricing team. It is also the work that is easiest to explain to a customer, a journalist or a regulator, because the reason for each price sits outside the shopper.

The risk is not that grocers are secretly building person-level pricing. Most mid-market grocers are nowhere near it. The risk is that they cannot prove they are not. If a reporter asks why the price of milk changed three times last week, "the algorithm did it" is the worst possible answer.

Five guardrails that keep store pricing explainable

Walmart can afford a public letter. A regional grocer needs something more practical: a pricing setup where the answer to "why this price?" is written down before anyone asks. Five guardrails get you most of the way.

1. An approved list of pricing inputs

Write down the signals your pricing is allowed to use: cost, competitor prices, stock, shelf life, category elasticity, store cluster. Then write down what it may never use: anything tied to an identified shopper. If a vendor or an internal model wants a new input, it goes through the same approval as a new supplier. This one page is the most useful document you can hand a regulator.

2. One price per store, per day, per channel rule

Decide how often a shelf price may change and stick to it. One Maryland bill proposed that shelf prices hold for at least one business day. You do not need a law to adopt that discipline. A predictable cadence protects price perception and makes ESLs a labor tool, not a suspicion.

3. Shelf, app and till must agree

If the app shows a logged-in shopper a different price than the shelf, you have built personalized pricing by accident. Any difference between channels should come from a named, published mechanism: a member price, a coupon, a delivery fee. Not from a model.

4. Loyalty prices that anyone can get

Every new law protects loyalty programs, as long as any shopper can join. That makes member pricing the legitimate way to reward your best customers. Keep it transparent: one member price, shown on the shelf, available to everyone who signs up.

5. A log of every price and its reason

Each price change should carry its trigger, the rule that fired, who approved it and when. That audit trail is what turns "the algorithm did it" into "cost rose 6%, our margin floor applied, the category manager approved it on Tuesday." It is also the core of the six-step surveillance pricing audit we published for pricing teams.

The mid-market advantage

National chains carry a trust deficit on pricing. Every algorithm they run is assumed to be the worst version of itself. That is why Walmart needed a public letter, and why commentators already see an opening for smaller retailers.

A regional grocer starts from a different place. Shoppers know the store. Staff know the shoppers. "We price the shelf, not the person" is easier to believe from a business with 40 stores than from one with 4,000. And it is easier to prove, because a mid-market grocer can show its whole pricing logic on a few pages.

The trap is waiting. Grocers who hold off on modern pricing until "the regulation settles" will spend that time pricing on gut feel and last year's spreadsheet, while costs keep moving. The grocers who come out ahead will be the ones who modernize now and build the guardrails in from day one. Then they say so, out loud, to their customers.

What a CEO should ask for in the next 90 days

You do not need a legal review to start. You need four documents from your pricing team.

1. The input list. Every signal that touches a shelf price today, and its source. If the team cannot produce it in a week, that is your first finding.
2. A sample price log. Pick ten SKUs whose price changed last month. For each change, ask for the trigger, the rule, the approver and the date. Gaps here are the gaps a regulator or journalist would find.
3. A channel parity check. Compare shelf, website and app prices for 50 items, logged in and logged out. Every difference should map to a named mechanism.
4. Vendor answers in writing. Ask every pricing, loyalty and e-commerce vendor three questions. What shopper-level data do you receive? Does any of it influence a price? Can you show us the log? A vague answer is an answer.

Then set the policy. Write a one-paragraph pricing principle the board signs off on, in plain language. Something close to: our prices respond to costs, competitors, stock and demand, never to who you are. Put it on the website. It costs nothing and it will age well.

What this does not change

None of this makes grocery pricing easier. It only makes the rules clearer.

Margin pressure is still there. McKinsey's State of Grocery Retail Europe 2026 is subtitled "margins under pressure, models in motion." The FAO food price index climbed to a near four-year high in September. Suppliers are still pushing increases through. The need to price precisely, item by item and store by store, has never been higher.

Static pricing is not the safe option. A price that never moves is still a decision. If it sits above the market, you lose volume. If it sits below cost after an increase, you lose margin. Freezing prices to avoid scrutiny swaps a regulatory risk for a commercial one.

Electronic shelf labels still make sense in most markets. New Jersey has paused them. Most places have not. They cut the labor of price changes and keep shelf and till aligned. The point is to run them on a published cadence, not to rip them out.

Personalization in marketing is not dead. Targeted offers, coupons and loyalty rewards remain legal in every law passed so far. What changes is the bar for transparency. A personal offer is fine. A personal shelf price is not.

Price the shelf, and show your work

Walmart's letter will be remembered as the moment dynamic pricing in grocery split in two. Person-driven pricing is heading toward disclosure rules and bans. Market-driven pricing is heading toward an audit trail. For most grocers, the second path is where the margin was anyway.

The grocers who win the next five years will move prices with the market and explain every move. That is what we build Retailgrid for: rules-based, explainable price optimization where every price carries its reason. If you want to see how a price log like the one above looks on your own data, talk to us. We are happy to walk through it.