The Same Cart, Charged Differently: How Instacart Turned Groceries Into a Surveillance-Pricing Lab

Dek: You already knew inflation was real. What you didn’t know is that the grocery store has decided – using AI and your purchase history – exactly how much you’re willing to pay. For each other person in the checkout line, the answer is different.

---

We all know the excuses for the grocery bill. Inflation. Supply chains. Tariffs. For years they’ve been the comfortable story – a 30% jump in the price of your groceries since 2020 explained away by forces “nobody controls.”

But what if something else is happening? Something intentional. What if one person is charged $2 for a carton of eggs, and the person next to them in the same store at the same moment is charged $2.40 – not because of any shortage, but because an algorithm decided the first shopper would pay more? And what if that system is quietly pushing everyone’s prices up, day by day, one penny at a time?

That’s the question a months-long investigation set out to answer. The result is one of the clearest public demonstrations yet of what’s called surveillance pricing – and it connects groceries directly to the ride-hailing blueprint that perfected it.

Four hundred people, one cart, many prices

The investigation – a joint effort by More Perfect Union, the Groundwork Collaborative, and Consumer Reports – started with a deceptively simple experiment. Eric, a former grocery-pricing professional, had been working with Katie Wells, Director of Research at Groundwork Collaborative, on a puzzle: Instacart pays its Washington, D.C. workforce different wages for nearly identical work. Her question was the natural one. If they do this to workers, what are they doing to consumers?

“We’re in the same room, the same address ordering, and we still had different prices.” – a participant in the first Instacart test

The plan was straightforward. Get people in a room. Open Instacart. Select the same store, the same items, the same time. Order pickup to strip out delivery calculations. Write down the basket total. Repeat with over 400 volunteers in four cities, shopping the same baskets at Safeway and Target, screenshotting everything.

The results were not subtle. Some people were charged about $114 for the same 20 items. Others nearly $124. And it wasn’t a fluke or an edge case – this was the lion’s share of shoppers, for the lion’s share of items.

“Nearly three out of four products had different prices during the experiments. Four different prices for Wheat Thins. Five different prices for eggs. Three prices for peanut butter.” – Eric, More Perfect Union

Worse, the differences weren’t random. The system had sorted shoppers into price groups – buckets where everyone paid the exact same amount for every item, all higher together or all lower together. In one test there were four price groups. In another, seven. That’s not noise. That’s design.

Consumer Reports’ own analysis put hard numbers on it: about three-quarters of products were offered at different prices to different customers, with variations ranging from 7 cents to $2.56 per item, and basket totals differing by as much as 23% between shoppers. For a household of four, Groundwork estimated that averages out to a swing of roughly $1,200 a year in grocery costs.

“We don’t set prices. Retailers do.”

When the investigators pushed back, Instacart’s answer was firm: retailers control all pricing. The different prices were just stores “putting out random feelers” to find the pricing sweet spot – not targeting anyone specifically.

Then the story cracked. The team contacted the stores. Albertsons-Safeway, an Instacart partner, went silent. Target responded – and their answer floored them:

“We don’t work with Instacart. We don’t set prices. We have nothing to do with any of this.”

That was strange, because Instacart had just told them the retailers control the prices. So they asked Instacart about Target’s claim. Instacart’s answer shifted: they admitted they actually do manage Target’s prices on the marketplace – and that they were running tests to figure out how much they could charge on top of Target’s normal prices.

That’s a pattern worth sitting with. Instacart – a delivery platform – controls the prices of a retailer that claims to have nothing to do with it, and runs tests to find the maximum markup. In 2022, Instacart bought Eversight, a company that optimizes grocery prices using AI. Eversight says it runs “constant experiments that are not made visible to consumers” – including testing how much you’re willing to pay for canola oil. Instacart markets it as “AI for everyday price performance,” promising partners 2% to 5% profit increases.

Then came the smoking gun. A Consumer Reports follow-up at Costco found price gaps – 15% on a bag of chips. When asked to explain, a Costco executive inadvertently forwarded an internal Instacart email describing a program called “Smart Rounding.” Instacart had defined it once, in a 2023 letter to shareholders: a “machine learning-driven tool that helps retailers improve price perception and drive incremental sales.”

“For some of our major grocery partners, this has led to millions of dollars in annual incremental sales.” – Instacart, 2023 shareholder letter

Translation: flex prices up or down based on what the algorithm says will maximize profit. As the report put it, this is “a very sophisticated, very deliberate, explicit attempt to maximize profits, and do so somewhat surreptitiously.”

The quiet threat of the electronic shelf label

You’ve probably heard the scary version of electronic shelf labels – the digital price tags that could enable surge pricing, raising the price of lemonade when it’s hot out. But one regional grocer, Schnucks, showed no price variation at all in the Instacart tests. That inconsistency made the team wonder: what if Schnucks had already moved past online testing? If prices are being tested in the store on digital shelf labels, and the app just mirrors the shelf, their experiment would never catch it.

The team’s hypothesis was subtler than surge pricing. Instacart’s patents describe an infrastructure for constant in-store price testing – clustering dozens of stores, testing different prices in different clusters, and adopting the most profitable price point.

“Would you rather charge $10 for a bottle of water during a single hurricane? Or 20 cents more every single day?” – Eric, More Perfect Union

The permanent 20-cents-a-day version is far more profitable – and far harder to notice. Asked about it, both Schnucks and Instacart denied using it. But Instacart’s own website told a different story: before the investigation, it listed a feature for “price optimization with Eversight via electronic shelf labels.” After being asked, the feature was gone.

The patents tell you what it’s for

Instacart’s patents aren’t abstract legalese. They describe exactly how the system could decide who pays more. The criteria that raise red flags – demographics – are illegal, and the data showed no correlation with the price groups. But behavioral characteristics are legal, and that’s what the patents actually list:

• Buying behavior
• Purchase history
• How frequently you shop
• Whether you’re shopping around
• Coupon usage
• Loyalty program participation

One promotional patent even describes calculating “shopper and product headroom” – how much more money a person could spend overall, and how much more on a particular product. It says the system can compute this for every individual consumer, but for efficiency, people are typically “grouped with similar consumers.”

Groups. Just like the tests showed. Every company practicing this, as one expert put it, “strenuously denies” that they’re doing so. Instacart denies segmenting anyone by personal or behavioral data. But the patents – the same documents Instacart files with the government – describe the mechanism in detail.

The blueprint was Uber

This didn’t start with groceries. The price side of the equation – and the wage side – were perfected by Uber.

In 2022, Uber rolled out algorithmic pricing for riders and drivers. In the three years that followed, the average rider price per mile went up, up, up. Uber denies any connection, but around the same time it went from burning money to becoming a cash machine. As Columbia Business School professor Len Sherman put it:

“That’s why more and more companies are trying as hard as they can to perfect this dark art – devilishly effective dark art, as I’ve called it. That’s how you turn grocery sales into a goldmine.”

Consumer Reports found the same dynamic on the ride-hailing side: for the same ride, at the same time, the lowest price was $41.21 and the highest $56.96 – a 38% difference. Uber insists it “does not engage in surveillance pricing” and doesn’t personalize prices to individuals. Yet the same algorithmic machinery – segmenting users, experimenting on them, extracting the maximum – powers both industries.

The warning from former FTC Chair Lina Khan, who reviewed the grocery findings, is blunt. This isn’t just another way to squeeze a few cents:

“This could ultimately end up being another big form of wealth transfer from ordinary Americans to massive corporations.”

And there’s a deeper, more dangerous twist. When one company like Instacart is setting prices for retailers who are supposed to be competing with each other:

“It could also ultimately allow firms that really should be competing against one another when setting these prices to actually engage more in something that looks like price fixing or collusion.” – Lina Khan

That’s the endgame. One platform – holding the pricing data for a whole category of competitors – becomes, in effect, the single brain setting all their prices. That’s not a market. That’s a cartel, administered by algorithm.

The fix is boring, and that’s the point

The FTC, under Khan, launched an investigation into surveillance pricing. The Trump administration claims it’s ongoing, but the new leadership stopped public comments. The good news, as Khan pointed out, is that states don’t need to wait:

“The steps we’ve seen, even from states like California and New York, to ban certain types of algorithmic price fixing, in areas like housing and rent, are extraordinarily important steps. We need more investigations. So I think we need a 360 view.”

In the wake of the investigation, Instacart actually stopped offering the technology that let retailers charge different prices – a rare, direct consequence of public pressure. And state attorneys general have subpoena power. They can demand to see inside the black box.

“State AGs have subpoena power. They can be requesting and demanding information so they could look under the hood and understand how firms are actually using this data.” – Lina Khan

The system exists – that’s no longer in question. Whether it’s operational, at scale, right now, we can’t fully see, because it’s designed to stay hidden. That’s the whole game. But the fact that it’s hidden is not a reason to trust it. It’s the reason to be suspicious.

The enshittification pattern never changes: bring people in with something good, become the only game in town, then use every bit of power – and every bit of data – to extract more. The new twist is that the extraction is now individualized. You don’t get the same price as your neighbor, because the machine has already decided, from your shopping history and your loyalty card and your willingness to click “buy,” exactly how much you can be made to accept.

Nobody votes for this. Nobody opens Instacart hoping the algorithm will price their Cheerios by their desperation. It happens because it’s profitable, because it’s legal, and because there’s been no federal consumer privacy law worth the name since 1988. The tools to stop it aren’t secret and they aren’t new. They’re called antitrust enforcement, price-transparency rules, and a privacy law that actually means something. We built them before. We can build them again.

---

Sources & further reading:
– More Perfect Union – We Had 400 People Shop For Groceries. What We Found Will Shock You. (the video this article is based on)
– Consumer Reports & Groundwork Collaborative – Instacart’s AI Pricing May Be Inflating Your Grocery Bill
– Groundwork Collaborative – Same Cart, Different Price: Instacart’s Price Experiments Cost Families at Checkout (full report)
– Consumer Reports – Instacart Stops AI Pricing Tests
– Consumer Reports – Different Prices for the Same Ride: How Uber and Lyft Use AI to Get More Money Out of You
– Reuters – FTC probes Instacart’s AI pricing tool
– Reuters – Instacart ends AI-driven price experiments after criticism
– Uber – Comment on Consumer Reports study

Related on enshitified.com: [Uber for Nursing: The Algorithmic Wage Discrimination Hiding Behind “Surveillance Pricing”] – the same machinery, aimed at wages instead of prices.

Watch: We Had 400 People Shop For Groceries (More Perfect Union)