Instacart says it’s done A/B testing your dinner. On Monday, the company said it’s pulling the plug on all item-level price tests — effective immediately — after its dynamic pricing practice drew public criticism and regulator scrutiny.
Instacart said in a press release that retailers “will no longer be able to use Eversight technology to run item price tests on Instacart,” and promised that “if two families are shopping for the same items, at the same time, from the same store location,” they’ll “see the same prices — period.” Instacart bought pricing software Eversight in 2022, and Instacart has said it began offering item-level price testing to retailers in 2023, limited to a subset of 10 partners — and said those partners already apply markups. The groups that flagged the practice said they saw evidence on major storefronts, including Safeway and Target $TGT.
The blowback was fueled by a Dec. 9 report from Groundwork Collaborative, Consumer Reports, and More Perfect Union, which recruited 437 shoppers across four cities to build carts simultaneously on Instacart and compare what they were shown. Every one of the 437 shoppers saw pricing experiments.
The researchers found that almost three-quarters of items in their experiment — citing examples from Safeway, Target, Costco $COST, Albertsons, Kroger, and Sprouts Farmers Market — appeared at multiple price points, with the platform showing as many as five different prices for the same product at the same store at the same time. One Safeway listing for a dozen Lucerne eggs in Washington ranged from $3.99 to $4.79.
In practice, that meant different shoppers loading the same items from the same store page at the same moment could be placed into different pricing “test buckets” — without once being told they were part of an experiment.
For items that the report flagged as being experimented on, the average spread between the lowest and highest displayed price was 13%, and the biggest difference observed hit 23% — for the same item, the same store, the same everything. On total baskets, researchers found that the same cart could vary by an average of about 7%, which they translated into about a $1,200 annual swing using Instacart’s stated grocery-spend figure for a household of four.
The company’s “effective immediately” wording may have been helped along. Last week, Reuters reported that the Federal Trade Commission had sent a civil investigative demand seeking information about the Eversight tool, and Instacart spent days arguing the coverage was misleading — including disputing the $1,200 extrapolation — before announcing this week that the tests were over. Instacart’s stock dropped about 10% after news of the FTC probe hit. And days before the pricing controversy erupted, Instacart agreed to pay $60 million in refunds to settle separate FTC claims that it misled shoppers about delivery fees, free trials, and a “100% satisfaction guarantee,” drawing sharper regulatory attention to its business practices.
And Congress is circling, too. House Oversight ranking member Robert Garcia has asked Instacart for details about its pricing practices and data use, with a response deadline in early January. And Senate Minority Leader Chuck Schumer wrote to the FTC urging clearer on-screen warnings when shoppers are placed into pricing tests.
This dynamic pricing episode also lands amid growing scrutiny of algorithmic pricing across consumer apps, where companies pitch optimization and regulators — and consumers — hear fairness, transparency, and consent. At the moment, grocery inflation is a top cost concern for families; online grocery sales have topped $10 billion in a single month, amplifying fears that opaque pricing could worsen already strained household budgets.
Instacart’s defense was that this wasn’t demand-based, real-time “surge” pricing and that it wasn’t “surveillance pricing,” either. The company says the tests were short-term, randomized experiments, not personalized pricing — and now the tests are over. Period. But Instacart’s exit still leaves an important caveat: Retailers will continue setting their own prices on Instacart, and the company says partners may still vary pricing store-by-store, similar to physical locations.
Shoppers may tolerate personalized ads and algorithmic nudges, but when the number on the receipt starts feeling experimental, trust gets expensive — fast.