Two people pick the same cereal, in the same store, at the same time. One pays $4.99. The other is charged $6.12. That mismatch — once the stuff of consumer gripes and urban legend — is exactly what a new participatory study found happening on Instacart.
Researchers from Groundwork Collaborative, Consumer Reports and More Perfect Union recruited volunteers to add identical baskets of groceries to Instacart carts simultaneously across four cities. The result: the app showed multiple price tiers for the exact same items in the exact same stores at the exact same times. The numbers are telling: 74% of items in the tests had more than one sale price, the average spread between the lowest and highest sale price was 13%, and some items varied by as much as 23%. Overall basket totals for identical lists of goods averaged about a 7% difference — a swing that could translate into roughly $1,200 a year for a family of four, the report estimates.
How the experiment worked
More than 400 people signed up; after cleaning incomplete submissions, researchers analyzed screenshots from 193 participants across Safeway and Target locations in Seattle, Washington, D.C., Saint Paul and North Canton. Volunteers joined video sessions (one test was done together in person near a Washington, D.C. Safeway) and added between 18 and 20 specified items to their carts. They took real-time screenshots of prices and basket totals, which the research team used to map how shoppers were grouped into distinct “price cohorts.”
Those cohorts mattered: many shoppers saw identical prices within a group and different prices across groups, not random noise but clear clustering. A dozen Lucerne eggs, for example, showed five different sale prices on the same Safeway page in D.C. The study also found cases where the sale price was identical for everyone, but the displayed “original” price varied — a presentation that can make discounts look deeper for some customers than others.
What Instacart says
Instacart acknowledged it runs pricing experiments and that retailers set prices on the platform, but told researchers that most customers see the standard price and that price tests are randomized and don’t use personal or demographic data. The company also noted that some CPG advertisers using its platform may use behavioral data to shape offers.
The report points to Instacart’s 2022 acquisition of Eversight — a pricing-tech company — and cites corporate materials that describe dynamic, AI-driven price optimization. Instacart’s own marketing describes the tools as ways to “continuously drive growth with dynamic pricing,” and its executives have mentioned using algorithms to learn where customers are more price-sensitive.
Why this matters beyond pennies and convenience
People expect grocery prices to be uniform and predictable; price tags and shelf labels created that expectation for a century. Algorithmic experiments reintroduce variability at scale without shoppers’ knowledge. When prices lose transparency, comparison shopping and budgeting become harder. For lower-income households that rely on predictable deals and coupons, the impact compounds.
This is also a data story. Price tests run on a platform that knows what you buy, where you are, and how you shop. Those inputs—alongside weather, nearby competitors, or whether someone is a first-time buyer—can be folded into optimization models. The same forces powering helpful features elsewhere in tech are at play in retail pricing; the line between convenience and surveillance pricing can blur. For broader context on how modern AI products tap into personal signals to automate decisions, see examples like Google’s AI Mode and its agentic booking features and how large models are being tied into personal accounts in ways that raise privacy and control questions, such as Gemini’s Deep Research integrations with Gmail and Drive.
The regulatory tug-of-war
Policymakers are paying attention. The Federal Trade Commission queried several firms about surveillance pricing in 2024, and states have started writing laws aimed at algorithmic price-setting and disclosure. New York passed the Algorithmic Pricing Disclosure Act, requiring companies to flag when an algorithm using personal data set a price — though critics say disclosure alone may not be enough. At the federal level, Rep. Greg Casar introduced the Stop AI Price Gouging and Wage Fixing Act to ban individualized prices set by automated systems based on personal data.
Legal theories that could be applied include the FTC’s authority over unfair methods of competition and deceptive pricing rules if “original” prices are fabricated to make discounts seem larger for some customers.
Retailer relationships and transparency questions
Target told researchers it doesn’t have a direct pricing relationship with Instacart; Instacart later acknowledged it scrapes publicly displayed Target prices and tacks on an additional charge to cover platform costs. Some major retailers did not respond to queries for the study. That raises practical questions for shoppers: when the price you see in-app diverges from the shelf, who bears responsibility — the marketplace, the retailer, or the advertiser?
Local news tests corroborate the basic complaint. A side-by-side check in Seattle found about a 10% markup on five items bought through Instacart versus the same in-store prices, not counting tips and delivery fees. For consumers, even modest markups can add up over months and across many households.
What shoppers can do now
Be deliberate. If a substitution, markup, or promotional discrepancy looks odd, compare the in-app price to an in-store visit before ordering, especially for staples or bulk buys. Use loyalty programs and retailer apps that sometimes carry exclusive in-store deals. And consider pickup options where available — the study focused on pick-up orders to remove delivery variables.
If you’re watching the wider technology debate, remember this story sits at the intersection of retail, AI and privacy: algorithmic systems are increasingly able to test and learn from consumer responses in real time. That same capacity powers convenience features across tech, but it also requires guardrails so pricing remains fair and legible.
The research group behind the study is calling for clearer rules and stronger enforcement to prevent opaque, individualized pricing that consumers can’t detect or challenge. Whether industry self-regulation, disclosure mandates, or statutory limits will be the right remedy is still being argued — but the evidence that shoppers are being sorted into different price buckets is now harder to ignore.
For readers who want the original research, Groundwork Collaborative and Consumer Reports published the full methodology and dataset; Instacart’s public materials about pricing tools are also available on the company’s site.