Instacart Price Testing Evidence Research Note

Instacart Price Testing Evidence Research Note

Question

What did the 2025 grocery-pricing investigation establish, what did Instacart say about the mechanism, and which later legal or policy events must remain separate?

Investigation design and finding

The Consumer Reports investigation was conducted with Groundwork Collaborative and More Perfect Union. It recruited 437 volunteers across four groups or cities for coordinated Instacart shopping sessions. Participants added the same 18 to 20 products from Safeway or Target and submitted screenshots.

Consumer Reports says roughly 200 error-free screenshot sets supported the main analysis, with additional retailers examined later. Around three-quarters of checked items appeared at multiple prices. Reported item differences ranged from seven cents to $2.56 and reached 23 percent. One basket ranged from $114.34 to $123.93.

The often-repeated $1,200 annual household figure is an estimate based on an observed basket spread and household-spending assumptions. It is not an annual charge observed for every participant or a finding that every family paid that amount.

Instacart response

Instacart's December 18, 2025 response calls the practice randomized, short-term A/B item-price testing. It says assignments used product category and store location and did not use personal, demographic, or user-level behavioral information to set item prices.

The company says retail partners control base prices and rejects the Consumer Reports annualized estimate. It also distinguishes the tests from dynamic and surveillance pricing.

These are company claims about the internal mechanism. The public investigation observed outputs but did not independently publish the full assignment code or data. Editorial treatment must preserve both the observed variation and the counterparty explanation.

Later developments

Consumer Reports later reported that Instacart stopped item price tests after the report while allowing partners to test promotions and discounts.

The New York attorney general demanded information in January 2026 and warned about possible noncompliance with the state's algorithmic-pricing disclosure law. That was an inquiry and warning, not a final finding that personal data caused the tested prices.

The FTC's December 2025 $60 million settlement announcement concerned alleged delivery advertising, refund, satisfaction-guarantee, and subscription-enrollment practices. It was not a settlement of the Consumer Reports item-price testing dispute.

Editorial rule

The package may state that coordinated shoppers observed different prices and that Instacart described the system as randomized testing. It may report the annual estimate with its assumptions and limits.

Do not say Instacart used personal data in the experiment without evidence of that causal input. Do not call the New York letter a violation finding. Do not connect the FTC settlement to item-price testing except to explain that it was a separate matter.