Instacart Didn’t Read Your Mind. It Ran an A/B Test.
On a Thursday in early September, more than 40 strangers logged into Instacart to buy eggs and test a hypothesis. They all selected the same store, the same brand, and the same pickup option. The only difference was the price they were offered: $3.99 for some, $4.79 for others.
We are, of course, supposed to not only be shocked by this, but also outraged. The Groundwork Collaborative, which organized the study, warned that companies like Instacart are jeopardizing trust in markets.
The New York Times ran it as the hook for a larger piece on algorithmic pricing: “the notion of a single price, offered to all customers for a predictable period, is breaking down.” The Times piece ends with Groundwork Executive Director Lindsay Owens saying: “This isn’t about managing scarcity or efficient markets… [It’s about] pushing to figure out the maximum amount you are willing to pay and squeeze it out of you.”
Both pieces blur together several distinct ideas: that prices vary, that prices are set by algorithms, that algorithms use your personal data, and that all of this harms consumers. But these are not the same thing.
Importantly, the study actually found no evidence that Instacart based prices on individual characteristics. As the Times piece correctly notes: “The Groundwork study found no evidence that Instacart was basing different prices on customers’ individual characteristics like income, ZIP code or shopping history.” Yet Groundwork uses the findings to call for Federal Trade Commission (FTC) action against “price discrimination not justified by differences in cost or distribution.” The Times frames the study as evidence of a trend toward algorithmic price discrimination.
To see why this leap fails, consider that there are at least four distinct reasons you might see different prices for the same product.