Same goods, different prices: Algorithmic pricing is spreading even to groceries
Price differences between stores, or even between branches of the same store, are common. It’s almost expected that the fancy store near the city center will cost more than the one in a suburb. But what about prices that differ not by location, but by time or even user? Welcome to the world of algorithmic pricing.
A recent months-long joint investigation by Consumer Reports and Groundwork Collaborative highlights how far algorithmic pricing has expanded. No longer limited to hotels and airlines, grocery stores and online retailers are now piloting it to increase profit margins by charging customers the highest prices they can afford, on a personalized basis.
In regular business settings, prices fluctuate primarily due to supply and demand, but can also be affected by retailer overhead and other factors. Algorithmic pricing changes the equation. Instead of a single price for everyone, the price may vary based on the customer’s spending habits and preferences. If you are a heavy spender or purchase a product regularly, you might pay a higher price than a frugal customer or someone trying the product for the first time. This fact is ruffling many feathers, with regulators, consumers, and policymakers finally beginning to push back.
The investigation focused on Instacart, one of the world’s largest grocery delivery platforms, and uncovered widespread price testing driven by artificial intelligence. Researchers found that identical grocery items were often priced differently across customers, with disparities reaching up to 23%. In controlled experiments involving 437 volunteers across the United States, every participant encountered some form of algorithmically determined price variation when building identical shopping carts.
These pricing differences were not limited to obscure products. Everyday household items such as breakfast cereal, snack foods, and eggs showed measurable variation, while entire shopping baskets in some cases differed by nearly $10. This could amount to approximately $1,200 in additional costs per year for consumers.
Instacart has acknowledged conducting dynamic pricing tests, describing them as limited and negligible. However, the findings suggest a far more extensive system, powered by machine-learning tools developed after the company acquired AI startup Eversight in 2022. Investor materials indicate that these tools are designed to optimize price perception, test consumer sensitivity through techniques such as “smart rounding,” and ultimately increase profit margins for retail partners by up to 5%.
Beyond variable pricing, the investigation also documented instances of so-called “false reference pricing,” in which the platform displayed different original prices for the same discounted item, thereby artificially inflating the perceived value of the sale. This behavior is illegal in many countries, yet it’s a common tactic used by sellers to drive urgency.
The implications extend beyond the United States. In the Middle East, where grocery delivery platforms, super-apps, and AI-driven retail ecosystems are expanding rapidly, similar pricing models could emerge as digital commerce matures. The region’s high smartphone penetration, centralized retail platforms, and growing use of data analytics create favorable conditions for dynamic pricing, particularly for essential goods.
Regulatory responses are slow, as most countries lack a robust foundation for pricing controls. In the US, the Federal Trade Commission warned that algorithmic price discrimination may violate fairness standards. Laws in the UAE, Saudi Arabia, and many other nations include provisions aimed at promoting “fair pricing” for consumers. However, it is difficult to define what constitutes fair pricing and whether algorithmic pricing is compatible with these rules. In the meantime, and in the absence of clear precedent for prosecution, more retailers will likely adopt algorithmic pricing schemes to the detriment of consumers.
