How AI Is Changing Prices for U.S. Shoppers

Pricing software is making it possible for businesses to vary what different customers pay. In one test, 400 shoppers used Instacart to purchase an identical set of 20 products, from matching stores, for pickup at the same place and time. Three-quarters of the products appeared at different prices across their carts, with gaps reaching 23 percent. Instacart stopped the test after it became public.

Lindsay Owens of Groundwork Collaborative says businesses can feed algorithms information such as a shopper’s location, purchase record, loyalty-card activity and app use. Her team also reviewed 315 patents, including Walmart applications involving customer-tracking technology and wearable devices. Walmart says it neither uses nor plans to use personal details, income, buying history, urgency or willingness to pay when determining an individual price.

Differences can also appear in fees. Owens, her staff and Paul Solman simultaneously ordered the same Uber Eats meal from one restaurant for delivery to one address. The food cost each person $13.59, but delivery charges included $2.49 and $4.49. Uber provided no response. Other examples included higher Expedia hotel pricing for people connecting from the San Francisco Bay Area and Target app prices changing as shoppers neared a store.

The technology may also shape apartment rents, gasoline prices and restaurant menus. McDonald’s supplies franchisees with algorithm-based price suggestions, though it says those optional tools do not automate or coordinate prices. Wendy’s has discussed testing dynamic pricing, daypart offers and A.I. For Spokane households, these systems could influence costs across groceries, delivery, travel, housing, fuel and meals.