Our experience on the web today is so personalized—our Google search results and Facebook News Feeds, our Amazon and Netflix suggestions. So it makes sense that online retailers would try to “personalize” prices they charge for products, too. Except that kind of deal doesn’t sound so great to most shoppers, especially when they don’t know if they’re the ones getting the short end of the stick.
A new investigation conducted by researchers at Northeastern University provides a unique glimpse at how some online retailers engage in subtle price discrimination (charging different prices for the same product) and price “steering” (showing different search results that show products at higher or lower prices). Everything from whether a shopper was on a mobile or desktop browser to their history of clicks and purchases on the given e-commerce site seemed to have small effects.
“Ahead of time, we had no idea what we were going to find,” says study co-author Christo Wilson. “We thought that companies may be shying away from this. Turns out that’s not true.”
Travel websites, the researchers found, were the biggest offenders. For example, Travelocity charged mobile iOS users an average of $15 less than others. Orbitz and Cheaptickets charged people who weren’t logged onto the site an average of $12 more per night for hotel rooms. For other sites, like Expedia and Hotels.com, the researchers weren’t able to figure out the cause, but found that some people whose browsers contained specific cookies were guided to more expensive results than others. Priceline was found to alter hotel search results based on a user’s history of clicks and purchases on the site as tracked by browser cookies.
Among the 10 general retailers surveyed such as Walmart, Staples, and JCPenney (this list did not include Amazon), only Home Depot steered mobile users to more expensive products, sometimes as much as roughly $80 more expensive. However, unlike with travel sites—where testers reserved hotel rooms but later canceled the reservations—the researchers did not generate “purchase histories” on the retailers’ sites to show how that might influence pricing.
The authors, who recruited 300 volunteers to browse the web and also created fake accounts to acts as controls, had to use an elaborate setup to conduct their research. It’s much harder than simply asking two people to shop on the same site at the same time, and seeing if they’re shown the same prices. They wanted to remove the effect of a shopper’s location (that’s already been shown to affect price) and account for other differences unrelated to personalization, such as which data servers process each search. read more..