I study how the information a search intermediary has about consumer preferences impacts the market. Consumers participate in costly search among different sellers’ products, relying on the rankings order provided by the intermediary based on their preferences. Better product targeting affects consumer search and purchases, which, in turn, changes the seller pricing incentives. I considered these aspects by modeling both sides of the market under various ranking algorithms used by the intermediary. On the demand side, I developed a joint model of consumer costly search and purchase. On the supply side, I considered the sellers’ pricing competition. To estimate the demand and supply models, I utilized a rich dataset provided by Expedia, which includes consumer search and purchase data and information on the hotels and prices they charge. I find that if the intermediary uses data on consumers’ preferences to provide them personalized rankings of products, consumers, on average, experience a 3.6% ($4.9) utility decrease due to increased transaction prices, a 0.8% ($1.1) utility gain due to a reduction in search spending, and 0.5% ($0.7) utility gain due to finding a better-fitted hotel.

The paper discusses markets with consumer’s search frictions and partially informed intermediary. The main finding is the increase in the intermediary information might decrease the average quality of the product consumers purchase and decline the total economic welfare and consumer surplus. The mechanism is if the intermediary makes better advise to consumers in average what product to explore first, all consumers have lower expectations about the next products and explore them less often, which decreases the quality of purchased good for consumers who got the wrong recommendation and might lead to a reduction of the average quality of purchased products. This effect appears in the case of low search cost, which makes it especially important in the analysis of online search platforms.