Identifying Consumer-Welfare Changes when Online Search Platforms Change Their List of Search Results

Online shoppers are guided by search platforms: consumers type a search phrase into the platform’s query box, and the platform chooses how products appear in response. While search platforms may choose responses that help consumers find products more efficiently, they may also have incentives to mislead consumers. For example, search platforms may organize responses to favor their own products over third-party products that better suit consumer needs.

This paper uses a search-platform experiment to determine how to measure the effects of search responses on consumer welfare. In my model, the search platform has two different search algorithms, and the algorithm used for each consumer is assigned randomly. I apply the model in a study of users of an online travel agency. The data—from a data-mining competition sponsored by Expedia—contain an experiment that fits my model’s framework: Expedia treated some users with randomly ordered product listings and others with products ranked as Expedia saw best. I measure how consumer welfare changes across the two algorithms.

I find that under the random listing, the welfare of the users of the online travel agency is lowered by an average of $8.84 per user relative to Expedia’s own ranking system. I also simulate the welfare harm from a counterfactual search response that removes the top five products from all search results and find the welfare lowered by an average of $20.51 per user. Thus, platforms may create value for consumers by providing useful product orderings, but they can also substantially harm consumers by hiding products.