Search with Refinement

Search with Refinement Preliminary. Please do not cite or distribute The development of online search technology has profound impacts on consumers, business, and the society. One important feature of online search is that consumers are able to refine the search results using tools such as sorting and filtering. Albeit such refinement tools have significant effects on consumer behavior and market structure, there is little empirical research documenting and measuring the effects. We propose a structural model of optimal consumer search that coherently integrates the decisions of consumer search and refinement as well as the structure of uncertainty about the products. The model is estimated using a unique data set of individual level search activities of hotels provided by a large travel website. We find that the refinement tools encourage 57% more searches and enhance the utility of the purchased product by nearly 20%. However, most websites by default rank search results according to the results’ qualities or relevance to consumers (e.g., Google). We find that if consumers are uninformed about such default rules for ranking, they may engage in disproportionally more searches using the refinement tools. As a consequence, the overall welfare surplus may deteriorate by up to 8%. In contrast, when consumers are informed that the default ranking of the results already reflects the qualities, they search less and the welfare surplus increases 3.3%. We also find the refinement leads to a less concentrated market structure.