Exploitation and Exploration: Improving Search Precision on E-commerce Platforms

Abstract

E-commerce platforms match online buyers and sellers using their search technologies. Although a more precise search algorithm may improve search targetability, it may also reduce cross-selling opportunities, as consumers spend less time exploring different products. We empirically quantify these tradeoffs through a collaboration with Alibaba Group. Specifically, we take advantage of a 2019 quasi-experiment on Taobao.com, in which the platform refined some product categories into finer subgroups in order to return more-targeted search results to online shoppers. Using granular data on consumer search and purchase behaviors across multiple search sessions and product categories, we find that the improvement in search precision leads to a 37.3% increase in consumers’ click-through rates and a 64.4% increase in gross merchandise volume in the product category we study. The immediate improvement in matching outcomes, however, comes at the cost of a substantial decrease in consumer engagement and unplanned purchases in a longer time horizon for consumers prone to spending more time searching. On average, these consumers conduct 5.5% fewer searches, spend 4.1% less time on the platform, and decrease their spending on related categories by 2.2% in the week after the search precision increases. Our findings illustrate the tradeoff between exploitation and exploration in e-commerce search design that has not yet been previously documented in the literature.