IDEI Distinguished Research Seminar Series

Data on individual consumers are a critical asset for online retail platforms, which enable them to use personalized query-based search algorithms to help consumers find the products they are looking for. Yet data privacy regulations have been scaling up to protect customers’ personal data, which may result in the de-personalization in search algorithms. To understand its impacts on consumer search and purchase behavior, we design and exploit a high-stake large-scale field experiment involving 4,189,498 customers with the collaboration of a world-leading online retail platform. We find decreases in customer search efficiency and market transactions due to the de-personalization of the search algorithm. Compared with the control group with personalization, customers in the de-personalized treatment group on average browse more products in the product listing returned by the search algorithm but make fewer clicks and purchases. Meanwhile, the clicks and purchases from the sponsored ads increase in the treatment group. Finally, we find evidence that customers adapt their expectations of the search results and accordingly adjust their search behaviors almost immediately upon the de-personalization of the search algorithm. The findings offer insights for platforms and regulators to understand the implications of de-personalization in search algorithms arising from the evolving privacy regulations.