Personalization in E-commerce Product Search by User-Centric Ranking
Authors: Lucia Yu, Ethan Benjamin, Congzhe Su, Yinlin Fu, Jon Eskreis-Winkler, Xiaoting Zhao, Diane Hu
Presented at: Knowledge Management in eCommerce Workshop, in conjunction with The Web Conference (WWW) 2021 (Ljubljana, Slovenia)
Abstract: E-commerce platforms offer the convenience of browsing through an entire catalog of inventory via a search bar. An unconventional inventory of unique products presents even greater challenges for product search, given that many of listings fall outside of standard e-commerce categories. With the potentially overwhelming number of relevant items per query, it becomes increasingly important for market places and platforms to help the user find items that best fit their preference and interest via a user-centric ranking model that generates personalized search results. In this paper, we demonstrate how we use a combination of learned content-based and session-based listing representations to build user profiles from multiple implicit feedback types aggregated over various time frames in order to create a personalized tree-based ranking model at Etsy1. Etsy is one of largest e-commerce marketplaces with millions of unique, handcrafted items being sold to shoppers around the world. In the proposed personalized model, we observe offline improvements in ranking metrics (i.e, purchase NDCG@10) and higher degrees of personalization measured by Kendall Tau coefficients, when com-pared to a non-personalized ranking model. We successfully deploy the user-centric ranking model across multiple platforms on live traffic to hundreds of millions of users, with thousands of search requests per second. With the results from three different online A/B experiments, we show that users spend less time searching and buy more items in the personalized variants compared to the baseline.
Bibliography: Lucia Yu*, Ethan Benjamin*, Congzhe Su**, Yinlin Fu**, Jon Eskreis-Winkler**,Xiaoting Zhao, Diane Hu. 2021. Personalization in E-commerce Product Search by User-Centric Ranking. In Knowledge Management in eCommerce Workshop with The Web Conference, April 16, 2021, Ljubljana, Slovenia. New York, NY, USA, 5 pages.