Consumer preferences in user- vs. item-based recommender systems for search and experience products

ABSTRACT This study investigates consumer preferences for user-based versus item-based collaborative filtering (CF) recommender systems when purchasing search versus experience products online, addressing consumer interactions with AI-powered recommender systems across product types. The study also explores the moderating effects of susceptibility to informational influence (SII) and susceptibility to normative influence (SNI) on the relationships between product types and CF preferences. Furthermore, we examine the mediating role of risk aversion in explaining why consumer perceptions of a product as an experience product influence their preference for user-based CF over item-based CF. Two experiments reveal that consumers prefer item-based CF for search products and user-based CF for experience products. SNI moderates these preferences, while risk aversion mediates the preference for user-based CF for products perceived as experience ones.