“…Recommender systems also suffer from start‐up problems in new applications, when the database of previous ratings is small, or sparse. In practice, the accuracy of recommender systems may also be bounded by their strategies to collect user ratings, suggesting questions for future work on the behavioral aspects of review elicitation (Avery, Resnick, & Zeckhauser, ; Berger, ; Burtch, Hong, Bapna, & Griskevicius, ; Yeomans, ).…”