Paper ID
2321
Description
The literature on recommender systems mainly focuses on product recommendation where buyer’s preferences are considered. However, for user recommendation in two-sided matching markets, potential matches’ preferences may also play a role in focal user’s decision-making. Hence, we seek to understand the impact of providing potential candidates’ preference in such settings. In collaboration with an online dating platform, we design and conduct a randomized field experiment and present users with recommendations based on i) their own preferences, ii) potential matches’ preferences, or iii) mutual preferences. Interestingly, we find that users are sensitive to the provision of potential candidates’ preferences, and they proactively reach out to those “who might prefer them” despite those candidates’ relatively lower desirability. This leads to a greater improvement in matching. The findings provide valuable insights on how to design user recommendation systems beyond the current practice of recommendations based on focal user’s preferences.
Recommended Citation
Shi, Lanfei and Zhang, Kunpeng, "Your Preference or Mine? A Randomized Field Experiment on Recommender Systems in Two-sided Matching Markets" (2019). ICIS 2019 Proceedings. 26.
https://aisel.aisnet.org/icis2019/economics_is/economics_is/26
Your Preference or Mine? A Randomized Field Experiment on Recommender Systems in Two-sided Matching Markets
The literature on recommender systems mainly focuses on product recommendation where buyer’s preferences are considered. However, for user recommendation in two-sided matching markets, potential matches’ preferences may also play a role in focal user’s decision-making. Hence, we seek to understand the impact of providing potential candidates’ preference in such settings. In collaboration with an online dating platform, we design and conduct a randomized field experiment and present users with recommendations based on i) their own preferences, ii) potential matches’ preferences, or iii) mutual preferences. Interestingly, we find that users are sensitive to the provision of potential candidates’ preferences, and they proactively reach out to those “who might prefer them” despite those candidates’ relatively lower desirability. This leads to a greater improvement in matching. The findings provide valuable insights on how to design user recommendation systems beyond the current practice of recommendations based on focal user’s preferences.