Hemant Bhargava
Graduate School of Management
Thursday, October 27, 2005
1131 Kemper Hall
3 :10-4:00 p.m.
Sponsored Search: Recommendation Bias by Information Gatekeepers
This talk examines preferential recommendation, observed in many information gatekeepers including credit raters, radio deejays, travel experts, movie critics, and Internet search engines, which are an essential entry point for many information search and decision making tasks. We develop and analyze a tractable model that covers gatekeepers’ choice of bias level, impact of competition, and the interplay between technological quality and recommendation bias. In general, a gatekeeper will choose positive bias level, in order to optimize the gatekeeper’s tradeoff between placement revenues and user-based revenues. If placement revenues are channeled towards technology improvement, then recommendation bias may benefit users in the long run because users’ benefit in visitng a gatekeeper depends on both its technological capability and extent of bias. Competition among gatekeepers will further moderate their propensity to bias results. Product innovations or better domain expertise give the gatekeeper greater flexibility in choosing bias level and responding to users’ demand sensitivity towards recommendation bias. Innovation by one gatekeeper has positive effect on consumer surplus as long as it does not create substantial imbalance or monopoly power. Consumer interest might also be hurt if gatekeepers coordinate in setting (higher) bias levels, negating the moderating effect of competition.