We would like to invite you to participate in Management Engineering(ME) Seminar.

1. When: April 4th (Thursday), 16:00~17:20
2. Where: Supex Building, Lecture Room 101
3. Speaker: Prof. Chul Ho Lee (KAIST)
4. Topic: The Value of Shopping for Review Information: What is the Impact of Rating Differences Across Sites?
5. Research field: IT Management
* Lecture will be delivered in Korean.
* Seminar materials: Abstract

[Abstract]
Given the availability of multiple sources, a customer processes online review information to form an aggregated quality belief. In this study, we first mathematically model how a customer updates her quality belief in a Bayesian manner. Then, we develop hypotheses around the insights obtained from the structural analysis of the model. We empirically examine these hypotheses using restaurant availability and rating data collected from Yelp.com and OpenTable.com. We used the regression discontinuity (RD) design---a proxy to a quasi-experimental research setup---to elicit the causal effects of review information on sales. Our empirical specification is based on the theoretical model we develop, and therefore, captures the dynamics of review information from multiple sites and also addresses the shortcomings of sales estimation under highly correlated information. In this context, we analyze the contribution of the information from a single site regarding its first-order (direct) and a differential (in comparison to the previously visited sites) impact on the customer's final quality perception. We find that when there is a lack of consensus on a single site on the quality of a good, the rating information collected from the site will have a lower impact on a customer's propensity to buy that good. Interestingly, our analysis also points to a trade-off in the differing impact of reliability (a measure of ambiguity in true quality) on site ambiguity (variance of ratings at a site) and perception ambiguity (variance of aggregated information across sites). Particularly, an increase in reliability decreases the both ambiguity measures, yet which of the two is more dominant determines the extent of utility impact of ratings.