A trio of business school professors -- in trying to understand the potential effects on consumers and the economy by fake reviews -- has identified which hotels are likely to generate or attract fraudulent reviews.
Authors Dina Mayzlin (USC), Yaniv Dover (Dartmouth) and Judith Chevalier (Yale and NBER) don’t make overly broad claims. But by looking at the “differences in differences” of two travel sites with extensive hotel reviews, they have shown that certain hotels distinguished by geography, ownership or management will post more fake positive reviews for themselves and more fake negative reviews for their competitors.
The hotels that are more likely to engage in review manipulation are those that...
- Have a neighbor
- Are owned by non-chain or small owners
- Are independent
- Are managed by a smaller management company
Their paper, “Promotional Reviews: An Empirical Investigation of Online Review Manipulation” cleverly exploits the organizational differences in writing reviews for two travel websites: Expedia and TripAdvisor. “While anyone can post a review on TripAdvisor, a consumer can only post a review of a hotel on Expedia if she actually booked at least one night at the hotel through the website,” they write. Since the cost of manipulation on Expedia is much higher than the cost of manipulation on TripAdvisor, they expected to find less review fraud on Expedia; so by finding real differences in numbers of positive and negative reviews for the same hotels on different sites, they derived many insights.
Does Professor Mayzlin think these differences are material for the average consumer? She wrote in a recent email, “The effects we find in our study are actually not that small. For example, the mean hotel in our sample has thirty negative reviews. We find that a hotel that is located next to an independent hotel owned by a small owner will have six more fake negative Tripadvisor reviews compared to an isolated hotel.”
Mayzlin continued, “ If we also consider the fact that previous research has found that negative ratings have a very strong negative impact on reviews, this could imply a substantial negative effect on sales due to review manipulation.”
An interesting observation the professors made about the primary cost of fake reviews is that “the activity will be publicly exposed.” So larger management companies or chains have much to lose if they are exposed – such as being thrown off of review sites.
The professors note that their work has implications for reviewer systems, in that "the policy of verifying reviews does limit" promotional or fake reviews, but there are far fewer reviews on verified sites. Orbitz and Amazon mark verified and unverified reviews, and they note that this is an "interesting compromise."
Has this research discouraged Professor Mayzlin from the helpfulness of reviews? “While review manipulation does exist, there is a lot of useful information in reviews. I am an enthusiastic user of online reviews, for all categories of products, including hotels.”
