A Learning Curve for Using Online Reviews

Opinions abound in cyberspace. But do online ratings and reviews actually help consumers make better decisions about doctors, hotels, schools, restaurants, movies and other purchase options? The answer may be no for inexperienced community members who have limited interactions in a specific domain. 

New research co-authored by marketing professor David Godes at the University of Maryland’s Robert H. Smith School of Business finds a learning curve in platforms like Amazon.com, Tripadviser.com and Ratemyprofessors.com. “Using another person’s opinions to infer your own expected satisfaction may be difficult, especially if you aren’t familiar with the environment,” Godes says.

The challenge is compounded by the range of opinions expressed among individual posters and the different scales used at various sites. Does a four-star hotel rating on one travel site, for example, align with a similar score on another site? And what does it mean if the scores are widely divergent?

“Complexity in interpreting opinions may arise due to significant heterogeneity in ratings across different online communities, which would require learning and expertise-building on the part of new community members,” Godes says. 

Over time, as community members connect the reviews they read online with their actual experiences, a type of mapping occurs. A “Fresh” rating at Rottentomatoes.com, for example, actually means something to moviegoers after they see several films and map them to their personal Tomatometer. “You learn how expressed opinions match up with your own preference structure,” Godes says. 

Community members may also map their preferences with individual posters, as they learn to discern which individuals most closely share their preferences. Initially, community members may benefit most from opinions posted by friends — an advantage in a platform like Facebook that connects people who already know each other. But over time, the viewpoint diversity that comes through interaction with strangers is increasingly helpful.

“Opinions from weak ties may be particularly valuable due to the novel information they make available,” the authors conclude. “Critical to their usefulness, however, is the extent to which consumers are able to learn how to interpret and utilize information from these ties.”

The study is co-authored with Smith School PhD alumnus Yuchi Zhang, who now teaches at Santa Clara University in California. The authors base their findings on an analysis of user experiences at Goodreads.com. 

“We follow the actions of new Goodreads users as they establish a social network and learn how to make use of the information their network makes available,” the authors write.

Read more: Learning from Online Social Ties is featured in Marketing Science.