Yelp on Fake Reviews: We Filter 25% of Reviews for Being Suspicious

YelpThe validity of Yelp reviews has always been called into question. Are those great reviews from the business itself? Are those horrible reviews from competitors?

It's always hard to know whether the online business review you're reading is authentic. Add to the mix the fact that New York fined 19 businesses for fake reviews, and it becomes clear that there's definitely false reviews getting through on various review sites.

Yelp has revealed that its automated review filter suppresses about 25 percent of all reviews submitted to the site, as they may be fake. However, Yelp admits that the system isn't completely accurate and there is the chance that some of the suspicious reviews they are filtering could actually be legitimate.

A recent study, Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud by Michael Luca of Harvard Business School and Georgios Zervasy of Boston University estimated that 20 percent of Yelp reviews in 2013 are fake, compared to only 5 percent in 2006. Part of this could be attributed to the fact there were not as many users on Yelp in 2006, so the stakes weren't as high financially for having either positive or negative reviews.

The study discovered that restaurants with a low number of reviews were more likely to post fake positive reviews, while chain restaurants were not as likely to do the same.

We have seen that positive fake reviews are more prevalent when a restaurant's reputation has deteriorated or is less established. In contrast, our results show that negative fake reviews are less responsive to a restaurant's recent ratings, but are still responsive { albeit to a lesser degree { to the number of reviews that have been left. In other words, while a restaurant is more likely to leave a favorable review for itself as its reputation deteriorates, this does not drive competitors to leave negative reviews. At the same time, both types of fake reviews are more prevalent when a restaurant's reputation is less established, i.e. when it has fewer reviews.

They also discovered that fake filtered reviews tended to be overwhelmingly positive or overwhelmingly negative.

Yelp continues to work to filter reviews are algorithm deems are fake. They also are using Consumer Alerts to alerts Yelp users when a business has been caught offering any kind of incentives for positive reviews.

About the author

Jennifer Slegg began as a freelance writer, and turned to search engine optimization and writing content for the web in 1998. She has created numerous content-rich sites in niche markets and works with many clients on content creation, strategy, and monetization. She writes about many search industry and social media topics on her blog, and is a frequent speaker at search industry conferences on SEO, content marketing and content monetization. Acknowledged as the leading expert on the Google AdSense contextual advertising program, she runs JenSense, a blog dealing exclusively with contextual advertising. She is also the founder and editor of The SEM Post. She is known by many as her handle Jenstar on various webmaster forums.