Are customers talking trash about your business on Yelp? It may be time to start packing.
Using a database of 130,000 Yelp reviews, researchers at the University of Maryland's Robert H. Smith School of Business have identified a way to use software to predict which restaurants are about to close. While it seems obvious that restaurants with lower ratings are more likely to close, the professors and Ph. D. students involved in the research indicate there's more to a restaurant's success than numerical ratings.
Looking at 2,000 restaurants that were open at the time of the study plus 450 that had closed down in the last few years, researchers found that words like "good," "like," and "friend" were more commonly used in Yelp reviews of restaurants that ended up staying open longer. By combining text analysis with restaurant rankings, researchers created software that can predict if a restaurant would close within the next three months with 70 percent accuracy, reports the Washington Business Journal.
The predictive powers of the software have not yet been tested in the real world, according to the authors of the study. However, if the model is picked up by the hospitality industry, it could serve as a key indicator of when businesses are skating on thin ice and need to turn things around – fast.