Skip to content

How Can You Predict Business Failure? Look To The Stars

Online reviews can predict the failure of seven out of 10 businesses, new research finds
Online reviews can predict the failure of seven out of 10 businesses, new research finds
  • Restaurants with a mix of good and bad reviews have a higher likelihood of failure
  • Restaurants with consistently bad reviews are more likely to stay in business
  • If reviews predict a restaurant to be in a risky state, managers can take small but effective steps to improve service and public opinion

Frank Sinatra once sang, “This is my quest, to follow that star”… who knew he was giving us sound business advice rather than romance?

While “Ol Blue Eyes” wasn’t exactly renowned for his business acumen, his lyrics predicted the sentiment felt more strongly by managers than ever before. The advent of online reviews has brought pros and cons to businesses the world over. One advantage is that, if the business reviews well its word gets around through the power of the net, turning a local success into a national or even global one. However, that power can also be used for evil, as the internet also has the power to elevate negativity beyond what was previously possible. Businesses can (and often!) go viral for all the wrong reasons.

But it’s not just any old reviews we are talking about here. In particular it’s the immense power of the star rating: a comprehensive summary of a business or service, usually with a scale from one to five, and which can be recorded, and understood, by users within seconds – hence their popularity.

Numbers give the connotation of objectivity, a precise measurement which is hard to argue against because we assume some form of calculation has churned out the result. It’s not hard these days to find the highest rated or “best” product or service because search engines like Google find it on the basis of arithmetic, saving us the time of reading lengthy written reviews from delighted or disgruntled patrons.

The inescapable influence of the star rating is no more apparent than in the restaurant industry. Customers want to be reassured of good food and are increasingly prompted to leave a review of their dining experience. However, it may be that star ratings hold far more power than persuading us where to go for dinner.

Research shows that star ratings can predict, with near pinpoint accuracy, whether a business will succeed or fail.

This idea was put to the test by researchers from Rotterdam School of Management, Erasmus University (RSM), who developed an AI model which could accurately predict a restaurant’s likelihood of failure, 70% of the time. Dr Markus Weinmann (now at the University of Cologne), Christof Naumzik and Prof. Stefan Feueriegel used data from Yelp, analysing the ratings of over 934 restaurants between January 2010 and December 2017. The AI predicted whether review ratings are consistent with the failure of restaurants.

The researchers found that there are three categories of restaurants in terms of their performance: “Well Running”, which had good ratings, “Bad Ratings”, which were still running, and “At Risk”. The restaurants in the “At Risk” category had wildly fluctuating reviews – many excellent and many poor.

Whilst these restaurants gained more positive reviews than the consistently bad ones, the researchers labelled them as “At Risk” because it became apparent that these were the businesses associated with the highest levels of failure. To put it in numbers, 19% of “Well Running” restaurants failed, compared to 25% of those in the “Bad Rating” category, and as many as 35% of those identified as “At Risk”.

But how is it that the restaurants with the highest number of bad ratings don’t have the highest chance of failure?

The researchers explained that reviews from the “At Risk” group revealed that customer satisfaction has been subject to considerable variance, which they believe is indicative of service performance issues. It is this which is likely to lead a business to collapse. Conversely, whilst reviews from the “Bad Ratings” restaurant highlight consistent issues concerning customer satisfaction, there are reasons why these restaurants remain in business. Aside of customers knowing what to expect, such outlets have benefits such as close proximity to customers and longer opening hours that help paper over the cracks. At the end of the day they may not be popular on paper but they still get a significant number of customers.

Further food for thought in the research is the disparity between the success of independent opposed to affiliated restaurants. over the seven year time period of the study, 101 out of 200 independent businesses failed in contrast to just five affiliated chains out of 200.  One reason for this may be, according to the researchers, is that many fast-food chains (e.g., McDonalds, Burger King) are operating a model of offering low-rated quality at a low cost.

Dr Weinmann says that providing a consistent service may take precedence over quality, even if such things stand to be largely improved. “The restaurant market is very volatile and dynamic. Even minor changes – such as updating a menu or a change of personnel ‒ can make customers switch their loyalty,” he says. “The success, or even the survival, of service businesses depends on their ability to satisfy their customers. Businesses are often too late to recognise slumping customer satisfaction and suffer the ultimate failure. So an early warning system for restaurants would help them to adjust their service offerings in time.”

The stars, the researchers say, provide that early warning system for those restauranteurs who take the time to keep track of their ratings. Dr. Weinmann believes the study should serve as an aid for restauranteurs wanting to maintain momentum and prevent business failure. “Managers can use our model to inform their decision-making,” he says. “​​Furthermore, they can use the model to predict business failures and, based on the estimated risk, plan timely interventions”.

The researchers also suggest the model could benefit customers: Assuming a rating platform decides to display a restaurant’s state on its site, customers could factor this into their decision.

Positioning ourselves as the customer we should take it upon ourselves to be vigilant in analysing reviews and making what we believe is an informed choice. Research has shown we too can bite off more than we can chew by reading too deeply into star ratings as definitive proof of what is good and bad. We are better off defining our expectations based on what kind of service or restaurant we want and go from there.

True to culinary form we should take such subjective measures of quality with a pinch of salt.

Leave a Reply

%d bloggers like this: