The power of reviews: a half-star difference can make or break a business
Love it or hate it, Yelp is the standard bearer for local business reviews. With 78 million uniques searching 30 million reviews in Q2 2012, a good review on Yelp is a vital marketing tool – especially for restaurants.
Two Berkeley economists have given Yelp – specifically 328 restaurants in Yelp’s most active market, San Francisco – the stastical treatment to determine how star rating impacts table availability.
The study, published in the Economic Journal, was constructed by taking 1-5 star reviews for each restaurant and merging it with availability of a table for four on Thursday, Friday, and Saturday evenings.
By combining the two data sets, the researchers found a direct correlation between star ratings and reservations: A half-star increase in rating is equivalent to a 19% reduction in available reservations.
There are various factors that might affect this number, including venue capacity and availability of external information about the restaurant.
In fact, restaurants that have little information available outside of Yelp – and therefore nowhere else for consumers to inform restaurant selection – sell out 27% more often with a half-star increase.
It’s important to note that Yelp rounds their review scores to the nearest half-star – a 3.74 average rating becomes a 3.5, while a 3.76 average rating becomes a 4. So, even though Yelp makes the actual rating available, it only shows stars in half-star increments; a small difference in average can have enormous impact on a restaurant’s bottom line.
From the report:
Figure 2 plots mean 7:00 pm reservation availability by Yelp rating. Panel A focuses on the window where restaurants have either 3 or 3.5 stars; Panel B focuses on the window where restaurants have either 3.5 or 4 stars, and Panel C focuses on the window where restaurants have 4 or 4.5 stars. Restaurant availability appears to respond primarily to the displayed rating, and not the latent average review score.
The median restaurant might see an increase of 6-8% in customer flows, which, although modest, can yield an extra $816 in weekly pre-tax profit for a median mid-to-high end restaurant seeing only a 6% increase in customer arrivals.
The analysis also includes a deep-dive into gaming the system with fake reviews – it’s a highly statistical analysis, but essentially concludes that while there are some desirable short-term benefits, the long-term impact is nil when it comes to increasing ratings via fake reviews.
The biggest flaw in the research is that the two independent data sources used do not match date-wise: the Yelp star reviews were taken as of February 2011, and the reservation data was from the period of July to October 2010. Restaurant ratings change quickly, and so the correlation of reviews-to-reservations is not as accurate given the 4-month discrepancy.
Other potential flaws include relying on online reservation systems that don’t account for walk-in traffic. Restaurants have the ability to limit or cease online reservations during peak periods, so the available information may be limited.
This academic study seemingly contradicts a recent study by retail researchers NPD Group that found the power of online recommendations and reviews to be relatively weak.
Only 6% of respondents were influenced by any form of online marketing, of which 27% were influenced by vague “general information,” and 14% by online reviews and recommendations.
Nonetheless, 6% still amounts to nearly 1 billion restaurant visits influenced by online marketing, and with a sizable boost from higher-rated reviews on Yelp, there is no reason for restaurants to not take notice. This also includes hotels that feature restaurants, and other travel-related businesses that rely heavily on word-of-mouth marketing that can be accelerated by sites like Yelp.
For the number geeks, check out the full Berkeley report here. It’s chock full of formulas and explanations that are fascinating-but-thick.
Nick Vivion was a senior reporter for Tnooz from August 2012 to July 2015.