Learning to Live with Zomato

Urbanspoon was sold by its parent company, InterActiveCorp, to New Delhi based Zomato in January for around 55 million USD. Zomato already had a strong global presence with 330,000 listings in 33 countries. Acquiring Urbanspoon added another 600,000 listings, mostly in North America, and will give them a stronger footing to compete head to head with American giants like Yelp.  

So, what’s different?

Doesn’t support categories from the search bar, which is how most first-time users of the site are going to use it. For example searching for ‘pub’ or ‘bar’ returns no results found for Saskatoon. Likewise, searching for ‘restaurant’ only returns businesses with the word ‘restaurant’ in their name. To search within a category, you must choose one from drop down menus in the left-hand navigation bar. Hopefully this functionality is cleaned up in the future.

Displays “popular” reviews by default. The popularity of a review seems to be calculated based on the reviewer’s credibility. Credibility in turn is calculated based on a user’s activity - which we can infer to mean not only how many reviews they write, but also how many other people “like” their reviews, and possibly even how much credibility their friends have. For a number of our clients this has pulled up reviews, good or bad, that are a number of years old. This simply doesn’t give users a good picture of the experience restaurant customers have had recently. A restaurant could have gone downhill, or be under excellent new management; however, only the opinions of the most “popular” reviewers are displayed.

Has switched over to a five-star rating system from Urbanspoon’s unique percentage system based on likes and dislikes. Zomato’s ratings are also distributed on a normalized curve. A normalized curve artificially adjusts each rating so that the majority of restaurants will cluster around a 3.5 rating. This means that only a small percentage of the restaurants in Saskatoon can have a rating over 4.5, while the majority will be lumped into the mid-range of ratings.

While we are not going to miss Urbanspoon’s strange “likes-it” percentage system, we do feel that normalized distribution is a very poor decision because restaurants tend to be rated relative to other restaurants of the same tier. For example, the best greasy-spoon in the city with get a good rating amongst its patrons, while an average fine-dining restaurant will rank lower. Having a forced distribution is going to exaggerate this effect, so a fine-dining establishment that is struggling against stiff competition could be forced to the bottom of the distribution of all restaurants. We can see this in effect on Zomato by comparing Earl’s, Humpty’s and Bottega. Zomato pushes Earl’s below Humpty’s and then pushes Bottega even further down. While Humpty’s might provide an adequate family dining experience, it seems to us to be a weakness of Zomato’s algorithm to give them an artificial advantage against Earl’s and Bottega.

Zomato also claims that Urbanspoon had a very simple rating system where the percentage of reviewers who “liked” a restaurant translated directly to the percentage rating of the restaurant. And that the “dislikes” had no effect on the rating. However, having helped launch a few brand new businesses here in Saskatoon, at OpenSail we know this simply isn’t true. On Urbanspoon, each reviewer’s rating was weighted by a hidden algorithm. Now this isn’t surprising, most review sites do this; however, it does make us wonder what other information that Zomato is providing has been “over-simplified”?

Only one review per restaurant is allowed now, rather than multiple reviews. Users are expected to update their review to reflect their latest experiences. This is a net neutral change. It will take longer for a restaurant’s rating to reflect a change in management, yet it will be easier to get an upset customer to change a negative rating once they have been appeased.

Users can comment on each other’s reviews. In general, Zomato is positioning itself more as a “Facebook-for-foodies” than a simple review site. Getting customers more involved with each other and Zomato is a good decision for them as a site. It will also be beneficial for the restaurants that are paying attention to and interacting with their customers on Zomato.

Can we tell Zomato to eff off?

We’ve had more than one client wonder if they could just turn-away from Zomato and focus on other review sites; however, we don’t believe that is going to be possible – at least in the short to mid-term. Urbanspoon pulled in almost half of the total reviews generated in Saskatoon, and we expect Zomato to start at about the same point. This means they get far too much traffic to ignore, and we suspect users will like the more social direction Zomato is heading in. However, for our clients that use review solicitation online or in-store, nudging their customers to leave reviews on other sites, like Google Plus, is certainly an option.

Finally, while we aren’t very excited about Zomato yet, we need to be fair to them and give them some time to adjust. Tripling your database over-night is a herculean task and data errors and server problems are to be expected. We hope in the coming months, once the database merge is behind them, and Zomato is pulling in fresh data, that some of the concerns we have raised above will be smoothed out.