Millions of people globally use Zomato every day to decide where to eat. While location and cost are factors a lot of folks consider, a restaurant’s rating also plays a role in one deciding whether or not to eat there. Given that it plays such a central role, we feel it’s important that a restaurant’s rating should be representative of the experiences people have had at that establishment.

On Urbanspoon, restaurants had an absolute rating system based on the percentage of people ‘Liking’ a restaurant. So if 80 out of 100 people ‘Liked’ a restaurant and 20 down-voted it, the restaurant’s score would be 80% – not taking into account the people who may not have enjoyed their experience there. This also makes restaurants vulnerable to spam ratings from unverified users, which can have an adverse effect on a restaurant’s percentage score.

Zomato uses a 5-point classroom-style grading model, where the distribution of scores in every city is normalized, resulting in ratings in each city being distributed over a normal curve. Additionally, a user’s credibility on Zomato – calculated by an algorithm based on the user’s activity – also affects how much their rating will affect a restaurant’s overall rating. This means a restaurant with a very high percentage score (let’s say 96%) on Urbanspoon won’t have a directly proportionate rating (4.8 in this case) on Zomato, because there are more factors at play than a simple ‘Like’.

‘Likes’ and ‘Dislikes’ from Urbanspoon have been migrated to Zomato as well, and appear as ‘Positive’ or ‘Negative’ labels on reviews, corresponding to users’ up-votes and down-votes. We’re encouraging migrated Urbanspoon users to add ratings to their votes, which will count towards the overall rating of the restaurants they’ve visited, so there might be a minor increase or decrease in restaurants’ overall ratings over time.

Using a normalized distribution method for restaurants also helps prevent restaurants in a city from getting clustered in the mid-range of the 5-point scale. It also lets people comparing two restaurants decide which one they think is ‘better’ – something we all do a lot of when picking a place to eat. The goal is to make things easy for the user to relate to – which we hope we have been able to achieve here.

We would love hear thoughts from Urbanspoon users and restaurants owners over at feedback@zomato.com