A problem that has plagued Zomato for very long has finally caught the attention of the restaurant-review major. The company is known to have started out-of-the-box approaches to acquire customers and has succeeded well in our country.
Biased reviews and comments on Zomato have been problematic not just for restaurants and cafes, but also Zomato, considering it is the largest and single-handedly the best restaurant-review site out there.
Zomato also has a set algorithm in place to weed out biased reviews from its portal, but some reviews keep cropping up over time and there needs to be a definite response to those. There will be businesses which bad mouth other businesses on the website, and try to improve their own ratings through paid reviews.
9th June marked the day Zomato rolled out a strong anti-bias algorithm to remove biased reviews retroactively. Titled Project Warp, the team at Zomato will also put sophisticated new bias checks on their website and app.
Features of Project Warp
While Zomato does not want to detail out the plan to extract biased reviews from its website, it has laid out some definition action plans –
1) Less deletions, more hiding – This is probably the smartest technique to cover biased reviews and not delete them altogether. While Zomato will be deleting the obvious biased ones, the less obvious ones will be hidden from their review section so that users of Zomato do not see the spam reviews. The one who has written the review will be able to view it, thinking the review has not been deleted.
2) User credibility scores – Every customer will have a score based on usage history on Zomato, and it makes sense to use this. This is very much like credit scores for credit card applications or loans, where a user’s history is judged to predict future actions.
The other two parts of the new algorithm cater to giving weightage to a customer’s moderation history, so that if the reviewer has been flagged a number of times before, the restaurant’s rating will not get affected by a larger multiple. The second part works to reduce the aging of the reviews, so that a deed done two years in the past, which does not hold true anymore, is slowly phased out of the review section.
Zomato also clarifies in their blog post, “Zomato ratings are not simple averages. For a city, we forcibly fit all restaurants and their ratings on a normal distribution curve. In short, there’s a lot that goes on under the hood to make sure that you get a true sense of what you can expect.”
Well it is out in the open now, a 3.9 rating of your favourite restaurant is low because of the normal distribution curve and not because other rated it lower than your expectation.
What this means for restaurants is that they keep working hard to improve their rating scores on Zomato and leave it to the company to figure out the scammers. A bad review could be genuine or biased, and we can rely on Zomato to decide that.
A lot of restaurant owners have started delisting themselves from the website because of the rating structure and reviews, and this is definitely a good sign that Zomato is taking steps to counter spammers.
Source: Zomato Blog
