Instead of trying to delete all biased reviews, Zomato will now delete some of the obviously biased ones and hide some of the others from most users, the company mentioned in a blog post. It will do this by rolling out an anti-bias algorithm that it claims will help clean up biased reviews even retroactively.
Deleting only the most biased reviews will refrain from alerting biased review providers, which would otherwise help in detecting what kind of reviews are flagged and deleted. The algorithm will also consider a credibility score based on the user’ behavioral patterns on Zomato over time, helping the company cluster users in various ‘bias’ categories.
Zomato will also start providing less weightage to reviews of users that have had their content flagged and moderated in other reviews on Zomato. Other than this, the company will also start providing less weightage to reviews that are old, as restaurants can improve or decline in quality over time. The company mentions that it has made other tweaks to the algorithm as well, but disclosing these would make it easier for fake reviewers to workaround.
Financials: Zomato reported operational revenues of Rs 184.97 crore for the fiscal year 2015-16, Info Edge’s financials showed. In FY 15, Zomato had reported revenues of Rs 96.7 crore, a growth of 91.28%. However, Zomato reported an operating EBITDA loss of Rs 492.27 crore for the period. In contrast, the company reported an operating EBITDA loss of Rs 136 crore last year.
Business streamlining: Zomato’s focus on streamlining its business is showing through multiple measures it took last year and this year. In January this year, it shut down online ordering operations in 4 cities; Lucknow, Indore, Cochin and Coimbatore citing that the market size in these cities accounted for less than 2% of its order volumes. Then, it claimed that its online order volume continues to grow by 40% month on month in the rest of the cities.
In October last year, the 8 year old company laid off around 300 employees, about 10% of its 3,000 strong workforce, mostly from the US content teams who collected data from restaurants.