Netflix is betting on its recommendation engine to get an edge over Prime Video, Hotstar

Shrabonti Bagchi April 5, 2017 4 min

Even though Amazon Prime and Hotstar, global entertainment streaming giant Netflix’s two biggest rivals in India, offer cheaper subscriptions, the one thing that Netflix demonstrably does better is recommendations. Its algorithm knows that since you watched Stranger Things and Black Mirror, you may like Brazilian sci-fi drama 3% or Cosmos: A Spacetime Odyssey, a science documentary series hosted by rockstar physicist Neil deGrasse Tyson.

This is also, of course, a factor of its broader and deeper catalogue of content compared to its biggest rivals globally, be it Amazon Prime or Hulu.

And Netflix knows one way it can capture more eyeballs in India is by betting bigger on its recommendation engine, which, let’s admit it, is leagues ahead of Amazon’s or Hotstar’s. The company has also crafted the algorithm keeping in mind what it calls “taste communities” — people who are separated by geography, but united by cultural preferences. This sets it apart from most other recommendation algorithms, which tend to go by ‘what people near you are watching’.

This sets it apart from most other recommendation algorithms, which tend to go by ‘what people near you are watching’.

Starting today, Netflix is retiring its five-star rating system and replacing it with a ‘thumbs-up’ and ‘thumbs-down’, which the company feels will be a more intuitive way to tell users to register their preferences. “A thumbs-up tells Netflix that you like something and want to see similar suggestions. A thumbs-down lets us know you aren’t interested in watching that title and we should stop suggesting it to you. You can still search for it, but we’ve heard what you were trying to tell us — you aren’t a fan — and it will no longer show up on your homepage,” a Netflix spokesperson told FactorDaily in an email.

The star ratings system also created ambiguity around taste compatibility, the company believes, because most viewers thought of it as an indicator of the quality of the content (which is how star ratings are traditionally used by film or TV reviewers) rather than taste compatibility — how much *you* liked a show. “Members were sometimes confused about how to think about rating a title and how it impacted their recommendations,” the spokesperson says.

The star ratings system also created ambiguity becasue most viewers thought of it as an indicator of the quality of the content  

Going forward, there will also be a personalized ‘% match score’. This score is a prediction of what Netflix thinks a viewer would enjoy watching and is based on its algorithms analyzing individual viewing habits and behaviour. “It is not a measure of overall popularity across the service,” says Netflix.

According to the spokesperson, one way that Netflix generates personalized recommendations for individual members involves identifying communities of other members with similar movie and TV show preferences, and then making recommendations based on what is popular within that ‘taste community’. Essentially, it’s using Big Data to help people discover content they never thought they’d like, whereas, as the company spokesperson puts it, algorithms are typically associated with creating echo chambers.

It’s using Big Data to help people discover content they never thought they’d like  

“Rather than looking at audiences with a segmented and narrow view, Netflix’s recommendations deliver the content an individual viewer will enjoy the most, wherever they are from. We aren’t looking at a member in the US versus a member in Japan,” says the spokesperson.

One example of this is the Brazilian Netflix original series 3%, which has travelled around the world, with more than 50% of viewing hours coming from international markets. The company thinks this might represent a real opportunity for Indian filmmakers and content creators to find new audiences around the world.

Does Netflix intend to use personalisation to get ahead of rivals like Amazon Prime and Hotstar in the Indian market? The company declined to answer directly. “We compete for a share of members’ time and spending for relaxation and stimulation, against linear networks, pay-per-view content, DVD watching, other internet networks, video gaming, web browsing, magazine reading, video piracy, and much more,” the spokesperson told FactorDaily.

Netflix CEO Reed Hastings, who was in India recently, said India was “hugely important” for the company in the long term, as one of the strongest internet markets. “Internet TV is the future over the next 10-20 years, given that linear TV will decline in viewing eventually,” Hastings told reporters.

Currently, most content streaming entities — from YouTube to Netflix and Amazon — are betting big on India, which is expected to become a billion dollar market for video on demand service by 2021. YouTube released its much-awaited made-for-India app, YouTube Go, today.


Disclosure: FactorDaily is owned by SourceCode Media, which counts Accel Partners, Blume Ventures and Vijay Shekhar Sharma among its investors. Accel Partners is an early investor in Flipkart. Vijay Shekhar Sharma is the founder of Paytm. None of FactorDaily’s investors have any influence on its reporting about India’s technology and startup ecosystem.