Netflix Execs: Our Recommendation Engine Saves Us $1 Billion Per Year

As we’ve written before, online video platforms can reap big-time benefits by putting proper resources into discoverability. No one understands that idea better than Netflix, which has worked hard to ensure its recommendation algorithm highlights as much of its library as possible. According to a paper published by two Netflix execs and shared by The Motley Fool, the streaming video on-demand (SVOD) service’s algorithm saves it $1 billion per year.

The savings produced by the Netflix algorithm, as laid out by the platform’s VP of Product Innovation Carlos Uribe-Gomez and its Chief Product Officer Neil Hunt, show up in two major areas. Netflix’s recommendation engine is known for the hyper-specific categories it produces, and those genres can match the titles in the service’s catalog to the exact subscribers who will be interested in watching them. By getting the most mileage out of its library, Netflix therefore justifies the $6 billion it spends on content each year. Strong recommendations also increase the average watch time among viewers, thus keeping the clip at which Netflix loses subscribers — known as its “churn rate” — as low as possible.

The financial savings produced by Netflix’s recommendation engine can’t be backed up with hard numbers, but even if Uribe-Gomez and Hunt’s billion-dollar figure is a bit overstated, there’s no denying the amount of effort Netflix puts into the optimization of its algorithm. A recent change to it, for example, was reportedly implemented and perfected by a team of 70 engineers.

Here’s the kicker: As Netflix’s user base increases, its recommendation engine will only get better. More viewer data means more insights about viewer behavior. The end result: A strong algorithm.