Over 20 years ago, Netflix.com launched and began offering subscription-based DVD-by-mail services. Today, it is the top streaming service for movies and television with over 200 million subscribers. What keeps Netflix on top? How do they keep their users interested? I believe the answer lies in Netflix’s advanced algorithm and recommender system. More than 80% of stream time on Netflix is made up of shows discovered through the recommendation system, so it is clearly pretty dang effective. To curate the best possible recommendations, the algorithm considers the user’s search history, viewing history, ratings of other titles, time of day they usually watch, data collected from other members with similar tastes, and so much more! Let’s dive into this system and find out what keeps us coming back.


Because Netflix is aiming to create a user experience that improves retention rate, they seek to make the interface as simple and engaging as possible. One example of this is the use of the two-tiered row-based ranking system. In simpler terms, the Netflix algorithm puts their strongest recommended titles for each user on the left of each row, and their strongest recommendations for categories towards the top of the page. For example, if the user watches a lot of documentaries, the Algorithm may place the “Documentaries” row at the top of the page, with the titles the user is most likely to be interested in on the left of the row. 

Netflix uses the technique of having rows with similar items to make the experience easier for the user, allowing them to ponder what they’re in the mood for. Other than traditional genres, some examples of these rows include “Because you watched___,” “Continue Watching,” Watch It Again,” etc. These categories may seem simple, but they are all calculated to increase the likelihood of the user watching a title. For example, the “Continue Watching,” section not only considers the time elapsed since viewing and sorts the titles by most recent, but it also considers the point in the movie or episode of abandonment, which device it was watched on, and more. 

(Netflix TechBlog)

Some parts of the algorithm are less obvious, like the tactic of changing the thumbnails for each user based on what they’re most likely to click on. For instance, if a user tends to watch a lot of romantic movies, they may put a picture of a couple kissing as the thumbnail for an action film, while another user who is interested in horror films may have a frightening image as the thumbnail. This technique attracts users to titles they may not have initially been interested in, which I think it’s quite clever!

Netflix doesn’t just want to help us make choices; it wants us to love these choices so much that we can’t stop watching their content. So what about the algorithm makes this stuff so binge-worthy? Netflix understands that people love to watch things they’re comfortable with, whether they’re rewatching titles or binge-watching every title in one of their favorite categories. Their system helps users discover what they like, and once they find a movie or tv show they love, they gain trust in Netflix and want to keep finding more. I know this from my own experience because I spend so much of my time in the “Watch it Again” section, or the “If you Liked ___, Then You’ll like ____” section. Nobody likes to be stressed as they figure out what to watch, so coming onto Netflix and having a variety of options you know you’ll like is what keeps users coming back!


Even if the main purpose of the Netflix algorithm may be motivated by money or maintaining a high number of subscribers, I think it’s a fascinating and pretty genius way to improve the user experience! Knowing all of the different factors contributing to the titles Netflix recommends, it’s easier to trust their suggestions!

Feature Image Credit: The National

+ posts