One of Spotify’s better features are its personalised playlists. These compilations of tunes have expanded beyond Spotify’s Discover Weekly to also include other selections, like Release Radar and your Daily Mix playlists.
This has got to be one of the ways Spotify retains it's competitive edge. While a few of the additions may seem minor (I mean, how often do you really visit the “Decades” section in “Browse” in Spotify’s app?) The draw of personalised playlists is one key way Spotify is able to keep many of its now 140 million total users loyal to the app.
Such features also work to increase engagement across the application. Again, take the Discover Weekly feature, for example, it was so popular right off the bat, it scored over 40 million listeners in its first year of launch. Today, Spotify says that it makes a Discover Weekly for every user who has been on Spotify for a minimum of two weeks. Translated to human terms, this means that Spotify is making millions on millions of these playlists every Monday, or so they say.
And of course, people aren’t simply using Spotify because they can stream what they want when they want it, it's a new form of radio – a playlist-driven-on-demand experience that caters to what you like to hear. So how does Spotify do this amazing job of choosing those 30 songs per person per week? Check out the linked article below explaining Spotify's 3 Types of Recommendation Models...
Spotify’s 3 Types of Recommendation Models Spotify actually doesn’t use a single revolutionary recommendation model — instead, they mix together some of the best strategies used by other services to create its own uniquely powerful Discovery engine. To create Discover Weekly, there are three main types of recommendation models that Spotify employs: 1. Collaborative Filtering models (i.e. the ones that Last.fm originally used), which work by analysing your behaviour and others’ behaviour. 2. Natural Language Processing (NLP) models, which work by analysing text. 3. Audio models, which work by analysing the raw audio tracks themselves.