Are you fed up with Spotify recommending music that doesn’t match your tastes at all? Well, you’re in luck, as the music streaming platform is testing a new option that would allow users to deactivate personalized recommendations.
According to MacRumors, the new option could allow users to prevent songs listened to on their accounts by others (like family or friends) from affecting their personalized recommendations. Although it’s not entirely clear how it will work yet, it’s likely to be well-received by Spotify users who are a bit tired of the app not recommending songs that resonate with them.
Spotify’s recommendation system combines various methods based on machine learning and artificial intelligence. The service’s algorithm analyzes the history of listened songs, favorite genres, and the time of day users listen to music to create their personalized recommendations based on these parameters.
This process is combined with content-based filtering, which analyzes the audio of a song a user interacts with, considering characteristics such as rhythm, key, and tempo to recommend new songs with similar features.
Spotify‘s recommendation system also utilizes something called “collaborative filtering.” For instance, if two users have similar musical tastes, Spotify may recommend songs that one user has listened to but the other hasn’t.
However, this can lead Spotify to recommend popular songs more frequently than lesser-known ones, creating an “echo chamber” effect, which renders it practically useless for new users without listening history.
It’s unclear when the new option to deactivate personalized recommendations will appear for Spotify users or whether the trial is region-specific. The company simply states that it is “evaluating the possibility” and appreciates users trying it out to help improve Spotify.