Every Monday Spotify selects new great music for you to discover. Many users will probably agree that Spotify’s taste in music aligns very well with their own. How is it possible that Spotify seems to hit the right note each week and provides you with songs you could repeat all day long?
“The function of background music is completely replaced by playlists with genres we are interested in and can play anytime. Moreover, we listen to music much more than we did to radio ever before”, wrote blogger Jo Caudron about his first experience with Spotify in 2011. By then, he already saw how powerful this music platform would become. Currently, Spotify has worldwide more than 170 million users, providing a huge amount of information to collect.
What does Spotify know about you?
So what does Spotify actually know of you? When you keep your basic settings: usernames, profile picture (from Facebook), people who are following you, the artists you listen to, playlists you made by yourself, music you play, the most listened songs of your profile, and the content you place and upload.
The perfect playlist for you
In order to create the perfect playlist, the music platform will not just simply look at which songs you listen to, but also uses algorithms that incorporate artist pages you’ve looked at and playlists you created yourself. Therefore, Spotify uses three kinds of strategies.
The first is collaborative filtering. Spotify categorizes their users based on their favorite songs. If you (person A) have some songs in common with person B you will receive songs that person B also prefers and vice versa. This could also work with playlists. If Spotify finds out that two of your most listened to songs are on a shared playlist, Spotify uses a third song from that playlist to add to your Discover Weekly Playlist.
Another strategy is natural language processing. Spotify searches on the world wide web to find out what people are saying about artist, especially: with which other artists are they compared? When both artists in these articles are described, Spotify suggests you would like both artists. A third strategy is using audio-analysis. New songs have probably not reached the news yet and may be quite unknown. Spotify uses the characteristics of these new songs to compare them with your most liked songs.
But is it the perfect playlist for you?
It is of course very handy that Spotify can provide you with personalized playlists and music through these algorithms. This way you know for sure that you only have to listen to music that you like. But is this really the case? If, for example, you share one account with your family, you will also get playlists of the music your dad likes. If this is totally not your piece of cake, this could be quite annoying. Another example, if you don’t share an account but your six-year-old wants to play some music, the algorithms will pick that up as well. So, considering this, there will be pros and cons to the Spotify algorithms.