How Spotify is using machine learning to offer you music you’ll love

Technology has long been a part of the music industry. When multi-trackrecording came into force in the 1950s, it was the musical equivalent of themoon landing. Before that, if you wanted to record a song, you had to do it allin one take with every instrument that you wanted on the track.

Then in 1966 Arbiter Electronics released the ‘Fuzz Face’, a key componentto Jimi Hendrix’s sound, which recreated the distortion originally found inbusted speakers. The concept of having pedals that could alter the sound ofyour instrument was not new, but Hendrix and Keith Richards bringing these‘stomp-boxes’ into the mainstream revolutionised the way that music was played.

Now in 2019, machine learning is revolutionising the waythat we find music.

Anyone who has used Spotify’s ‘Discover Weekly’ playlist hasseen how eerily accurate it can be in suggesting new music to you. This is noaccident; Spotify has invested heavily in its machine learning capabilities aspart of its strategy to assert market dominance over heavy competition fromcompanies such as Pandora, Apple Music, Tidal, SoundCloud, Amazon and Google. Spotify’sIPO prospectus said as much with it promising to “continue to invest in ourartificial intelligence and machine learning capabilities to deepen thepersonalised experience that we offer to all of our users”, and that “thispersonalised experience is a key competitive advantage.”

To achieve this, Spotifyemploys three types of machine learning: collaborative filtering, naturallanguage processing (NLP), and raw audio models. Collaborative filteringinvolves recommendations based on those with similar tastes to yourself. WithNLP, Spotify analyses a plethora of data sets from blogs to articles, bandprofiles, song metadata and more to find which artists are commonly mentionedalongside each other.

Raw audio models then are the pièce de résistance ofSpotify’s machine learning strategy. While you may be aware of other artists ina similar genre, or may know who has collaborated with who, this is the part ofthe system that allows Spotify to throw you a curveball that you may never havethought of but end up loving. By analysing the musical elements of a song(tempo, time signature, key, etc.), Spotify is able to recommend you songs thathave similar features to what you listen to most from artists who aren’t in anyway associated with those you already listen to.

This is fundamentally changing the way that we find musicand the way that artists are discovered. A recent example of an artist who wasbrought into mainstream consciousness through Spotify’s algorithm is DermotKennedy. One of his songs ‘Glory’ was picked up by the algorithm and was placedon users Discover Weekly, and soon after his songs were performing so well thatthe CEO of Spotify, Daniel Ek, was notified. Kennedyhimself recently emailed Ek to thank him for all that his platform has donefor him. The success his EP saw from streaming allowed Kennedy to make a livingpurely from royalties and gave him great leverage when it came time to signwith a label.

However, the attitude towards streaming platforms is not always aspositive as my colleague Markrecently mentioned in his blog post. Many legacy artists are resisting theurge to transition to streaming platforms due to the low royalty rates (reportedlybetween $0.006 to $0.0084 per stream on Spotify). Thiscontentiousness may be due to the fact that revenue from streaming is only afraction of what the revenue from physical sales used to be, and these artistsare reluctant to embrace the changing landscape of the music industry.

With the age of DJ’s dictating musical taste to the massescoming to an end, all that’s left to wonder is who will be the next artist tobe crowned by Spotify’s algorithms.