Feature
Spotify Unwrapped
Designer: Norman Zeng (Graphic Design)
How the Spotify Algorithm Creates Passive Listening 2020’s Spotify Wrapped, the computer-based year end review that analyzes and presents listeners’ musical habits, took many by surprise this year. While in the past listeners may have received “Indie Pop” or “Neo-Soul” as their Top Genre, this year one of Spotify’s main goals was to pinpoint more specific genres: this total tallying up somewhere around 5,000. Whether it be Swedish techno, Gypsy Punk, Deep Psychobilly, or Neo-Jazz Soul, Spotify’s newest set of goals is around defining every possible sub-genre (or sub-sub-subsub genre) that their audience may take an interest in. But where does this broad list of individualized genres go? How many people actually listen to Kentucky hard metal, and why doesn’t Malaysian Mandopop show up on your discover weekly playlist? Rather than using this analytical tool to introduce listeners to various sounds, Spotify uses it to further box in and segregate listeners’ tastes, ensuring that they never have to leave their comfortable corner of musical space.
Meeting the Algorithm I usually reference “The Algorithm” sarcastically, when a friend is in awe that Spotify’s Discover Weekly feature predicted their new favorite song. The algorithm, or more specifically multiple smaller algorithms, create personalized playlists for every unique Spotify user based on previous listening habits and one of the prized algorithmbased playlists is the Discover Weekly. Every week, a playlist is released for each spotify consumer based on a computer-generated analysis of their listening habits. The algorithm does this by analyzing groups of songs, and searching through playlists (both official and
Spring 2021
8
nonofficial) to find overlaps within these song groupings. For example, if you listen to Phoebe Bridgers’ Kyoto and Mitski’s Strawberry Blonde on repeat for a month, the algorithm will then search through thousands of playlists that feature both songs, and then choose another title that is frequently placed in playlists with the former two, then adding it to your playlist- why Soccer Mommy’s latest single being at the top of your Discover Weekly isn’t nearly as coincidental as it may seem. In addition to creating personally catered playlists, Spotify is also responsible for creating broad playlists with huge audiences. Some are formatted around genre- like the widely successful “RapCaviar,” while others are catered towards specific moods: “Down in the Dumps,” and still others are curated around a single artist: “Pop Smoke Radio.”Then, of course, there are those like the algorithm-created “The Most Beautiful Songs in the World” playlist. These playlists aren’t created through personal parameters, but rather broad terms: analysis of tempo, instrumentation, electronic influence, acoustic-ness, and ‘timbral emotion’. These parameters are understandable for artist and genre based playlists- Rap Caviar seeks certain intense rhythmic choices and lack of melodic structure, while the algorithm searching for additions to the Rock playlist searches for guitar and drums. However, these parameters become more vague when considering something like “Most Beautiful Songs in the World.”. A computer-based algorithm takes listener ownership out of deciding what ‘beauty’ consists of and defines something as seemingly subjective as “beauty” within the bounds of computerbased parameters.