There are several reasons behind charts and playlists that do not represent our musical diet. From imprecise algorithmic recommendation criteria to inattentive listening. Some depend on us, others are only linked to the platform. We tried to figure out what’s not working.
“Spotify Wrapped is out,” says a colleague sitting next to me. He has his phone and shows me the colorful screen on the streaming service. We are on our lunch break sitting around the table, and a few minutes ago the annual overview calculated by Spotify was made public. We take the smartphones and open the app. There are those who keep it low so as not to show it to others, those who scroll through the stories packaged by the streaming service with a perplexed look. Something doesn’t add up.
Almost all of us are surprised by the rankings. They are not what we expected. There’s a reason Spotify Wrapped is one distorted mirror of your musical diet. There are trivial elements that can affect the listening data such as forgotten playlists in the backgroundshared accounts or friends who use our Spotify on car trips. Beyond the contingencies, however, there is a much more complex underlying problem.
Original sin is the algorithmic distribution which immediately distinguished Spotify from other music streaming platforms. Songs can be badly categorized, placed in playlists where genres and artists are crammed without a common denominator. And then there are the unclear criteria recommendation algorithmic. We therefore risk adapting our listening habits into “taste profiles”, which are measured and then re-proposed using a set of parameters.
So our brain can rot while we watch social media: the psychologist explains what brain rot is
This tendency is reinforced by passive or background listening, induced by endlessly playing playlist. It seems like the inevitable drift of digitalized capitalism, where the consumer is trapped in its own feedback loop. But let’s clarify things and then some practical example.
Four banal reasons that turn your Wrapped upside down
Let’s start with the simple things. Simply put, if you let your friends choose the songs during a trip or at an evening, it’s not that strange to see genres and artists that aren’t part of your listening list appear on your Spotify Wrapped. Not only that, even if you have decided to sacrifice your Spotify like musical background during a party (perhaps lasting hours) it is probable that some anomalies will arise. Inside playlists like “Greatest Hits of the 2000s” there is everything. Often artists and songs you don’t listen to.
Not only that, there is a second, more incisive degree of influence. Let’s take an example, if Spotify records four plays of Britney Spears, to remain faithful to the case of the “Greatest Hits of the 2000s” playlists it is likely that the following week it will offer you something similar in your Discover Weeklythus strengthening the influence process. Spotify Wrapped, then, does not consider the last month of listening. It has never released an end date for the analysis, but it is estimated that it collects data for its Wrapped up to October 31st, everything you listened to in November, so the music you hear as most representative, is automatically excluded.
Finally, there is a fourth “trivial” element: the free account. If you don’t pay you can listen to all the music you want, it’s true, but under conditions. For example, there is a maximum of six skips an hour (skip the track), it is therefore not strange to stumble upon more than one artist (present in a chosen playlist) who we don’t want to listen but we can’t skip it. And even in this case there is a risk of falling into the algorithmic recommendation loop. I’ll give you an example, in the playlist 90s garagethere are Kiss (strange but true), I don’t like them, but if I don’t skip them because I’ve run out of skips available, the algorithm will not only record them as listening data but will re-propose them to me through its recommendation mechanisms. A paid account therefore ensures a more faithful photograph than the free version, where often the randomness dominates listening choices.
How we get stuck in the hamster wheel
Basically, however, there is a structural problem linked to what we can define algorithmic playlists. Spotify tracks your music, categorizes it, measures your listening habits compared to other users, and uses this information to suggest songs you might like. In other words, it compares listening patterns to make recommendations. Then based on our data it determines which songs fall into categories similar to our preferences. In this way it generates flavor profiles for personalized recommendations.
These flavor profiles are based on both explicit feedback (saved or skipped songs), both implicit (listening time and repetitions) and are used to create for example your “Discover Weekly“. We thus remain stuck in the hamster wheel. The playlist that we start every week out of curiosity or laziness then forms the basis for our Spotify Wrapped; if a song is repeated enough, it then appears in your top 101 lists and Spotify.
Fluid market and inattentive listening: this is how our Wrapped was born
Let’s go even deeper, because algorithmic reproduction has a flaw in its premises: the genre. Even before Spotify, the music market exploited the concept of genre to build an industry, intercept fan bases, and set up clearly identifiable record labels. The genre, however, is one conventional category which identifies and classifies artists and songs based on affinity criteria. With the exception of a few cases, to give a popular national example, Nirvana and grunge, it is very difficult to include a project in the genre box. Lana Del Rey is defined as “pop, indie R&B, classical, chamber pop, synthpop”, she’s a bit of all of that, but at the same time she’s not really any of those genres. Not only that, the music market of the new millennium is churning out above all post-genre artistsTherefore musical hybrids which slide on different influences.
Paradoxically in this one post-genre rhetoric Spotify is also involved the entire digital music market. Millennials or ‘digital natives’ are the first generation to have the entire world’s music at their fingertips and this has influenced both artists and listeners. It is clear that defined genres still exist, it is difficult to transform a reggaeton rhythm into something else, yet the tendency, also given by the countless featuring, is to fluidize. However, algorithms do not move well in fluid. On the contrary. They need well-defined categories to apply the formulas that return our annual statistics.
The result is that to make algorithms coexist within a fluid market, the hand is forced. So they slip into a playlist or pop rock recommendations, projects that maybe aren’t pop rock, too, but not only. Finally, it is also worth considering countless playlists created by users, from those based on mood (happy, sad, melancholy) to those built on genre, which are often a chaotic drink of artists and songs that we descend without paying too much attention.
Because at the end of this long reconstruction we must also point the finger at us. The incorrect statistics are also the result of our own inattentive listeningto the point that we are surprised if Billy Eilish appeared among the most listened to artists of the year. Spotify has become a sound treadmill that flows and takes us nowhere. It’s an undercurrent driven by algorithms. And one wonders: “Do we really listen to what goes in our ears every day?”.
