Are you curating your own music tastes or are you getting curated by the algorithm?
Looking further into how data driven editorial and algorithmic playlists shape our listening experience, and whether we should continue to use them?
When I first started to really get into music, a good chunk of my new discoveries came from Spotify’s personalized algorithmic recommendations like Discover Weekly and Artists Radio. I used to listen to Discover Weekly almost religiously just because it was so convenient, with a single press of a button I’d get a mixtape of recommendations tailored personally to what I liked, and that dopamine hit of “discovering” new artists kept me coming back for more.
Spotify has since continued to implement more and more data-driven AI functions into their platform to give the user as personal of an experience they can get with functions such as the;
Daylist - a mood based playlist that changes and updates throughout the day that reflect the change of your preferences at different times of the day. Often titled with some bullshit made up microgenres
AI DJ - a personalized DJ which uses an AI voice as your radio host which cycles through music you’ve listened to as well as new music
The new tech was admittedly fun and intriguing at the start, but the more I listened to it something felt kind of off. While I was being fed new artists and songs almost daily, it kind of felt I was listening to the same things I was already listening to, almost as if I was confined to a certain genre or vibe, I was in a rut for months and even though i was listening to all these new artists, it never felt like I was discovering something new.
So I decided to dive deeper into how Spotify are creating these playlists, and I came across a YouTube video of Anthony Fantano interview the journalist and music writer, Liz Pelly, the author of the book Mood Machine (A book which I will be referencing throughout this article, you can purchase it here).
But before we go into WHAT Spotify is doing that is messing with your listening habits we have to first understand WHY Spotify is doing so in the first place.
▸Spotify wasn’t made the “save” the music industry, it was a business opportunity
Spotify was founded in 2006 Stockholm, Sweden where at the time piracy was running rampant in the country, where in that year 1.2 million of the 9 million citizens took part in piracy. Sites like Napster and Pirate Bay were easily accessible avenues for consumers to pirate music which was a huge headache for large music labels.
Daniel Ek and Martin Lorentzon, the founders of Spotify often tout themselves as “saviors” of the music industry from pirating but that couldn’t be further from the truth. The two men had realized that a lot of the major labels had considered a Sweden to be a “lost market”, immediately pounced on the opportunity and stroke up deals with majors. In 2010 the biggest daily newspaper in Sweden, Dagens Nyheter reported that the majors were getting paid a rate up to six time higher than the independents. So while the majors like Universal, EMI, Sony and Warner which at the time owned 18 percent of Spotify were reporting streaming as one of their largest income, the independents were left questioning whether they would get to see any of the money.
So lets be clear, these guys aren’t music lovers nor do they care about the welfare of artists. They’re businessmen who prioritizes shareholders and profits over anything.

To be richer than someone like Taylor Swift? Surely something’s up.
▸The Push of Lean-back listening
In 2012 Spotify was looking for a way to grow into the mainstream, as such they conducted a study aimed at marketing “moments”; like studying, cleaning and going to the gym in search for a breakthrough. Having done the study they realized active listening was a small fraction of the experience on the app, there were way more hours using music as background filler, people who wanted a soundtrack to fill the silence. I guess it’s a similar phenomenon where you leave the TV on in the background while you doomscroll on your phone, people don’t really want to go through the bother of choosing music that they aren’t going to pay attention to and would rather leave that to Spotify. This was such a breakthrough for them they started to find ways to optimize the experience for the less engaged listener to make their music consumption as convenient as possible.
Spotify flooded their “Home” and “Browse” pages with playlists such as “Mellow Morning” and “Evening chill” which were filled with easy to digest, inoffensive music sorted by context and mood, making sure these playlist were as little clicks away as possible. With these chill playlists, Spotify are trying to sell the user the idea that listening to their playlists, it can help stabilize their moods. With titles like “Peaceful Piano”, “Chill Vibes” and “Calming Classical” serve as streambait to users, a form of sonic clickbait as Liz Pelly calls it. The sheer amount of chill playlist on the platform is just crazy, it’s almost impossible to avoid. And it doesn’t contain itself to genres like ambient or classical, to Spotify, anything can be chill-ified. Dance, Pop, R&B, you name it, Spotify has probably already gone and done it.
This push isn’t just on the consumer side, but also the artist side as well. Since monetization is shaped by how many streams you get, to get the most out of it, your song has to have that repeat value but more importantly, playlist friendly so it doesn’t get skipped and gets placed on those money-making playlists. It’s just not rewardable to put out challenging or experimental records.
In 2013 Spotify launched a marketing operation aimed at artists “Spotify for artists” or S4A where artists are encouraged to view their work through data-orientated lens. The app included stats that tracked streaming stats and playlists adds while the editorial side provides videos on how to utilize such data. These videos not only demonstrated how to use the given data, but more importantly nudges musicians in a direction more profitable for Spotify. That direction often being to slow down and mellow out your music to suit the chill agenda.
In 2017/18 era of pop music, music that was playlist friendly was rewarded handsomely. Take Billie Eilish for example, her muted, mid tempo and sorrowful songs were a big hit with the playlists thanks to the mellow elements, eventually catapulting her to stardom. As such frankly uninteresting and uninspiring musicians like Lauv who targeted these playlist placement thrived. Harder, aggressive music just was not as effective, leading the platform to be oversaturated with “chill” music as even indie artist started to lean into this softness to fit into the appeal of these mood playlists. To be honest the thought process of “oh the song that got playlisted did better than usual, i should change my style to fit this genre” is a very human thought process and could happen to any artist, you can’t really blame them for falling for Spotify’s agendas.
▸Ghost Artists
In 2017-18 there was an influx of artist with no bio, no information and zero digital footprint yet they were featured on a number of Jazz, Classical and LoFi Spotify playlists. Some speculated that Spotify was bloating these playlist with their own cheaper creations, while Spotify denied creating these fake artists, they did not deny adding them to the playlist. Turns out the speculations were half true, rather than making such artist, they were paying other corporations to make them (lol).
Spotify has a name for incorporating these fake artist, the Perfect Fit Content (PFC). Essentially a program to link up with a web of other companies to make low quality, mass produced slop which is sold to Spotify at a lower price, which Spotify then allocates them to the matching playlist, preferably one where no one is really paying attention so Spotify can benefit financially.
Managers would often pressure their employees to include the PFC songs into the editorial playlists and by 2023 more than 150 playlists with titles like “Deep Sleep”, “Cocktail Jazz” were almost entirely populated by PFC. This obviously doesn’t help the artist being replaced by PFC but it doesn’t benefit the artist making the PFC as they almost never retain the rights to their creation. With AI music starting to get more prominent, I won’t be surprised if Spotify started using that as a means to further boost their profits.
As a consumer, I sure as hell don’t want to be fed fake artists that are created for the sole sake of increasing Spotify’s profit margins.
▸Data and algorithms kills culture
Spotify tracks EVERYTHING on its platform, from what your listening history and songs you skipped to what you searched. All in an effort to prioritize and optimize tracks that has the most replay-ability. Spotify doesn’t care why a genre was created or the history behind it, the LoFi Hip Hop genre is a perfect example.
LoFi Hip Hop was originally just a bunch of musicians talking about legendary hip hop artists such as Madlib, J Dilla and Nujabes and how they push boundaries in the music industry and unorthodox sampling methods. Inspired by Dilla’s off the grid, dusty drum style, home based producers started posting their tracks on SoundCloud, where they tried to outdo each other by flipping samples. A reportedly fun activity for hiphop enthusiast, devoid of commercialism. Once the LoFi Beats culture moved from YouTube to Spotify, the genre changed into a play and forget, deep focus-orientated music thanks to the push of “lean-back listening”
With all the data pointing to tame beats being more profitable, LoFi hiphop lost all its character, no more crazy drum chops or sample flips, no experimentation, no boundaries being pushed. It was now all about making your music as consumable and inoffensive as possible to fit your music into these huge money-making playlists. Label curators like Lofi Girl became more influential than artists themselves and artist struggled to develop fanbases in a culture where they were essentially rendered anonymous. With PFC and AI coming into the fray, LoFi Hiphop is as good as dead.
Spotify doesn’t just kill genres, they make them up as well. Take “genres” like escape room or braindance. These aren’t real, they’re made up. Hyperpop, a niche internet micro-scene turned full blown pop phenomenon originally derived from a YouTube series called Hyperpop Origins by music journalist Noah Simons who interviewed dozens of artists from small scale labels and collectives. Hyperpop wasn’t a genre, it was a creative impulse and a community for queer and trans alternative people. In August 2019, Spotify rebranded its “Neon Party” playlist which featured PC Music artists into Hyperpop and it just took off. The word “hyperpop” was chosen because the word appeared in metadata collected by Spotify and not because they knew anything about the scene. Spotify started grouping artists from other genres which had no correlation to the hyperpop scene, completely misguiding listeners about the roots of hyperpop.
The official Spotify playlist had essentially flattened the scene, transforming it into music that could be sold to the masses, making sure it was commercialized and palatable. They could not care less if original voices were left out in the process.
Smaller, independent musicians trying to make a living are constantly being forced to dance to Spotify’s tune and were forced into a decision. Either conform your music to fit the data driven algorithms or retain your original style and culture and get stuck in algorithm hell.
▸ Are algorithmic recommendations really “discovery”?
With algorithms, there’s such a focus on the amounts of minutes listened and the replay-ability of a track. To say that your taste are solely represented by the things you listen to the most is probably really misguided, but that’s how this algorithm essentially thinks. By tracking your listening history, it aggregates your taste to feed it back to you. Back when I still listened to Discover Weekly and Release Radar, I sometimes found myself absorbing the music without really paying attention to any of the artists because of just how many recommendations there were. It’s basically quantity of quality, the algorithm tries to gauge what you like, throws a bunch of similar sounding music in hopes it sticks.
Some of the records or albums that I consider to be high up my list perhaps doesn’t have a lot of replay value but still holds emotional and cultural value in it. Take Bjork’s album “Debut” an album that helped me open my music taste to the more experimental side of music, compare that to perhaps newjeans’s “Super Shy”. While I probably stream Super Shy a lot more than Debut, it doesn’t necessarily mean I prefer it over Debut, but the algorithm doesn’t understand that.
Another issue is your taste will rarely diversify if you rely just on algorithms, the algorithm isn’t daring enough to take a chance and recommend you something new and fresh, in fact it’s in Spotify’s best interest to keep you in your comfort zone to keep you streaming and to keep skips to the minimum.
I guess this is just a personal take, but as someone who enjoys music, I want to be curious about as much different music as possible. Having “recommendations” confine you to a certain genre/space just seems restraining, and just boring after a while. I want to listen to things that pushes boundaries, I want to listen to whole new different genres and algorithms just can’t really give you that.
Conclusion
This isn’t me saying “Oh AI is terrible! This is why AI should never be used and banned in our country!” Truthfully algorithmic recommendations does have its uses and I have benefited from it quite a fair bit, finding a good chunk of talented artist through discover weekly, daylist or even the “Fans also like” section. I just think it’s important to be aware that Spotify is a business, financially it makes sense to tweak their algorithm to benefit them, it makes sense to pump in ghost artists an AI slop into their playlists (still really fucking unethical). Without acknowledging it, it’s easy to fall into the traps of lean-back listening or constantly being fed similar music over and over again with these discovery tools being a literal click away.
So not to say Spotify’s algorithmic and editorial playlist are completely worthless, I think they do have a place when trying to discover new music. But if you truly want to discover new music and genuinely find new artists or music, you’ve got to take the time to look for them, research about them, actually go through their discography. Challenge yourselves with new genres, see what you like and what you hate (you gotta know what you dislike to know what you like after all). So find out what you like by your own means, through your own discovery, wouldn’t that feel a whole lot more authentic, a whole lot more you?
Hopefully I shed some light on how Spotify’s algorithms work, and I’m going to work on a post on how to discover new music, fingers crossed its done in a few days. Once again big thanks to Liz Pelly’s book Mood Machine, it seriously blew my mind how badly these artist were getting fucked over and how our listening habits were carefully shaped to their benefit. If you want to purchase her book you can do so here! (It’s available on Amazon as well)
I’m so glad you mentioned the ‘ghost artist.’ I’m not sure if it’s the same thing, but on multiple occasions, I’ve seen songs labeled under an artist I listen to, and when I played the track, it was nothing like their usual style almost as if their account had been hijacked or falsely associated with the track. I wasn’t sure if it was due to poor security practices or if Spotify was up to something shady. This article definitely shed light on Spotify’s worst practices.