Filed under: art, economics, music, Social Media, technology | Tags: Adventurous, algorithms, Elvis Costello, Indigo Girls, International TV, music, Netlfix, playlist, popular, shuffle, songs, spotify, Top Artist, Top Song
At the end of the year, Spotify sends everyone (with an account) a summary of their year on the app/website. They’ll tell you your Top Song and your Top Artist – that is, the stuff you listened to the most. Sometimes it’ll assign you a personality based on this information. One year they told me I was Adventurous because I “listen to non-mainstream artists 100% more than the average Spotify listener”. Aside from the suspect percentage, I liked this personality assignment. I like to be seen as adventurous. (Handily, this year, Spotify told me that the personality of my podcast listeners was The Adventurer, so I guess we all have something in common.)
But I’ve realized that this yearly accounting of my listening habits is only an effect of the algorithm that runs the thing. It’s given me some insight into how a lot of algorithms work and, given how dominated our lives have become by algorithms, how our cultural moment tends to work. I’ll explain.
The last couple of years, I’ve been creating a playlist at the start of the year into which I put all the music I want to make sure I remember to listen to. I called last year’s playlist “New Moment” – probably because I called the playlist the year before “The Moment.” “New Moment,” by the end of 2022, had 42 hours and 30 minutes of music in it. (I was a little overly liberal with the “Add to Playlist” button in 2022.) I generally just hit shuffle and let it make me a randomized radio station full of things I was interested in hearing. The 2022 list began with four Lake Street Dive albums, three Stromae albums and a LOT of bossa nova. It also featured seven Indigo Girls albums because I realized I wasn’t really up to date with their catalogue so I wanted to catch up. Sprinkled around all these albums were single songs I wanted to get to know.
However, even with 42+ hours of music, I noticed that the shuffle function tended to play me the same handful of Indigo Girls songs. I found it odd, though I did appreciate that it got me to learn the lyrics of some of their more recent songs. But I did wonder why this supposedly random shuffle function returned again and again to similar material, especially given how much there was to choose from.
It wasn’t that I had more Indigo Girls on that list with those seven albums. There were nine Elvis Costello albums (for similar reasons) and in my year-end sum-up, he didn’t even make the top five artists. Who was number one? The Indigo Girls, of course. Now, I love the Indigo Girls and have done since I first heard “Closer to Fine” in 1989 but in 2022, I hardly ever purposefully pushed play on their music. (Exception: “Prince of Darkness” because I was learning it for the podcast.) The fact of the matter is, Spotify chose to play me the Indigo Girls over and over, even when I started skipping them, and then Spotify told me they were my #1 Artist in 2022. Did Spotify select them, purposefully, to be my top artist? I doubt it. I suspect the algorithm, which, like most algorithms, privileges popular content, played some Indigo Girls and played it again because it had been played before and before we knew it, these songs were the most popular on my list. Spotify created that popularity, probably out of the algorithm’s formula that continuously weights popularity. Meanwhile there are so many songs that I would like to have heard in this playlist, songs by people who never had a top 100 hit or a top ten album. And those songs that Spotify never played me might never become popular since they weren’t popular in the first place. Spotify is unlikely to be promoting those unpopular songs anywhere else, given how “adventurous” I am. Spotify likes popular music and makes it more popular.
Our social media posts work this way, too. The way to have your post be a hit (be it on Facebook, Instagram, Twitter, TikTok and beyond) is to have it already have a high number of clicks. The way to have more people see your news is to have had people see your news before. Those algorithms weigh previous engagement more heavily than anything else.
I think this makes for an increasingly less interesting world. If we can only see what is popular, we are missing some of the most interesting songs, people or art. It’s how we end up with a highly polarized art world. It’s why there’s no middle class left in music (or art, or theatre, or film, or dance or…). You’re either Taylor Swift or no one. Bands who had hit songs years ago, who get a million listens on Spotify don’t make any real money for those million listens – because to get real money for Spotify, you need at least ten million listens. 300k listens is about $80. And the system encourages the imbalance. I feel fairly confident that if I’d had Taylor Swift on my New Moment playlist, the algorithm, leaning toward its most popular songs, would have quickly made her my number one artist in 2022. These algorithms are actually pretty stupid, even though they seem smart.
For example, Netlfix’s algorithm has worked out that it can get me to watch its International TV programming without much effort. They flash me a group of older women doing heists in Poland and I don’t need to be advertised to twice, no sir. But despite its having figured that out, it’s only ever as smart as the previous show I watched so if I’m watching a Korean romantic comedy, it’ll show me nothing but Korean romantic comedies for a while. I like to switch around to try and confuse it but it never works. And, of course, it’s only showing me the most popular of its offerings, thereby reinforcing its own patterns and hits. I don’t have another way of finding International TV so my viewing is entirely dominated by what Netflix has already found to be popular. It’s essentially self-canceling its unpopular shows by only pushing forward what’s already popular. And the thing is – I’d love to watch some unpopular stuff. I’d love to see something from countries that I don’t know a lot about. I want to see things that are beloved by small handfuls of people – not just the megahits of wealthy nations. But the algorithms aren’t built that way and over and over again, they show us the things that other people like, leaving interesting beautiful things to languish in obscurity. A world of only popular things is very dull and I think a little dangerous. Please, technologists, if you can’t break us out of the popular stuff, can you at least make a setting where we could choose to try less popular stuff? I think it would help tip things back to a more reasonable world. Next year, I want my top artist to be someone none of us has ever heard of before.
To listen to me read THIS blog on Spotify, click here.
This post was brought to you by my patrons on Patreon.
They also bring you the podcast version of the blog.
It’s also called Songs for the Struggling Artist
You can find the podcast on iTunes, Stitcher, Spotify or wherever you get your podcasts.
Every podcast features a song at the end. Some of those songs are on Spotify, my website, ReverbNation, Deezer and iTunes
*
Want to help me stay adventurous?
Become my patron on Patreon.
Click HERE to Check out my Patreon Page
*
If you liked the blog and would like to give a dollar (or more!) put it in the PayPal digital hat. https://www.paypal.me/strugglingartist
Or buy me a “coffee” (or several!) on Kofi – ko-fi.com/emilyrainbowdavis
Leave a Comment so far
Leave a comment
Songs for the Struggling Artist 