no. 18: algorithm fatigue.

on spotify's viral feature launches.

over the weekend i went to the jungle concert in los angeles. it was arguably the best concert i have ever seen. when i talked to many of my friends about jungle, i realized that only the ones who had eerily similar taste in music to me knew who they were. when i probed further, i came to realize they only discovered jungle because of spotify’s recommended music for them. we quickly compared our daylist playlists and our recommended music, and in that moment i realized our suggested artists were nearly identical. music discovery is no longer what it used to be - i know that. listening to the radio, shuffling through your friend’s music - whatever i was doing before i joined spotify - is long ancient. whether technology as a whole is to blame or spotify is, i don’t know, but this became a larger question for me. so, i have decided that the music discovery process is broken, thanks to spotify.

reddit agrees with me.

yes, i am not the first person to claim this. and i am not the only one who is tired of spotify. as of june 2024, spotify has over 626 million monthly active users, including 246 million paying subscribers. spotify is the world's most popular audio streaming subscription service, with users in more than 180 markets. given that spotify is the leading provider for music streaming in the world, you would think we would be able to access music outside of our genres we like more easily. in my experience using the app, i get recommended the same song in a different font. music discovery should not be as passive as it has become: “as we grow accustomed to the convenience of shuffling a pre-generated playlist, we have completely forgotten that discovering music is supposed to be an active exercise.”

if you have seen the daylist feature utilizing ai that they rolled out, it is absolutely ridiculous. it has been a hit because it’s easy to laugh at and immediately share your “delulu situationship tuesday morning” with your friends.¹ 4 out of 5 spotify users actually point to the platform’s personalized offerings as what they like most about the brand — not because any of the music is actually exploratory. my biggest gripe with all of this is “the algorithm assumes no risk, [it is] simply offering what’s mathematically sound.” this all just feels like an experiment by spotify to say that they are using ai in a novel way when its really not that revolutionary. spotify is king at making instagram viral-worthy features (spotify wrapped and now daylist), with no real value add. and it’s working for them really well as a company. but, when they rolled out daylist, i wondered does my music discovery even need to be this curated?

a daylist example. they are usually more insane than this.

these are some examples of features i think would be really great to see from spotify, to beat my algorithm fatigue. the product person in me emerges yet again (i know, its annoying):

• a heavier emphasis on community and organic discovery - i remember i used to be able to see my friends activity a few years ago. while i still think this is a feature on the desktop view, i think this would help me discover music in a less algorithm-forced way. it allows for actual organic and non-forced discovery.

  

• utilizing ai in a way i can discover on my own, instead of what it is now - i wish i could ask the app “what is my top listened to artist right now?” “what are other artists that are emerging in this genre?” and get real responses based on your data. i don’t like that this information is forced upon me in the app. why can’t i have controls like i do when i use chatGPT? this might not be for everyone, but i personally wish i could ask the app questions so i could explore on my own. it’s essentially letting me dig into the data in the way i want to, not in the form of pre-conceived and unoriginal daylists.

  

  spotify dj feature, where you can shuffle music using an ai dj.

maybe, we just don’t care to beat the algorithms. maybe, we just want to see the same music in a different font. and if that’s the case, spotify is doing all the right things. a similar phenomenon is happening with instagram - where algorithms are trying to predict what you want to see, all the time. to the point where i can’t have a logical chronological view on my instagram feed. ai and hyper-personalization doesn’t need to infiltrate of every part of every technology experience. sometimes things are best in moderation. if you’re anything like me, how are you beating the algorithm fatigue?

as a side note, i am an active spotify user and i enjoy using the app. they have a brilliant product marketing team from my perspective. these are just notes on what i think could be done better and facilitate a more organic discovery process.

a little research for me:

---

📚 what i have been reading lately, linked here for you:

1. see a great piece by the technology review by mit on how to “break free of the spotify algorithm”

2. the daylist feature review by forbes

3. a good post on how searches for spotify's daylist spiked 20,000% after trend took over instagram

4. nytimes on algorithms

1. this is real playlist title that my friend got a few weeks ago.