Why You’re Using Spotify Wrong
And how you can hijack the recommendation algorithm to improve your mood.
Have you ever had the eerie experience of typing just one word into a search query and having the rest auto complete perfectly? That has a happened to me even while searching the most seemingly random and esoteric of topics. It’s… uncomfortable.
The age of machine learning and big data have paved the path to create algorithms that know what you want even before you do. While it is unfortunate that most of our smartest technology seems geared to sell us more stuff, many of the same algorithms can be used deliberately to your advantage.
Spotify is the perfect example.
When I became a Spotify user back in 2013, my music taste was as diverse as Trump’s cabinet; which is to say, not very. My music taste ranged from the Anti-Flag and Bad Religion to Yellowcard and Zebrahead. Yet in the years since then, Spotify has identified what I liked about punk rock music and slowly introduced more variety to my recommendations.
The Spotify recommendation algorithm is amazing and has been written about extensively elsewhere. I would recommend this piece on Medium’s OneZero if you’re interested in learning more about it.
I don’t want to talk about how it works. I want to show you how to use it. The amazingly accurate recommendation algorithm can be your personal tool to happiness, relaxation, or whatever emotion you would like to classically condition yourself to feel.
For me, I built a mood boosting playlist. 2020 was a rough year and having a playlist that I could listen to that almost guaranteed I felt better was a lifesaver. While building a playlist is easy, curating one that harnesses THE POWER OF THE MACHINES is a bit trickier.
Here’s how I did it
Trending AI Articles:
1. Preparing for the Great Reset and The Future of Work in the New Normal
2. Feature Scaling in Machine Learning
Step One
First you need some variety. If your whole playlist is punk rock, that’s mostly what you’ll get as recommendations. Create a temporary playlist with the express purpose of tricking Spotify into thinking you have a wide variety of music that you like. You want the algorithm to “wake up” and get to work trying to figure out what you’re doing.
I started with some songs from Vini Vici, Roberto Cacciapaglia, Solomon Burke, Canned Heat, Alice et Moi, and Joe Cuba. The idea is to have as varied a playlist as possible that Spotify can start to analyze. Look for songs you like in genres you don’t typically listen to. If you don’t listen to this new playlist, you wont get new recommendations based off of the diverse songs.
Step Two
Create a conditioning playlist. This is where you will put the good recommendations. It could be named anything. Mine is “Relaxed Music”. It’s empty now, you’ll add songs that make you feel your targeted emotion. Not a lot to say about this so we’ll move on.
Step Three
Use the Discover Weekly playlist. You have to listen to the songs that Spotify will start to recommend. Do this regularly. Think of it as the inputs to your conditioning playlist.
This step can be tricky to get right. Really? Listening to a playlist is tricky? Yes.
Here’s why.
You’re not trying to analyze the songs. You’re trying to get lost in the moment and then temporarily take note of the soundtrack. For me, I listen to my headphones while I clean the kitchen at night. I’ve done it almost every night for a few years now (adulting) and so it is a task I don’t have to think about. I can almost meditate while I’m cleaning on auto-pilot.
Now when you’re lost in that moment but you can feel yourself feeling happy, relaxed, or comforted, quickly save that song to your conditioning playlist.
Step 4
Now listen to your conditioning playlist. Sometimes when I find a song that has me feeling my conditioned feeling, I’ll jump to my conditioning playlist while I finish cleaning. This will have the duel effect of telling Spotify that this is the type of song you’re looking for, as well as telling your brain that this is the playlist that you listen to when you feel happy. Classical conditioning.
Additionally, listen to this playlist when you’re doing something you really enjoy. For me, I like to have it on when we have friends over. Again, you want your brain to keep associating this playlist with a feeling and if you can create that feeling, use the playlist to reenforce it.
Step 5
Keep listening to your Discover Weekly and building your conditioning playlist. Over time, just like a slobbery Pavlovian dog, you’ll find that you can shape your mood with your playlist.
Conclusion
Artificial intelligence works much the same way as non-artificial intelligence. It makes connections and then reenforces connections that get used often. By being deliberate with the recommendation algorithm and to your own listening habits, you can quickly leverage one of the best recommendation tools in the world along with the largest library of music ever compiled to have fresh songs delivered to you regularly.
It will boost your mood and you might even diversify your tastes as a side benefit.
What do you think? Are there other algorithms that can be deliberately used to actually help you feel better? Are there non-machine recommended songs you would suggest for boosting your mood? Leave a comment. I would love to hear from you.
Don’t forget to give us your 👏 !
