There is One Feature That is Killing Your Enjoyment of Music

Nobody wants the music to end. From automatic CD changers to early DJs seamlessly transitioning songs for hours on end to the first person to shout “Encore!”, we’ve always wanted to hear another song. Now, that desire is a reality. 

I work for the streaming service Audiomack, and like nearly every other streaming service, we have an autoplay feature. When autoplay is enabled, the service will use some sort of algorithm to play songs related to your previous queue. It will do this until whatever device you are listening on dies. These recommendations are usually based on some combination of two types of algorithms: content-based filtering and collaborative filtering.

• Content-Based Filtering: This methodology looks at the actual sonic characteristics, or metadata, associated with a recording – like BPM, loudness, genre, instrumentation, and release date, among others – to make a recommendation. This is the machine equivalent to recommending things that sound the same.

• Collaborative Filtering: This methodology looks at user listening habits rather than sonic characteristics. For example, if I listen to a ton of Bruce Springsteen and John Mellencamp, it’s going to find other users that listen to them and recommend whatever else they are listening to. If those users are listening to Tom Petty, then it will serve me some of his music.

These algorithms are important from both a business and listening perspective. If you are listening to music without a paid monthly subscription, then the service and artists are only getting paid when you click play and ads are being served. Algorithmic recommendations and autoplay keep you listening to music, consuming ads, and generating revenue for the service and artists.

Business considerations aside, these algorithms are important because they help people painlessly find music they are going to enjoy. As told in Tarzan Economics, a recent book by Will Page, the former chief economist of Spotify, over 100,000 songs are now onboarded to streaming services daily. That’s more music than was released in an entire calendar year in the 1980s. In short, it would have been possible to listen to every major rock release in 1982. That is no longer possible.

Algorithmic recommendations make it so listeners can find music they are interested in without having to spend hours searching. Like your average listener, I enjoy many of the algorithmic recommendation features built into contemporary music streaming services. Nevertheless, I turn off autoplay recommendations wherever I am listening. Let me explain why.

Silence has been built into the listening experience since the dawn of days. If you were sitting at home listening in the 1970s, you’d eventually have to get up and turn the record over. If you were enjoying some pianist at a local saloon in the 1870s, you’d eventually have to head home and go to bed. It’s in these moments of silence where we get to contemplate what we just heard, where we get to appreciate the unique nature of each song that we listened to. Here’s an example.

I enabled autoplay on my Spotify account the other day while I was listening to The Beach Boys’ classic album Pet Sounds. When the final track wound down, Paul McCartney’s “Junk” started playing. Then George Harrison’s “If Not For You”. Then The Left Banke’s “I’ve Got Something On My Mind”. Then John Lennon’s “Oh Yoko!”. By the time “Oh Yoko!” ended and Buddy Holly’s “Words of Love” started, I’d nearly forgotten that I had even listened to Pet Sounds. The vibe that Spotify’s algorithm had curated after Pet Sounds made complete sense, but it made my experience listening to the album worse.

Pet Sounds is an album that deserves contemplation. It’s a deep melodic and lyrical meditation on the difficulties of growing up, falling in love, and figuring out who you are. Because the album immediately transitioned to a Paul McCartney track, I never had time to think about that. It was just onto the next song. In other words, Pet Sounds’ final track “Caroline, No” was not the conclusion to a masterful album. It was just a musical commodity, an input to keep the 60s pop vibe going.

As I’ve said, algorithmic recommendations have their place. If you’re not sure what to listen to or want the equivalent to an endless radio station based on a specific feel, then they are great. I’ve discovered countless songs at the suggestion of an algorithm. Nevertheless, we should make sure to take a break from listening and bask in silence to really appreciate the hours of music that are available at the click of a button.

You can find Chris Dalla Riva’s weekly musing about music and analytics at his newsletter, Can’t Get Much Higher. You can check out his latest EP You Know I Can Be Dramatic wherever you stream music.