Dear Spotify: Stop trying to read our minds
Feel free to administer a Homeresque slap of the forehead if you’ve ever heard yourself saying these words: “Oh, I like all kinds of music.”
Truthfully? (Oh and ouch.) You like bluegrass and death metal and free jazz and opera?
How about polka? Nah. “Everybody hates polka,” David Pierce winked this week on Wired.
Pierce was writing about Discover Weekly, a new feature launched by Spotify.
It’s the latest attempt by a streaming-music service to transform itself from the place where you go to play a couple dozen songs over and over, and dutifully save a couple hundred others you’ll probably never listen to, to the first place you go to find new music.
“It’s a lovely idea, to think that someday Spotify or Apple Music might be so smart that the service knows exactly what you want to listen to or watch, and all you ever have to do is open the app and press a giant play button,” Pierce continues. “But that doesn’t work. People are finicky, especially with music.”
And even if someone does find the secret to knowing what we want before we know we want it, there’s still a huge stumbling block to that level of personalization: it’s based on the assumption that our capacity for not just new music but new artists is as deep as we like to think it is.
The truth is, there’s a relatively short list of musicians to which we feel a profound attachment, and there’s everybody else. If you love, say, David Bowie or Kendrick Lamar or Taylor Swift, wouldn’t you rather be able to find a whole bunch of stuff by them that you’ve never heard before than simply be directed to an artist who supposedly sounds like Bowie, Lamar or Swift?
It’s an argument that Berklee College of Music professor and big-time R.E.M. fan George Howard made convincingly in Forbes in a piece titled “Why Music Services Are Wasting Time Recommending New Music.”
What is the value of being presented with artists “similar to R.E.M.,” Howard asks. “No other artist is similar to R.E.M. in my brain. That’s why I love R.E.M.”
So let’s try a little experiment: What “related artists” does Spotify nudge you toward if you’re listening to Taylor Swift? Carrie Underwood, Kelly Clarkson, Shania Twain and Justin Bieber, among others. How useful is that?
Using R.E.M. as bait, meanwhile, brings up the disparate likes (and not especially helpful suggestions) of U2, the Lemonheads, Crowded House and Crash Test Dummies.
When you think about it, music recommendation is kind of like matchmaking, which perhaps explains why one of the smarter ideas is billed by TechCrunch as “Tinder for music discovery.”
Still in beta, the app lets you sample and swipe through songs till you find one you like, and adds your friends’ new favourite songs to the mix.
In the end, it’s worth remembering that the inexorable rush to streaming music negates one of the implicit functions of music recommendation.
“We’re not listening to streaming-music services so we can find the next song to buy,” Jason Notte noted on thestreet.com.
“We’re listening and subscribing to them so we don’t have to buy songs anymore.”
“It’s a lovely idea, to think that someday Spotify or Apple Music might be so smart that the service knows exactly what you want to listen to or watch . . . But that doesn’t work.” DAVID PIERCE WIRED WRITER