New Spotify feature seems to miss the entire point of music

0
11

Can a classically trained violinist enjoy Metallica? Of course!

There’s something really, truly special about flipping from a band called Oozing Wound to, oh, the second act of Sergei Prokofiev’s Romeo and Juliet ballet. It’s why Spotify is one of the greatest products of the App Age, and it’s also why we have some issues with the service’s new "Spotify for Brands" page.

Visit spotify.me, which launched during the Cannes Lions Festival last month, and you’ll be treated to an assessment of your music-listening habits. We use that word carefully, because this is absolutely an evaluation of who you are as a consumer, as measured by what music you enjoy. Like any widely used online product trying to make a buck, Spotify collects a ton of data about its users and generates insights based on that data.

For example, according to this new site, 74 percent of the tracks I’ve listened to recently "are energetic":

And what that translates to in Spotifyese is this:

"You are high energy!"—a patently, hilariously false statement, which will be relevant later. Scroll down a bit further, and:

"Based on the playlists in your library, we’re guessing you like to soundtrack your soufflés. (And BBQs.)" Hmm, alright. (I don’t cook.)

At this point, it should be noted that Spotify absolutely does understand what kind of music I listen to. The top three genres on my spotify.me page are alternative rock, the hilariously specific post-doom metal, and indie rock, which I definitely like. But something’s clearly wrong with the generalizations the algorithms are making about my character.

So what? Well, a couple of things.

First of all, Spotify is trying to parlay its insights into meaningful relationships with advertisers. The full Spotify for Brands website hints at the possibilities: According to an analysis of listening habits and survey data from February 2017, Spotify concluded that millennials are more likely to "listen to music while running errands" and "pay for brands they like." Two overall categories of music-listeners are suggested: reliables and explorers. The latter will "stick to their favorites" but "recommend a brand or company to a friend," while explorers "seek out unfamiliar songs" and "try new restaurants."

Far be it from us to assert that Spotify’s research is absolutely wrong, per se, but it seems like a leap to connect listening habits to personality traits. I listen to a lot of really loud, fast metal music, and I couldn’t say that means a single thing about my personality: No doubt, I would trade every Mastodon album for the John Prine song my wife and I bonded over years ago.

And sure, it’s really unlikely that Spotify is approaching, like, Heinz Ketchup and selling an ad campaign based on 10 million metalheads who are maybe also sous-chefs because they listen to fast music. It’s besides the point technologically, because we are learning yet again that, in 2017, algorithms do not understand human emotions or thought processes no matter how desperately major tech companies insist they can.

What do the reams of data we generate using services like Spotify and Instagram amount to? Some worry that the answer is a lot, that so much can be gleaned from our subconscious behavior on these platforms that savvy organizations could shape our actions through machine learning in the very near future. "Because Person A has done X, Y, and Z on all of these platforms, we can get them to buy a cherry soda by showing them a specific type of advertisement through very targeted venues," the logic goes.

Right now, the reality is a bit closer to tech companies misunderstanding who their users are and what they want. Spotify is okay at making playlists for me based on what I’ve listened to, but come on:

That’s absurd! And yet, it’s telling: This is how consumer technology companies understand us. They force conclusions from big data because it’s what they’ve got—there are so many of us using their products that it’s the easiest, savviest way to figure out who we are.

Or at least try to. Because while it’s very clear that a robotic revolution is well underway, it’s a safe bet that it’s happening in labs and think tanks, not on our social apps, much less our streaming services.