That age is — drumroll, please — 42.
(Forty-somethings in the audience: This one goes out to you.)
Admittedly, the exact age of the average midlife crisis isn’t exactly mind-blowing research. But there’s more where that came from: By analyzing user data, Spotify has also nailed down the listening patterns people adopt over their lifetimes.
Young adults and 20-somethings, for instance, tend to stick to what’s on the top charts: your T Swifts, your Wiz Khalifas. During their 20s, however, most people — particularly men — start listening to less and less mainstream music as they explore artists and genres outside FM radio and as they return to the artists they liked growing up. By 35, most adults have stopped keeping up with whatever the hot new thing on MTV is.
But at 42, there’s a sharp regression: a return to pop music that doesn’t quite dissipate until the mid-40s. And it can’t be blamed on parents listening to their kids’ music, either — parents actually never have that little backwards-looking interest in pop music.
This is all a very good reminder not only that the Cool Dad archetype has some basis in reality, but also that Spotify is less a music company than a data one: Like all of the other major music-streaming platforms, its future lies in its ability to measure and predict your tastes better than anyone else.
To that end, Spotify acquired the “music intelligence” service Echo Nest in March of last year — this latest research was actually conducted by a former Echo Nest-er. Spotify can now pinpoint, among many, many other things, the exact sections of songs that listeners like most, the differing schedules of fans in New York and Laredo, and the exact songs a 55-year-old will want on his running playlist, versus a 25-year-old.
Spotify is far from the only company vying for data dominance here: Pandora employs a team of more than 80 curators, analysts and scientists solely for the purpose of perfecting personalization; Songza, which Google acquired last year, mines everything from the user’s location to the current weather when recommending playlists. (Wrote the Guardian’s Stuart Dredge last July: “Never underestimate the importance of data in a Google acquisition.”)
With all this research and data, perhaps the big question is: Why isn’t music personalization, well — any more accurate? You’d think that, if Spotify can predict what jazz tracks a 13-year-old girl will like most, they could also figure out that I never, under any circumstance, want to hear Jason Derulo.
It turns out that, for now, at least, algorithms can’t perfectly replicate the whims and weirdnesses of human taste. There’s just too much that data doesn’t capture: your mood, the relative loudness of the Metro, that bad teenage association you still have with “Bright Eyes.” Brian Whitman, Spotify’s “principal music scientist,” once mocked the very idea of a computer understanding complex human feelings: It’s “postmodern insanity,” he claimed.
And yet, thanks to the efforts of Whitman and his colleagues, that insanity becomes more real with each passing day! Just ask all the 42-year-olds jamming, on Spotify’s recommendation, to Ed Sheeran and Ariana Grande.