There’s a myth going around that music geeks are essential to good music curation.
It’s an attractive idea. Just look at BuzzFeed writer Reggie Ugwu’s glowing feature on the “small teams of anonymous, hardcore music fans [racing] to solve the record industry’s toughest problem.”
Indeed, Pandora, Apple, Spotify, Tidal, Google Play, and the rest bend over backwards to make clear how much they rely on music experts, and there seems to be no worse slander than saying a competitor is dependent on algorithms.
But do you really need a human to pick your music? In a world when we increasingly trust AI to pick our movies, dating options, email replies, scheduling, and much more? Don’t believe it.
Consider the range of human involvement in streaming music:
— On radio (e.g., Beats 1), human experts have total control over what you hear.
— On handmade playlists at Apple Music and Spotify, human experts choose tracks and track order, while AI recommends different playlists
— On handmade stations at Google Play, humans experts choose tracks, while AI recommends different stations and personalizes track order.
— On custom radio at Pandora, humans experts help categorise music, while AI chooses what you hear on any given station. On custom radio elsewhere, human experts have little-to-no involvement.
— On algorithmically generated playlists like Spotify’s Discover Weekly, humans experts have little-to-no involvement.
So how important are human music experts?
Well, radio can be fun, but it isn’t blowing anyway away, which is why everyone is moving to streaming music services in the first place.
As for the rest, there’s little evidence that handmade playlists are better than handmade stations or that either are better than custom radio or algorithmically generated playlists. There’s personal opinion, of course, and some people love Spotify or Apple Music’s handmade playlists, but plenty of people, including many music geeks, love the other options, too.
Personally, I subscribe to Google Play Music and listen primarily to custom radio and albums, recommended for me by AI. After testing all the options, I firmly believe it’s as good or better than the alternatives. Indeed, I’d rather listen to an infinitely personalizable feed of music than a limited set of handmade playlists.
The only data BuzzFeed cites for the popularity of handmade playlists is that 50% of Spotify users listen to them: so what? I wouldn’t expect anything else in an app that puts those playlists front and center. No doubt Google Play users listen to a lot of stations and Pandora users listen to a lot of custom radio.
Notably, the most popular product from the past year was Discover Weekly, which amazed millions of users by understanding their preferences on a deep level. Discover Weekly does all that with little-to-no human involvement, predicting what you’ll like through big data analysis of how your listening habits compare to other people’s listening habits.
All signs point to more stuff like this in our future. Apple’s big new feature this year is a Discover Weekly knockoff called Discovery Mix. Pretty much everyone is talking about getting deeper into personalised recommendations, which inevitably means recommendations that depend more on AI. Even Pandora has steadily shifted away from human-powered labelling in favour of big data analysis.
Music geeks still help, adding a meaningful personal touch here and there and making sure the algorithms are running smoothly, but they matter less and less.
Business Insider Emails & Alerts
Site highlights each day to your inbox.