2013-08-12

Co.Labs

Music Discovery UX Is Broken--Here’s How To Fix It

Music recommenders have failed to convince a decade’s worth of listeners to purposefully visit, pick an artist from a recommended playlist, and choose what to listen to. It’s time to fix that user experience. Here’s how to do it.



Despite what you may have heard, the state of music discovery is still in shambles. Even with Spotify and Rdio’s recent enhancements to their discovery features, they’re still competing for a very small group of deeply interested music fanatics. Online streaming services of all kinds are still pursuing discovery with a brute force mentality: If we can just suggest a few hundred more songs, maybe we can satisfy everyone! Unfortunately, the problem with discoverability isn’t actually discoverability: It’s the user experience that surrounds it.

The issue with music discovery sites used to be that they just weren’t very good at picking music the listener enjoyed. Everyone’s now much better at understanding the listener, but in trying to improve discovery algorithms, the algorithms themselves inadvertently became the sole focus of most music discovery. The result is essentially someone spouting off a thousand really good recommendations in a five-minute span, which is just as ineffective as if a friend gives you a single bad recommendation. It’s a terrible user experience: Being overwhelmed by too many choices, however good or bad, is not an efficient way to get average, non-enthusiast listeners to find their next favorite band.

The problem is similar to the one Apple foresaw when they developed the App Store. How many new apps per month can a typical user add to their daily workflow? At most, maybe one new app per week. To solve this problem, Apple promotes roughly 8-10 new and noteworthy apps each week. People who are enthusiastic about cool new apps can find them on their own. The new and noteworthy features are for the much larger group of users who only open the App Store once a month.

The same is true about music. A typical person just can’t sustain a new favorite artist every single day over the course of a year, which is why discovery playlists and stations just aren’t very useful. What if, instead of bombarding users with suggestions, discovery services followed the Twitter ads model? In place of more stations and playlists, these services could slip a recommendation based on the people you follow naturally into your stream at a comfortable pace that doesn’t overwhelm you. Users could check a few boxes in their profile to manage frequency and other customizations to music recommendations slipstreamed into their feed.

Spotify and Rdio’s discovery features are pretty good for constantly engaged music fanatics, but that’s not the majority of users--if it were, paid subscriptions to these services would be growing much faster. There’s no doubt that Rdio and Spotify have good suggestions for everyone, but the hardest part is actually getting people to take advantage of these recommendations. Music recommenders have failed to convince a decade’s worth of listeners to purposefully visit, pick an artist from a recommended playlist, and choose what to listen to. It’s time to fix that user experience.

[Image: Flickr user Kara Harms]