Latest market-cap update: Pandora is a $6.7-billion company as of this writing, trading at all-time highs.
But that’s not the mains news. Today Pandora announced the addition of personalized recommendations in its iOS and Android mobile apps. Whether that justifies a stock spike is arguable, but the new feature does help solve a usability problem for Pandora, and could increase listener hours while retaining active users — two of Pandora’s key metrics.
The usability problem might best be described as paralysis. Pandora positions itself as a reinvention of radio, but as an interactive service it doesn’t just turn on like radio does. Pandora works best when you tell it what you like. In fact, for a new user, it doesn’t work at all until that step is taken. Even long-standing advocates of Internet radio get stuck in that paralysis moment, rummaging their thoughts for an artist or song to build a station around. Pandora’s now takes responsibility for suggesting new stations. Naturally, it works best for active users who have already personalized the platform to some extent through station-building, voting on tracks, and skipping tracks.
How well does it work? At first blush, in an active test account, the recommendations seemed both obvious and incomplete. The account we used was stocked with stations in the classical, blues, and Celtic genres. Pandora recommended three new blues stations, appending them to the account’s station list. Each station was seeded with a star blues performer, well known to anyone who follows the genre. So — useful, but not valuable music discovery.
However, clicking More Recommendations lengthened the suggestions and added detail, listing bands that would be heard within each station. That’s helpful. Even so, the rampant overlap of artists appearing in multiple recommended stations conveyed a message that Pandora lacked an adventurous repertoire, or concluded that the account owner has narrow taste.
Throughout the testing, classical and Celtic stations were never recommended, despite the account’s existing stations in those genres.
The new feature would really come to life in accounts seeded with song stations, not artist stations. For listeners who know what they like when they hear it, but aren’t genre specialists, Pandora’s Music Genome can identify musical characteristics and recommend cross-category stations that play those musical qualities.
At any rate, the update brings new choices to the app, gets listening started faster, and is clearly an activity and user-retainment tactic.