Tuesday, November 25, 2008

Thing 18: Web 2.0 Awards

I've played around with Pandora before, but never thought to compare it to other sites like it. The Web 2.0 Awards has Pandora ranked #2 in the music category, with lower ratings than winner Last FM in most categories.

Both sites start with the user declaring a favorite band or song, then branching out with recommendations. In Pandora, recommendations are based on the Music Genome Project, an analysis of more-or-less identifiable qualities of music (like major chord structures, tight harmonies, or repetitive rhythms). As the site plays a new song, you can indicate that you enjoy it or that it's not to your taste. The site adds that information to it's data, makes an adjustment, and suggests another song. You can create multiple "stations" based on artists or songs and can combine playlists so you don't end up listening to the same stuff over and over (and over and over).

Last FM, as I understand it, uses a different methodology. Rather than try to isolate what you like, they assume you'll like what other people who also like your picks like. You can still tell them you don't like a song, so you aren't chained to their playlist. Last FM is more keyed in to the Web 2.0 ethic, with space for comments and other feedback, links to concert calendars, videos... Last FM has a more robust website than Pandora, but I still like Pandora better.

Both sites promise to turn you on to music you might not have thought to try before, but Last FM assumes that you are - more or less - like everybody else. They pigeonhole you. Naturally, it works most of the time. Most people are like most people. But if it works as a way of identifying which group you belong to, it isn't likely to work as a means to introduce you to new music. If I actually fit in the group that predictably, the group isn't likely to have many musical interests I don't already have myself. Pandora's method actually makes more sense. If you like a sound, find out what it is about the sound that you like, then find other pieces of music with that quality in common. You aren't being stereotyped. You're being shown other ways of looking at the universe of music.

It's like Readers' Advisory. If someone said they love mystery books, I could run through a list of the most popular mysteries and probably give them something to read. But if I really wanted to give them a book that suits them, I would need to ask some questions: which are your favorites? What stand out to you about them? Do you prefer hard-boiled crime mysteries or cozies? Does the gender of the protagonist matter to you? and so on. The why is more important than the what in readers' advisory; and in listeners' advisory.

No comments: