Attention data is the stuff people pay attention to (and don’t pay attention to). All we need to do is record what people do. Of course this sounds a lot like spyware, so better to get people to spy on themselves and share the data… that’s myware. People will want to do this if they get a benefit out of it. And in this case they do– in the form of musical recommendations and “personal web radio” broadcasts.

They created AudioScrobbler, a set of plugins for many media players that records each song one listens to and submits that data to their servers. A single song is a submission– they get about 8 million submissions per day (over a billion in 2005).

Using this database they can algorithmically determine other artists you will probably like, but may never have heard of. Users can browse what others are listening to, using these connections, or listen to personalized streams that they can interact with by skipping, banning, or indicating their affection for particular tracks.

Musical taste is complex and diverse, so they introduced tagging which lets the users self-organize and make sense of the data while Last.fm uses the tags to make sense of the aggregate data.

Back to the question of why users participate: Last.fm provides a valuable and compelling service. But users remain in control. XML feeds are provided from profiles, so you can export data, integrate into your blog, etc. Everything is CC licensed. If you don’t like something you can edit it out, or nuke your account.

It’s time to put the listener in charge– that’s the real music revolution.

[ruminate etech etech06 lastfm myware]