Last.fm

I'm going to take a slightly different approach from the other posts and simply write about new technology that I think is pretty cool.  Last.fm is a website that aggregates music play across the internet using interestingly-titled Audioscrobbler technology (the service was first available on audioscrobbler.com).  Part Friendster/Myspace, part Nielson Ratings, Last.Fm essentially tracks what music its registered users listen to on their computers via a plug in that works with Winamp, Windows Media Player, iTunes, etc, tracks that users play through a user profile, tracks artist and song play across the network and links "similar artists" based on usage rather than editorial opinion.  You can search for your favorite band, see how many plays they have across the network, see which of their songs get the most play, see who their biggest fans are (by usage), see who has similar musical taste to yours (listed as your "neighbours") and add those folks as friends.

The value of this application to record companies and bands is pretty tremendous now that unit sales and radio play are such incomplete metrics.  Last.fm is especially useful as a buzz tracker for blog hype-driven new acts like the Arcade Fire, Bloc Party and Clap Your Hands Say Yeah! whose success is difficult to capture in record sales, video rotation or conventional radio play.  There are a bunch of forums although the main community action seems to be based mostly in the charts and the "similar artists" lists.  For users, especially new music early adopters and obsessive music geeks, it's a clean, uncluttered (a la the Google school of UI design), unspoiled (so far there doesn't seem to be commercialized "placements" of songs or artists to ruin the party) and utterly engrossing community technology/site.  Like usenet newsgroups (rec.music.pavement anyone?), internet radio, blogs, Myspace, & podcasts before it, Last.fm is a cool usage of web-based technology for devout, adventurous listeners.
 
As a newish technology/website there are obviously some warts: the site is often slow or offline as the company presumably rushes to upgrade its infrastructure to handle the traffic and data crunching and the data integrity is reliant on properly tagged mp3, which most people don't know or care about much less maintain.  Plus, with no indication on what the technologies penetration level actually is, its hard to project how popular an artist or song is in absolute terms.  Still, it's awfully cool and engrossing (to me anyway).

I remember how exciting it was to surf through Friendster (oh summer of 2003, where have thou gone?), silently judging what people listed as their favorite music, movies and books.  With Last.fm I can check actual usage metrics and see who's listening to what and how often.  With Friendster, you can claim that you're into Sonic Youth or Neu! in an attempt to gain hipster brownie points but with Last.fm you can't hide your listening habits.  There needs to be a literary equivalent just to see how many people actually read Infinite Jest vs. how many people make the claim to appeal to the NPR/record store clerk set.

Comments

[Kayvaan Ghassemieh]

Great post, Jonathan - I love Last.fm -- I think it's a great "web 2.0" place for music aficionados. Last seems to be a prototypical "wisdom of the mobs" type web 2.0 app. It allows me, as a consumer, to find music I like and the poeple who like that music. It also, like so many other current internet apps, allows me to create and "market" (so to speak) a 'net identity.

Last provides great word-of-mouth marketing for businesses.

Also - if you note in their T&C - Last reserves the right to SELL your listening habits data! Wow - they've really got a potential goldmine there. They've basically disguised a commercial data-gathering service as consumer product. How often someone else listens to Wilco's "Poor Places" is MUCH more valuable (in terms of $$$) to Nonesuch than it is to me -- although it's interesting to me. ;)

Last also uses a very currently typical pricing model of:

FREE: Ads, some functions limited
PREMIUM: 3 Euros -- No ads, premium functions

My big problem with "taste" driven engines is that they tend to give you recommendations of stuff that's all really similar. It still seems limited to "genre" -- even if the genre has been segmented down to like 10 bands. But I think this is a fundmental problem of music that may be insurmountable. How can you really compare one song to another other than a) comparing it to other songs (the "genre" trap) or b) using adjectives that mean nothing (see AllMusic's attempt at categorizing songs or bands by allowing reviewers to categorize them as "moody", "wistful", "angst-ridden", etc.... wtf?).

When it comes down to it, trying to describe music is like trying to describe a color. The only thing you can do is state what primary colors make it up. Those primary colors in music might be analagous to a bands influences, etc.

But of course, these are all just tools to *help* us make decisions -- not make our decisions for us, right? :)