There are a number of music recommendation services out there, and even a decent amount that use your iTunes library as the starting point for recommending new music. But Goombah does things a little differently, due to its distributed architecture. Users install a free application (Mac or PC) that analyzes their iTunes music collections, grinds the data, and links them to the 20 Goombah users who most closely share their tastes.
This morning, the site announced that it will allow users to browse each other’s profiles, seeing which tracks they have in common and viewing their favorite artists and tracks. Users can also now add bios and links to their profiles, which can be browsed to give you another way to find people on Goombah.
That’s the ‘new’ news, but in case you’re not familiar with italready, Goombah also lets you browse similar users’ entire musiccollections (and then go out and acquire the music in whichever way yousee fit), as well as getting music recommendations of varying adventurousness (there is anactual Adventurousness slider you can tweak to cast your net wider ornarrower) based on the over six million tracks the service has already analyzed.
If there’s a particular playlist in your iTunes collectionthat you want to add to, you can ask for specificrecommendations based only on that playlist. Finally, Goombah generates recommendations for free, un-DRMed MP3s it has acquired from labels; double-click one of those and it gets sent straight into a special playlist in iTunes (double-click most songs and you get redirected to Amazon, iTunes, or Napster – of course, you could almost as easily type the album information into the music acquisition tool of your choice if it’s not one of those).
All of this happens within an actual application, rather than on theGoombah website, which is the key to Goombah’s approach, according toDiane Sammer (CEO of Emergent Music, LLC, which develops the application).
The distributed processing architecture lets Goombah crunch song analyses and member/song recommendationdata on the users’ own PC or Mac (indeed, as I write this, my harddrive is humming from all of the analysis going on during the initialinstall), which would apparently be prohibitively expensive to do on the serverside. Sammer told me that Goombah’s distributed approach allows them to offer moredetailed, accurate recommendations than server-based services that analyze similardata can.
Browsing the collections of the members Goombah connected me to,
it’s apparent that I have a lot in common with the 20
people the app connected me with. It’s working well so far – really well, actually – but the question remains… do people have the patience to download an application in order to find music recommendations, no matter how good they are? Only time will tell.