Chief Strategy Officer and Founder
Companies like Amazon have tried looking at buying preferences—people who bought CDs by the Shins, say, often bought CDs by Death Cab for Cutie as well.
But that system only rewards bands that are already known. And if there's one thing that Tim Westergren learned from 10 years on the road playing keyboards in rock bands, it's that there are a lot of talented groups out there that you've probably never heard of.
"I spent a lot of years pursuing a musical career, so I came face to face with the challenge that all musicians face—to get noticed among the tens of thousands of bands trying to do the same thing," says Westergren, 40.
"I became very interested in how to solve the problem, from a musician's standpoint, of how to find an audience"—and, from a fan's standpoint, of how to find music you like.
Westergren also had experience scoring films in Hollywood, where he had to interview directors about their musical tastes and divine from that what other music they'd probably like in their films.
With online advances making music more readily available to consumers, and with the Human Genome Project, a monumental effort to decipher mankind's genetic code, hitting critical mass toward the end of the millennium, Westergren's wheels began spinning.
He and a few colleagues began brainstorming about the Music Genome Project, a mammoth undertaking to analyze hundreds of thousands of songs and catalogue them according to their constituent parts—different elements of vocals, instrumentation, melodies, harmonies and other characteristics—creating a database that could draw links between songs on the basis of these 400 or so musical "genes."
"We just went through the process of dissecting music. It's something that any trained musician could do if they had to," Westergren says. "There's an accepted set of descriptive musical criteria that go into giving a song its sound. We just codified that and turned it into a working genome. We borrowed lots of years of theory and existing research on our own."
There's also, not surprisingly, a lot of old fashioned hard work involved. Each song takes about 15 minutes to analyze—and the Genome Project so far contains about 500,000 songs.
One thing is clear. Wikipedia this is not. Westergren's company, Savage Beast, employs about 40 analysts, all professional musicians who have undergone 100 hours of training in song analysis. And the Genome's potential growth is exquisitely infinite: More music is produced and recorded each year than Westergren's team is capable of analyzing.
The challenge has been exploiting the proprietary database's commercial potential. Deals with AOL, Barnes & Noble, Tower, Best Buy and Borders have helped anchor Savage Beast financially. But Westergren's baby is Pandora, a free, online radio program that creates personalized playlists of songs with similar "genes" to music the listener already likes.
So far Pandora has over 2.5 million users. Feedback from listeners, and commercial indicators such as increased iTunes sales or concert ticket sales for musicians, indicate that the service is fulfilling Westergren's goal of bringing bands and listeners together.
Harnessing the Genome could help provide navigational solutions to what Westergren calls a "tyranny of choice": There now are so many different ways to access music online, but precious few tools to direct a consumer through the chaos to music of his or her liking.
But Westergren has even bigger goals: Pandora's ability to harness the Music Genome Project, he believes, could fundamentally change the industry by helping legions of unknown bands build a following.
"This has the potential to turn the music industry upside down: We can move it from being a $10 billion industry of a few hundred artists to a $100 billion industry from thousands of people," Westergren says. "If we keep going as we are, I think the music industry is going to look very different down the road, and I think we'll be able to say that we were a significant part of it."