Last night I had the opportunity to sit in on a fascinating conversation between the founder of Pandora.com, Tim Westergren, and sixty or so loyal and curious Columbus users of the internet streaming music site.

Westergren has been traveling the country holding town meetings to take the temperature of Pandora's users. He began the four million-member site in 1999 on a shoestring (for this industry) budget of $1.5 million as an attempt to find a way to use a listener's established tastes to predict new music he might enjoy.

This isn't a new idea, but their approach is. Most companies attempt to accomplish this by using the audience's current taste to deduce similarities; if Ted likes Jack Johnson, and Lisa likes Jack Johnson, then Lisa might enjoy other artists on Ted's play list.

Pandora's approach, called the Musical Genome Project, is inductive. They employ a bevy of musicians who listen to each song and define its DNA according to almost 400 different characteristics. In Tim's words, they believe this provides sufficient 'granularity' to allow them to build a music genome. In Pandora's world, a Black-Eyed Peas song and a Barry Mantilow tune may end up side by side, if the criteria match.

When the listener registers for the free streaming-audio feed from Pandora, he/she selects a seed band or song. He is then fed songs from this core artist and songs that Pandora's research suggests are similar. The site has a catalogue of almost ½ million songs already analyzed, to which they are adding 15,000 songs a month, so a close match to most tunes seems likely. They also have staff assigned the enviable job of mining for new music.

Westergren was very forthright about the business aspect of Pandora. They've burned through $20 million in venture capital so far, along with a new infusion he couldn't reveal. The site is not yet profitable, although he claimed they were on schedule per their plan. They have evolved from their initial hope of profiting from music sales to now planning to make their money via advertising, with the possibility of a subscription service for elevated content.
Viewed in the light of success stories such as Youtube, which recently sold for $1.65 billion, such an investment seems reasonable in the insane e-business world. However, to hit the home run that Westergren envisions, I think Pandora has significant challenges to overcome.

The first is convergence. I believe that the winner in this stakes race is the company that offers me a device that plays my owned music on demand and feeds me new music that is prescreened to my taste, anywhere, any time. It would also be nice if the device operated my garage door. (Joke). Pandora is restricted by the Digital Millenium Copyright Act (DMCA) from allowing listeners to access specific songs on demand. I asked Westergren if Pandora was poised to address this challenge. His response suggested it's not in their near future.

Another problem is the ubiquity of wireless service. For Pandora to grow exponentially, as Westergren hopes, access will have to extend beyond stationary PC's into cars and iPod-like portable devices. I have no doubt such service is coming, but will it arrive in time?

Pandora is not without competitors, either. The battle for your ear is fierce, with contestants including Apple and Microsoft, cellular phone companies, terrestrial radio, XM Radio, cable television, and satellite-based feeds. Direct competitors for internet-based personalized streaming audio feeds also exist.

The most appealing part of Westergren's presentation to me was when he talked about his vision for Pandora as a way to "create a musician's middle class." Pandora pays artists for each play, and happily includes any musician's work in their catalogue, regardless of label or lack thereof. The only qualification is whether their listeners might enjoy it. Westergren hopes that this might result in a revenue stream to those local artists struggling to make a living. I see it also freeing musicians from the tyranny of the label.

In any market, there is a need for someone to connect the provider with the consumer, and in the music world, the conventional marketing mechanisms have crumbled. Pandora has an interesting approach, and if they can make it work, the site stands to benefit us all. I'm going to plug in and see where my Radiohead station takes me today.