Let’s start at the beginning

When I moved this blog over from blogger the rationale was to gear up to tell the Pandora product story — at least my perspective on it. With our new product launch out the door, and things generally starting to quiet down here at Pandora central, I suppose I should start to make good on my threat to tell the tale. Here goes. Screaming into the void (hi mom!) and all that…

The Beginning

This story begins for me back in May of 2004.

I’d been the VP of Engineering since late 2000 at a company called Kenamea that was building a kind of AJAX framework designed to help people build rich internet applications. This was of course before these kinds of solutions had a name (Jesse James Garrett didn’t use the AJAX term publicly until Feb of 2005, nearly a year after I left Kenamea). Kenamea had been a tough road; trying to get enterprise customers to see the opportunities created by rich internet applications.

I was very ready to go to work on a consumer application and work on solving some problem that I had in my own life. As I started to think about what it was I wanted to invest myself in, I quickly came to the conclusion that it had to be something to do with music. Like many, I’d spent an unspeakable number of hours ripping and tagging my CD collection in iTunes and was left disappointed with how little that iTunes (at least in 2004) did to leverage that wealth of musical information. It seemed to me that there was a bunch of interesting opportunities around helping people connect with artists that they’d like by tapping into the iTunes-collected metadata.

So, I set to work cooking some ideas of my own about a consumer music service built around recommendation and tight integration into the iTunes “view” of the your musical taste. In the middle of this I took a trip down to Indio, CA for the ’04 Coachella music festival (The Cure, The Pixies, and Radiohead!). In a very lucky stroke I bumped into a high school friend and I started telling him about some of my music recommendation ideas. He said “One of my good friends just went to work for a little company in Oakland that’s working on some of those same problems. You should call her.”

I did.

Pandora started out as a company called Savage Beast. The company got its start in January of 2000 founded by Tim Westergren, Will Glaser, and John Kraft. Tim was the music guy, Will was the technology guy, and John was the business guy. By the time I arrived on the scene in 2004, John was long gone and Tim had taken on the mantle of both “business guy” and “music guy.”

When I first met the Savage Beast team, there were 5 engineers, the two remaining founders, a just-hired VP of Business Development and a handful of music analysts. The were spread across two floors in a pretty miserable building in Oakland. Crammed into the tiniest spaces you can imagine.

The company had been busy building the Music Genome Project for the last 4 years and was focused on selling access to the Genome to pretty much any large consumer company that played in the music space. Tower, Best Buy, Borders, and AOL were all customers. The company had just raised its first real round of venture funding and the technical founder and CTO Will was looking for a VP of Engineering to build out and manage the engineering team as the company took on new projects post fund raising. In all honesty they were calling the position VP of Engineering, but what they really needed was an engineering manager. So goes it in the early stage startup world.

I came in for a round of interviews — it was in fact the very first company I’d talked to after deciding to leave Kenamea — and immediately fell in love with the team. A great group with a fascinating approach to the music recommendation problem. None-the-less, I was reluctant. These were the first people I’d talked to, and the actual position they were hiring for wasn’t a perfect fit for me. I’d played a strategic role at Kenamea and I wasn’t entirely sure I was ready to take a job that was 99% engineering tactics (hire the team, keep them happy, hit the release dates as promised, repeat). I was also looking for something that was a direct-to-consumer play, and they were pretty resolutely in the B2B space.

They offered me the job, and I decided to keep looking.

About six weeks went by. I went to Alaska and kayaked for 10 days with my Dad. I met with a bunch of cool early stage companies. I thought more about starting my own music recommendation company. But for some reason my thoughts kept going back to the little rag-tag team in Oakland.

Then my phone rang; it was Will again. While they’d been talking to other candidates they kept coming back to the idea that I was the right guy for the job. They made me another offer and this time I accepted. I was the new VP of Engineering at Savage Beast.