The Future of Radio

20 04 2008

Tim Westergren

I had the pleasure of seeing a lecture at Columbia University by Tim Westergren, the founder of and current Chief Strategy Officer. started in January of 2000 as The Music Genome Project, in an effort to characterize all music by attributing hundreds of musical ‘genes’ to songs. These include things like type of vocalist, guitar sound, rhythm structures, etc. A typical rock song has around 150 ‘genes’, while jazz songs mainly have upwards of 350 tags associated with them. was created to use this system and an additional algorithm to provide its users with a personal radio station.

Basically, a radio station is created when a user searches for a certain song or artist, and Pandora cues up a playlist that is made up of songs that match the musical ‘genes’ of the original query. Then, every time a song comes up, a user has options to give it a thumbs up or thumbs down to tell the algorithm whether he/she wants to listen to the song or songs like it. This type of recommendation system has done the company well, with over 7 million users and over $22 million in venture capital and debt financing. Also according to, Pandora has increased the total number of people visiting their website by 126.2%. The lecture was primarily about the business tactics of the company, and how they struggled and fought to finally make it profitable(for two years most employees didn’t even work for a salary). It solidified my ideas that traditional radio and satellite are both dead. My reasoning after the jump!

1. Terrestrial Radio Sucks. I stopped listening to the radio years ago, because simply there was nothing even close to what I liked. I’ve been finding music ever since through MySpace, word-of-mouth, and surfing the web. Everyone I talk to seems to be tired of traditional radio. As Tim said, it is estimated that only about 40,000 different songs have ever been played on conventional radio.

2. Satellite Radio Is Dead. I asked Tim why they never thought of using satellite technology to distribute their radio because of its wide coverage band. He replied that it’s simply a bandwidth issue: satellites can only handle up to 120-150 simultaneous data streams, while Pandora serves up about 90,000 individual ones every day. It’s further proof that satellite radio has no expandability.

3. The Future Is Interactive and Personalized. Saul Williams says it best when he describes a “Nike air force fleet, custom made, unique”(check it). With manufacturing so cheap, everything can be customized, and when it’s software, it’s even easier. Radio is one of the only things besides perhaps soda bottles and TV sets that isn’t customized these days.

So what’s the future for Pandora? From answering some of my questions about adding type functionalities and music videos, it seems there isn’t too much changing at Pandora for the time being. I asked if they were considering changing the Pandora recommendation scheme to include possible scrobbling functionality in order to create a better and more accurate musical profile, and Tim replied they were very comfortable with the ‘thumbs-up and thumbs-down’ approach. While disappointing for me, I think Pandora is on its way to becoming a very powerful player both on the web and in people’s ears.



Leave a comment

You can use these tags : <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>