Tim Westergren

While terrestrial radio seems to have enough challenges trying to solidify its budding turnaround, a relatively new and different competitor is gaining strength. Pandora, originally called the Music Genome Project, started out as online tool to help expose indie bands to the public. Despite several near-death experiences, it has grown into a legitimate, profitable online radio station for 48 million users who "tune in" an average of 11.6 hours a month.

However, a Pandora "radio station" is quite unlike a traditional one. The website boasts a library of 700,000 songs that have been categorized -- using 400 musical characteristics -- by a staff of musicians. Thus, when a listener chooses a song or artist, Pandora's computer program will "play" a succession of songs that boast similar qualities, be it vocal style, melodic construction or rhythmic attack. Odds are that one or more of those songs will be by artists the user has never heard of, satisfying Pandora's original goal.

After almost folding due to onerous royalty demands, that issue was resolved and thanks to the introduction of an iPhone app, Pandora has grown exponentially. Last year, it was the first online radio station to turn a profit. Now it's fielding a sales staff to serve local advertisers and has pacted with Ford to be part of its voice-activated Sync system. That will enable drivers to simply say, "Launch my U2 station" -- and the users' personalized "station" comes out of the car speakers.

A recent New York Times story reported that Pandora has become so admired by the financial community that it may go public or be acquired by a deep-pockets suitor. In this Power Player interview, Pandora Co-Founder Tim Westergren dismisses those rumors and offers a more detailed portrayal of Pandora's past, present and future.

When you and your colleagues first conceptualized the Music Genome Project, were you thinking about it as a business or merely as a novel way to spread new music?

I had been a musician with indie bands in San Francisco in the early '90s. My colleagues and I started the Project as a tool to help working musicians bring their music to new audiences, which was something indie bands have always needed.

When did you realize that Pandora could become a viable profit center?

I founded the company in 2000, but it took five years before we became Pandora Radio, both in name and in an actual product. When it did launch in November 2005, it took off like a rocket ship. In the beginning, I certainly had a sense of optimism of what it could be, but then we went through some severe trials and tribulations over royalties for several years that almost killed us. Even so, shortly after launch I did think it could be something special.

Has your vision of how the Genome Project and Pandora changed over time ... and if so, how?

I originally didn't think it would turn into a radio; that came a lot later. It was more of a licensing business at first, but that wasn't panning out. It was a tough time for retail and our customers were really struggling. We were chasing around for a new business model when we hired Joe Kennedy to be CEO. He pointed us in the right direction.

Was the original radio incarnation pretty much the same as it is today?

We first came up with the radio design in the Fall of 2004 and we had the prototype internally about six month later. As soon as people saw it, they got pretty excited about it.

The original radio version is very similar to what it is now; we did make some tweaks to make it more user-friendly, but it's substantially the same.

Things seemed very tense during the CRB negotiations with SoundExchange; in 2008 you were quoted as saying the company was on the verge of collapse should you be forced to pay the Copyright Tribunal's royalty rate. Was there a turning point that led to the resolution?

That original rate hike shocked everyone ... including folks on the other side. If you read the ruling and the Copyright Tribunal's rationale, it's apparent they fundamentally misunderstood our business. Something you need to understand about it is that the DMCA Copyright Act essentially provides language for the rate setting guidelines, but not the rate per se. It's the standard mandated by the DMCA called "willing buyer and willing seller" that needs to be changed. The Copyright Tribunal came up with ruling bound by this standard, but to another extent, they also just missed.

How did you marshal the forces necessary to defeat the original rate and convince SoundExchange to accept a more reasonable royalty structure?

We learned on the fly; we'd never done anything like this before. We started with the belief that maybe our listeners would be passionate enough about this to get involved. We already knew their e-mail addresses and zip codes, so we were able to communicate to them in a targeted way. We got advice from various people, banded with other web companies and put together a plan. The big catalyst and the reason it worked well was that Pandora had an enormous audience who was super-passionate about us ... and we tapped into it.

There were two phases to our efforts. The first phase was to bring as much political pressure to bear on the negotiations. We persuaded Congress to intervene and demand that we renegotiate for a more reasonable rate.

The second phase was working overtime with [SoundExchange's] Jon Simson and his constituents ... and in time they began to understand the economics of our business. They got comfortable with what we were arguing. Also, while SoundExchange's focal point is rights holders, they are actually a pretty diverse group of big labels and small labels - and most small labels and their artists are huge champions of web radio and diversity. They need web radio to give them exposure, and their opinions on the other side of table ultimately helped swing momentum towards a deal.

Is the royalty rate as it is now something that Pandora can grow with over the long term, or are you concerned that SoundExchange may want to jack up rates even more at some point in the future?

It will be renegotiated in 2015, but personally I think we'll be fine, so it's not a big concern.

Once the agreement was reached, Pandora started offering a two-tiered system -- a free subscription with ads with a 40-hour cap on listening ... and a pay subscription (at 99 cents a month) without ads or any listening cap. How has the customer response been to each?

It's actually been really good. When we did the 40-hour cap, which was the most significant move we made, we got a large number of people complaining bitterly about us baiting and switching. They were upset about it, but at the same time, we were very transparent. We really explained why we had to do it. A lot of people came to understand that Pandora actually pays royalties and had pretty substantial expenses, so the net result was that a very small number of people left and didn't come back. The vast majority would either pay the 99-cent monthly subscription or wait till the end of the month. We lost a very small number of listeners, while the growth of our subscribers increased dramatically.

To be sure, even now a very small percentage of our listeners are paid subscribers. That's why out of our 180 employees, 90 do nothing but sell ads. We work as hard as we can to make it work.

What kind of tweaks do you see making to each to optimize use and revenue?

We're not totally sold out yet, but we're getting there. The big challenge for us is figuring out how to sell effective audio ads. We're putting a lot of effort into that.

Have you settled on an optimum number of ads you can run between songs and on the site, or does that number change during certain times of the day or week?

We can target ads based on age, gender, zip code, style of music and time of day. We can do all those things -- and we put a lot of the same thinking into this as traditional radio companies, which is figuring out how to most effectively deliver ads. We're just able to do much greater targeting of certain demographics.

How has the advertising community's reaction been to your effort to sell local ads?

They think we've done a great job on the website; they're just beginning to exhibit some movement into audio ads. Our job is to keep working to help them understand how to translate this new medium in the context of what has historically been terrestrial radio ads. We spend a lot of time evangelizing.

How has the success of Pandora's iPhone app changed your perspective in terms of Pandora's potential growth?

We thought it would be a good thing for us, but we had no idea the magnitude of its impact. It has doubled our growth rate -- and not only that, it has also redefined our entire category. It's taken Net radio from "computer radio" to something you use everywhere a traditional radio can be. People are taking their iPhones, plugging them into their car dashboards ... taking them to the gym ... or docking them at home, streaming it to their stereo. It has thoroughly transformed Pandora.

In terms of numbers, what kind of impact has the iPhone had on your customer base?

We have 14 million iPhone users; overall iPhones make up 70% of our smartphone audience. Mobile listening now comprises close to 40% of total listening; the majority is still on the PC, but that percentage is shrinking.

Has this new accessibility changed the way you let your customers "program" their stations?

I wouldn't say that. We don't do anything differently on our end when it comes to the music, but it does definitely impact the mode of advertising. You can't take as much advantage of visual stuff on a device as small as an iPhone. It has made the development and use of audio ads more important than ever.

Has the profile of your typical customer changed as Pandora changed over the years? If so, how - and what are you doing to accommodate the changing clientele?

I would say Pandora has gotten to be more of a mass-appeal medium. Our audience is now more evenly divided between men and women. In the beginning, it was predominantly men. Our age demos have spread out -- both younger and older. We're looking more like the general population instead of early adopters.

Our engagement rate has stayed the same. Our customers come back to us six to seven times an hour. There's plenty of activity out there.

You've recently hired George Lynch to work the auto side and you just did a deal that will get Pandora on Ford's Sync platform. What kind of presence would you want Pandora to have in autos?

It's going to be huge for us. Cars are where half of all radio listening happens, so we have to be in the car. Fortunately, with smartphones, it's much easier. We don't have hardware that needs to be installed because it's controlled through the iPhone, which is patched into the car stereo.

We did do a deal with Pioneer and Alpine to embed Pandora in their products for aftermarket stores; with these implementations you can control Pandora's app through your stereo.

Pandora has spoken out for the artists and labels on the performance royalty issue. Is your stand based on a "fairness" angle, in that since Pandora pays, terrestrial should pay?

You can't defend a playing field that's not level. It doesn't make sense to have an outdated law. Whatever happens, there needs to be some degree of parity. What's the answer in terms of a number or a royalty rate, that I won't speculate, Pandora can't truly compete with other forms of radio if we have a royalty and they don't. The royalty inequity is like having both hands tied behind our back

How does Pandora fit in the playing field with terrestrial and satellite radio ... do you see yourself as a direct competitor or a complementary media choice?

We're all certainly competing for the listening hours. We're not currently getting ratings measurement, which will be a huge issue, a legacy issue, with us. We're certainly looking for some improvements in how Arbitron and others do their ratings. There needs to be a more accurate means of measurement across formats, so the buyer can compare, head to head, Pandora vs. satellite vs. broadcast. There needs to be some sort of translatable standard.

Have you talked to Arbitron about it?

Yeah, and their response has varied. They ultimately want to do the right thing here, which is produce accurate data through an effective measurement system that serves the advertisers. I don't think they've found that yet.

What's your take on the PPM?

It's a flawed system; it doesn't capture web radio. I don't think it's accurate when it may not pick up when you're listening by yourself, but then pick up broadcasts when you're walking down the street and not paying any attention to them.

You once said, "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 artists ... 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." Do you still believe that?

Our goal is really to put together a great radio for each listener and as a consequence of that, expose them to a very deep and broad catalog of music they may not be familiar with, but whose form, structure and style is something they already appreciate. The biggest difference between Pandora and traditional radio is that while we both want to reach the largest audience, we want to do that by providing them the widest selection of music that suits their tastes. We don't want to reach them through a small playlist. We believe that the more our audience becomes engaged with Pandora, the more it can become a launching pad for literally thousands of bands.

So it's not a high priority that Pandora becomes a place where bands can get the kind of exposure that can lead to Gold or Platinum sales status, like radio has done for years.

We're not consciously trying to make as many "big hit artists" as possible. If it happens, great, but we're more interested in creating a rising tide where far more bands gain exposure. We're not trying to favor or focus on specific records or bands. If an artist becomes successful through exposure on Pandora, we want that to happen organically.

I assume you have data on what records get played and how often they're played. I would think there'd be a lot of interested parties - the labels, to name several -- who'd love to know how popular their records are.

We have shown spin information to some labels in the past, but not as a matter of course. I'm not sure we want to think about Pandora in that way.

What's wrong with thinking that way? I would think that not just major label bands, but indie bands - especially indie bands -- would give their eye teeth to publicize the fact that their song was the #1 most-played on Pandora. The potential of using Pandora success as a major stepping stone to considerable mainstream success ... what could be wrong with that?

I don't ever like to say any record is #1 on Pandora. It's really unimportant. What's important is the customer ... that each individual listener enjoys the Pandora experience.

Obviously they are, because Pandora is the first Net radio entity that has actually reported a profit. The perception on Pandora is so bullish that now we're seeing stories about Pandora going public and inevitably selling itself to the highest bidder. Any comment on those "going public" rumors?

Those stories aren't coming from us. People may think about us going public and/or selling out, but we aren't. Pandora is very focused on execution; we keep our heads down and our focus sharp. I assure you that any dialog you hear about us isn't coming from Pandora.

So where do you see Pandora in five years?

We have our hands full, getting the most out of the U.S., but we do have designs on going global. We want to get that ball rolling, but the licensing situation outside the U.S. is badly broken. It needs to be fixed before we can move forward; we'd like to play a constructive role in facilitating that.

Did you ever think that one day you'd become some semblance of a global politician?

That's not a totally inaccurate description. There's no reason we can't go there and replicate a centralized licensing system, but it would involve different legislative bodies, so it surely won't be easy.

Wouldn't creating a "YouTube Pandora" for videos be easier and offer more profit potential?

I do think someone might take the "genomic" approach to solving the video navigation and discovery problem, but it won't be us. Perhaps for music videos, but otherwise, we have our hands full!