The Scoop on Pandora for the iPhone and Other Platforms: Tim Westergren Speaks at Boston’s Apple Store

11/18/08Follow @wroush

At least two of us here at Xconomy—Rebecca and myself—are huge fans of Pandora, the Oakland, CA-based streaming music company. My enthusiasm has only grown since July, when Pandora released an iPhone application that, I think many users would agree, is the single most useful and enjoyable third-party app available for the device. (It’s currently #30 on the “Top Free Apps” list at the iTunes App Store, but I think that’s only because everybody has already downloaded it.)

So when I heard that Tim Westergren, Pandora’s founder and chief strategy officer, would be speaking at the Apple Store on Boylston Street in Boston on Monday, I made like any other self-respecting Pandora fanboy and cleared my schedule. The table-full of listeners who joined me to hear Westergren’s talk starting at noon Monday gradually grew to a crowd of perhaps 50 onlookers.

This being the Apple Store, Westergren spent part of his time explaining how the Pandora iPhone app works. But he also gave an interesting look behind the scenes at Pandora, explaining how the company’s musicologists analyze songs, how the company itself survived nearly three years on the edge of extinction, how it monetizes its free streaming service today, and how it dealt with the threat of changes in performance-royalty rates that threatened once again this year to put the company out of business.

Tim Westergren, Founder and Chief Strategy Officer, PandoraIn a nutshell, Pandora lets users create personalized, streaming radio “stations” that play tunes by artists they like, along with music from other artists with similar styles. Its system decides what to play based on the “proximity” of one song to every other song in the company’s database, where songs are scored based on hundreds of attributes such as genre, key, tempo, the gender of the lead vocalist, and the timbre of the various instruments. Users can give a thumbs-up or a thumbs-down to every song that’s played; Pandora’s algorithms use this information to adjust the playlist for each station over time.

The iPhone version of Pandora—which wasn’t possible until Apple allowed third-party apps onto the device last summer, and wouldn’t have been practical, in any case, until the broadband iPhone 3G came along—does essentially everything that the Web-based version does. It’s by no means the only mobile version of Pandora—the company has actually built about 60 different versions of its software for various phones sold by Sprint and AT&T, and is also working on a version of the app that will run on RIM Blackberry devices—but the iPhone version is the first one that has really demonstrated the potential of the Pandora service on an untethered device, Westergren says.

“The iPhone has really changed the trajectory of the company,” he says. Since July 11, the day Apple launched the iPhone 3G and the iTunes App Store, 1.8 million iPhone and iPod Touch users have downloaded the Pandora application, and people listening on these devices are now consuming 10 percent of all music streamed by the company, he says. And whereas the company’s peak listening times have traditionally been 9:00 a.m. to 5:00 p.m.—when people are at their desks at work—that period has stretched out by an hour on either side of the work day, as iPhone owners tune into Pandora during their commutes.

(As for Android, the Google-led, open-source phone operating system, Westergren says it’s still “up in the air” whether Pandora will build an app that works on that platform. “We want to see how much it gets adopted first,” he says.)

Westergren shared some hopes for the mobile versions of Pandora that will make owners of commercial broadcast radio stations either shiver or snicker. “Our objective is to completely replace music radio with this,” he says. And that includes listening to music in your car. With a cheap adapter, Westergren notes, it’s easy to stream Pandora music from an iPhone to a car stereo. “I encourage you to try that,” he says. “It’s weird—I hadn’t actually used it in the car until a few months ago, and one day I was driving along and I said, ‘Hey, weird, I really like this music.’ Then I realized ‘Oh, it’s my Pandora station.’ Then a song came on that I didn’t like, and I just skipped it. It’s amazing how emotionally different it is. In a car, you’re used to being captive to broadcast stations, and all of a sudden you have this personal listening experience.”

The feeling that Pandora knows you and your musical tastes is only possible thanks to the years of work the company has put into its catalog of music ratings, called the Music Genome Project. Since Pandora’s launch in 2000, the company’s staff of 50 trained musicologists has classified 600,000 songs, and continues to add about 10,000 songs per month. It’s an incredibly labor-intensive process, Westergren says. “For a simple, three-minute pop song, it might take 15 minutes” for a Pandora musicologist to rate a song along all 400 parameters. “A symphony might take two hours.”

Having recently visited and profiled The Echo Nest, a Somerville software company with music-recommendation software that automatically—and very quickly—assesses a song’s acoustical qualities, I asked Westergren how Pandora can keep up with all of the music coming out these days, given its slow, meticulous, human-dependent classification method.

“It’s an absurd approach,” Westergren agreed jokingly. “But it’s the only way to do it, we think. It also has one really distinct and unique advantage, which is that it’s blind to popularity. If a garage band’s music goes into Pandora, it gets an equal shot to Bruce Springsteen. Every other system is really about collaborative filtering–that people who like this also like this. You don’t get recognized on Amazon if you’re not already somewhat popular.”

Actually, The Echo Nest’s software is an exception to the collaborative filtering phenomenon—but I didn’t get to pursue that with Westergren. He followed up, however, with another important point: while automated rating software may churn through more music, not all of that music is worth a human’s listener’s time. “Our goal is not to be all-inclusive,” Westergren says. “We want to try and find just the best music. I can say this as a musician: most music that’s published is not very good. Maybe 5 or 10 percent is actually ready for prime time. That’s the stuff we want to find, and I think we have an operation big enough to do this.”

Pandora makes money on the whole operation by selling ads alongside its Web-based console; the ads are customized according to users’ gender, age, zip code, and musical taste. The company also provides links from its songs to Amazon and iTunes, and earns a small commission on all song purchases. But while they’re small, the commissions can add up—especially given what Westergren calls the “crazy” conversion rate on the site. “Forty percent of listeners end up buying music,” says Westergren—making Pandora the number-one referring affiliate of both iTunes and Amazon’s MP3 store.

The company’s recent climb toward profitability comes after a long drought. From the company’s founding in January 2000 to the launch of the streaming radio service in November 2005, the startup had almost no revenues, Westergren recounts. With the Music Genome Project, “we were building what we thought would be a music recommendation technology that we would license to other companies,” he lsays. But there weren’t any takers, and by the end of the end of 2001, the company was broke.

For the next three years, most of the staff worked without salaries. Then, in 2004, after 347 consecutive unsuccessful pitches to venture capital companies and other investors (which has got to be some kind of record), the company got lucky, winning $7.8 million in funding from Labrador Ventures, Selby Venture Partners, WaldenVC, and angel investors. (That was followed by a $12 million Series C round that brought in Crosslink Capital and a Series D that involved Hearst Corporation, JP Morgan and Piper Jaffray; the company has collected more than $22 million all told.) But 2004 was when “we kind of got resurrected,” Westergren says. By that time, “broadband had grown so much that streaming audio was becoming viable mass market opportunity, and that’s when we turned into a radio service. So the last couple of years have really been a Cinderella story for me, because we were for all intents and purposes out of business for a long time, and we shouldn’t have survived, but now we’re all having a whale of a time.”

But the good times almost came crashing to a halt earlier this year. Westergren shared his version of the recent battle in Washington, D.C., over changes in music-royalty rates that threatened to put an end to Pandora and most other Internet radio stations. “It’s a bit of a long saga, but Internet radio stations are required to pay both a publishing and a performance fee for every song we stream,” Westergren explains. “The publishing fee is about 3.5 percent across all forms of radio—broadcast, satellite, and Internet. But the performance fee is actually determined by an arbitration panel at the copyright office in Washington. And a year and a half a go, this committee met and ruled to almost triple our performance rates. Overnight, it basically blew up our business model.

“So we were getting ready to shut the whole service down, as it would have been too expensive to run. We decided as a Hail Mary to e-mail everyone and ask them to contact their Congressperson and urge them to stand up and prevent this from happening. Capitol Hill received 400,000 faxes and e-mails in three days and eventually over two million.

“It was so overwhelming that Congress intervened and forced a renegotiation, which, as we speak, is continuing. That’s going to result in a much-reduced performance fee that will allow us to continue…It was a pretty scary time for us, but thanks to this grassroots mobilization, we survived that. So, thanks for calling! It made a huge difference.”

One person in the audience at the Apple Store asked Westergren to name his favorite Pandora station. “Ben Folds,” he said unhesitatingly. “Heavy-hitting piano stuff with interesting vocals and harmony. That’s sort of my sweet spot.”

So I listened to some Folds yesterday while I was writing this piece. I can’t say his whimsical style really struck my fancy. But that’s okay: everyone has a different musical genome, and Pandora has 600,000 other songs and counting for me to choose from.

Wade Roush is a contributing editor at Xconomy. Follow @wroush

By posting a comment, you agree to our terms and conditions.

  • http://www.HearWhere.com Pete

    Great article Wade,

    One note, though the Echo Nest is currently the darling of the ‘acoustic matching’ world, they are not the only ones who are looking more at the acoustic attributes rather than collaborative filtering.

    MusicIP has been doing this for years, and their music recognition algorithm (fingerprint) is used by MusicBrainz.

    Check them out http://www.MusicIP.com, and if you’re a fan of live music, be sure to check out my site http://www.HearWhere.com