Internet radio, Fuzz style: Where humans upstage the algorithms
Fuzz is tackling one of the hardest startup categories around -- yet it's not necessarily doomed. And that's high praise.
Any entrepreneur taking a crack at a digital music startup must either be super determined, completely crazy, or a both. The chances of legal run-ins with the labels are high. And even when you play by the rules, the rights payments are so steep that making a profitable business is all but impossible.
Yet this isn't deterring Jeff Yasuda, a 40-year-old venture capitalist turned entrepreneur who's been doggedly running music startups since 2006. His latest, Fuzz (the same name as his first startup), comes out of stealth mode today after a year of building a team -- one that includes former bigwigs from MTV, Clear Channel and Rhapsody -- and a product from the basement of his house in San Francisco.
Fuzz bills itself as great radio made by real people. What that means is that it's a site where anyone who participates becomes a DJ. You upload music you own (legally, of course), and it all lives on Fuzz's servers; Fuzz, in a sense, becomes a digital locker that let's you showcase your music. You create playlists, people follow you, and they give you "props," all of which ups your social standing and surfaces your playlists. The listeners decide what's most popular and which DJs are best in any number of genres -- from "indie madness" to 90s Hip Hop" to "ear pounding bass" and "Rare R&B." Unpopular DJs essentially fade into the background.
"Our thing is that everybody can be a DJ," says Yasuda. "But just because we let you be a DJ doesn't mean you're good."
In short, while other Internet radio services -- think Pandora or the lesser known Slacker -- are designed to let software help you discover music, Fuzz is about finding the people who share your musical tastes. At the core of the business is Yasuda's belief that algorithms just don't cut it when it comes to creating great radio.
"Give people a voice, give people a platform, and they'll do tremendous work for you," he says. "It's a lot more egalitarian, and it's a lot more interesting. The best recommendations -- for anything -- often don't come from people who aren't professionals, whether it's for music or wine."
Yasuda and his team have been building the product with the advantage of 4,000 early testers, and some of the early stats are impressive: They use the site for 30 minutes a time, although hard-core users often stick around for two hour sessions. Half his users are returning to the site nine times a month, and a quarter are coming back more than 50 times. Users have created 1,700 stations. "These stats blew my mind," says Yasuda, "but it's obviously a small number of users."
In other words, the hardest test awaits.
Yasuda launched his first music startup in 2006, after he was burnt out as CFO and managing director with Redwood Ventures. He was determined to work on something he was passionate about, and since he'd been playing guitar in bands since he was 13, music fit the bill. He launched Fuzz Artists, which was a set of online tools for bands. The service attracted a strong user base -- 25,000 bands at its peak -- but ultimately Yasuda couldn't make it fly as a business. He closed down the office, let workers go and regrouped.
Then came Blip.fm, which was often described as "Twitter for music." Instead of tweeting about what you were doing, a Blib.fm user would 'blip' a 150-character message that included a link to a band or song you're listening to. The blip included a widget that you would click on to hear the song. Blip.fm quickly amassed more than one million registered users, and a core group of loyal fans. "People frickin' loved it," says Yasuda.
One of the big labels didn't, however. Yasuda won't -- meaning he legally can't -- go into details, but in a blog post to users, he summed it up by saying, "The challenges involved in running a start-up in the music space are immense. To be honest, it's completely nuts..."
He altered the product to appease the labels, traffic slipped -- from more than two million monthly uniques to around half a million -- and today Blip.fm continues as what he characterizes as a mom and pop business. It makes money from ads, but only a little more than enough to cover costs. To get the new Fuzz started, Yasuda raised $500,000, from a new angel fund called Core Ventures, hedge fund Kilowatt Capital and a few individuals.
"The DJ culture is strong and lot of people want to share," says investor Mika Salmi, who is the former president of digital for MTV Viacom and now serves as executive chairman of Fuzz. "The human touch is very different."
As Fuzz goes for a bigger audience, Yasuda knows he's in for a hard climb.This isn't a service that can go viral like a game. It takes time for people to upload their music and to build up followings. If it works, he expects the behavior on Fuzz resemble YouTube: A tiny fraction of people create content -- in this case, become active DJs; a slightly larger number interact (comment and give "props"). And the vast majority will be passive, using Fuzz simply as Internet radio with a variety of distinct stations that you can't find anywhere else.
Other startups are trying to ride the wave of the DJ culture as well. Songza, for instance, offers playlists for certain moods - "an energy boost," "working (without lyrics)" -- that are created by "experts." Another startup, called DJZ, is set to launch this fall as a hub for electronic music; it's run by Seth Goldstein, the co-founder of the much-hyped Turntable.fm, now a shadow of its former self. Even Spotify now lets users with their Facebook friends.
So what about the business, and potential legal minefields? Yasuda's co-founder and longtime business partner, Brian Venneman, is an attorney, and the company has a ton of time making sure it's doing things by the book. And some hard-core skeptics give Yasuda better odds than most. Dalton Caldwell, who after to MySpace publicly swore off music startups, said Yasuda has a "big leg on most music startups because... he understands all of the tech and licensing after his years of experience working on this."
Yet the Internet radio model is fraught with challenges. Fuzz, like Pandora or others, has to pay SoundExchange and publishing rights organizations such as ASCAP. Early on, Yasuda pays a flat fee (he won't disclose) since he doesn't yet have any revenue or many users. That makes it manageable. The challenge comes when you get big, and are required to pay a royalty for every song that's played. It's a bind that newly public Pandora knows all too well: The more people listen, the more you pay, and that usually means the more you lose. Your job is to make that challenge feasible by boosting revenue, and no company has yet made it work. Pandora, with losses growing, is now trying to get legislation passed that would lower the royalties it has to pay out.
And don't even talk about ads. Yasuda knows that's a losing path, that the numbers don't work. In some ways, Yasuda's bigger plan to move away from this model entirely. In the near term, he intends to add premium subscription services, but then he wants to move aggressively into mobile and the app business.
The app business? "Those are the breakout opportunities," he says, pointing to the success of SongPop, a name-that-song game that thanks to Facebook now has more than 18 million monthly players. Eventually, says Yasuda, he wants to take what he learns on Fuzz, create apps -- maybe games, maybe something different -- that he markets to his users.
"We're not just selling music," says Yasuda, as if to emphasize that he isn't, as he wrote, "completely nuts" to be taking another crack at another music startup.
And if nothing else, he's certainly determined. "I failed with the first Fuzz. I kind of failed Blip.fm, and I'm now taking another swing at it," he says. "But I'm still alive. I've always said you just need a long enough runway for you to screw up a bunch of times so that you can take that one more chance to try something new. And maybe that's the one that makes it."