It was when the San Francisco police burst into his apartment, guns drawn, that Justin Kan first found cause to question his business model.
The incident—later immortalized on YouTube—occurred one night in March 2007. The cops found Kan working at his laptop, hip-hop blasting through a living room littered with clothes, paper coffee cups, and other boyish debris. Startled, the then 23-year-old Yale graduate threw his hands up. He was wearing a small cylindrical video camera strapped to his head. "Did somebody get stabbed in the chest in here?" an officer shouted. "No, no, no!" Kan replied. "Shit." As the police stalked around with flashlights, radioing with a dispatcher about the distress call that had led them to the apartment, Kan silently identified the likely perpetrator of this prank—his audience. Nervous, he blurted out to the officer: "Uh, tech company." Earlier that week, Kan and his three roommates had launched Justin.tv, a business devoted to broadcasting, on live-streaming video, every single minute of his life. The visit from the police was not a planned part of the show. Some of Kan's viewers, bored with his quiet evening, were making mischief. Kan and his partners had started off with the idea—possibly dystopian, but certainly naive—that they could create a new form of entertainment: reality television that captured the totality of a person's existence. But the late-night prank provided a sobering lesson. On the Internet, the audience can control the plot. The next night, six fire trucks showed up at the apartment building.
"We freaked out," Kan recalls. But the founders of Justin.tv stuck it out, at least for a while, because they truly believed that "lifecasting," as they called it, would become a popular pastime, once Kan had proved the concept with their custom-designed portable camera. Eventually, though, it became clear that the pranks were a symptom of a deeper flaw: People were not that interested in the unabridged life of a postcollegiate entrepreneur. To survive as a business, Justin.tv would have to change—but into what? In the mythology of Silicon Valley, there's no shame attached to starting in the wrong direction, because every failure is just another opportunity for inspiration. The fashionable jargon for this is the term pivot. "This is very common in the history of startups: You get 9 out of 10 things right," says tech investor Paul Graham, who provided the seed funding for Justin.tv. "Then you get the last one, and bang!" A floundering website called The Point, founded to promote social activism, pivots to marketing discounts, and bang: It turns into Groupon. The designers of a check-in app called Burbn decide that what their users really like to do is share pictures, pivot to become Instagram, and bang: 18 months later, they sell to Mark Zuckerberg for a billion dollars.
Justin Kan's experience, though, presents a different kind of founding myth, one in which each turn through the labyrinth leads back to the same monster. As Justin.tv pivoted and pivoted again, it would continue to struggle with the Internet's untamed id. There would be other brushes with the law, with Kan vilified as a "high-tech pirate," his company hauled in front of a Congressional committee and confronted by the country's most brutal professional-fighting organization. It was only out of desperation that Justin.tv finally hit on its potential path to salvation—in serving a most unexpected market.
At 28, Kan carries himself with the self-assurance of someone old enough to be considered a veteran in the precocious world of San Francisco startups. He likes to code while listening to dubstep and wearing shades, habits that have led some to cite him as a prime example of a species known as the "brogrammer." One archival Justin.tv clip, entitled "Justin at 118mph," shows him traveling at ridiculous speed down a highway at night. Kan is cocky—on Twitter, he often prefaces news with the words boom! or kapow!—but people don't seem to hold it against him, because he's good company.
Though Kan no longer wears a camera, he still shares a lot on the Internet: business tips, pictures of himself in a motorcycle helmet, hot-tubbing status updates. "I've always been someone who's, like, more of an exhibitionist," he says. Kan has an athletic physique, which he's worked hard to perfect through weight lifting, and he's not shy about showing it off. During his senior year at Yale, he posed for a charity calendar he created, naked save a dollop of shaving cream covering his naughty bits. Justin.tv grew out of a half-serious conversation between Kan and Emmett Shear, a childhood friend from Seattle and Yale classmate. After graduation, they both went through Y Combinator, Paul Graham's tech-entrepreneur training camp. To hear them talk about it now, the experience sounds a bit like Treasure Island, with Graham playing the role of Long John Silver. Companies that graduated around this time include famous successes like Scribd and Reddit, but Kan and Shear's initial venture, an online calendar, died when Google introduced a similar product. Graham urged his protégés to come up with something else: the more buccaneering, the better. So they told him Kan wanted to walk around wearing a camera on his head.
Graham loved the zany ambition of Justin.tv. "He thought it was hilarious," Shear says. "We were going to enable this new form of reality TV based on streaming people's lives 24/7, and that was going to be the business. We were going to be reality-TV moguls." Experienced producers know it takes lots of editing to make real life entertaining. These were not experienced producers. Shear had never watched much reality television. Graham doesn't even have cable. Over dinner with his father and another Yale friend, Michael Seibel, Kan explained the concept. "We both thought it was a horrible idea," Seibel says. But Kan and Shear were about to pile all their stuff into a Honda Civic and drive to San Francisco, and since Seibel was between jobs he joined the cross-country road trip. They crossed the Bay Bridge just in time for Fleet Week. "It was nice, it was sunny, there were fighter jets flying around," Seibel says. He was ready to join the show. "I guess I had a little bit of faith in it," he explains, "because we were funny, popular guys at school." In retrospect, Justin.tv turned out to be a record of a lost historical moment, when many of today's tech tycoons were still living off pizza and beer, striving by day and partying by night. The three guys rented a two-bedroom apartment in a high-rise in North Beach, which was populated by so many Y Combinator alumni that it was nicknamed the "Y Scraper." There was Steve Huffman, who had just sold Reddit, and Drew Houston, who was preparing to launch Dropbox. They soon crammed in a fourth cofounder, Kyle Vogt, an MIT undergraduate who dropped out to build their portable camera system, which fit into a backpack and used Sprint's 3G mobile network. "It was like a dorm," Seibel says. "We could go out together, get drunk, hit on girls, work on startups—awesome."
While their neighbors toiled away, building unglamorous businesses, Justin.tv's March 19, 2007, launch became an immediate sensation. The San Francisco Chronicle did a front-page story. Ann Curry, in ]an excruciating Today show interview, lectured Kan. "Fame, I have to tell you, Justin, has a price," she said. But it was all fun, at first. Kan took the camera with him to the park, to business meetings, even to bars, where it made for awkward small talk. When one young woman took him back to her place one evening, he left the camera in the dark outside the bedroom; the gang back at Justin.tv headquarters overdubbed the video stream with audio from a porn movie. On his walk back to his apartment, Kan encountered a group of cheering viewers.
"If this doesn't scare the shit out of TV networks, it's only because they don't understand it yet," Graham told the San Francisco Chronicle. On NPR's All Things Considered, he declared: "Their ultimate plan is to replace television." It was fortunate for television, then, that Kan and his friends knew very little about running a business. "We had one week's worth of a plan," says Kan, laughing at what he describes as his youthful folly. "Today, I have an understanding of the world, and of the entertainment and media industries, of how people consume content," he tells me. "But at the time, I had no idea." Justin.tv was utterly unprepared to handle all the traffic created by the global press barrage. "The website would always break around 4 a.m., for some reason," says Vogt, who handled the engineering. Graham's initial investment had been $50,000; the first web-server bill was $40,000. Since Seibel was considered the most financially savvy of the four—thanks to a brief stint as a fundraiser for an unsuccessful Senate campaign—he took over as CEO. But product-placement deals he arranged with the likes of Zipcar, Bawls energy drink, and the Shia LeBeouf thriller Disturbia, didn't go far in offsetting the site's mounting expenses. Kan says he switched on his camera thinking that he would "just do it forever," but very quickly, he became exhausted by the continual pressure to perform. Kan tried to insulate himself from the trolls, but his efforts—concealing his phone number, speaking in codes—had a limited deterrent effect. The commotion around the Justin.tv apartment angered the building's landlord, who sent an eviction notice. Viewers, meanwhile, were frustrated with the lack of action. "My first thought was that he spent an awful lot of time sleeping," remembers Graham.
"We kind of realized, this isn't working," Seibel says. "We're not as interesting as we thought." Recruiting more vibrant folks didn't help. A handful of personalities were equipped with costly backpack camera systems and given their own Justin.tv channels. But these "stars" were fickle: Some just stopped broadcasting, while the most popular of them, an attractive blond web designer who called herself iJustine, deserted Justin.tv after less than a year to start her own website, and became a YouTube sensation. Needing to raise more capital, Justin.tv made its first big pivot. The founders had made considerable progress with their technology for streaming video. Their software had brought the cost of delivering an hour of video down to half a penny, cheap enough to serve constant video to a mass audience as an ad-supported business. They decided to enter the marketplace for live user-generated content.
The shift sounded good to some venture capitalists, including Stewart Alsop of Alsop-Louie Partners. He paid a visit to the Justin.tv apartment in the summer of 2007. "It was pretty gross," he says. "There were people asleep in the middle of the day, and I sat on the couch with Michael, the designated CEO of the company, to negotiate a term sheet." Undaunted by appearances, Alsop's firm invested $2 million. "We had this basic feeling about these guys," he says, "that they would do something important." So, just eight months after he'd launched the supposed lifecasting revolution, Kan proclaimed what his new mission would be: "democratizing live video." Justin.tv would now air whatever was just in. Promotional materials suggested that lifecasters could produce cooking shows, or deliver coverage of protests in Myanmar. There were categories for music, animals, "divas and dudes," and even a strange breed of users watching one another play video games. Other than a prohibition on sexual content, the site was a free-for-all. In effect, Kan issued a challenge to the Internet masses: Can you find a compelling use for this technology? The answer turned out to be, yes, there was at least one. The problem was, it wasn't exactly legal.
In 2008, a teenaged football fan named Greg Punzo was on a Chicago Bears message board, reading complaints from readers who weren't able to get the team's upcoming game on television. Punzo decided to deliver the game on Justin.tv, under a category labeled "sports," where people were supposed to offer Little League games and the like. But with a little technical savvy and store-bought components, Punzo was able to run the signal from his cable box into his computer, and start broadcasting the NFL. To executives with the networks and the major sports leagues, this was terrifying. They started to notice a bazaar of pirated game feeds on Justin.tv and other live-streaming sites. Between October 2008 and October 2009, Justin.tv more than tripled its traffic, reaching 21 million monthly unique visitors, according to Quantcast. Although copyright infringement is a perennial problem for sites that rely on user-generated content—in 2007, Viacom sued YouTube for $1 billion over the issue—free and easy streaming poses a particular threat to sports, whose broadcast rights are so valuable, and so perishable. The usual remedy, taking down a channel, was useless if the feed stayed up long enough to show the final score. In December 2009, Seibel was called to testify at a House Judiciary Committee hearing on sports piracy. "Do you think it is morally right for your firm to continue to engage in the use of copyrighted material?" Democratic Congressman Hank Johnson asked him. Seibel explained the company's legal stance: Justin.tv removed pirated material when it received complaints, but made no attempt to manually monitor the site, because that would "create an unfair expectation among copyright holders." He said that given the enormous number of Justin.tv users, it would be impossible to police a small minority of wrongdoers. "I also want to be clear," he added, "that the business model of Justin.tv is not based around supporting that." To rebut Seibel's claims, Lorenzo Fertitta, chief executive of the Ultimate Fighting Championship, held up a guide to streaming video games from the Justin.tv site, instructions that could easily be adapted to pirating television broadcasts. "It is kind of like going into a jewelry store and robbing it, but not making a mess when you are in there," said Republican Representative Tom Rooney.
Piracy did not appear incidental to the growth of Justin.tv. Traffic spiked dramatically on the weekends, especially during football season and events like the NCAA basketball tournament. Though the site introduced an automated takedown system, complaints continued to roll in from as far away as England, where many soccer matches are blacked out due to league regulations, and Australia, where weekend football games are shown on tape delay. Eventually, a boxing promoter and the UFC filed separate lawsuits in federal courts, going after Justin.tv for lost pay-per-view revenue. A UFC executive told Bloomberg Businessweek that he wanted to see companies like Justin.tv "obliterated." As if the lawsuits were not enough, last year Congress jumped in with the proposed Stop Online Piracy Act (SOPA), which sought to toughen copyright enforcement by making sites like Justin.tv legally responsible for the content they host. Kan was one of many tech entrepreneurs who protested. "The Internet is a place that fosters creativity," he says. "It would be a terrible thing to lose at a time when the economy's benefiting from so much innovation." Kan's mentor, Paul Graham, recently wrote a provocative essay arguing that the entertainment industry is "using a definition of property that doesn't work" at a time when data moves freely. Internet libertarians point to the fact that piracy flourishes most when broadcast regulations, such as regional blackouts, create artificial scarcity. "I'm obviously against stealing," Greg Punzo tells me. "I would stop doing this if the NFL broadcast every single game online with whatever ads they needed." Punzo's bootleg Bears channel has grown in popularity over the past four years, despite Justin.tv's efforts to drive him away—annoying but circumventable measures like removing his streams and banning his IP address. Now 20 years old, he spends football season running a donation-driven operation that streams many games in high definition. Earlier this year, I watched most of a Saints-49ers playoff matchup on Justin.tv courtesy of Punzo. When the channel was pulled down in the fourth quarter, he immediately sent another link to his 20,000 Twitter followers. Still, Punzo will be moving to another site next season. "Justin.tv is basically unusable at this point," he says. Representatives of the NFL and other sports leagues say that Justin.tv has become more responsive to their complaints, while the UFC, which declined to comment, recently settled its lawsuit. As federal law enforcement has cracked down on Internet piracy, much illegally streamed content has moved offshore. Coincidentally or not, Justin.tv's audience, which Quantcast pegged at around 20 million visitors in March 2010, began to decline, falling about 20% over the following year. By last summer, its once-thriving community had dwindled to a ghost town populated by puppy cams, eccentrics—like the stoner who filmed himself taking bong hits—and a few lonely lifecasters. Mostly young women, they tend to sit, static, in front of a webcam, rambling about their interior lives to a tiny audience of leering commenters. The line between this reality and tragedy can be thin: A few years ago, a Florida teenager named Abraham Biggs streamed live video of his suicide, by overdosing on pills, on Justin.tv. Though the Biggs incident was an isolated occurrence, it was the ultimate manifestation of a business challenge Justin.tv could never overcome. Even when the site was thriving, advertisers were wary of the unpredictability of live user-generated video. So were potential investors and buyers for the company. "Because of the gray-market status of a lot of the content, no one was going to touch that with a 10-foot pole," says James McQuivey, an analyst with Forrester Research. And the fun was gone: By last year, says Kan, he had "grown a little flat." Justin himself struggled to find anything worth watching on Justin.tv.
After four years of twists and turns, the Justin.tv guys knew a couple of things: Their business was stagnant, and people loved to watch streams of live games. What if they could find a sport that didn't belong to anyone, one that would actually appreciate their attention? They realized that a potential answer to their problem had been sitting there all along, out in the audience. The open platform of Justin.tv—the very freedom that the pirates exploited—had also allowed video gamers to build a niche on the site. But the company had always considered them to be something of a bandwidth-clogging nuisance. "I didn't get the gaming streaming," Kan says. "I wasn't a fan and I didn't understand it." Why on earth would a viewer choose to spend his time watching some Halo player blow away grunt after grunt? But traffic numbers suggested that a substantial number of people were doing just that. "Anytime you say to yourself, 'Really, people want to do blank?'—that means you've discovered something," Graham says. Emmett Shear, an avid gamer himself, argued that with some encouragement and added features, the audience could grow. "It's advertiser friendly," he says. "When you have a webcam, anything can happen. Gaming is much more controlled." Last June, the company launched a new gaming site called TwitchTV. Since then, it has built an audience of 17 million monthly unique visitors, according to company figures. Twitch has aggressively worked to turn the best gamers into sports stars. The professional scene is divided into subcultures, the most developed of which revolves around the game StarCraft II, a complex and fast-paced strategy game, in which two players build and coordinate armies that fight on an alien landscape. At this year's South by Southwest festival, hundreds of spectators—mostly young men—clapped thundersticks as two top Europeans, Aleksey "White-Ra" Krupnyk and Ilyes "Stephano" Satouri, played a match that was streamed live on Twitch. In Korea, home of the world's best players, StarCraft II tournaments fill arenas and are shown on TV. David Ting, general manager of e-sports for IGN, a Fox subsidiary, oversees one of several loosely organized U.S. organizations hoping to duplicate the Korean model, with Twitch as a partner. At a SXSW panel, he predicted that online viewership for his league's upcoming Las Vegas tournament—also to be broadcast on Twitch—would match ratings for MLS or the NHL. "Is playing a video game considered a sport?" he said. "Yes." Past efforts to popularize professional gaming have failed on television, but the new model—low overhead, few salaries, lots of audience interaction—is well-suited to the Internet. Twitch recently formed a revenue-sharing partnership with CBS Interactive, and it has signed deals to broadcast the two biggest American leagues, which offer both advertising-supported and pay-per-view events. (This puts the company, somewhat ironically, in the position of protecting its content against piracy.) Between big events, Twitch allows players to set up channels where they can broadcast their practice games, receiving a substantial cut of the ad revenue. While no American professionals make as much as the top Koreans, whose winnings amount to hundreds of thousands of dollars, some of the best have been able to quit their day jobs. "To me, e-sports is like poker," Kan tells me. "Twenty years ago, if someone told you that poker was going to be a massive spectator sport, you'd be like, 'What are you talking about,' right?"
Though Justin.tv still exists, when I visited the company's offices in March, all corporate enthusiasm had shifted behind Twitch. Seibel, meanwhile, had taken a couple of employees and moved to an incubator space across town, where he was working on Socialcam, a mobile application that repurposed Justin.tv's video-streaming technology. When I visited him there, on a Friday afternoon, he broke out a bottle of rum and we had a leisurely conversation in the incubator's shared lounge. "In most cases, Justin.tv would have died," he told me. "But it didn't, and we traded it into these two new entities." The four founders had gone their separate ways. Shear is now Twitch's CEO, while Kyle Vogt is one of two employees who maintain the rump Justin.tv. Kan stepped away from day-to-day involvement earlier this year, going back to Y Combinator to launch his third startup. Called Exec, it helps users hire assistants to perform short-term tasks. Kan says there are no hard feelings; he knows he is just one of those people who prefers the beginning of things. Kan and Shear are still roommates, in a much nicer pad than the Y Scraper, though part of him misses those days. One afternoon, he showed me a recent text from his old neighbor Drew Houston, of Dropbox. "You bros out?" it read. "What's he worth," Kan asked me, "like a billion dollars now?" Recently, Kan wrote a column for TechCrunch, entitled The Rat Race. "I've had friends whose startups have grown much larger and more successful than mine, made orders of magnitude more money, and I've felt envy at times," he wrote. But he advised his readers to be content with more modest accomplishments. "Don't do it because you expect that there is something magical waiting for you at the other end, some state of nirvana for the rich and successful." In Silicon Valley's current hothouse investment climate, it's sometimes difficult to figure what distinguishes the bright ideas from the billion-dollar ones. The Justin.tv guys didn't transform television, but they do seem to have stumbled, almost by chance, into some promising opportunities. After Facebook's surprise purchase of Instagram in April, the crowded competition to become video's version of the photo-sharing application heated up, and Socialcam suddenly turned into a lucrative property. A recent financing round for the spin-off company attracted high-profile investors such as TechCrunch founder Michael Arrington, Hollywood figures—Ashton Kutcher, agent Ari Emanuel—and a handful of current and retired basketball players. As of early May, the application had a reported 44 million users. As Socialcam roared up the list of Apple's most popular free apps, Kan touted its sudden success on Twitter with a triumphant "Kapow!" Twitch, meanwhile, seems to have found its own agreeable niche. On a spring Sunday, I met Kan and Shear at a bar in the Haight. Half the place's TV sets were showing Kansas and North Carolina playing for a berth in the Final Four, while the rest carried the broadcast of a Major League Gaming tournament in Columbus, Ohio. The cheering for StarCraft II was louder; the place was packed with video-game enthusiasts. Kan told me he now enjoys watching the sport. "StarCraft is a head game, a mind game," he said. "It's time!" the play-by-play man said, as the tournament finalists entered glassed-in isolation booths for their showdown. They were top-ranked Koreans, known to the audience by their screen names: DongRaeGu and MarineKing. Through the thrust and parry of a best-of-nine series, the fans hooted for animated explosions and picked-over tactics. "He was just getting a little too greedy in building up his economy," Shear said, after MarineKing dropped Game 3. In the climactic game, though, MarineKing sent an overwhelming wave of soldiers at DongRaeGu. "Oh, he's gonna get it!" Shear shouted, and then the screen flashed "Game Over." Around the world, viewers—Twitch and tournament organizers wouldn't disclose how many—watched as MarineKing, wearing parachute pants and a red-dyed Bieber haircut, emerged from his booth and did an awkward, couch-flopping celebration dance. Kan wasn't part of the audience, though—he had left the bar early to get back to work on Exec. This new startup, he said, is serious business.