Page Comparison

The story of Twitch starts in Seattle. That’s where its founders, Shear and board member Justin Kan, grew up, childhood friends living three blocks away from one another in the city’s Capitol Hill neighborhood.

But it was as classmates at Yale University that the two first worked together, launching a web calendar startup called Kiko in 2005. It didn’t last. One month after launch, Kiko — designed to combine the power of Microsoft Outlook with the modern web sensibilities of Google’s then-new email product Gmail — was buried by a different solution: Google Calendar.

Google’s clout and resources alone probably would have been enough to stop the startup. But as Kan recently said at Startup Grind, a startup event in San Francisco, the young co-founders weren’t heavy calendar users themselves and didn’t have the savvy to find any power users for feedback.

“One of the things I’ve learned the hard way over the years is that you should use your own product,” said Kan, now a full-time partner at Y Combinator.

In an interview with Re/code, Shear said needing to use calendars in the professional world, for more than college classes, has made him appreciate what he and Kan could have done differently.

“Now I’m an extremely heavy calendar user and I have all these ideas for things that are annoyingly wrong with Google Calendar, that I use today,” he said. “Now I have all these ideas, but then, I didn’t know. We didn’t talk to anyone, and it was a huge mistake.”

After 14 months, Kiko did find an exit, however … it was sold on eBay, for $258,000.

“We thought we had hit the lottery,” Kan said, having expected no more than $50,000 from the auction. “Maybe startups were not so bad after all.”

That money kept them going, and they learned from a series of further fizzle-outs like a Facebook-for-families social network and a company that would sell glow-in-the-dark gene-spliced roses. Then the duo landed on one idea that Y Combinator founder Paul Graham was willing to invest $50,000 to see: A reality show devoted to broadcasting Kan’s life on the Web, 24/7.

Kan recalled another seed investor saying, “I’ll fund that just to see you make a fool of yourself.”

Justin.tv began broadcasting in early 2007, and quickly scored some major headlines; Kan even wound up on NBC’s The Today Show and ABC’s Nightline. Shear said that part of what made the Truman Show-esque experiment work was that this time, the product was something its creators wanted for themselves.

“If we had any success with Justin.tv, it’s because we were our own user,” he said. “We knew what was important because it’s stuff we wanted.”

In time, though, viewers began complaining about the generally drama-less show (“We did not understand reality TV,” Shear acknowledges) and, more crucially, asking how to make their own live video shows. Kan said outsiders often assume that turning Justin.tv into a broadcasting platform for others was part of the plan all along. Those outsiders, he clarified, are wrong.

“The honest truth is, we weren’t that smart,” Kan said. “It was only later on that the lightbulb went off.”

The site opened up to other broadcasters in October 2007. Three years later, Justin.tv had raised $7.2 million in venture capital and was claiming some 31 million unique users per month.

It was at this time, in late 2010, that the company began working on two “skunkworks” projects. One project, a Justin.tv mobile initiative led by Michael Seibel, would spin off into an Instagram-for-video startup called Socialcam, which sold to Autodesk for $60 million in 2012. The other, a taskforce led by Shear to grow the audience for videogame content on the site, would also become its own product: A new site called TwitchTV.

“It’s kind of counterintuitive, right?” Shear said. “You want to be as big as possible as a startup, why would you pigeonhole yourself in just one kind of content? It took us a long time to realize that was a good idea.”