Even as he wrote his rules, he was ignoring them. He wasn't even podcasting. As Odeo sputtered, struggling to gain new users, Williams began to see his problem as one of corporate structure. He had accepted millions of dollars in investment capital, built a team, and worked the media before he knew what his company was. Odeo needed to experiment--to play, even. "If we were just two guys in a garage, we could say, 'I don't know about that idea, but let's see where it goes,' " he says. His solution was to organize what he called a "hack day." He broke the company into small groups and told them to spend a day experimenting--not just with podcasting, but with anything that struck their fancy. It was Dorsey's project that struck Williams's. Dorsey had long been fascinated by the status function on instant message programs: the short, pithy postings that allow you to tell your online friends what you are doing. He built a prototype of Twitter in two weeks.
"Thinking twttr is the awesomest," Williams Twittered in March 2006. With little fanfare it went live in July. Like Blogger before it, Twitter was introduced as an experiment, a fun little side project. Nonetheless, Williams was excited--more excited than he'd been about anything that had happened at Odeo. This got him thinking about the hack day that had led him to Twitter--and then about the two years in which he had struggled to build anything, despite having plenty of money and all the hype in the world.
How had a single experiment succeeded where an entire company couldn't? And more important, how could he do more of them?
On October 25, 2006, Williams blogged his answer. He was buying Odeo, taking the odd--to some, almost unbelievable--step of returning his venture capitalists' money. It cost him $3 million out of pocket, plus all the cash Odeo still had. It was a lot to pay for a failing Web company and an unproven prototype.
He called the new endeavor Obvious, a nod to a lesson learned from the success at Blogger--that seemingly silly and trivial ideas often look like great ones in retrospect. Obvious would be a workshop where Williams and his cohorts could experiment with ideas in an environment free from financial distractions. If an idea worked really well, he could spin it off into an independent company using outside investment. Otherwise, he could either keep it for Obvious or throw it away. "I don't want to have to worry about getting buy-in from executives or a board, raising money, worrying about investor's perceptions, or cashing out," he blogged. The move was widely seen as heroic. "Odeo Buys Back Soul," read the headline of gossip blog Valleywag.
Shortly after buying Odeo, Williams wrote a blog post that announced his intentions to sell the podcasting part of the company--a New York start-up paid a reported $1 million for the service--and focus on Twitter. The text messaging service had its coming-out party at the South by Southwest technology festival in March, where conference attendees eagerly began Twittering one another. From there it grew rapidly, reaching a hundred thousand users in a matter of weeks and garnering nationwide media coverage. In July, Williams formally spun off the company, raising several million dollars from Union Square Ventures, a New York City VC with a hands-off reputation. (Managing partner Fred Wilson, who, judging from his Twitters, really, really loves to eat at Murray's Bagels, had been using the service for months.) Williams appointed Dorsey CEO and told him to focus exclusively on fixing Twitter's reliability problems. Though Williams remains the single largest shareholder, he has taken pains to stay out of Twitter. The business model, he says, can wait until millions of people are using it.
Beginning on the first day of this year, Williams started working in earnest on Obvious. His work area is a small nook under a lofted conference room in Twitter's San Francisco office. The building has served as a private home, a snowboard factory, and an underwear store. The soiled carpet is a sort of puke-green color, and the only natural light comes from a few skylights far overhead. To date, Williams has hired two contract engineers to build small software products; they are building an application that will allow users to write "notes to self." Obvious isn't particularly counting on this product--"It's almost not worth talking about," Williams says--but that's the point. Williams wants to make product development less risky and more prone to the kind of spontaneity that created Twitter.
At the same time, he's trying to find early-stage start-ups to roll up into Obvious. He says he would like to invest roughly $100,000 in each company. Everyone will work in the same office, which means he will eventually have to look for additional space. He's also trying to hire an assistant: The job description warns that the candidate will be paid hourly "until you set up the payroll system for the company, and then we can discuss salary and insurance (once you set that up, too)."
The goal is to separate the creative environment of the start-up process from the regular work-a-day of running a business. "It's all theory for now," Williams says. "But we're hoping that by setting up an environment with multiple projects at once, these happy accidents can occur." If this sounds unbusinesslike, then that's the point, too. Obvious is, in the broadest sense, a company founded on the idea that it's hard to predict which ideas will work and which won't. "It's almost like a theater troupe," says Stone. "The idea is to tinker around and to be willing to come up with flops."
Like most good theater, Williams's new company is at once disruptive and self-indulgent--an ambitious challenge to the Silicon Valley rule book and a test for all of those blog-worn theories. The company of little experiments is itself an experiment, and a chance for Ev to do something grand on his own terms.
