But it isn't much good for an entrepreneur who has a promising idea--and who would prefer odds that are better than 20 to 2. Spolsky believes that working with a VC imposes a level of risk that someone prepared to invest his life--not to mention his life savings--in a single enterprise simply should not tolerate. "An entrepreneur would rather have a 100 percent chance of owning an $80 million company than a 10 percent chance of having a $800 million one," he says.
Friendster never felt like a long shot to Abrams, who seemed to understand Silicon Valley as well as just about anyone. He came to Netscape in 1996 as a software engineer, having worked several years for Canadian telecom giant Nortel (NYSE:NT). He spent a year and a half at Netscape, writing code for the Navigator Web browser and immersing himself in the culture of the time and place, becoming a regular at meetings of the Silicon Valley Association of Startup Entrepreneurs and the Software Development Forum.
Abrams left Netscape in 1998, and nine months later started HotLinks, an early foray into what is now called "social search." Abrams's idea--to organize webpages based on users' favorite sites--was prescient and would eventually appear in the form of successful ventures like Digg and Del.icio.us. Over the course of a year and a half, HotLinks attracted 500,000 registered users, but it ran out of money in the wake of the technology collapse. In the spring of 2001, HotLinks merged with a British software company, and Abrams left to work for another start-up.
As he suffered through the dot-com bust, Abrams began mulling a new idea: software that would somehow integrate one's online and offline identities. "The way people interacted online was either anonymous or through aliases or handles," he says. "I wanted to bring that real-life context that you had offline online--so instead of Cyberdude307, I would be Jonathan." Abrams was also mindful of Gary Kremen's Match.com, which eight years after its founding was finally coming into its own. Now part of IAC, the site was booking $76 million a year in revenue--roughly one-quarter of the $300 million online dating industry in 2002. Abrams saw that the cultural perception of online dating had changed. "All of a sudden, people who I would not think of as strange and desperate--normal people--were talking about using Match.com," he says.
Friendster crystallized in the summer of 2002 while Abrams was walking with a friend in a Santa Clara park. They were chatting about the online-offline problem and out popped the idea: Each person would have a standardized homepage, à la Match.com. But instead of simply advertising their interests and good looks, users could link their profiles to those of their friends, creating a network of connections that would mirror those that existed in the real world. His friend liked the idea, and Abrams started work immediately. Three months later, he had a prototype, which he posted on the fallow server of a friend's failed dot-com. He sent invitations to about 20 of his closest friends, unsure of what would happen next. "The least likely thing in my mind was starting another company," he says. "I wasn't sure what I was going to do."
Abrams's invention--which would be awarded patent number 7,069,308 four years later as a "system, method, and apparatus for connecting users in an online computer system"--was far more enticing than he initially imagined. As an online dating tool, it represented a potential improvement over Match.com because users could figure out if they had acquaintances in common with a potential mate, thus bypassing the awkward unsolicited e-mail in favor of an introduction by a mutual friend. But beyond its applications in dating, Abrams's software was compelling as a pure idea.
The beauty of Friendster was its exhaustively complete network. Every time a homepage loaded, Friendster's servers calculated a single user's connection to other users within four degrees of separation, which could mean hundreds of thousands of individuals. Because the network was constantly changing as new users joined and connected with one another, these calculations had to happen on the fly--in what would eventually amount to trillions of rapid calculations. The effect was to give users a vivid sense of how they fit into their social groups as well as into the larger world. Abrams, it seemed, had created a piece of software that could tell us who we were.
Prototype in hand, Abrams began looking for seed funding. He delivered his first pitch on Thanksgiving Day in 2002 to a former HotLinks vice president, Melissa Lloyd, over dinner with Lloyd and her husband at their home in Sun Valley, Idaho. Abrams's hosts had no idea what he was talking about but agreed to invest a few thousand dollars anyway. "We believed in Jonathan," Lloyd says. "So we said, 'Here's your money; we don't want to hear about it again." The Lloyds sent Abrams home with a check for several thousand dollars and eventually invested tens of thousands more. What happened next, says Lloyd, who now lives in Seattle, was "one of the most exciting times of my business life."
Over the next few months, Abrams rounded up $400,000 from a dozen investors. He opened Friendster in March 2003. The site grew virally as Abrams's friends invited their friends, and by June it had 835,000 registered members. Four months later, there were more than two million, generating some 10 million page views per day. The growth presented immediate engineering headaches. In theory, Abrams's intricate network was a beautiful thing. In practice, the constant calculations, which were being continuously served on millions of homepages, required more than a terrabyte of expensive RAM memory. By late 2003, load times regularly clocked in at over a minute and users were beginning to complain in blogs and forums. Abrams's software would need to be scaled somehow. "We would fix one problem, and then a few days later there would be another bottleneck," recalls Ian McFarland, a software developer who joined Friendster in April 2003. Simply buying enough servers to keep up with the growth was a major challenge.
