How to Kill a Great Idea!
Published June 2007
PREV
EMAIL THIS ARTICLE
PRINTER FRIENDLY
COMMENT ON THIS ARTICLEKleiner and Benchmark were, in fact, so eager to grab a piece of Friendster that they agreed to a highly unusual condition: a $4.7 million cash payout for Abrams. Nonetheless, Abrams believes he made a critical mistake in negotiating the deal. He kept about a third of the company's stock but no longer had control of the five-person board. The deal specified that "preferred" shareholders--the VCs--would pick two board members, that Abrams would pick two, and that there would be a tiebreaker who would be mutually acceptable to both sides. Abrams says he didn't pay much attention to the issue because he had resolved to let the experienced VCs take charge.
As Friendster entered 2004, Abrams tapped Tim Koogle to join the board; he took a seat himself. Roger Lee, a partner at Battery Ventures, took the mutual seat, while the preferred seats went to Kleiner's John Doerr, a director at Amazon and Google, and Benchmark's Bob Kagle, the VC who discovered eBay. Abrams agreed to let Koogle run the company as interim CEO while the two men focused on building a professional management team. To fix Friendster's engineering woes, they hired Jeff Winner, who'd co-founded Collabra Software, a business collaboration tool sold to Netscape in 1995 for $108 million in stock. John Briggs, a seven-year veteran of Yahoo, was hired as VP of product management. Mary Lou Song, employee No. 3 at eBay, was charged with managing the rapidly expanding user base. The CEO job went to Scott Sassa, who had been president of NBC West Coast, overseeing hits like The West Wing. He was a well-connected entertainment executive who could credibly strike content deals with traditional media companies.
Each of the new hires came to Friendster with strong ideas about how to make the company as big as possible as fast as possible, with an eye toward a big exit for the investors. With new rivals--most notably, MySpace and Facebook--emerging, they wanted to move fast. But agreeing on a game plan turned out to be a problem. "There was this leadership battle on top that was like a war in Valhalla," says Chris Lunt, who joined Friendster in 2003 and took over as director of engineering when Winner left in late 2004. "Everybody had their own agenda." The result was a kind of corporate schizophrenia. Rather than improving the software, Friendster went on a partnership binge, resulting in a hodgepodge of incongruous and poorly integrated features: blogs (with Six Apart), video sharing (with Grouper), personalized searches (with Eurekster), VoIP (with GloPhone), and Internet radio (with Pandora).
The tenor of the board meetings quickly deteriorated, with Abrams becoming increasingly isolated from the board, which now also included Sassa. "We had an inexperienced founder and a lot of experienced and high-powered board members," says Kleiner's Siegelman. "There were too many cooks in the kitchen." Abrams, the board's chairman, hardly considered himself inexperienced and felt ignored by his five colleagues, who, he says, generally sided with Doerr. He was particularly vexed by the company's apparent obsession with partnerships. "At the board meetings they would say, 'We should do a deal with AOL,'" he recalls. "And I'd be like, 'Guys, the site is not working.'" He never got anyone's attention, and in 2005 he was stripped of his chairmanship. He stopped coming into the office regularly.
In hindsight, the decision to marginalize Abrams, an experienced engineer, probably was a bad move. Rather than address the problem of too many calculations, Sassa opted to make massive investments in hardware and software in 2004. Under Winner's leadership, a team of engineers completely rewrote Friendster's code into a different programming language and spent more than $1 million on a Hitachi (NYSE:HIT) storage area network, effectively halting business development for six months. Although Winner claims the rewrite was successful, load times continued to be a problem as late as 2006, according to Chander Sarna, Friendster's current vice president of engineering. "The ex-Friendster people are not going to like me for saying this, but there was a lack of spending discipline," Sarna says. "There were very basic problems that good code writers should have fixed to begin with."
Whether or not it was a success, Winner's program was divisive. Many of the software developers considered the rewrite unnecessary, while the product development team complained that they needed to add features to the site in order to compete with MySpace. The result was constant bickering between cliques and side projects that went nowhere. Sassa, a Hollywood deal maker who had never closely managed engineers, lacked the technical expertise to moderate those disputes. "It was the most frustrating place to work," says Lunt. At the board level, Siegelman says, there was a realization that the management team was "dysfunctional," but the board was loath to micromanage. "The pot kept getting stirred, but nobody said, 'I'm turning down the flame because this isn't the right recipe,'" he says.
