The real money in social media might not reside in the ads that sit on Web sites like Facebook and Twitter, but in the data produced by users' frantic friending and sharing.
The rationale is simple. Internet users are now spending 22 percent of their time in social media, and Internet activity leaves behind a trail of data: what people like, what they share, and who is connected to whom with similar tastes. For publishers and application makers, licensing all that social data is a no-brainer -- found money for what is in essence a waste product of their services. For advertisers, social data is a potential boon: a way to find likely customers based on their sharing and communication habits.
Facebook, the obvious key player in the social data game, is currently taking a backseat to scrappier startups, both to avoid overstepping privacy sensitivities and to focus on other growth areas. That's left an opportunity for new companies like Media6Degrees, 33Across and RadiumOne, which are licensing large amounts of data from instant-messaging clients, sharing applications and blog services in order to piece together customized social networks for ad campaigns. For instance: A campaign for Nike based on social media data will show relevant ads to people connected to Nike customers, on the supposition they're likely to have similar tastes.
“It's your real social graph, not people you went to high school with 20 years ago,” said Eric Wheeler, CEO of 33Across.
Other companies are collecting data directly in order to target ads outside social media sites.
ShareThis, which has content-sharing buttons on over 1 million Web sites, collects as much as a terabyte of social data every day. Earlier this year, it took the search terms from skin care brand Mederma -- stretch marks, dry skin, scar tissue and others -- and identified people who had viewed related content, shared it or received it. ShareThis then built an audience segment of 12 million with the DoubleClick Ad Exchange. Frequent sharers were seven times more likely to view a Mederma coupon than the less-specified audience of previous campaigns.
ShareThis CEO Tim Schigel said that such methods could help crack the code on brand advertising online. While current Web-targeting methods are geared to direct response, social characteristics are more suited for awareness and consideration -- the critical backbone of brand advertising budgets locked up in TV. “There's a huge potential that's not tapped,” he said.
Still, collecting social data is a tricky business. One social data outfit, Rapleaf, was rocked after a Wall Street Journal story questioned its tactics in collecting user information. As for Facebook, if and when it does enter the data-mining market -- which several industry observers have predicted it will do, as the site marches toward an inevitable IPO -- it will need to be careful because of its history of igniting concerns over privacy.
“The really big question is whether anyone but Facebook can make a big business out of it,” said Terry Kawaja, CEO of investment banking firm Luma Partners.
CEO: Tom Phillips
Differentiator: “We look for your customers in the corners of the Web. We find the places that have a signal.”
CEO: Eric Wheeler
Differentiator: “We use social data to understand the people that are most likely to share interests and purchase behaviors.”
CEO: Gurbaksh Chahal
Differentiator: “We’re extending where behavioral targeting left off. We’re finding look-alikes of your customers and the people who have direct and indirect connections with them.”