Splicing relevant ad messages into the social graph was supposed to be a cinch, given all the personal data profile creators leave lying around. But the big networking platforms and their ad reps have struggled to find the right formula. Google admitted in January it has searched in vain to monetize its vast inventory on Orkut and partner sites including MySpace. MySpace has its own data extraction project, dubbed hyper-targeting, yet still sells inventory in bulk rather than by niche segment. Facebook is in a similar conundrum.
What’s the problem? For one thing, it can be hard to separate out useful data in a person’s profile. For another, the meaning of “friend” is relative, making it hard to “[get] into the conversations that already happen between people,” as Facebook’s Mark Zuckerberg grandly forecast marketers would do last November. Put another way, when someone claims a connection to 400 people, it’s tough to tell who the real buddies are.
A behavioral targeting firm called Media6degrees said it’s engineered a workaround to the dual problems of data and relationship glut. Rather than look at a person’s friend list, the company uses a combination of cookies and ad server logs to pinpoint a person’s interests and generate anonymous profiles of her real friends.
The New York-based start-up has toiled in obscurity for much of the past year, employing a technology development team of about 13. With its hire this week of a senior Microsoft ad executive as CEO, it hopes to burst onto the consciousness of the ad community.
New Chief Joe Doran was general manager in Microsoft’s digital advertising solutions unit and was a principal architect Microsoft’s acquisition of aQuantive. Doran was involved in other acquisitions and strategic initiatives during his nine years at Microsoft, including the purchases of in-game ad network Massive, ad exchange AdECN and European mobile ad firm ScreenTonic.
Doran said Media6degrees has based its platform in part on 1990s Bell Labs research on marketing and social theory. Bell Labs analyzed individual phone customers’ call records to determine recency and frequency of contact with other individuals. It then found marketing offers directed to those “friends” saw three to five times greater response rate than a control group.
“The team looked and said, how’s this going to look, to replicate Bell Labs’ work on the Internet,” Doran said. “Can we do that in a consumer friendly, privacy friendly way… looking at interaction on the social graph and getting rid of the voyeurs and stalkers? We want to see people that really have rich connections.”
He said Media6degrees found an Internet analogy in referral information stored in ad server logs. Those logs can be made to cough up specific social networking profile pages an individual has visited recently by analyzing referring URL structures, which is potentially actionable information when juxtaposed with an action taken by the ad-clicker. For instance, someone acting on a Miami timeshare or scheduling an F-150 test drive may have friends and relations who would be receptive to a similar offer. Affinity groups of presumably anonymous individuals are created by grouping individuals who have visited the same set of profiles.
But how does the firm get access to those server logs? In a manner similar to a traditional behavioral ad network such as Tacoda, it partners with a network of advertisers that each volunteer to place a cookie on the browsers of Web users visiting its sites. Participating advertisers and site owners receive a small CPM to distribute its pixel.
“If you put up a media pixel on your site, we can not only help you retarget an individual who came to that site, but we can help you identify micro-affinity groups,” said Doran. “The data we used to capture this is data that’s owned by the advertiser today. We get a piece of data with every ad served.”
Doran argued the owners of social networking platforms are hurting themselves by focusing on the content people share in their profiles.
“The most important thing is not to look at the content but to look at the interactions between individuals,” he said. “I’m defined not by my interactions on MySpace or on Facebook. I’m defined by my interactions with my friends.”
The company has worked with about 30 “alpha” advertisers to test its system. These marketers get inventory at cost. It’s now preparing for a wider debut with a number of advertisers that Doran declined to name.
Media6degrees is funded by Contour Venture Partners and Coriolis Ventures. It’s a member of the Network Advertising Initiative, a coalition of behavioral targeting firms.
So what makes content go viral? And what makes people participate in these phenomena?
Brands have been upping their investments in new ad products from popular social media services, but are they getting their money's worth?
Instagram is determined to introduce as many new features as possible in 2016 and that's why it has launched Live video on Stories, as well as ephemeral posts on direct messages.
Audience targeting can be challenging in social media, especially when brands make quick assumptions about their target users. How can you avoid generalisation and what are the real benefits of it?