The Good, The Bad and The Ugly of Using Social Media for Ad Targeting
[Social Media is a major buzz phrase these days. For Media Planners it's even more real than that, as clients demand it be a part of your plans. But how should you go about ad targeting using social media. In this premier, Rebecca Lieb, writing in our sibling publication, Advertising Age, provides some helpful guidance.]
By Rebecca Lieb
Get whiter teeth! Special discount for alumni!
Hasn't social-media-audience targeting gotten better than that? How helpful are social relationships when it comes to targeting? Will ads be the answer for Twitter, Facebook, Foursquare and other social networks, or will broader forms of marketing, e-commerce or virtual currency ultimately be the model? The answers: yes, somewhat and likely If these questions sound familiar, they should. Ten years ago, we were asking the same questions about the relatively non-participatory Web 1.0.
Many social-network-targeting methods are deceptively simple -- and simplistic. Is there a point in targeting all the friends of someone who "likes" a brand on Facebook, particularly if that person has thousands of friends in nonintersecting circles (family, colleagues, personal friends, the kids' piano teacher)? Not likely. No wonder a rumor surfaced at SXSW that soon to bow was Google Circles, a social network that shares content with those contacts who actually matter, not with everyone in the proverbial phone book.
We may see it yet.
Meanwhile, Facebook is the ground zero of social-network targeting. With more than 600 million users, it's not hard to understand why. Blinq Media is one of the dominant players in targeting Facebook audiences on Facebook, in no small part because the platform is integrated with Facebook's own API. Companies such as Turner and agency clients including Group M, Omnicom, OMD and Havas use the platform for what CEO Dave Williams calls socially endorsed ads. In other words, you can see which of your friends likes a specific brand or product. Blinq also can run campaigns to build fans, RSVPs or install applications.
Williams, who earlier founded 360i and SearchIgnite, calls social-audience targeting "a branding opportunity we hadn't seen before in digital advertising or in search." He cites a Nielsen study indicating performance and brand impact compound when 'likes' are streamed into news feeds, and an increased ability to see which demographics are most responsive to specific brands and campaigns. "It's the difference between targeting fish floating by vs. fish in a pond," he claims.
Motivity Marketing CEO Kevin Ryan sounds a note of caution, however. He sees advertisers, "Dialing up the targeting with a gazillion different options: audience, attitudinal, behavioral -- but they use same creative because all that targeting increases their cost. It's a young platform without a lot of stability. It's really hard to invest heavily in it because it's in a constant state of change."
Datran Media Senior VP Chris Gaebler cites other limitations of Facebook's platform, such as the inability to apply cookies to identify specific users or groups. "We've made many campaigns better over time by using smarter targeting. But you have to drive them off of Facebook to apply cookies."
TARGETING SOCIAL INTERACTIONS
Companies such as 33Across, Media6Degrees and RadiumOne target ads based on social interactions. Anonymous sharing of content, posting to a blog, emailing from a publisher site or commenting are some of the interactions that make the platforms assume people are connected, if not socially than at least in terms of affinities or interests. RadiumOne offers consumers the ability to "like" or "share" its ads with friends.
The new social data is fertile ground sourced outside the trough of generally available behavioral data that all networks, exchanges, agencies and DSPs have access to.
"Behavioral data has become commoditized, hence the rise of DSPs," RadiumOne CEO Gurbaksh Chahal said. "Everyone has access to the same data. Our model looks at everything that happens outside of Facebook that allows you to have that social experience. When you share I don't know what kind of connection you have, but I do know there's an influencer and someone being influenced."
But here's something to consider: Just because consumers are talking about a brand, or sharing stuff or even "liking" it, it's not necessarily an indicator of positive sentiment. Put more bluntly: They might be hating on you.
"Working with networked targeted audiences requires one, possibly two extra steps," said Pauline Ores, who until recently headed social-engagement strategy at IBM. The first determines that a discussion is taking place. The second should determine if customers "are actively disparaging an aspect of your offering or service."
THE SOCIAL-ANALYTICS BUBBLE
But how easy is it really to monitor and analyze social conversations? Social-listening-software solutions are proliferating. Radian6, SocialCast, Lithium Social Media Monitoring, Sysomos, Meltwater Buzz, Jive, Attensity, Brandtology, Omniture and Visible Technologies account for just part of an increasingly crowded landscape.
There's no doubt monitoring social conversations can help enormously when it comes to targeting audiences. San Francisco digital shop Questus was able to identify an untapped market of urban Latino and African Americans who liked to customize sport bikes, which led to Suzuki's Busa Beats campaign.
Yet its clear these are early days; having test-driven some of these solutions, it's obvious plenty of kinks are still to be ironed out. One product I played with differentiated between "editorial" and "blog" product discussions. But where do you draw that line? I mean, if your brand is mentioned in a CNN article that happens to be posted on the network's blog platform, there's a difference, right? Yet this solution lumped those mentions into the "blog" category. I've also seen social-link reports that are unable to screen out glaringly obvious (as in based in Cambodia) linkfarms.
Even when such kinks are ironed out, there are tougher nuts to crack when it comes to listening.
"Precise? No. Useful? Yes. It's really complex stuff. They're functionally valid for business today, but they're not going to get exact," said web-metrics guru Jim Sterne. "Sentiment analysis is a nut we've been trying to crack from the beginning of computing. On the whole, sentiment polarity is deemed good, bad, or neutral with just enough accuracy to be useful. 'I hate AT&T!' is pretty clear, and can be monitored and trended over time to determine if the T-Mobile acquisition is really for the better or not."
DON'T TARGET ME, BRO
Social-media-audience targeting is clearly here to stay, and methods will only become more sophisticated with time. In the current climate of privacy protection and do-not-track, there will be continued pressure from advertisers, agencies and vendors for more access to data and more-granular forms of targeting.
Forgetting for a moment about whether these indeed violate user privacy, paramount is that users don't perceive that their privacy is being invaded, or that advertisers have cross-hairs squarely on their social-media profiles.
Consumers are spending more of their lives on social networks. The challenge to advertisers is to target them with ads that don't feel too targeted, but that might be better than the alternative: low-rent, generically creepy acai berry/skin-care miracle/lower-your-auto-insurance-premium dross.#