This is the subhead for the blog post

One of my favorite panels at last month’s AdSpace conference was the discussion on social media advertising. Jeremy Liew – the panel’s moderator – wrote a great summary of the learnings from that conference, which I recommend reading. What really stood out to me in this session was the incredible degree of hyper-targeting available to advertisers on social media. For example, Jeremy summaries an example from Ro Choy, CMO of RockYou, about a local yogurt store:

Ro related the example of Rachel’s yoghurt, an advertiser that targeted coupons to women living within 5 miles of Whole Foods in 10 cities through Rock You. The campaign delivered 0.20% CTRs to the Rachel’s Yoghurt site, with a 35% coupon download rate.

And Tim Kendall from Facebook gave a cool example of a city counsel candidate in DC:

Tim described a very detailed campaign that a politician, Patrick Mara, ran on Facebook to defeat a 16-year incumbent in a DC city council primary last year. Mara was in favor of allowing gay marriage, so he pushed information about his stance out to DC Facebook users who’d listed their sexual orientation as gay. If Facebookers had kids, he targeted them with ads about the school system, and if they were Republicans, he hit them with information about taxes, school vouchers and similar conservative favorites.

I’ve noted in the past on this blog that Facebook and other social media sites represent the first opportunity to truly get honest demographic and psychographic information from users. Tim and Ro’s examples show how this data can really be used to a marketer’s advantage. And yet, most advertisers who have dipped their toes into social media marketing have not had the same successes. Most people I know tell me that the conversion rate on social media is very low, to the point that they’ve completed dropped their social media initiatives.
The disconnect, I think, comes down to targeting. Tim and Ro are describing marketing campaigns that are so granularly targeted that it’s hard to imagine them as any but successes. My marketing friends, however, took the “mass market” approach and attempted to drive huge volumes of clicks from entire geographic regions, or entire demographic groups. If you try to market a “California mortgage” advertisement to California residents over the age of 18 on Facebook, you’ll fail, but market a “California refinance bankruptcy loan” campaign to people in Stockton who are homeowners who make less than $50,000 a year and who have expressed an interest in debt consolidation, and you’ll be successful.
The difference in cost between these two hypothetical examples is enormous, probably along the lines of $20,000 a month versus $200 a month. This is a problem for Facebook. If success only comes in $200 intervals, the path to a $15 billion valuation is a long way off. Trust me, if Google’s advertisers only spent a few hundred dollars a month on average, Larry and Sergey would have gone back to grad school by now.
The other problem with social media that I discovered during the panel was the lack of fungibility. Keywords on Google are fungible – we can buy the same keywords across multiple search engines, we can get nice reports that show which keywords work and which don’t, and there really is a pretty limited universe of potential keywords (especially thanks to the ever-expanding broad match algorithm!).
Social media advertising, however, isn’t as fungible. When you combine demographics with psychographics and geography you end up with an almost infinite number of possible combinations. Should you be marketing to black, married, women, over 45, in San Francisco, with a passion for motorcycles, or Hispanic, gay, men, under 20, in Cleveland, with a passion for gardening?
And how do you create an interface that makes it easy for the average marketer to understand this web of choices? Big, tech-savvy, advertisers could use an API to access and test this data, but the little guys – who make up quite a bit of Google revenue base – are already confused enough about keywords to come close to understanding this sort of marketing.
Facebook can’t grow into an advertising juggernaut if it relies on small, hyper-targeted advertisers to spend $500 to win their city counsel election, mass generic campaigns don’t work, and the complexity of figuring out targeting on a mass scale will limit the number of advertisers who can make large budgets work. Too much targeting, too little fungibility. I think it’s going to be a while before Facebook and other social media networks can really drive a lot of marketing revenue.