Facebook may be about to cut their own throats.
According to Business Insider, they will soon be launching a product that will...
"...prove to marketers that there is a type of online advertising besides Google search ads that is worth spending large amounts of money on."(By the way, that bold "prove" is not from me, it's from the gee-whiz article at Business Insider.)
Facebook better be careful. If their methodology is any good, the only thing they're likely to prove is how ineffectual display ads on their site are.
According to the article, the basis for Facebook's new product is a research tool from a company called Datalogix that allows them to correlate the delivery of ads on Facebook with the purchase of products. So, for example, they can tell Coke that of the 1 million people who had a Coke ad delivered on Facebook, 300,000 bought a can of Coke in the next 3 days.
Sounds pretty good, huh? Except there's a little problem. There's a big difference between correlation and causality. Correlations can be very misleading. For example, of the 1 million people who get a flat tire today 300,000 will probably also buy a can of Coke in the next 3 days. Does this mean getting a flat tire causes the purchase of Coke? I don't think so.
Most ad agencies and advertisers are so dumb and so addled by impenetrable "metrics" that they'll probably buy a correlation scam, if that's what Facebook has in mind. But if some smart people start asking the right questions, Facebook could find itself in deep snow. If they think they can prove substantial causality between the crappy little ads on their site and substantial purchase increases of major brands they're delusional.
Right now, small scale advertisers who don't need significant reach (dentists, weight-loss scammers, rehab clinics, and other crappy direct marketers) are willing to buy Facebook ads because they only pay for clicks, and on a cost-per-click basis they can project a positive ROI. But if Facebook thinks they can prove a positive ROI to brand marketers who aren't buying clicks and need massive reach they're asking for trouble.
The problem for Facebook here is that if this new methodology cannot prove a positive ROI to brand marketers as it promises to do, they will seriously damage their push for a reach-and-frequency based revenue model. And they will remain in the thrall of crappy little direct marketers.
They better be pretty certain they can prove what they think they can prove before they launch this thing.
3 comments:
Really the only meaningful test of Facebook ads for brand marketers is to stop advertising on Facebook. All other things being equal, did your sales go down? If so, advertise on Facebook. If not, don't.
@Stuart: But isn't that more or less the basis for the logic of how we know Facebook ads DON'T work. Because we're still in recession; and despite all the millions (billions?) spent on "social media" sales in fact have NOT gone up (and instead in the vast majority of cases, have gone down)?
That might be the most awesome comment spam I've ever seen. Utterly surrealistic text. Any chance you can leave it up?
Post a Comment