Banner Ad Research

There is another set of research out about the demographics of banner ad clickers. This comes via AdLab. The ComScore press release it talks about is here.

I can't really say strongly enough how bad this piece of research is. Specifically some of the conclusions it draws from the finding that people who click on banner ads aren't representative of the US population at large.

I think I have linked to this before, but Seth Godin made the best point about this in a blog post a while back. In a nutshell, there is no reason they should be. Banner advertising caters to free-offers, gambling sites, one-time promotions, coupons, 'You've just won $1,000,000', type advertising. Who do you think clicks on these types of links?

A further comment by Starcom's Director of Research, Grant Prentice:

“Natural Born Clickers shows us that we can’t count on click-through rate as our primary success metric for display ads; Starcom is more reliant on shifts in brand attitude metrics and analytics tying on-line exposure to sales as the true measures of online advertising efficacy.”

I beg to differ. I've done research on the 'residual' brand effects of banner advertising and finding a clear brand effect is like trying to squeeze blood from a stone. Besides, surely a display ad that has any impact at all would overcome the 'click' barrier. It's not as if placing your mouse over an image and clicking is a laborious task.

It's very hard to argue that a residual brand effect occurred via a banner ad but that effect wasn't large enough to encourage a very simple behavioral response. If they don't click on it, it didn't do its job. There is no large, hidden brand effect that won't be strongly correlated with click-through rates (although the opposite could be true, high click-through rates with no brand effect - i.e, some sort of forgettable promotion).

The whole issue reminds me of some work I once did for a bank where their DM team was convinced credit card offers in the mail must have left some type of residual brand effect that could be measured (yes, the ones you open and throw away). Like banner ad makers, they saw response rates in the <1% range and were thus convinced their hard work must be having an outcome they couldn't see.

I think I lost the initial argument to try and find this effect, but the research got stopped by someone higher up the chain. I think they realized that the 'brand effect' of credit card offers served everyone's best interests as a myth.

You produce better work when you believe its impact is meaningful. Sometimes myths are more useful than facts.

Sphere: Related Content