Are we really blind to Internet banners?
It's a line of research that Google doesn't want you to know about. Many studies suggest people have a habit of simply ignoring web banners on Internet sites - a phenomenon known as banner blindness. The evidence for this ad avoidance is based largely on tests of people's explicit memory of ads after they've browsed a site. Of course that doesn't mean that the participants hadn't looked at the ads, nor does it mean that the ads hadn't lodged their message subconsciously.
Now Guillaume Hervet and his team have attempted to address these points in an eye-tracking study. Thirty-two participants read eight web-pages about choosing a digital camera. On the third, fourth, seventh and eighth pages, a Google-style rectangular text ad (180 x 150 pixels) was embedded in the right-hand side of the editorial content. The second ad was different from the first, and then the same two ads appeared on the seventh and eighth pages, respectively. Also, half the participants were exposed to ads that were congruent with the camera topic of the web-pages; the other half to incongruent ads. All advertised brands were fictitious.
The results may be of some consolation to Google and their advertisers. Eighty-two per cent of the participants did actually look at one or more of the ads. Or put another way: of the 128 ad exposures, 37 per cent were looked at once or more. Had the ad content made a lasting impression? To test this, after the browsing phase, the participants attempted to read the same ads presented in varying degrees of blurry degradation. Their performance was compared to a new group of control participants who hadn't done the earlier web browsing. If performance was superior among the participants who'd earlier been exposed to the ads, this would suggest they had a lasting memory of the ad content. In fact, performance was only superior for web-browsing participants who'd earlier been exposed to ads in a congruent context.
Another aspect to the results is how the participants' behaviour changed over the course of the web browsing. The first and third ads were looked at for longer than the second and fourth ads. This is probably because the second and fourth ads appeared on pages that had been preceded by a page with an ad on it in the same location - the participants seemed to have learned to ignore that area of the page. On the other hand, it seems a couple of pages without ads was enough to restore ad-looking behaviour.
The lessons for web advertisers are clear: don't advertise on every page, vary ad location, and make sure the ad topic is congruent with the web-site content.