For some time now Internet marketers have believed that banner ads are slightly better than useless (only slightly better mind). The accepted wisdom is that web users are blind to banner ads, that they're so used to seeing them on websites that they unconciously blank them out.
This wisdom is usually based on studies asking participants to recollect the ads they've seen after browsing a particular website. Guillaume Hervet has recently investigated the effectiveness of banner ads using an eye-tracking study.
Participants were asked to read 8 web pages. On the 3rd, 4th and 8th pages Adsense style ads appeared to the right of the content. For half of the participants these ads were related to the content, the other half not, with all ads for ficticious brands.
Interestingly the study found that 82% of people looked at one or more of the ads. What's more, of 128 ad exposures, 37% were looked at more than once.
So did the ad make an impression?
This was tested by asking the participants to read the ads after they'd been blurred, ie made difficult to read. This group was then compared to another group asked to perform the same task, only not having read the web pages from earlier. The theory being that if group A could read the ads better than group B that they would have absorbed it during their earlier browsing.
Context is everything
The study found that recollection was high, but only for ads relevant to the content being read. Likewise, location of the ads was important. If browsers had already seen an ad in a particular location then they tended to ignore subsequent ads placed in the same location. It only takes a couple of ad'less pages however to get people back to normal.
Take away lessons
- Keep ads relevant to the content
- Vary their location
- Don't have ads on every page.
Adsense ads aren't really banners though are they? Sure I've seen research showing they're clicked on more often than traditional banners are. This must be especially so when banners are generic ones with no relevance to the content.
With that in mind though, I wonder how newspaper ads perform, as usually they have no relevance to the content at all, it's all done based on demographics.