Do NOT use banners on your own web pages

It’s a BIG difference in how you should use banners for external marketing and banners on your own web pages.On other people’s sites you don’t really have to worry that much about it, just measure and optimize and you’re good. But if you forget that effect and use banners on your own sites, you’re sure to be in a mess…

I think Jakob Nielsen was the first one to address the effect of banner blindness. In its original form it said something like “people see it, but don’t recognize it” and it relates to banner on a website. Later the theory has evolved, by whom I’m not sure, to say something like: if you use banners that are “selling” something, and if the user recognizes this, the user will reduce the area of the screen to scan. The user will see less of your site’s content.

In essence… if you use banners that are “selling” something, if you use animations or sound, the more visible the banner is, the less visible is the rest of your page. It’s very difficult to see this effect if you don’t look for it, because the banners itself are clicked on. And it’s usually the most clicked on object on your site. But the fact is that those clicks are fewer than what you would otherwise receive…

I haven’t seen this hypothesis tested, but since it seems like a plausible theory, I decided to do my own test. At dnbnor.no/bedrift we did the following test, and found that the total number of click was reduced by 20% when using an animated banner on our site.

On test day #1 and #2 we tracked the following page, with four different objects. The top object is an animated banner with the three selling points numbered 1, 2 and 3 bouncing all over the place. The wanted effect was to attract the attention of the users:

On test day #3 and #4 we tracked this page. It was the same page, less the banner, with only three objects.

After correcting for general volume of traffic, we found that the total number of clicks on days with four objects, including the banner, was 20% less that the total number of clicks on the days with only three objects, and no banner. We believe that this is a practical case of banner blindness in a broader perspective than the original findings.

As a consequence…
We have reduced the use of banners to a minimum on our sites. In the case where we still use it, we try to minimize the appearance of the banner in such a way that it blends in with the general background and design of the site. The only reason we have found to use banners, is when we need more freedom and flexibility than our CMS system can give us, in cases like tests/quizzes, calculators or product navigators - when it comes to complicated products.

If you have any ideas or experience with banner blindness or the use of banners on your own sites, please feel free to share your wisdom, with comments to this article.