Banner blindness

Many have heard about the “banner blindness effect”. But such a large number of various rumors and interpretations are wound around this term that often under this term people understand completely opposite things. Without claiming to be absolute truth, I will try to describe my opinion on this subject, based on several years of experience in IT tracking research.

To begin with, we need to remember how vision works in people. Vision can be divided into lateral and central. Lateral vision is approximately 190 degrees horizontally and 135 vertically. But the central one is only 1-2 degrees. At the same time, we sharply distinguish exactly what the central (it is also foveal) vision is aimed at. Try to concentrate on a word on this page and see what you can read without moving your eyes. You will find that you can distinguish between text in a spot about one and a half to two centimeters in size, and everything else begins to blur. This small spot is your central vision. And it is precisely his movement on the page that is tracked during usability research using eye-tracking technology.

Now let's move on to banner blindness. People do not like advertising, and when they come to some page of a site for information, they try to find and perceive this information, but do not pay attention to advertising. The question arises: how to separate the necessary information from advertising. And here the previous experience and lateral vision come to the aid of the site visitor. Having noticed with a side vision a colored spot of approximately the same form as a banner and in the place where the banners meet, the visitor subconsciously classifies this spot as an advertisement and does not direct his foveal vision in that direction.

Actually this very effect - the visitor’s non-viewing of banners - is called banner blindness. This effect has one unpleasant consequence, which we often encounter during our research: if some elements of the navigation of your site are similar to advertising, then visitors will not notice them either.

Since the effect of banner blindness is statistical, it is not always easy to calculate. If at all no one paid attention to the elements prone to this effect, then this would be easily tracked in the statistics on the complete absence of transitions, but the low number of transitions usually falls out of sight (a pun, however). Moreover, if this quantity has always been low since the site was launched ...

For example, this is how the heatmap for viewing one of the sites looks like:



At first glance, nothing criminal is imperceptible. But if you select the areas of interest (the top block with announcements, the menu, the color announcement in the middle of the page and the beginning of the content) and see the perception of these blocks in numbers, you can see the following picture:

	Top	Menu	Middle	Content
% of page area	1.4	1.1	2.1	2.9
Time to 1st fix	4.5	6.2	1.8	5.7
Previous commits	6	10	2	eleven
Number of fixations	4	9	6	17
Commit duration	1.2	2.1	2.1	4.4
Number of views	3	4	4	5
View duration	1.2	2.3	2.1	4.6

You may notice that despite the fact that the Top and Middle areas, due to their location in the center of the page, are noticed quite quickly (the parameters “Time to 1st fix” and “number of previous commits”), they attract much less attention. This is clearly noticeable when comparing the adjacent areas: “Top” with “Menu”, and “Middle” with “Content”.

Most often, banner blindness besides, in fact, banners, affects various announcements and advertisements of internal pages. Moreover, during our research, the difference in the perception of such elements by regular visitors of the resource and newcomers is quite clearly noticeable. For example, when researching one of the major news sites, the following results were obtained:

		Average		Median
		Announcement	Menu	Announcement	Menu
The number of respondents who noticed the area (people)	Readers	7	7	7	7
	Newbies	5	9	5	9
Number of fixations	Readers	3.4	3.0	2.5	2.0
	Newbies	0.9	8.1	0.5	6.0
Time to first fix (s)	Readers	7.7	7.3	7.5	8.2
	Newbies	12.5	2.9	12.3	1.1
Duration of fixation (s)	Readers	1.6	1.1	1.4	0.5
	Newbies	0.3	3.0	0.1	2.4
Number of previous commits	Readers	26.7	15.0	16.0	13.5
	Newbies	61.2	9.5	26.5	5.0
Time from commit to click (s)	Readers	2.8	35.5	3.5	35.5
	Newbies	-	18.9	-	12.1

You can see that regular readers perceive the block with announcements almost exactly the same as the menu, but new visitors ignore these announcements.

From the foregoing, the following conclusions can be drawn:

• It’s better to use large elements on banners so that they differ in lateral vision - there’s a higher chance that they will somehow attract the conscious attention of the visitor
• Everything that may be like advertising when designing your site should be avoided.
• It is better to make internal announcements textual and place them in the text of the articles or immediately after the article - so that the visitor read them “by inertia”
• Do not neglect testing your sites. Even if you don’t have the money or the desire to order testing “on the side”, spend half an hour and ask someone outside (not connected with you or your company, for example, a visitor to an Internet cafe) to find some information on your site or perform some other actions. Even such simple observations are enough to catch the most serious usability errors that you simply don’t notice, because your eyes are blurred ...