Essays & Articles
Banner Blindness, Human Cognition and Web Design
ORIGINALLY PUBLISHED IN INTERNETWORKING, MARCH 19991
Benway and Lane have studied “Banner Blindness” (ITG Newsletter, Dec. 1998: 1.3) – the fact that people tend to ignore those big, flashy, colorful banners at the top of web pages. This is pretty interesting stuff, for the entire reason they are so big and obnoxious is to attract attention, yet they fail.
Evidently nobody ever studied real users before – they simply assumed that big, colorful items were visible. This paper, shows once again the importance of observations over logic when it comes to predicting human behavior. People behave the way they behave, not the way our logical analyses and wishes would have them behave. People follow their interests, their needs, their customs. They are driven by curiosity, boredom, emotion. And the “they” refers to “we”: us.
Benway and Lane showed that if something is too obvious, too big, too powerful, it is overlooked (a point well known to Sherlock Holmes, by the way – or perhaps more precisely, known to Conan Doyle). What Benway and Lane found was this. Suppose a designer wants to make sure that people browsing a site can find “important-information.” The designer carefully makes the link to “important-information” big, bold, colorful. Nobody could possibly miss it. Lo and behold, Benway and Lane showed, it was the rare individual who noticed that this bold and salient banner contained the information being sought.
Why this behavior? Why does something so big and central escape notice? Well, there is seeing and then there is seeing. I suspect the banner was indeed noticed: I suspect that people carefully scrolled past it to get it out of the way so they could search the less bold, less colorful, less salient places. The fact that the searchers missed the supposedly salient information has nothing to do with big, colorful and salience: it has to do with schemas, frameworks, and expectations.
People are lawful. If they weren’t, we couldn’t have science of cognition. And as a cognitive scientist now applying what I have learned in a manner I once labeled “cognitive engineering,” I believe that behavior does have its discoverable causes.
Why banner blindness? Here is my theory. Note that this is still theory, informed, but nonetheless, still untested. If the banner blindness paper tells us anything, it is not to accept theory until it has been tested through observation. Still, the theory seems pertinent, relevant, and consistent with a large body of existing knowledge, so let me try it out on you. Once again I invoke schema theory.
I go to a web page to look for a link to “important-information.” My first task is a meta-search: I must discover where to look, where to focus my attention. Do I want a link? I must look in link-like places. Web pages are now a reasonably established genre, one in which lists of links have a standardized appearance. I therefore quickly scan the page for areas that look promising, for things that look like lists, often one item per line, often indented from the surrounding text. All other places are ignored, for they are unlikely to have lists of links. This behavior comes under many names in cognition: frames, schemas, scripts. But all point to the same point: people use their deep understanding of the situation or the genre to guide their search. When I lose my glasses, I do not look on the ceiling: I look on the floor, or the cabinet top. So too is it with links –- I do not “look” at that big banner glaring at me –- that doesn’t fit the link-positioning structure of my web-page schema. So I scroll past it and focus on regions that fit the schema.
When one puts critical links in a banner, that breaks with established convention, breaks the schema. The human attentional system is superb at homing in on areas of importance. This is human efficiency at its best, using very little information to decide where best to focus. Usually this behavioral pattern works well. But when something doesn’t fit the established schema, it is apt to be missed. This is the cause of numerous perceptual tricks and illusions, of numerous errors. I have used it as an explanation for some type of human error. Designers who break out of the established genre thereby do so at considerable risk: information outside of a person’s schema is apt to be overlooked.
Think about it. Do you read the headings in papers? Authors spend a lot of time formulating them, wording them, formatting them. I suspect most readers skip over them without even noticing them, just as we do not notice page numbers when captured by an absorbing novel. Our attention moves directly to the location most likely to contain the information of interest, most likely, that is, in terms of previously established convention, most likely in terms of the mental schema held by the reader.
So what do you do if you want something to be salient, to stick out? Morkes and Nielsen (1997; Nielsen, 1997) discovered that people read web pages very differently than printed pages: They start at the beginning, but are not apt to get very far. As a result, Morkes and Nielsen urged web designers to follow the “inverted pyramid” style of writing, in which key points and conclusions come first, less important matters and background material last. Recently, He, Gupta, White & Grudin (1998) found the same result with the viewing of online video presentations: they made the same “inverted pyramid” argument for the construction of on-line videos.
The same philosophy should apply with lists of links: put the important ones at the top, least important last.
Moral: if you want something to be salient, follow conventions. Violate the conceptual model, even if the violation seems perfectly sensible, and you are apt to discover that readers miss critical information.
In summary, the “Banner Blindness” findings are very important. They reaffirm the rule of consistency, coherence, and the following of established conventions. People use cognitive schemas to guide their attentional focus. So, web designers, be creative and original at your own peril. You would do better to follow a clear, coherent conceptual model, one consistent with established conventions and genres. Use the principle of primacy to control saliency: put the important things first, on top.
The Banner Blindness paper should awaken all of us to the importance of continual observation and test of even the most obvious of ideas –- including the ones I have just put forth in this commentary. The more we know about real human behavior, the better our designs will fit the people they are intended to serve
Together with Jakob Nielsen, Don Norman is a User Advocate, teaching companies how to develop human-centered product development. Norman is Prof. Emeritus at the University of California, San Diego, former VP of Advanced Technology at Apple, and author of “The Invisible Computer.” His company is the Nielsen Norman Group, at http://www.nngroup.com. Norman’s personal website is http://www.jnd.org.
Benway, J. P., & Lane, D. M. (1998). Banner Blindness: Web Searchers Often Miss “Obvious” Links. Internetworking: ITG Newsletter, 1.3 (Dec. 1998). http://www.internettg.org/newsletter/dec98/banner_blindness.html
He, L., Gupta, A., White, S. A. & Grudin, J. (1998). Corporate deployment of on-demand video: Usage benefits, and lesson. Technical report MSR-TR-98-62, Nov. 11, 1998. Microsoft Research.
Morkes, J. & Nielsen, J. (1997). Concise scannable and objective: How to write for the web. Sun Microsystems. http://www.useit.com/papers/webwriting/writing.html
Nielsen, J. (1998). Guidelines for multimedia on the web. Sun Microsystems. (Original no longer available, but see his 1995 paper with the same title (http://www.useit.com/alertbox/9512.html) and other papers on Nielsen’s website at http://www.useit.com)
1. Copyright © 1999 Donald A. Norman. All rights reserved. Originally published as Commentary: Banner Blindness, Human Cognition and Web Design in Internetworking, March 1999.http://www.internettg.org/newsletter/mar99/commentary.html