Don’t Look Now

In preparation for some work I’ll be doing this summer, I read Nielsen and Pernice’s Eyetracking Web Usability over the past few days. The book reports on the results of a massive eyetracking study that they performed to analyze how people use the web. My primary interest, actually, is in their methodology, which is not part of the book but is available free online as a separate report, “Eyetracking Methodology: 65 Guidelines for How to Conduct and Evaluate Usability Studies Using Eyetracking. But I thought I would start with the publication of the results, both to see what they had to say and as a bit of a crash course in web usability, which I know woefully little about.
Being a bit of a novice there, I thought the book was quite good. As someone with quantitative leanings, it was compelling to have design principles interspersed with heat maps or gaze charts illustrating the data they used to derive their principles. Much of it was common sense; I was not shocked to learn that most of us don’t even notice banner ads or aspects of design that look like banner ads anymore. But seeing how dramatically we ignore them was fascinating. The short section of the book about differences in what men and women attend to in images was also quite fun – as was the fact that everyone likes looking at dog genitals!
There were a few places where I felt that the authors were stressing their assumption of the purpose of a website or the intended audience too strongly, and critiquing sites for not maximizing their usability for the specific tasks being tested. Particularly when measuring how quickly a fact can be found on a fact-filled page, not everything can be the most prominent piece of information. But if the examples are taken as illustrations of what you should or should not do in presenting what the site authors perceive as the most important information, the principles are valuable.
One thing I am curious about is the focus throughout the book on not making users read – to the extent that textual precision should be sacrificed in some cases in order to save the user reading a single extra word. The specific example I am thinking of involved a webpage with a variety of bolded headings, one of which was a two word phrase. The contention was that the adjective within the phrase should be omitted since users might reach that adjective, not have it be specifically the word they were looking for, and move on before seeing that the second word of the phrase matched their goal.
Now, partially this goes back to my previous point about the intended users of the site, and their expected goals and familiarity with the language of the site. But as an educator, it troubles me to see evidence that if the information someone is looking for is not apparently within the first word of a heading, the instinct to move on instead of continuing to read is so strong. I can concede that the web is filled with so many competing information sources that sites must design themselves for easy perusal and navigation. But I had no idea that the pressures against reading text were so strong. It is distressing.