Eyetracking Web Usability

** May202012

Eyetracking Web Usability, by Jakob Nielsen and Kara Pernice is not a thrilling read (in that sense, you can judge this book by its cover). While there are some “take-aways” that web developers and user interface designers can apply to their own work, it may prove more valuable to those preparing for their own eyetracking studies. More than anything, it provoked the question: what are eyetracking studies good for?

The basics of eyetracking

The first chapter covers the basics of how eyetracking works and basic terminology. It also foreshadows a few fundamental questions about the worth of using eyetracking data to understand usability:

Why do users not see something? (p. 7)

Looking at and thinking about are certainly not always intertwined. (p. 9)

…it is not inherently good or bad for a certain design element to attract fixations or be ignored. Any judgement of the design implications of eyetracking data must include konwledge of users’ intent and the usability impact of their behaviors. (Under the heading “Are Looks Good or Bad?”, p. 10)

Vague and usupported conclusions

The various eyeplots and accompanying walkthroughs in this book often do little to support design recommendations or general conclusions; or, the conclusions are so general as to be unactionable. For example the plots on pages 54 through 64 show that users often exhibit vastly different patterns of fixations even while executing the same task. (As a side note, they also show a major usability flaw of this book: black numbers printed on very dark blue dots indicating each fixation, and the order in which they occurred. Pretty much illegible.) The conclusion the authors draw is:

In sum, a combination of layout and content almost always dictates what draws or repels users eyes. (p. 58)

Let that sink in.

One premise that chapter starts with is that “Web pages are not bank accounts: Full is not better”, and a heading on page 66 states “Light Pages Encourage Looking.” As a designer, I was salivating at the prospect of having cold hard evidence to throw my clients’ way when they wanted to cram more and more stuff on their homepages. Tragically, the conclusion a few pages later is that “density only determines 8% of the variability in how much people look at a page.” This contradiction goes back to the question of “Are looks god or bad?”. Simply analyzing the number of fixations on a page doesn’t tell you much, because you don’t know if the person was exploring with interest, or scanning in desperation.

Take-aways from eyetracking

That is not to say that Eyetracking Web Usability is devoid of useful conclusions that do seem well-supported. There are many that website creators should be aware of (though some are already well-established best practice), including:

  • “Banner blindness” can cause users to overloook the very elements you sought to highlight (p. 75).
  • Place the most important and relevant words at the beginning of a headline or or label (p. 391)
  • Images with high contrast and low complexity invite the most looks (p. 204)
  • In photos (as in real life) people have a tendency to look at what others are looking at (p. 264)
  • Don’t put columns of form fields side by side (p. 193)


This book is certainly useful for someone undertaking her own eyetracking research: covering the set-up, potential pitfalls, and terminology. It also shows though, that eyetracking is not the first place to turn for analyzing the usability of a website. When web analytics, customer complaints, or other forms of usability testing reveal a specific problem with a site, eyetracking may be the right tool to discover the problem and look for solutions. For the most part, though, it’s simply too difficult to abstract away from specific users tackling specific tasks on specific websites to draw useful, general conclusions from eyetracking data.

Respond to this review

Author(s)Jakob Nielson, Kara Pernice
PublisherNew Riders Press

No comments yet…Be the first!

Comments are closed for this article.