blog
Eye tracking finds your blind spots - but it’s all in how you use it
April 21st, 2008 by JoEye tracking has been around since the 1990s, but it holds a science fiction-like fascination for many people (it certainly does for me). The concept is incredibly cool; the results are striking and visual; and the studies make clever use of technology. The idea behind eye tracking as part of usability testing is to determine where users actually look when they use a website. Using specially designed technology (sometimes mounted on a helmet worn by the user, sometimes embedded in the computer screen), testers can produce a “heat map” showing where the user’s eyes travel, and how much time they spend there. Just as telling, the heat map leaves dark any area that the user doesn’t even see; if these blind spots contain information that you need your users to see, it may be time to rethink your design. These are the sort of concrete-looking test results that anyone would love to plug into a PowerPoint presentation; they appear to speak for themselves.
A sample eye tracking “heat map” (www.useit.com)
Like any usability test, however, eye tracking doesn’t tell you much on its own; it is most effective when the tester asks good questions and interprets the results in context. Task-oriented tests are especially helpful because the heat map suggests how the user went about completing the task, and may reveal unexpected barriers. For example, one study asked participants to find the current U.S. population on the Census Bureau website. Though the number appeared in large text on the homepage, only about 14% of participants saw and recognized it, due to design choices on the page - perhaps because the information was formatted like an ad.
In many cases, eye tracking studies confirm what we’ve always known or believed about good web design principles: users scan web pages rather than reading; key words and bullet lists get the most attention; most people look for search or navigation first; hardly anyone looks at ads.
The studies also suggest new practices to make sure that users find what they’re looking for. For example:
- Usability expert Jakob Nielsen has found that most users tend to scan pages in an F-shaped pattern: first they scan across the top of the page, then scan across further down, then scan down the left side of the page. The first part of each line gets the most attention. This pattern varies depending on the user and the site, but holds true for a range of tests: the top and left of the page get plenty of attention, while the rest of the page gets a briefer scan.
- Images relevant to the page content attract the user’s attention, while images that are chiefly decorative are ignored. Images of people’s faces are especially good at drawing user’s eyes. (The implication is that extraneous “pretty pictures” should be avoided, but from a design standpoint, I wonder if that’s true. I would guess that peripheral vision plays a strong role in aesthetics, and I wonder if users are enjoying these decorative images on a sub-conscious level, without looking directly at them. Sounds like a subject for another test, if one hasn’t already been done.)
- Ads, especially overly styled/highly graphical ads, are rarely seen by users. Text ads are much more likely to be seen and receive eye focus from the user, especially when positioned close to relevant content.
- Text size affects how closely people read. Large text encourages quick scanning, which helps the user find things quickly, while small text is likely to be read more thoroughly and take up more time - if the user is interested enough to read it at all.
- Concise writing and liberal use of whitespace mean that users absorb and remember more information in less time.
While these tips are helpful for site and content design, they lead advertisers into a quandry. The data shows clearly that web users have been trained to ignore ads and focus on navigation and content; this behavior is sometimes called “banner blindness.” Does that mean you should place ads in often-viewed areas, like the left sidebar or the main content space? Probably not; users may get annoyed if ads appear in obtrusive places. Similarly, large colorful ads that stand out may be seen by the user, only to send them straight to the “back” button. This is a good example of why eye tracking data should always be taken in context with the test and the user’s emotional experience. Some studies suggest that unobtrusive, text-based ads placed near relevant content (a la Google) are more likely to be seen and elicit positive emotions from the user.
Ultimately, eye tracking studies are most useful as an integrated part of usability testing, where their role is carefully considered and the results are interpreted in context. They can help identify a design’s strengths and weaknesses for particular tasks, but beware of using them to make massive changes without thinking through the implications. As with so many tools, especially those with such a high shininess factor, use with care.
