What is Eye-Tracking Good For?
Our recent tests showed us that eye tracking does some things well—and others not at all.
Eye trackers bounce infrared light off a user’s eyes and follow the reflections to determine where the eyes are looking. They make it easy to collect specific data on user behavior, but interpreting the data can be an issue.
Eye Trackers Can…
- Tell whether users are even looking at the screen. Without an eye tracker, it’s difficult to determine exactly where users are looking. But with the eye tracker, observers can tell whether users are looking at anything while they’re waiting for a web page to load.
- Tell whether users are reading or scanning. It’s easy to differentiate reading—a user’s orderly fixation on word clusters—from scanning for particular words or phrases.
- Learn the relative intensity of a user’s attention to various parts of a web page. By dividing the screen into distinct content areas, such as the left panel or banner ads, we can determine when a user gazes at each area, and for how long.
- Determine whether a user is searching for a specific item. Pupil diameter appears to increase when users are not sure what words they are looking for. Typically, this occurs on unfamiliar sites or pages with groups of general categories, such as search engine home pages.
- Compare user scan patterns. We planted ads of interest to some users (such as tango lessons) and compared their scan data to that of other users. Counting how long each user looked at each area, and in what order, let us compare user strategies. We found few differences among users’ overall scan patterns, so we believe we need to test only a few users to learn how scan strategies apply to a page’s design.
Eye Trackers Can’t…
- Let you know whether users actually “see” something. Users can aim their eyes at an area for a short time without any result they are aware of.
- Prove that users didn’t see something. Users acquire useful information through peripheral vision—such as the location of the scrollbar controls—even when the scan data shows that they haven’t focused directly on it.
- Determine why users are looking at something. Several users glanced at an animated baseball in an ad we created. We don’t know whether the baseball, the rotation, or the combination caused the brief fixation, or whether this had anything to do with the task (finding information about a baseball star). The test had no controls for any of these factors, so we can only say that the baseball attracted some users’ eyes.
- Test everybody. Eye trackers don’t work well under all conditions. We had problems measuring some users who wore eyeglasses or hard contacts, had small pupils or wandering eye, or who smiled frequently (this can cause them to squint).