Ask first, think later

Eye tracking is a powerful tool because it can identify behaviours that users cannot articulate, or may not even be aware of. However, one of the most common ways of implementing it in usability testing can render its results meaningless.

Research tells us that the way the eyes flit over a webpage reveals what the user is thinking. For example, repeated fixations on an object usually mean the user is trying to understand it or work out how it fits with the page or the task. Long dwell times could mean something is taking time for the user to understand. Organised scan paths show that the user’s attention is efficiently distributed, and therefore so is the information.

With this knowledge, a usability consultant can combine eye tracking data with other data – such as task completion time, error rate and backtracking rate – to answer a wide range of questions: Which of two designs is more efficient? Why are users not using a button? What elements in the page do users process first? Do users look at irrelevant parts of the page?

But one of the most common implementations of eye tracking in usability testing involves asking participants to think aloud, and explain their actions throughout the test. It provides the facilitator with an insight into the reasoning and motivation behind participants’ actions, but it can also affect the user’s attention and inflate task completion times. As a result, the eye tracking data loses its reliability.

The answer is to use retrospective think-aloud protocols, where users can explain their actions after they have completed them so that the eye tracking metrics are not biased. To get the best from eye tracking, test participants must be allowed to act first and think later.