Eye Miles: Measuring Mental Effort and Design Quality with Eye-Tracking

While eye-tracking is incredibly useful for understanding how humans interact with computers, and websites in particular, it's something of a holy grail to be able to instantly interpret from eye tracking data whether a design is good or bad.

There's a solid amount of prior art here, but no widely practiced & easily obtainable top level metrics. Check out Table 1 from a HCI2007 paper by Ehmke & Wilson titled "Identifying Web Usability Problems from Eye-Tracking Data" for a complete review of metrics considered.

While analyzing the data for the Scrutinizer Click Fu video for a study we did on our Tobii eye-tracker, we computed mouse miles for a couple participants. Here's what we came up with:

This is the distance the eye traveled during a Question/Answer study in which users were asked to choose a search result that had or led to an answer to the question provided.

This "eye miles" metric in concept, if not name, seems to have had it's first appearance in 2002: Goldberg, J. H., Stimson, M. J., Lewenstein, M., Scott, N., and Wichansky, A. M. Eye tracking in web search tasks: Design implications. In Proceedings of the Eye tracking research and applications symposium (ETRA 2002). ACM Press, New York, NY, 2002, 51-58. . Personally, I released a analytics system for measuring mousemiles in 2001. A more robust metric would, to carry the analogy further, distinguish between highway and city miles. For eye tracking, where major and minor saccades (eg. short and long) indicate very different underlying cognitive functions, this might be especially informative.

We'll keep on searching for the best diagnostic metrics from eye-tracking. Perhaps some of them will apply to user activity with our gaze simulating web browser as well?