Usability Analysis with EyeTracker

I just got back from a live demo of the Tobii 1750 Eyetracker at the Usability Labs here at the University of Minnesota. EyeTracker is basically a fancy monitor which can track and position the movement of your eyes while using any type of computer application. This data can be used to determine how users navigate through your software, what is being missed, and what is placed well.

The software demo went a little something like this: A man stares at a computer screen while an operator uses a basic program to calibrate the scanner. The calibration is then tested on an image from “Where’s Waldo” while the operator asks the man to look at different activities going on in the picture. Once the calibration is consistent, the real analysis begins. The user is directed to U of M Onestop where he begins to surf and browse naturally.

Now this is the cool part: On the large 50″ plasma display in the conference room all of the attendees watched a little blue ball skitter and track across the web page, stopping briefly and growing in size as the man fixated on blocks of information within the page. The little blue ball grew in size when the user stopped to look at something, creating a fingerprint on the fixated block of data. This information was sequenced and compiled into a visual information map, as seen below.

The usability group was really emphasizing the fact that people do not want to think when using web applications and web sites. Things should work how they expect, work the first time, and work the same as similar products. I couldn’t agree more.

When doing usability testing for Plogger, our group used a lo-fi method of achieving the same type of results. The method is known as a “cognitive walkthrough”, and it is pretty much what it sounds like. We sat the user down in front of a computer screen and asked them to “think out loud” while trying to accomplish several different tasks within the application. We could then keep track of mental stumbling blocks in the software, things which appeared to be obvious to us but actually very confusing to the user. This helped us to refine our prototype before going all out on our implementation.

Simple usability testing like this can save us from having to make assumptions about our users. Usability testing will highlight areas which may need added documentation or training. The important thing to remember is that you must employ usability testing early on in the development process, with lo-fi prototypes if possible. Even if you can’t employ expensive techniques such as the Eyetracker, simple methods will still give you enormously helpful results. The earlier you catch a usability problem, the more time and money you will save in your overall project.