Eye-tracking: State-of-the-art usability software

The Usability Lab in Building 6 in Mountain View has acquired a new Tobii eye tracker. Eye-tracking refers to the techniques that allow measurement and recording of eye gaze. The Tobii unit permits unobtrusive recording of eye movements. The sensors are built into the base of the monitor, so there is no bulky headgear or glasses for the user test subject.

The Tobii unit looks just like a regular computer.

Eye-tracking adds another dimension to usability testing. Normally, we ask the test subject to use the "talk-aloud" method, during which a user is encouraged to say what he or she is thinking during the process. Additionally, we watch mouse movements, keyboard movements, and facial expressions as the user steps through tasks in the usability lab. Now that we have the eye-tracker, we have reduced the need for the user to speak aloud, since we know what they are seeing as they see it.

To try out the eye tracker, we invited a number of VeriSign people to come to the lab to go through a session in which they looked at 10 pictures and performed two simple tasks on a Web site. Although the session was meant to be fun, we discovered some interesting patterns. For example, there was a big difference between the eye gaze patterns of those from the technical organizations versus those from the marketing side of the house. (See the illustrations below.)

This shows the collective recorded gazes of our technical participants.
This shows the collective recorded gazes of our marketing participants.

Even the fun eye-tracking sessions helped validate some previous user testing. In earlier usability testing sessions, the team discovered that target users for some products did not veer off task once they had started on the enrollment path. The users told us that they did not want to look at anything on the side and would not look at distractions, although they were willing to learn more later. When we did eye tracking, we discovered that this was true. Our technical users stayed within the task, never looking deliberately at ads, sidebars, or color distractions. This has since been re-confirmed in user testing done since, using more rigorous methodology.

Full participation by Bill Analyst team makes for full understanding

The User Experience Design team recently conducted usability tests for Bill Analyst. The testing occurred in the Mountain View usability lab, but all of the stakeholders from the Seattle office participated fully in observing the usability tests, either remotely or by coming to Mountain View and watching from the observation room, as Greg Valazza did.

Watching all of the sessions is a great way to really understand the issues that the users face. The UXD team arranges to test enough users so that we are confident of our results; however, that does not mean that every user will have the identical same experience during user testing! Often a product manager or engineer will observe a single session of user tests and just happen to see the one user who confirms their own beliefs. It is then difficult for that observer to understand that 10 other users didn't have that experience at all.

Trust the reports that come from the usability tests—they are based on full observation of all test subjects, with analysis of their eye movements, mouse and keyboard actions, facial gestures, and conversations. Adequate number of test subjects ensures that the results are reproducible.