A very simple way to illustrate what kind of additional information you get in a usability study by using eye tracking is to watch the video below video (http://bit.ly/gazemouse) that I just uploaded on YouTube. In this video you observe a person completing a task on a website (www.spotify.com) the task was to sign up for the paid service (called Premium). First you will see the person completing the task without seeing the persons eye movements, only the mouse movements can be seen. In the second video you can also see the eye movements in addition to the mouse movements.
As you can see in the video the eye tracking data does provide you with more information about the person’s behavior while completing this task. In this study we found a clear pattern: even though several participants in the study did click on the right button (Premium) they hesitated and compared all the other available sign-up options (4) on the web page. Based on this insight the web page has now been optimized and only has one sign-up option (see www.spotify.com), instead of four on the old web page.
In the same study we also compared the value of different cues when using the retrospective think (RTA) aloud method in web usability testing: an un-cued RTA, a video cued RTA, a gaze plot cued RTA, and a gaze video cued RTA. The findings suggest that using a gaze plot or gaze video cue stimulates participants to produce the highest number of words and comments, and mention more usability problems during the interview. Read the paper here (pdf).
Returning to the Premium problem described above, none of the participants in the un-cued group mentioned anything about this problem during the interview. They just said the web page was easy to use and that they had no comments about the first page. In other words without the additional information that eye tracking provided we would not have spotted this issue on the Spotify web page.