Usability test with eye-tracking glasses

Eye Driven Company
5 min readDec 20, 2022

Among various options on how to conduct usability studies, I will share with you one example with eye-tracking glasses from Pupil Labs. Our core idea behind this article is to introduce practical and simple tools, as well as possible measurements beyond well-known heat maps.

Many studies reveal that in response to increased cognitive load the number of fixations should rise, while, in contrast, the average fixation duration decreases.

Moreover, attentional load affects blink rates. Therefore, when we measure the number of fixations and blinks during the completion of the usability test, we could identify timeframes and corresponding parts of the webpage, that required the highest cognitive and attentional load. We are aiming to reduce the number of such peaks for observed values (blinks and fixations) and make the webpage user-friendly and convenient for the user.

Usability test with eye-tracking glasses Pupil Labs. Kaspersky careers webpage.
Picture 1

For instance, we will showcase to you one short usability test with eye-tracking glasses Pupil labs. We tested one webpage: Kaspersky Careers (https://careers.kaspersky.com/). To perform the experiment and track our eye movements we used the wearable eye-tracker Invisible by Pupil Labs. The picture above showcases the monitor overview.

During the performance of the experiment, we focused our attention on the pupillary response for a simple observation and research task: finding a predefined job offering on the Careers webpage. Among eye tracking data that we have analysed were the number of blinks and fixations, as well as gaze location. The initial idea was to support the data from users’ interviews on their perception of the webpage and strengthen it with the cognitive analysis of pupillary response. The base assumption is that cognitively demanding tasks or visual data affect the number of blinks and fixations, as well as their average duration.

The raw data was segregated into several parts corresponding with the elements of the webpage: 1) header -project start; 2) video about the company; 3) constantly moving line with numbers; 4) list of products and services; 5) courses; 6) various pages apart from job lists; 7) completion of the task, finding the desired job offering. Figures below (1,2,3,4) showcase the results — the number of blinks and fixations; the average duration of blinks and fixations.

Figure 1
Figure 2

Specifically, attention was paid to three segments on the graph, representing the peaks for observed values. Those timeframes correlate with visually demanding and destructive elements on the webpage, that are not directly connected with the topic — job offerings.

One example, even though the time spent observing the moving line in the middle of the webpage (see Picture 2 below) was relatively short the cumulative number of blinks increased considerably due to the nature of this object. Concurrently, the average duration of fixations and blinks decreased, indexing the increased cognitive load. Moving lines in both directions, text on top that is hard to read, and playing background video altogether add extra distraction and lead to attentional overload.

Picture 2
Figure 3
Figure 4

Lists of products and services (Picture 3) are also attention-demanding and result in clicking on one of those links, taking the user apart from the desired task — job offerings. These parts of the webpage lead to the highest blink rates as well as fixations. However, they do not provide any data about further careers but instead, sell products.

Picture 3

In summary, it would be reasonable to review the Kaspersky Careers webpage (https://careers.kaspersky.com/) and ease the user experience by reducing the number of information that is not directly connected with the topic of the webpage — job offerings and careers. Among the most cognitively demanding parts are: 1) the moving line with numbers and data about the company's performance; 2) the product line section with many links, that lead the user away from the initial page of interest.

Returning to the practical implications of the usability test with eye-tracking glasses. Based on our experience with many eye-trackers, Pupil labs is the only company that has free software and storage cloud, which is a great benefit for data sharing and analysis. In our case, we recorded the experiment on Invisible Eye-tracker Glasses with the connected phone and analysed the data on the laptop. It would be beneficial to measure pupil diameter as well. However, this version of the wearable eye-tracker doesn’t allow this option. Therefore, if you look for these measurements as well, another eye-tracker glasses (Core Pupil Labs) would benefit you more.

For more articles and information on usability studies, please follow our Facebook page.

All backup video files could be assessed using this link: https://drive.google.com/file/d/1kUANGEo5030xKHx6s9UsT95q_cGMYVH4/view?usp=sharing

Raw data and calculations: https://docs.google.com/spreadsheets/d/1hZrpBw1Jfm-U8Hp_SsRI9XUmtf7LJi3h/edit?usp=sharing&ouid=102158153009152126021&rtpof=true&sd=true

The study was performed, analysed and written by Julia Zhigulina, PhD scholar and researcher from Turku, Finland. The report showcases our own opinion and didn’t receive any funding from public or commercial companies.

--

--