What Do Your Customers See That You Don’t

By BenSteadman

12.04.2016

  • Eye tracking is a powerful tool for helping with conversion optimisation projects where you need quality usability data
  • Eye tracking is accurate and it is the only way to understand where a website user is looking and what they ignore in real time
  • It captures real-time user interaction with webpages and helps identify actual and potential money losing problems and how to fix them
  • You don’t need a large sample size to run eye tracking tests

Eye tracking is one of the most powerful tools in the conversion optimiser’s toolbox that supports other essential research, including polls, surveys, analytics, etc., that together provide a mass view of website user behaviour and feedback.

Eye tracking gathers real-time visual information on how visitors interact with websites, allowing you to review exactly where they looked on a webpage, when they looked at it, and for how long they looked at it. You can also see what was missed. So, from a very valuable usability perspective, we can see precisely how they engaged with it. No other research tool can show this. (And, where relevant, we can compare click, scroll and algorithmic models for the participant, so we can map those activities to see whether there is any correlation.)

How it works
The eye tracking system we use is based on infrared technology and cameras that capture high-frame-rate images of the test participant’s eyes and eye movement patterns as they view the webpage. Image processing software makes sense of these images, finding specific details in the participant’s eyes—it is even possible to measure eye dilation, which could be used to gauge excitement aroused by advertising they see on the webpage. The system then calculates the position of the eyes and gaze point.

Together with capturing accurate real-time images of a participant’s eye-interaction, we also video the session. And, we don’t just perform these tests on the desktop because we eye track on smartphones and tablets also—which is vital with the growth in mobile commerce and the fact that some websites are 50 per cent mobile!

This gaze plot shows fixations—the circles—and saccades—the lines connecting the circles. The numbers are the order of fixations and the larger the circumference of a circle, the longer the length of the participant’s fixations.

A quick look at the eye

The lens focuses on an object and the retina captures the image, resulting in chemical and electrical events that send nerve impulses to the brain’s visual cortex through the optic nerve.

To get the most out of eye tracking we need to have a basic appreciation of the eye and how it works when viewing a website. In the above diagram, note the fovea centralis in the centre of the retina. This captures direct line of sight images from the point of fixation (visual gazing on a single location), processing 2° (foveal vision) of our entire 180° field of vision. Surrounding the fovea is the parafovea (parafoveal belt) and beyond that the perifovea (perifoveal region). The fovea centralis, therefore, sees images in detail, and visual acuity lessens across the parafovea and perifovea. (This is what we generally understand as peripheral vision.)

Because they can’t look at two things at the same time, your visitors will use their parafoveal vision to locate the next most important item to look at, and the fovea will follow if there is something interesting. It is this movement that is of interest in learning how visitors view a website. We can use this data to understand and then guide us about how the visitor sees the website and the impact elements of each webpage for conversion. All of this is captured accurately by the eye tracking system.

How we test with eye tracking
The design of eye tracking sessions is informed by other research we conduct during our conversion optimisation process. When we test websites we look for blockages and any other issues that cause friction. The data we’ve gathered will highlight the problems. For example, there could be a problem with shopping basket drop offs at checkout and so we need to know whether they are seeing what they are supposed to/want to see, are they being distracted by something else, does the design and content of the website help them move through the process easily without effort, and so on.

So, we’ll devise a real time eye tracking test to see how visitors see and use these key areas of the website to identify the cause of the friction. For example, the test participant would be asked to buy a product and proceed to the checkout while we eye track them. The system allows us to see exactly what is happening by capturing their eye movements, facial expressions and interactions (clicks, scrolls, and comments). We can also replay the video recording to help answer any questions we’ve noted as they’ve used the website.

Work with smaller samples
For eye tracking we can work with between five and 10 participants (up to a maximum of 15), so it is very cost effective. There are very good reasons for only requiring such low numbers, as described by  Nielsen Norman Group (NNg).1

In summary, five people will find around 85 per cent of the problems, 10 about 95 per cent and 15 should find them all. Your first test participant will provide almost 33 per cent of what we need to know, the second user more and so on. However, there will be overlaps so you get to the point where you aren’t learning anything new. (Indeed, NNg suggests it’s better to spend the test budget on five participants and more tests.)  So the first five find 85 per cent of the problems, you fix those issues and then you retest with another five participants as a QA stage and to find the remaining 15 per cent. Those problems will be fixed and then a final round will mop up anything that remains.

What we learn from the data
We gain a lot from eye tracking sessions, including:

  • heatmaps of what the participants saw and didn’t see on a webpage
  • gaze plots of what they saw on the webpages and features of the website in the order in which they saw them.

All of the above are supported by an extensive list of time-stamped metrics. Using this information, we can breakdown the various webpages into areas of interest (AOI) and drill down into the specific data behind how these are seen and interacted with. The list includes:

  • Fixation count (number of fixations on each AOI)
  • Total visit duration (total time each AOI was viewed)
  • Average visit duration (average time of fixation per AOI)
  • Time to first fixation (time from when each participant first saw a specific stimulus until an AOI within the stimuli was fixated on for the first time)
  • Percentage fixated (percentage of participants that have fixated on a AOI)
  • Mouse click statistics (time from fixation to click, time to first mouse click, mouse click count)
  • First fixation duration

So, the eye-tracking system provides a wealth of useful information that we can use together with data we’ve extracted from other sources like Google Analytics and qualitative studies. Used together, we can build a very detailed picture of a website’s pros and cons, importantly we can identify where there are problems that are causing it to lose money and help fix them.

Eye tracking is, therefore, a very powerful real-time data collection tool. It helps gather valuable insights about user interaction on your website and so we use it throughout the conversion optimisation cycle to find the best tests and continually improve them. Used with other forms of conversion optimisation research, eye tracking will help with triangulating the problem and enabling you to create good hypotheses, leading to better treatments that convert more visitors into customers. 

Reference

1 J Nielsen, “Why You Only Need to Test with 5 Users”, Nielsen Norman Group (2000). https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/

 

BenSteadman
Hi, I’m Ben. I lead the analytics team here at Endless Gain. I’ve been analysing data and fixing analytics since 2001. Data plays such a vital part in today’s decision making world. The days of making business decisions based on opinions are disappearing fast and data is replacing it. Analytics is pivotal to this data led world.

Our Clients