OPINION11 July 2013

Can you see what eyes see?


Eye tracking technology is one of a number of new additions to the qual researcher’s toolkit, and Tpoll’s Steve Mellor has been putting it through its paces. He offers his verdict.


There’s an irresistible trend in qualitative research towards methods that use technology to observe or monitor people’s behaviour.

At its most extreme, scanning technology like functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) lets us understand how the brain works as we view media and content.

Technologies such as galvanic skin response (GSR) wrist bands can sense micro adjustments in the temperature and moisture content of the skin when exposed to stimulus.

All of these approaches remove the moderator from the mix and ‘let the body do the talking’. One such technology that I’ve started to use is eye tracking.

Eye tracking has been around for a while but has gradually moved out of the lab into more commercial use. 60% of use is still academic, but very recent advancements in the portability of the hardware and user friendliness of the software mean that qualitative researchers can, with training, use eye tracking within qualitative research.

I recently conducted an eye tracking study in three markets and literally carried the hardware with me from country to country. The Tobii X2 unit (pictured above) is as small as a Nintendo Wii bar and plugs into a laptop to power itself. It means that, for the first time, you can conveniently eye-track people in their own homes. This is perfect for website development and can alleviate many of the disadvantages of current interviewing practice.

‘Think aloud’ interviewing
Traditional depth interviews that are used to assess website usage tend to rely on the moderator instructing the respondent as follows: “As you look at the website please tell me what you are thinking.” This is a bad move.

Our short term memory can deal with seven elements at once and these elements tend to be used up during a depth interview at a computer, so asking someone to explain their actions as they move through a website – a further task – means they will stop behaving naturally.

“Eye tracking can supplement most studies that involve a detailed understanding of visual stimulus – website and app user experience, pilot testing of TV shows, TV advertising, outdoor and press, are the most common types of projects”

Steve Mellor

Steve Mellor

This ‘think aloud’ approach results in misleading output: problems with the website that may not exist; irrational mouse movements and clicks; a lack of respondent clarity and rational feedback.

But eye tracking allows the moderator to supplement what people are saying with some evidence about where they are looking – or, to be precise, where they have looked.

It’s standard practice in eye tracking for the interviewer to let the respondent look at the website, or app, without the moderator asking for feedback.

This means that the interviewer is no longer impacting on the respondent’s short term memory and the respondent views the stimulus naturally. The eye tracker tracks their gaze and the interviewer will record and follow their patterns – usually a sample of six depth interviews is reliable.

The interviewer can then do two things. They can start a verbal interview about the stimulus and explore it in the knowledge that they have some evidence about where people were looking – testing some considered hypotheses that emerge over the course of the interviews. Or they can conduct a retrospective think aloud (RTA).

An RTA is very useful for website or app development. It involves the interviewer replaying the video of the route the respondent took through the website or app, with gaze data overlaid onto the video and the respondent providing the narrative of what they were thinking.

In doing this, you’ll find that a respondent will recount their actions and explain their thoughts as they watch themselves clicking/tapping/looking around the content. Listening to this narrative can be hugely insightful because any preconceptions about what the gaze data has shown can be clarified.

This approach is very enlightening when compared to a classic ‘think aloud’ interview. The main advantage is more realistic, natural and reliable data to use for analysis.

For example, when a respondent browses an app or a website without the moderator nudging them for their feedback, the whole visit can be over in a matter of seconds, or minutes – as it would be in a ‘normal’ situation. The analysis of the eye tracking data and RTA approach is therefore grounded in the respondents’ normal, natural browsing behaviour.

The benefits of eye tracking
As qualitative researchers we are trained to value what is said and what is not said by a respondent, and what it means – and so we just need to keep that in mind when using eye tracking technology. We need to observe what they see, what they miss and the patterns that emerge.

These insights are captured by the technology and presented back to us in the form of gaze data that can produce heat maps, opacity maps and gaze plots. Video output can be edited to demonstrate typical events very effectively.

In recent projects, I’ve been able to supplement a qualitative report with additional insight regarding ‘dead areas’ of a website, cross selling opportunities being missed within ecommerce processes, specific colours and visuals monopolising gaze, and important buttons and messages being missed due to placement.

The qualitative researcher’s toolbox is growing yearly and we are now spoilt for choice when thinking about ways to interact with people – physically, virtually or in an ethnographic sense. Eye tracking can supplement most studies that involve a detailed understanding of visual stimulus – website and app user experience, pilot testing of TV shows, TV advertising, outdoor and press, are the most common types of projects.

It’s not something to enter into lightly. It takes time, patience, IT skills and a lot of thought. But it’s no longer expensive and can add real value to most studies when used thoughtfully.

Steve Mellor heads up the qualitative unit of Tpoll, a full service research agency

1 Comment

11 years ago

Nice piece Steve, very useful. I'd only add that the RTA is post-rationalisation. Very often we're not aware of why we fixate on items. There are 'rules' of perceptual hierarchy, for example, that Joe Public would not know so an explanation would be a contrivance. Similarly, attention can be driven, top-down, by things that meet our 'goals' but this is a non-conscious process (wine drinkers look more often at ads for wine is a simple example). There is no substitute for actual behaviour. It's in understanding the 'why' behind the 'what' that we need to be careful.

Like Report