Overview

You can demo the Busara Attention Measure here.

Brilliant minds, like Cicero, William Shakespeare and Daniel Kahneman, have long claimed that we can gain insight into the inner workings of a person’s mind by monitoring the behavior of their eyes.

There is now a wealth of empirical evidence in support of this claim.

Focusing on the context of consumer behavior, eye movements can, among other things, tell us how likely consumers are to buy different products, which attributes capture their attention and which attributes influence their final choice.

Based on this wealth of evidence, leading companies, such as Microsoft, L'Oréal and Proctor and Gamble, are increasingly using eye tracking technologies to better understand their consumers’ behavior.

At Busara, we are keen to apply eye tracking technologies to improve consumer protection, particularly the presentation of product information.

Based on the existing body of eye tracking research, we believe that eye tracking technologies can help answer many questions pertinent to consumer protection, including:

  • How much information do consumers process before choosing a product?
  • What is the relative weighting of different attributes when choosing a product?
  • Can we design alternative information layouts such that consumers process the minimum amount of information required to make an informed decision?

In pursuit of this aim, we have designed an online tool – Busara Attention Measure – which uses two techniques to proxy eye tracking – mouse tracking and keystrokes – to monitor how consumers browse product information and how this affects their final choice.

In this page, we provide a brief overview of Busara Attention Measure, the data it generates and the questions it can help answer.


Busara Attention Measure 

Busara Attention Measure approximates where on the screen users are attending to using two different techniques: mouse tracking and keystrokes.

Mouse tracking

With the mouse tracking technique, participants see a blurry screen with product terms and conditions underneath and can read the underlying text by moving the mouse across the respective sections of the screen (see Figure 1A and Figure 1B).

Figure 1A. Terms and conditions with mouse off-screen

Figure 1A. Terms and conditions with mouse
off-screen

Keystroke

Figure 1B. Terms and conditions with mouse on-screen and revealing text

Figure 1B. Terms and conditions with mouse on-screen and revealing text

With the keystrokes technique, consumers see two loan products, each of which consists of three attributes – loan amount, loan period and repayment amount – and must press three different keys – W, S and X, respectively – to reveal the value of each attribute (see figures 2A through 2D).

Figure 2A. All attributes hidden

Figure 2A. All attributes hidden

Figure 2B. Loan amount revealed by pressing W

Figure 2B. Loan amount revealed by pressing W

Figure 2C. Loan period revealed by pressing S

Figure 2C. Loan period revealed by pressing S

Figure 2D. Repayment amount revealed by pressing X

Figure 2D. Repayment amount revealed by pressing X

Each of these techniques – mouse tracking and keystrokes – provide a wealth information on how consumers browse product information and how this affects their final choice.


Data

The mouse tracking technique generates a heat map showing how long consumers looked at each section of the terms and conditions as measured by the position of their mouse (see Figure 3).

    Figure 3. Heat map for product terms and conditions generated by the mouse tracking technique

    Figure 3. Heat map for product terms and conditions generated by the mouse tracking technique

We can use this data to better understand several features of consumers’ browsing, including:

  • Do consumers browse all the required information?
  • Are consumers skipping over key information which is required to make an informed decision?
  • Can alternative information layouts nudge consumers to browse all the key information?

We can also use this data to better understand how the information presented affects consumers’ final choice. For example, by structuring the presentation of the terms and conditions such that different sections of the screen include different pieces of information, we can record how long consumers spend browsing each piece of information and estimate the effect each piece of information has on the probability that consumers choose the product. An example of how this would look is provided in Figure 4.

 Figure 4. Example screen structure such that different sections of the screen contain different pieces of information

 Figure 4. Example screen structure such that different sections of the screen contain different pieces of information

The keystrokes technique generates two sets of data: the total time spent looking at each attribute (Figure 5A) and the evolution of consumers’ browsing over time (Figure 5B).

Figure 5B is slightly complicated, so let’s walk through how to interpret it. On the x-axis, we record the length of time the consumer has been browsing. On the y-axis, we record the cumulative time spent browsing each attribute. So, looking at the lines in Figure 5B corresponding to amount (blue), period (black) and repayment amount (green), we can see that this consumer browsed as follows:

  • First, they looked at the amount (blue) for approximately one second;
  • Then, around one second in, they looked at period (black) for approximately two seconds;
  • Next, around three seconds in, they looked at repayment amount (green) for approximately two seconds;
  • Finally, around five seconds in, they looked at amount (blue) for a second time for approximately two seconds.
Figure 5A. Total time spent looking at each attribute

Figure 5A. Total time spent looking at each attribute

Figure 5B. Evolution of consumers’ browsing over time

Figure 5B. Evolution of consumers’ browsing over time


Key Questions & Further Extensions

Using the keystroke data, we can develop a better understanding of key features of consumers’ browsing, such as:

  • How much time do consumers spend browsing each attribute and does this align with what’s required to make an informed decision?
  • Do consumers browse all the key attributes required to make an informed decision and does this change with the number of attributes presented?
  • Can alternative information layouts nudge consumers to invest time in browsing each of the key attributes required to make an informed decision?

As with the mouse tracking data, we can also use the keystroke data to better understand how the information presented affects consumers’ final choice. For example, using the data in Figure 5B, we can estimate how the length of time browsing each attribute, the number of times browsing each attribute and the order in which consumers browse each attribute affects their final choice.

At Busara, we are continually iterating and testing our designs to improve them. Please do get in touch with us if you think this tool could be of use in your research or your work, or if you'd like to learn more about the future direction of this work.