Testing assumptions can be challenging; it can be difficult to put some of our strongly held beliefs under a microscope to evaluate and question. However, being able to discover a wrongly held belief and subsequently improve design or user experience because of this new information is extremely valuable. 

Eyetracking data visualization

Recently we tested our and our clients’ assumptions. We conducted a study on how consumers develop their perceptions: what features do they look at and evaluate as they consider a product on a given dimension, such as size or functionality. In this case our client was interested in:

• perceive the width of a product.
• how consumers perceive the height of a product.
• what aspects are looked at as a judgment is made.

Participants were asked to evaluate a hamburger on the dimension of” height”. The assumptions were that users would scan the hamburger from top to bottom; that makes intuitive sense. If you want to determine how tall something is, wouldn’t you look from the bottom to the top and evaluate the distance? A fair assumption, but in this case, incorrect.
*We simulated the desktop eye-tracking portion of the study with the hamburger images.

 
Here is a heat map from our hamburger study; it is very similar to the output from the actual study. The areas in red indicate a high concentration of viewing attention; this is what users spent most of the time looking at. The areas in yellow, or those that have no color, were looked at less. This heat map clearly suggests users spent their time evaluating the center of the hamburger, not the top and bottom.
A closer look shows that there are a few hot spots near the top and bottom of the bun, but they are small and likely due to viewing attention of only a small portion of participants. This is confirmed by a different type of heat map. This output plots the number of participants who looked at a particular element. Notice the concentration of participants who looked at the center of the burger, rather than the top or bottom.

 These surprising results prompted us to start asking questions; we followed up with participants and asked them why they were looking at the center of the burger. They explained that without a reference to compare the height of one to another, they instead evaluated each component – the height of the patty, the thickness of the tomato and layer of lettuce, the scope of each bun half. A fat patty would suggest a taller burger.
A few caveats and new questions (good research always raises new questions). These are considerations to investigate further; some also underscore the importance of thorough pilot testing (another topic for another time).
1). Did the wording of the question somehow inadvertently guide them to evaluate the center of the burger more than the overall height?
2). Should the image be presented differently so that the center of the burger isn’t in the center of the screen (thus automatically grabbing attention)?
3). How would viewing patterns differ if burgers were presented side-by-side and how could that output be interpreted?
4). If we had asked participants, without eye tracking, how they evaluate the height of a hamburger, how would they have responded differently? That is, were they aware of their own viewing patterns?

What do think? Have you ever been challenged to test an assumption, what were the outcomes?

This entry was posted on Monday, May 17th, 2010 at 2:09 pm and is filed under Marketing. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.