Heatmap overlayed over pump clips

Visual Attention and Product Choice

The Tram Depot POP
Point of purchase display used in the experiment


A real-life study was designed to test the relationship between visual attention generated by pump clips and product choice.


The real-life setting was The Tram Depot Cambridge where the pub operator wanted to understand the ‘bar-liveliness’ of the house beer pump clip. Respondents who described themselves as non-ale experts were asked to choose a beer whilst wearing eye-tracking glasses. After making a product choice respondents were asked to explain reasons for their choice.


Gaze data, product choices and qualitative feedback were analysed.

The heatmap shows aggregated visual attention from all respondents. Artigianale (a seasonal beer from Everards) generated most visual attention and was chosen by 70% of respondents. Overall, 90% of respondents chose the beer they looked at the most.

Visual attention to pump clips at the bar
Eye-tracking heatmap – visualisation of aggregated gaze data from all participants

As illustrated below, Artigianale received more than 4.7 seconds of attention per respondent, compared with the closest competitor Everards Tiger at under 3 seconds. Tram Light which was the focus of the study performed disappointingly being chosen by no respondents and generating only 2.6 seconds of attention per respondent.

Mean fixations time
Mean fixation duration in seconds

Respondents gave multiple reasons for choosing Artigianale, including a design that hinted at flavour characteristics through depiction of citrus fruit. 70% of respondents who preferred Artigianale mentioned lemon (or lime) present on the pump clip and described the product as being fruity and less bitter, more resembling cider.


The study suggests a very strong relationship between visual attention and choice. Brand name is predictive of neither visual attention nor choice. The product informativeness of pump clip design may serve to influence visual attention and choice. This study suggests the interesting prospect of being able to predict choice from visual attention.