Skip to main content

Metrics Guide

Understanding our attention metrics and when to use them

Iwan Keymer avatar
Written by Iwan Keymer
Updated over a year ago

Once you have built your report you can select from several metrics designed to provide you with the most relevant data for each study.


Switching between metrics:

Select the pie-chart button in the top toolbar above the image tray to display the available metrics. Then, click the metric you desire for your analysis.


Probability of Perception (POP)

This metric estimates the percentage of viewers who have their attention drawn to an area of interest or hotspot at first glance. Our brains can process up to 5 distinct objects at any given time, therefore it is possible for multiple areas of interest or hotspots to achieve a score of 100%.

Use POP to determine the probability of your audience seeing what you want them to see at first glance. For example, if you have 3 strategically important elements in a banner ad, you can use POP to measure the likelihood of viewers seeing that ad within its context.


Share of Attention (SOA)

Share of Attention (SOA) metric estimates how much of the total attention within a scene is attributed to a given hotspot or area of interest (as a percentage). The sum of all elements in a scene should be 100%.

Use SoA when you want to compare the relative performance of multiple elements within a scene.


Attention Power Factor (APF)

The APF metric tells you how well your region is performing in relation to its size. If a drawn area is achieving a score above 1, the area is attracting more attention than the area it covers.

For example, if an area is taking up 20% of an image, and is getting 20% Share of Attention, the APF score will be 1x. However, if the area is only taking up 10% of an image, but scoring a 20% Share of Attention, the APF score will be 2x.

We know that a larger area is more likely to grab attention, but sometimes you may be limited to how large you can make your key areas. Using APF allows you to ensure your chosen area is grabbing the most attention it possibly can - optimally utilising it's space in the scene.


Local Attention Score (LAS)

This metric provides the average level of saliency for a given element i.e. on a scale from 0-100, how attention-grabbing is a certain region.

LAS represents the average heatmap score obtained for your area, after scoring each pixel on the page with a saliency score of 0 - 100. It is best used to quantify the strength of the heatmap for a region, with a higher LAS representing an overall stronger heatmap in your chosen region.


Need more information? Please contact [email protected] or any member of the Dragonfly AI team.

Did this answer your question?