Predicting what users will notice first:  3M’s Visual Attention Service

from John Wooden, Director of Usability Services

on February 08, 2011

3 comments

Last week I saw a demo of 3M’s Visual Attention Service (VAS), a web-based application that applies an algorithm to predict with greater than 85% accuracy what users will focus on in the first 3-5 seconds of exposure to images. The predictions are based on 3M’s 30 years of research into the science of vision and take into account factors such as color, contrast, edges, size, and the presence of human faces.

If you design in-store displays, outdoor signage, or websites, VAS is a useful tool to add to your kit. Say you have put together three alternative design mockups for a website home page, and the home page needs to convey three key messages. Because website visitors tend to be impatient and make decisions very quickly, it’s important to know what they are likely to notice within 3 to 5 seconds. With VAS you can highlight three key areas of each home page design alternative and determine how successful each one is in quickly attracting user attention. You can then tweak your design and see what difference that makes.

In the examples below from Target’s home page (image taken on February 8), you can see what VAS predicts users will quickly notice.

So how is VAS different from eye tracking software?

The big difference is that eye tracking requires human subjects to determine what people actually look at. VAS makes a prediction based on an algorithm. So the main benefits relative to eye tracking are speed and low cost. (VAS is not free though. You will likely want to buy credits if you are going to be an occasional user or a subscription if you plan to use it frequently.)

The limitations of VAS are pretty obvious. It tells you what people are likely to focus on in the first 3 to 5 seconds, and that’s all. It does not tell you whether users will like and respond to those messages or images, whether users will stay on your site, or whether your site is easy to use. It doesn’t give you a lot of insight into how your users will think or behave. For that, usability testing is still the best option. In this way, VAS is a useful tool in the same way that spellcheckers and color contrast analyzers are useful tools. It has real value during the design process, but it can only tell you so much. It doesn’t replace getting feedback directly from real users.

Let me give you one quick example that helps illustrate this point. Humans are wired to focus on other human faces – we are drawn to look at them. (Think of LinkedIn profiles – you’re more likely to look at the ones with photos.) In a similar way, we are attracted to look at anything red. The example from Target demonstrates this well. But using red or including photos with faces is no guarantee of sustained attention and interest. I ran the home page of another site that we usability tested recently (which I can’t show you) through VAS and it predicted what I expected – that the two parts of the page with images of faces would be more likely to attract attention in the first few seconds than the other parts of the page. But regardless of whether our usability test participants noticed these images in the first few seconds, their comments and behavior indicated that they were not very interested in the sections associated with these images.

Despite our inherent interest in human faces and attraction to red, it goes without saying that the answer to drawing a user’s attention is not always going to be to display a photo of a woman in red. Users are sophisticated, and though they might initially notice certain images, if they associate them with advertising, or fluff, or if they consider them to be obvious stock photos, they are likely to actively avoid them. Large text, subtle color, and good contrast can also effectively attract attention and convey key messages – it doesn’t always have to be a face photo.

Give VAS a try and see what you think. Just bear in mind what precisely it is telling you, and what it isn’t telling you.

Comments

I’m out of legaue here. Too much brain power on display!

by Ollie
on May 01, 2011 - 6:43

That’s fascinating!

by Steve Nguyen
on May 26, 2011 - 8:25

De Pleasure naar uw blog te begrijpen. De bovenstaande artikelen is vrij uitzonderlijk, en ik heb echt genoten van het lezen van uw blog en punten die u geuit. Ik hou echt van om terug te verschijnen op een typisch basis, een veel meer in het onderwerp te plaatsen. Bedankt voor het delen van ... blijven schrijven!

by appliance repair
on March 21, 2012 - 12:48

Add Comment