Seeing isn’t believing

This clip from the people at Quirkology that illustrates the difference between vision and perception. It was mentioned during a discussion of eye-tracking technology in the IxDA forums, and it has an obvious parallel in what we’re now calling “banner blindness” – a phenomenon where web-savvy users unconsciously ignore content they believe to be banner advertisements. This phenomenon isn’t only bad for business, it’s also bad for designers whose editorial content somehow also triggers this behavior.

Were you fooled? I was, which probably means I need to work on my visual awareness if I want to improve as an interface designer. Boo.

I think the video raises some other interesting questions, especially when gathering feedback for a design. Though a user might not notice every change within the video, that doesn’t necessarily mean the changes don’t affect the user in some way. Oftentimes, subtle usability differences lead users to say “I don’t know why, but this is better.” Or, more commonly, the preference is attributed to a more obvious but unrelated factor.

During the early part of the video, I thought the production quality made it difficult to follow the magic trick. But was it the production quality or the variable environment? Having learned the video’s “real” trick, I now suspect it was the latter – the changing clothing and background colors disrupt their role as framing elements for the card trick, perhaps creating a subtle dissonance that hampers the overall information flow. I’d be curious to see the magic trick performed with and without the other changes, and test viewers’ reaction to the magic trick itself. Was it more interesting or entertaining? Easier to follow? This has a direct parallel in the navigation, support information, and other framing elements of a web interface.

This confusion also flows in the other direction – oftentimes, a production team spends a lot of time and energy deciding on a factor such as what color the magician’s second shirt should be, and then assume that the user is just as aware of this decision.

I often feel like this is happening when I view eye-tracking “heat maps” that cover a whole page. In those situations, it seems like the tester has said “look at this page and tell me what it’s about” so the user scans every part of the page. This behavior isn’t useful because we know that users don’t come to a page and think “alright, before I do anything, I should orient myself, so I don’t miss anything.” Instead, they satisfice – they scan a page until they see something that seems good enough to meet their needs, and stop (or click), regardless of how much time was spent by the copywriter wrangling over the footer links.

The video was introduced in the IxDA thread to support someone’s argument that vision doesn’t equal perception, which is certainly true, but it undermines the larger argument that this renders eye-tracking tests (and, from the tone of the thread) other empirical testing techniques useless when someone could trust a good designer and a quick conversation with some test users. It is exactly because users and designers are often biased and unaware of their actions that empirical testing techniques remain useful – they provide transcripts of unnoticed and unintended behaviors that point designers to subtle factors, problems, and solutions that would have otherwise been missed.