In my last post I introduced the idea of the “shelf life” of an observation that comes from UX research–how long I can expect some behavioral finding to continue to emerge from subsequent studies. Users’ expectations and behaviors change over time as they adapt to new technology and designs, and after a while a certain finding from research may “expire,” or no longer be useful. In part 2, I present the important flip side: Certain types of observations do not expire.
Users Do NOT Change
I can hear you now: “Wait a minute, Devin! You said users are ever-changing with the break-neck speed of innovation and the shifting of interaction paradigms. Now you’re saying they don’t. What gives?”
Perhaps a better way of putting this is that human beings do not change, or at least not fast enough for us to worry about having to throw out all that we know about human cognition any time soon. When it comes to interface design, our understanding of human perception, memory, and learning gives us a pretty well-researched set of dos and don’ts long before we get to the point where we’re putting a design in front of users to test it out. In cases where we have inadvertently violated one of these rules, we will usually see this come out in the form of usability issues.
In doing some testing on the checkout for our mobile site, we discovered people were doing some unnecessary closing and opening of the form for adding credit card information, and getting really confused.
Here’s a snapshot of the page in question:
Users arrived at this page, hit “Enter a New Credit/Debit Card” and didn’t know what to do when they didn’t see where to put in their credit card number. It turns out they didn’t realize the form was already open by default and they could just go straight to putting in the information. Instead they (unknowingly) closed the form. Why do you think that was?
You might argue that it’s a bad idea to have a button with the user’s most likely current goal (entering credit card information) written on it NOT be the first thing they need to touch on this page—That in our effort to cut down on clicks and have the form opened automatically we end up shooting ourselves (and the user) in the foot a bit. I wouldn’t disagree with you there, but there’s something else major going on here.
Users had no clue that the button was tied to the fields below it. Did the word gestalt pop into your mind? In psychology, gestalt principles such as proximity, closure, continuation, and symmetry (see: Design Principles: Visual Perception and the Principles of Gestalt) describe visual attributes that cause us to perceive separate items in a unified way. In this case, users are missing a visual cue—sufficient proximity, or a bounding box of some sort, for example—to let them know that the button is related to the items below it. As a result, their expectations of how this page works end up misaligned with our intentions.
The new design of this page—which isn’t live yet, so I will refrain from showing it to you now—addresses this issue directly through the use of accordion-type visuals, and has tested much better so far.
The Point Being…
Gestalt principles explain the way people perceive things on a basic level. If I observe an issue during testing that seems to stem from two elements not being close enough to each other to be interpreted appropriately, I know that this is not the sort of problem that’s going to go away if I test it again in a few months. This finding has a long shelf life.
There are many great resources out there geared towards UX folk which summarize some of the most important things we know about human brains, and why they matter to design. A popular one, which covers gestalt principles and also happens to be one of my favorites, is Jeff Johnson’s Designing with the Mind in Mind: A Simple Guide to Understanding User Interface Design Rules. Also see Susan Weinschenk’s 100 Things Every Designer Needs to Know About People.
So…We know there are some things about humans that will not change and that we need to account for in designs. Research findings related to these things can be pretty sure to be reliable for a long time. We also know that behavioral findings from some study I did in, say, 2009 could be invalid today (and that would be totally fine and expected).
Now we get to the difficult stuff.
If my observation is not related to some low-level cognitive limitations or principles of perception, and is more of the “This interaction is new to me” variety…what do I do about it? How long will this interaction be “new?”
Are you growing weary of hamburger menu discussions? Too bad. Abstract menu icons and more await us in Part 3.