I’ve lost track of the number of times I’ve heard someone in the web industry utter the words, “Well, a user would never do that.” Or “What do you mean? This makes perfect sense, anyone could figure it out, I don’t think it will be a problem.”
These statements are problematic for a few reasons. First of all, “users” are people. And people are weird, wonderful and come in a variety of flavors, shapes, and sizes. No two are alike. It's a pretty incredible statement to say "someone would never" do something. If you’ve ever had to walk your elderly neighbor through using a website, you know they do strange things. They use the mouse weird. They don’t use the same (or any) keyboard shortcuts. They open a new browser window for everything and they minimize everything (their dock bar, don’t even get me started).
The beautiful thing and frustrating thing about building things for people is, well, people. They don’t use your thing the way you wanted them to use it, they can’t find what they’re looking for, they can’t figure something out that seems so obvious! But at the same time, they find shortcuts and workarounds and new ways of using your thing you never even imagined. But you’ll never know any of this unless you actually talk to them and watch what they do.
When I was a Flash developer (it was a web animation platform, kids), my testing process before sending it off to the QA team was to click like a bat out of hell all over the page. At complete random, clicking as fast as I could click, dragging, banging on keys; I probably looked like I’d finally gone over the edge and needed to take one of those month long sabbaticals I kept hearing about. But I did that because I’d seen people do that. They would get frustrated when something was laggy, or they accidentally clicked on something, and in a panic, start clicking all over the place to stop what was happening. Was it rational? No. Would you uncover this bug if you only ever used the app as it’s “supposed” to be used? No. Was it understandable? Yes.
To assume that people are perfectly rational, come to the same conclusions that you do, want the same things, have the same goals, and will assuredly read the entirety of the homepage top to bottom, in order, is out of touch with reality.
Yet these are all very real and easy assumptions to make; after all, it’s much harder to design something when you assume the audience is nothing like you. The best way to see this in action is by talking to people, and by watching them fumble through your perfectly rational design. When we get feedback on products there is always a moment of surprise that uncovers something useful, and we never would have seen it if we hadn’t gone out to ask.