The Usability Mindset, Part 4: "Self-Explanatory" is Not Good Enough

The crux of this post is so simple that I wrestled with whether it was too obvious to be a blog topic. But then I thought back to some of the heated discussions I've had with web designers over the years... and remembered that it's not so obvious after all.
The principle is the central tenet of Steve Krug's excellent usability primer, "Don't Make Me Think!" And it deserves repeating, as it's a core concept in web usability. Here it is:
A website's functions must be more than self-explanatory; they must be self-evident.
We're not talking semantics. There is a fundamental difference between "self-explanatory" and "self-evident". Understanding this distinction is critical to good web design.
"Self-explanatory" simply means that the website contains all the instructions needed for visitors to "figure out" how it works. "Self-evident" means it's so obvious that no instructions are necessary. (Think of a door. If it's properly designed, it needn't be labeled "Push" or "Pull". How it functions should be obvious.)
As site owners, we sometimes have to be reminded of this. We're so familiar with our website and its functions, that we find it hard to imagine anyone struggling with it. Also, we tend to think our site is just so important — so wonderful — that visitors will be willing to stick around and "figure it out".
They won't. People don't want to read instructions. They don't want to think. They just want to be able to perform their task, whether it's finding a specific piece of content, making a purchase, or whatever. On a good website:
- Labels, including navigation, will be unambiguous and avoid insider language.
- Expected content (Contact Us, About Us, etc.) will be placed where expected.
- All pages will contain clear, concise titles.
- Copy will be clear and concise, with good use of headlines, sub-heads and bullet points.
- Links will obviously be a links. There's no need to tell visitors to "click here".
- The search box will be right where expected and its function will be obvious.
- Shopping and check-out will be a seamless experience with no surprises or roadblocks.
If all this sounds obvious, good! You get it. At least in theory. In practice, only your customers can tell you whether your website passes the "don't make me think" test. So make sure you run some user testing sessions, centered around realistic tasks. They can be very eye-opening...

I often wish I had a "turn objectivity up to 11" knob when reviewing my own site as it really is all too easy to forget how to look at it from an outsider's perspective.
*Insert reminder for another round of usability tests.*
Thanks for the observations.
I should add that we sometimes tend to over-think usability testing. That is, we put it off because we aren't able to run "perfect" tests (with subjects that exactly match our target audience, with recording software, with a usability lab...)
When you're just testing for usability, any test is better than no test. Use friends and neighbors if you have to. You're bound to uncover at least a couple of important issues.
Testing for persuasion is another matter: test subjects should closely match your target group, as I'm sure you know.