User Experience Stack Exchange is a question and answer site for user experience researchers and experts. It only takes a minute to sign up.
Sign up to join this community
I've run a few online click tests to find out where people click on my design screenshot to figure out what type of button for a specific task works best. Turned out about half of the people clicked in totally different places.
Does anyone know research that indicates how well people perform on first time tasks? I can imagine lots of interfaces used daily by a wide range of people also perform bad if they're evaluated only by their first time use.
This is something that varies a lot between interfaces (and people to boot). It's not very useful to compare results from other people in this case; unlike Fitts' Law et al there's no artificial yet valid way to test this, thus no useful comparison.
Click tests are about user confusion, and a lot of things factor into that. The visual layout including clutter, color, size of objects, the clarity of your copy; words that are semantically or visually similar...it's not something you can isolate and determine how many incorrect clicks is bad, or how long is too long for time to correct click.
You're always going to have to interpret click test results manually; they tell you what, but not why. Maybe 10 people clicked area 4 instead of area 10; why? You'll have to look; does the button look similar to the intended target? Does the accident make sense? Sometimes people just accidentally click. You'll have to determine why people made the error and evaluate how serious the problem is based on how long it took and why they made mistakes.
For an example of why you can't easily compare or objectively evaluate results: Google's Home Page might do extremely well in click tasks and your app might do "poorly" but you can't just compare results like that. Your app might necessarily be much more complex than Google's Homepage. Two different click tests are very often Apples to Oranges comparisons. Of course cleaner designs result in better click tests results and of course you should strive for simplicity, but some things simply must be more complicated than others. No one argues Google Drive should be as simple as Google's homepage!
The only meaningful click test results are A/B test results on functionally similar designs (pretty much always your design A vs your design B). A/B tests are vital here because it will show you when confusion is resolved or mitigated. Take your first click test to find what the problem is (Oh, the cancel button looks like the default action!) and retest redesigns. You can also test multiple redesigns or test multiple designs from the get-go, if you have an idea of what might already be a problem. Just make sure you counterbalance A/B tests if you're showing multiple designs to the same participant.
