The people using your website can show you where they get stuck better than you can. In my original article I described Usabilla: you gave visitors a task, such as showing where they would click to sign up. Their clicks and comments made the problem visible.
Test an action, not just an opinion
Asking 'Do you like this page?' gives little direction. Try 'Where would you click to make an appointment?' or 'Can you find the delivery cost?' Let the visitor act first, then ask what they expected. If several people click the same wrong place, the page's labels or structure may be misleading.
You do not need a large research program to start. Pick one important task and see whether people can finish it without help. Note where they hesitate and the words they use. Make one focused change and test again. This grounds design choices in observed behavior.
The tool changed
SurveyMonkey describes how Usabilla was integrated into GetFeedback after acquisition. The old free accounts and prices in my article are therefore not current information. The measurement principle remains: combine what people do with what they say about it. An opinion form alone often misses the real obstacle.