I like to take a playful attitude towards things. I try out the concept in various elementary or even trivial contexts at first, and see what happens. This is particularly useful when the result disagrees with what I had initially expected for the concept, since it leads me to revise my intuitions about the concept. But when it conforms with my initial expectations, then it lends support to those initial intuitions, building up my understanding. Then I can move on to more substantive examples. Gradually, by exploring the concept in a variety of these more substantive contexts, one arrives at a deeper understanding of the concept.
So eventually, it becomes clear, and you know it.
Nevertheless, I believe also that ultimately, one is never fully certain that a concept is completely understood. Surely we all at times learn new things about concepts that we thought formerly to be completely understood. Perhaps we have unknowingly considered examples only of a certain type, and therefore did not fully appreciate the possibilities and perhaps inadvertantly built incorrect intuitions that way. And while this sometimes happens, fortunately, for most of us it is rare.
So press on when you feel that you understand the concept and some principal examples, but be willing to reconsider things from the beginning when strange new examples come to light.