It’s hard to imagine there being a “best” tool because everything is in the context of a specific problem to solve, but if I could only teach someone one tool and the concept behind it, I’d have to go with control charts. In one tool you are able to assess the atability of the process and inherent variation, and compare the process before and after. Further, that stability forms the basis for so many other tools.
I agree with Joel the idea of a “best tool” really doesn’t even make any sense. A hammer is great for driving nails. You can use it with screws too but it is the wrong tool.
As much as I hate to disagree with Joel If I had one tool it would be MSA which is actually a few different tools. I really have no idea what my dat is telling me until I understand the quality of the data I am dealing with. My experience with engineers is that they are good decision makers with good analytical skills. If you give them bad data it doesn’t matter how good you are. The quality of the decision is a crap shoot based on what the measurement system was doing that day.