What is the best way to identify the root cause (major cause) of variation in the process creating a defined problem?
Some use Pareto Diagram, others using a questioning method (ask "why" 5-7 times), others use brainstorming, voting, etc.
Mike over at Got Boondoggle offers this recommendation:
Because all these methods are widely used to help identify or organize potential root causes or "the most likely cause." However, none of them actually can identify the real root cause.
The only way and best way is by testing and verifying. Without data to prove that it is a root cause, you only have a list of possible root causes. The root cause will be verified when you can pull it in and out of the process resulting in the problem appearing and disappearing. Think of it as a light switch that can turn the light on and off on command.