“Variable Analysis 1. Define the problem. a. What problem are you trying to solve? b. Determine the primary variable • Closely inspect the point of failure or failed output • What measurable property do you want to change? • Is it possible to define the problem using a more specific variable? 2. Describe the problem: describe the problem in detail. a. What does the problem look like? b. When did the problem start? c. How often does the problem occur? d. Where does the problem first occur? e. When where don’t you see the problem? 3. Create a variable tree: develop each layer of sub-variables by understanding how the process works a. How is the process designed to control the primary variable? b. What else determines the value of the primary variable? c. Can we combine any of the sub-variables? 4. Eliminate sub-variables from the tree a. What should the value of each sub-variable be to prevent the problem from occurring? • What is the relationship between each sub-variable and the primary variable? b. Eliminate sub-variables that do not contribute to the problem • What is the actual value of the sub-variable during failure? During non-failure? • What tests could you use to eliminate variables that are difficult to measure? • What does the pattern of failure tell you? c. Expand sub-variables that have not been eliminated (iterating step 3) d. Start with sub-variables that the pattern-of-failure suggests are most likely to contribute to the problem e. Continue to expand and eliminate sub-variables until you have found the out-of-spec variable(s) that are directly in your control • Can you explain exactly how the out-of-spec sub-variable(s) contribute(s) directly to the problem? f. If you get stuck . . . • Have you eliminated a sub-variable that you should not have? • Have you missed a sub-variable? 5. Implement the solution a. Implement the solution b. Verify the solution”