08. Cause & Effect (C&E) Matrix Overview The Cause & Effect (C&E) matrix is a mechanism to narrow down the number of Xs by using the existing knowledge of the Team. The tool is not used to rank the highest priority Xs, but to remove some of the trivial Xs with low priority. Construction requires the whole Team to take part. Logistics This is a team sport and should include representation for all the key areas affecting the process. Choose those individuals who live and breathe the process every day, not just Process Owners who might not be close enough to it to know the details. The activity typically takes a morning to complete. It is usually best to create the matrix directly in Excel, projected up on a screen using an LCD projector, rather than creating it by hand on a wall chart and then transferring it to a spreadsheet later. A common concern here is that the Belt wishes to create a strawman version prior to the working session with the Team. This is a mistake because the Team members usually just nod and agree to what the Belt has done and the matrix truly doesn't have a Team input. It is best to create it from scratch in the Team meeting. Roadmap If there are more than six process steps, consider using a two-Phase approach as listed in "Other Options: 2-Phase C&E" in this section. The stages to the construction are as follows: | | Step 1. | Starting with a blank C&E matrix (see Figure 7.08.1), enter the process responses (measures of performance) critical to the Customer (both internal and external). Each should be accompanied by a rating score of 1 to 10, because not all responses carry the same weight in the eyes of the Customer. Two or more responses might carry equal ratings if appropriate, but not everything can be scored a 10.
Figure 7.08.1. Single Phase Cause & Effect Matrix. 
| Step 2. | Enter the list of process steps and their accompanying inputs (Xs) from the Process Variables Map. Do not remove any Xs in the transfer; let the C&E matrix perform the task.
| | | Step 3. | The Team rates the strength of the relationship between each input and each response using the scale:
0: No relationship 1: Weak relationship 3: Medium relationship 9: Strong relationship
A common mistake here is to use a different scoring mechanism, say 0,1,3,5. This causes the matrix to give poor discrimination between Xs, and it is difficult to find a suitable cutoff point in Step 6. Also, don't try to second-guess the scoring to make an X more important, let the matrix do the work.
| Step 4. | The rating of each input is calculated by summing the multiples of relationship X response rating, for example, for the X called Budget in Figure 7.08.1, 291 = (9 x 10) + (1 x 8) + (1 x 7) + (9 x 9) + (3 x 5) + (9 x 1).
| Step 5. | The whole matrix is then sorted by the Rating column to give the ranked list of Xs.
| Step 6. | Team judgment is made on where to make a cutoff line, below which the inputs are considered unimportant and, therefore, not carried over to the next step in the roadmap. Obviously the trick here is to find the right cutoff point. If the cutoff is drawn too high, then potentially important Xs are eliminated from further investigation. If it is drawn too low, then many unimportant Xs are carried forward, and it will take time and effort to investigate them. Clearly the former would be much more of a problem than the latter.
|
One solution is to work slowly down the matrix from the top row until the Xs are deemed by the Team to be noisedraw the cutoff line above the first noise X. Then move slowly up the matrix from the bottom row until the first possibly important X is found and draw the cutoff line just below it. Now reconcile the two lines. A major mistake that is often made in this step is to assume that the C&E matrix is used to prioritize the top few Xs. It is not. The tool is used to eliminate unimportant Xs from the bottom. Interpreting the Output The C&E matrix provides a reduced list of Xs to carry forward to more powerful data-driven tools. The most important Xs (those that have the strongest effect on the Customer, based on the strength of the relationship between the Xs and the Ys and the corresponding rating of each Y) drift to the top. The least important drift to the bottom. The individual ranking of Xs is irrelevant; the cutoff line is the primary element here. Xs below the cutoff line are deemed unworthy of further investigation. Other Options: 2-Phase C&E For processes with many steps, it is usually best to reduce the number of steps first (Phase 1) and then use a second matrix to reduce the Xs (Phase 2). The stages to the construction are as follows: | | Step 1. | Starting with a blank C&E matrix, enter the process responses (measures of performance) critical to the Customer (both internal and external). Each should be accompanied by a rating score of 1 to 10, because not all responses carry the same weight in the eyes of the Customer. Two or more responses might carry equal ratings if appropriate.
| Step 2. | Enter only the list of process steps from the Process Variables Map, not the Xs (see Figure 7.08.2). Do not remove any steps in the transfer; let the C&E matrix perform the task.
Figure 7.08.2. Phase 1 of a 2-Phase C&E Matrix. 
| Step 3. | The Team rates the strength of the relationship between each process step and each response using the scale:
0: No relationship 1: Weak relationship 3: Medium relationship 9: Strong relationship
| Step 4. | The rating of each step is calculated by summing the multiples of relationship X response rating, for example, for the step called Place Order in Figure 7.08.2, 137 = (3 x 8) + (1 x 4) + (1 x 10) + (9 x 3) + (9 x 5) + (3 x 9).
| Step 5. | The whole matrix is then sorted by the Rating column to give the ranked list of process steps.
| | | Step 6. | Team judgment is made on where to make a cutoff line, below which the steps are considered unimportant and, therefore, not carried over to the next step in the roadmap.
| Step 7. | Starting with a second blank C&E matrix, enter the responses and ratings as before.
| Step 8. | Carry over from the Phase 1 C&E matrix only those process steps above the cutoff line.
| Step 9. | For this reduced set of process steps, list the associated Xs from the Process Variables Map (similar to Figure 7.08.1 but with a reduced set of Xs).
| Step 10. | As a safety net, check the omitted process steps from Phase 1 for any obviously important Xs and carry those over to the second Phase C&E also. Do not carry many over, just the obvious ones.
|
Continue as per the 1-Phase C&E from Step 3 of the 1-Phase C&E roadmap onwards to reduce the number of Xs. |
