Jonathan D. Port, a quality management expert and owner of Beacon Quality Services LLC (www.beaconquality.com), wrote an article on understanding variation in the March 2012 issue of Quality Progress.
“Organizations with quality management systems compliant to ISO 9001:2008 are required to take action to eliminate the causes of nonconformities,” says Port. “Clause 8.5.2 defines steps required for corrective action (CA), including determination of the nonconformity cause, along with determination and implementation of necessary action to prevent recurrence. Similar wording is also present in clause 8.5.3 regarding preventive action (PA).”
A cause of nonconformance should coincide with a variation occurrence. Is the source of the variation a common cause or a special cause? In the interview below Port discusses how this question must be answered accurately to identify the proper path for CA, and how determination of variability type is often absent from problem-solving methods, which leads to ineffective actions.
It seems like it should be easy to track down the cause of a variation—is it?
If it was, the world would be a lot better place. Identifying and reducing variation is often the main focus of many engineering disciplines across the globe. Fortunately, in the last several decades there have been widely publicized tools to help tackle the problem.
What is the difference between common cause and special cause?
Common cause is inherent variation in the system. To change it, the system has to change either by reducing total variation or shifting the mean to have more of the distribution within acceptable limits.
Special-cause variation is introduced from outside the system. Identifying and eliminating special cause is the most common variation-reduction method. It is often employed erroneously, when the true cause is common-cause variation.
In Six Sigma, common cause or special cause may be discovered in the analyze phase, if the data collected in the measure phase includes some control charting. Control charting is one of the easiest ways to discern common cause from special cause.
What are the most common mistakes companies make in approaching variation?
Not getting to root cause is a big problem. There are more tools now available to help address root cause, and yet the term “root cause” is becoming one of the most abused. Root cause is the fundamental source of the problem. Removing the cause removes the variation. Validation of corrective action helps ensure the right root cause was identified and that the resulting action was appropriate to remove the cause.
Not understanding measurement system variation is also common. It is important to know what contribution of variation in data is the result of variation introduced through the measurement system. If what you are reacting to is measurement system variation, you will be hunting a ghost. Measurement-system variation can be a large portion of observed variation. Several things can contribute, including calibration, correlation, poor method instructions, repeatability (ability for operator to produce same result under similar conditions), reproducibility (ability for several operators to produce same result under varying conditions), and environmental conditions of the measurement area.
The other big one as mentioned before is misdiagnosing a problem as special cause. Treating a common-cause variation as special cause may temporarily make it go away, but it will be back as it is an inherent part of the system.
Please share an example of identifying and responding to the wrong kind of variation.
There is a term used in Statistical Process Control (SPC) called “process tampering.” This is when someone incorrectly identifies common-cause variation as special cause. When this happens, often total variation actually increases, not decreases because the distribution starts to shift around with each change. This can often happen in processes where SPC is not in place to help differentiate the two types of variation. It can also lead to false rejection (throwing away a good part) or false acceptance (accepting a bad part).
For example, a part has a diameter tolerance of +/- 0.05 in. It seems like every 10 or so parts, the inspector finds a reject and adjusts his production process to compensate. However, it was found that the tool the inspector used to measure with only had an accuracy of +/- 0.03 in. The measurement system in this case is not capable and the inspector may have incorrect results. The operator assumed special cause and corrected when the real issue is measurement variability (common cause).
Please share an example of the correct way to respond to a variation.
A company notices unacceptable performance of the Material Review Board process for dispositioning non-conforming material (NCM). The effects that were noticed included 1) some NCM occurrences were executed well and quickly; 2) other NCM occurrences had a long cycle time; and 3) reviewing some NCM records indicated the process was not performed correctly. An is/is-not analysis helped clarify the problem statement, which boiled down to variation in the performance due to variation in the way the process was executed.
Root cause analysis (RCA) revealed that the process was sound (released documents and flow charts looked good), but there was a difference in performance outcome and process execution between MRB leaders. This seems to be “special cause” in that a few MRB leads seem to be “bad apples;” however, furthering the RCA showed that the training that was given for MRB leads was not effective. This is a common-cause problem. It worked for some, but not all. The performance of the MRB leads is a natural result of the training process. The MRB process was deployed in a rush and training did not include enough examples or practice. For corrective action, the training for MRB leads was re-designed and re-deployed and MRB leaders were able to successfully complete their processes. For preventative action, other training programs were also evaluated, redesigned, and redeployed for other processes where risk may have been present.
What can companies do better to understand and respond to variation?
This is the whole premise behind Six Sigma—identification and reduction of variation. One of the tools that can be used to help identify areas of variation is the fishbone diagram. The most common points on the diagram are contributions to the issue by material, method, manpower, measure, environment, and machine. Companies can better understand and respond to variation by shifting to a paradigm that work operations are a series of processes, and those processes will have variation in the mix.
Beacon Quality Services LLC is based in Fort Collins, Colorado. Jonathan Port earned a bachelor’s degree in engineering management from the Missouri University of Science and Technology in Rolla. He is a senior member of the American Society for Quality (ASQ) and an ASQ-certified Six Sigma Black Belt, quality engineer, quality auditor, and manager of quality/organizational excellence.