How important is statistical process control to an organizational quality management system?
(Note: This entry is available as an audio interview on the Quality Magazine website). That is a great question, because I think it focuses on some key issues that are sometimes forgotten by quality managers. I see this often, because we sell both QMS software as well as SPC software, and I am always amazed that there is not more interest in SPC from some of our QMS customers. The short answer is: SPC is extremely important for a successful quality management system; let me explain why.
A quality management system is often focused on a few key areas: Corrective Action & Preventative Action (CAPA) to identify, correct and prevent the reoccurrence of non-conformances; auditing to ensure processes are using the quality systems effectively; and continuous improvement of the quality system itself. The system encompasses the full supply chain from your suppliers through to customers, as well as training of staff on the systems and processes.
With regard to CAPA, when we talk about preventative action we are referring to actions that will ensure the detected nonconformance will not reoccur. I suspect some practitioners think they can inspect their way out of this problem, that is, they feel by increasing inspection or monitoring of the process output they can prevent non-conformances from being delivered to the customer, or received from a supplier. I discussed the fallacy of this argument in the April Quality Magazine article, as did Deming in his Out of the Crisis text nearly 30 years ago. The problem with an inspection-focused approach is that after-the-fact sampling from process output is only credible if the process is in statistical control. In other words, you cannot take a sample from a bucket of bolts and expect that sample to be representative of the bolts in the bucket unless the process that generated the bucket of bolts is in statistical control. The problem is, if the process is not in control, the bucket contains multiple distributions of bolts. The statistics of a sample from the bucket will assume the bucket contains a single distribution, not multiple distributions, and provide misleading results.
The fact is, without evidence of process control, you have to apply 100% inspection to the bucket, inspecting each and every bolt in the bucket. So the inspection-focused approach is actually a very costly method for preventative action. And it is really not preventative at all! At best, it is reactive, at least when a process is out of control. Even for a process that is in-control, it shows poor foresight, in that we could predict for the in-control process the percent of product exceeding requirements. Failing to address those issues before shipment is simply a poor quality system.
The economic approach to preventative action is process improvement to prevent the occurrence of the nonconformance. Here again, in Out of the Crisis Deming discussed the need for a control chart to achieve process improvement, since only a control chart can differentiate between a common cause of process variation, which is built into the process, and a special cause of variation. Deming discussed how reacting to common cause variation as if it were a special cause increases process variation. I also discussed this in the April Quality Magazine article. The point is that you cannot do meaningful process improvement without a control chart. Failing to recognize that is one of the reasons non-conformances reoccur at some organizations, and their quality department is constantly fighting fires!
Finally, when you talk about improvements to the quality system itself, you focus on internal KPIs (key process indicators) that estimate the system responsiveness to problems. Here again, you need SPC to differentiate between the expected common cause variation in response and the special causes, since the special causes often provide insight into the dynamics of your systems, and thus the potential for improvement. Quality America has developed an interface between our SPC and QMS software to help our customers take quality systems analysis to this next level and improve through systems feedback using effective dashboard display of their KPI.
Learn more about the Quality Management tools for process excellence in The Handbook for Quality Management (2013, McGraw-Hill) by Paul Keller and Thomas Pyzdek or their online Quality Management Study Guide.