Process Stability
The terms Process Control and Process Capability describe different concepts requiring different approaches. We often need to evaluate if our process is in control and behaving consistently over time. In another post, we will discuss Process Capability, which determines whether or not the output from a process meets customer specifications. When discussing Process Control, it’s important to remember that process performance always varies from one time period to the next. When we plot a performance characteristic over time, we are trying to determine how a process is behaving.
Process Stability is an important topic in Six Sigma & Lean Six Sigma. When a process distribution remains consistent over time with the outputs fall within the range of the process width, the process is said to be stable (or in control). On the other hand, if the outputs are spread across outside the limits, the process is said to be unstable (or out of control).
Another important question: Is the process stable and predictable from one time to the next, or not? In other words, does the process exhibit variation due to common cause or special cause? Let’s discuss below to explain the difference.
If we plot a key process performance characteristic over a period of time, we will get a certain distribution. We may find that if we plot that same characteristic on a different day, the distribution looks similar. If we have a stable process that is in control, each time we plot the distribution it will look very similar.
This suggests that this process and, specifically the key metric we’re plotting to assess process performance, exhibits what is known as common cause variation. This is the ideal case as long as the output being produced by the process is within specification. Let’s say we have another process that we pull data from, and that its distribution plot looks like this. When we plot the same performance characteristic over time, we may see shifts in the mean of the distribution, changes in the shape of the distribution, and changes in the spread of the distribution.
This process exhibits special cause variation, which occurs when something has changed or is different in the process. These processes, with special cause variation, make great improvement projects. Early in our improvement projects, we need to evaluate the process behaviour to determine if our processes are stable or if they exhibit special cause variation.
Dr. Walter Shewhart developed the theory of Statistical Process Control, SPC, at Bell Labs in 1924. SPC provided a theory behind process variation and control charts. The first part of his theory is that all processes display variation. That may seem obvious, but it is an important concept that he built on. Second, since all processes display variation, Dr. Shewhart suggested variation can be broken down into two types: the common and special cause variation we just discussed. He stated that common cause variation is the variation component inherent to the process; it is always there.
According to Shewhart, this variation is caused by random or undiscoverable phenomena which are part of the process. The second component of variation he referred to as assignable cause variation, now commonly called special cause variation.
Shewhart explained that special cause variation is intermittent to the process; sometimes it is there, sometimes it is not. Unlike common causes, this type of variation is due to special causes that are discoverable and removable. Lastly, Shewhart showed that by sampling output from the process over time and plotting the Y, or key process performance, metric in the same order it was sampled, we can estimate distribution parameters of that process. Using this method, we can detect whether a process is exhibiting common cause variation or special cause variation. By identifying which type of variation our process is displaying, we can determine if our process is stable and predictable or unstable and unpredictable.
Control charts are a great tool to help us assess process stability for multiple reasons. First, they rely on statistics to determine when special cause variation exists. You don’t have to use your untrained eyes to determine if a process has changed or not; statistics will verify your conclusions. Second, they are used in real‐time, making them an active measure of your process. Finally, control charts are an easy‐to‐use tracking tool for measuring process inputs and outputs. At this point, our focus is on the outputs of the process.
Using the outputs, we can measure how the process is behaving and determine whether or not the process is stable. Beyond assessing stability there are many reasons why control charts are a useful tool for process improvement. They support a data‐driven, decision‐making culture. Process improvement is about using data to drive better business decisions. Control charts are alsoa relatively low‐cost control method to implement. Unlike mistake‐proofing or other expensive measures, control charts don’t require the purchase of additional equipment or higher‐end IT solutions to implement. Considering all these benefits, control charts should be an integral part of any process improvement effort.
Related Page:
Check out our Statistical Process Control Training, Lean Training, Six Sigma Training, Lean Six Sigma Training, Continuous Improvement Training or the full range of Training Courses for relevant courses on how to streamline & improve your business processes.