Statistical Process Control (SPC) is a mathematical quality control model that tracks output variations. It is a toolkit for both lean manufacturing and Six Sigma methodologies.
SPC is prevention-based quality control using statistical variations in output
14 charts & tools ensure systems are reliable and capable
Process flow variations can be assigned either common or special causes
The creator of Statistical Process Control is Walter Andrew Shewhart, a statistics engineer at Bell Laboratories in the 1920s
The main premise of statistical process control is that manufacturing operations don’t provide the same type of data as natural processes like biology. Shewhart noticed this and developed the system of SPC, which caught the attention of manufacturing legend Edwards Deming.
Rather than following a bell curve, Shewhart noticed, manufacturing operations created data along a static line of statistical control.
An example of SPC in history occurred in the 1930s with World War 2, when Shewhart deployed his process to manufacture munitions for AT&T at a facility in New Jersey. SPC attracted the attention of generals and mathematicians alike. After the war, Deming traveled to Japan to introduce SPC to Toyota, the first and original lean manufacturing operation.
Nowadays, SPC is very common within the Six Sigma management philosophy, which is known for relying on many quantitative tools for in-depth analysis.
Literally anything – and that’s the problem SPC charts solve.
If you have a manufacturing assembly line creating an output of 5 finished products per hour and then suddenly only 3 finished products are produced per hour, you know something is definitely wrong. The cause of the problem can be anything from a worn-down mechanical part to improper setup to an unnoticeable temperature fluctuation.
That’s why SPC charts are so beneficial – they are the first line of defense for your quality control even when you have no idea that there’s any variation in the first place.
Some defects in finished products may not be visible to the naked eye, or products may look fine but fail in testing.
The only way to see whether defects are due to worker error or faulty machinery or something else entirely is by collecting quantitative data through statistical process control.
Common variations are ones that are expected to occur within the confines of the experiment or process. Common causes are non-assignable variations, in that they are not due to anything out of the norm.
For this reason, it is often difficult to tease out all common variations. Many occur due to regular life occurrences that cannot be solved.
For example, a common variation for a process would be a slight delay in time for a completed task. Perhaps the employee assigned the task is new, and requires an extra hour to complete it, or perhaps the paperwork was backlogged because of a printer error, pushing the job down the queue.
Special variations are ones that occur out of specific deviations from the norm. Special causes are assignable to errors or machine malfunctions that can be plotted as a deviation on an SPC chart.
For example, a special variation for a process would be if a newspaper printing press had a feed error and the stamp printed the front page’s content half on, half off the paper. This special variation is assignable to the feeder malfunction.
There are hundreds of applications for SPC measurement tools, but the benefits of them all can be summarized into the following areas:
Ultimately, SPC is a preventative measure against deviations in quality control. The importance of statistical process control is its multifaceted approach to identifying “non-normal” quality deviations specific to manufacturing.
Moreover, the mathematical background of each tool allows for eventual connected automation in the modern Smart Factory. This is what separates SPC from other, less precise models of workflow or process control.
Statistical Process Control is more of a toolkit than a single chart – that’s why the quantitative tools are grouped together. You can choose which elements to include based on your project specifications.
There are 7 Quality Control tools for SPC, and these are the real meat of statistical analysis due to their quantitative nature and mathematical emphasis.
There are also 7 tools that are labeled “Supplemental” because they are less quantitative and more descriptive, often including visuals to provide a new angle of understanding or perhaps to argue for a certain executive strategy.
If you understand the point of SPC charts but don’t know where to access programs to generate them, the best option is to make use of a digital platform.
VKS provides easy-to-use SPC charts within our work instruction software so that you can take control of data-driven quality control for higher profit and efficiency.
SPC is a multistep process, and some sources cite anywhere from 6 to 10 “steps.” In reality, while it’s debated exactly how many steps there are, there are indeed some basic rules to follow. We’ve simplified them here to a big-picture, 3-step process:
You can add additional steps if you prefer, and should do so by splitting the 3 steps above into sub-categories. (For example, the process outline and the specification limits can be two different steps.)
If you look closely, you might notice that the process of SPC is similar to the scientific method, where a scientist defines the area of concern, identifies the control and the variables, and then tests their hypothesis. Think of SPC as the scientific method for manufacturing processes!
Likewise to the 3 steps, there are two phases of an SPC analysis. Most of what we’ve already gone over falls into the first phase, and the second phase is when changes are assessed and either institutionalized or modified to undergo further analysis.
This first phase includes the 3 steps above including the introduction of any new process variables and variations.
This second phase includes an assessment of the changes made during the initial time period, as well as a decision to be made according to the specific following factors: