SPC or Shewhart Charts as Operational Definitions
The form of control chart which we use today was first developed in the 1920s by Walter Shewhart to increase reliability in the production of telephone equipment. He recognised that if a process was stable it was also predictable; i.e., once the natural variation of the process has been determined, it is then possible to predict future performance. This natural variation of the process does not alter over time unless action is taken to change the parent system.
A process is defined as being stable if its natural variation is due to common causes. The process is then said to be under statistical control. If a process is unstable, that is because unusual factors are operating on the process. These factors, known as special causes, result in the process being out of statistical control. Shewhart recognised that we all make mistakes at times, in that we take action when we should not do so. His chart helps avoid such mistakes.
Equally, we sometimes let things drift, assuming the process will right itself, when in fact we should react at the first sign of trouble. Shewhart was therefore aiming to devise a rule which would be sensitive enough to pick up a special cause, but not so sensitive as to react to extremes in terms of common causes.
Walter Shewhart circa 1926
Shewhart's first design for a control chart as submitted to his manager.He higlights 'trouble' (red circle) and the +/- 3 sigma warning limits (green elipse).