Every repeated process produces variation, but not all variation has the same interpretation.

Common-cause variation is the background variation produced by the many small effects built into the current process: ordinary material variation, ambient changes, small equipment fluctuations and other routine influences.

A special cause is a distinguishable change not belonging to that stable background, such as a wrong setup, failed heater, damaged tool or shifted fixture.

The distinction matters because the appropriate response differs.

• If a stable process has too much common-cause variation, improving it requires changing the process itself.
• If a special cause appears, the immediate task is to identify and remove that abnormal condition.

Suppose a measured output normally fluctuates around a stable centre with small random scatter. After an equipment adjustment, every measurement shifts noticeably upward. Treating those observations as merely more of the same random variation would miss a process change.

Conversely, repeatedly adjusting a stable process after every tiny high or low measurement can increase variation by reacting to noise.

Statistical process control therefore begins with a conceptual question: is the process behaving like one stable system, or has its behavior changed? Control charts formalize that question.