What Is a Control Chart? Definition and Uses

A control chart is a time-ordered plot of a process measurement with three lines added: a center line at the process average, and upper and lower control limits calculated from the process’s own historical variation. It exists to answer one question, continuously: is this process still behaving the way it usually behaves, or has something changed? That distinction — routine variation versus real change — is the foundation of statistical process control, and arguably of all modern quality work.

How it works

The chart was invented by Walter Shewhart at Bell Laboratories in the 1920s, and the logic has not aged. Every process varies. Most of that variation is routine — the combined effect of many small causes that are always present. Occasionally something new enters the system: a worn tool, a different supplier lot, an untrained operator on the night shift. The control limits, conventionally set three standard deviations either side of the center line, mark out the range that routine variation should stay within nearly all of the time.

Reading the chart is therefore simple. Points scattered randomly inside the limits mean the process is stable: leave it alone, or improve it by deliberately redesigning the system. A point outside the limits, or a non-random pattern inside them, is a signal that a specific, findable cause has appeared. The correct response is to investigate — not to nudge a dial out of habit, which usually makes matters worse.

A worked example

Picture a bottling plant that samples five bottles every hour and plots the average fill volume. For three weeks the points wander unremarkably between the limits, and the team correctly resists the urge to tinker. Then, on a Tuesday afternoon, a point lands above the upper control limit. The supervisor checks the shift log: a filler nozzle was replaced that morning and set slightly rich. Without the chart, the overfill would have continued quietly for weeks, paid for in product given away. With it, the cause was found while the trail was still warm and corrected within the hour.

• Do not confuse control limits with specification limits — the first describe what the process does, the second describe what the customer will accept.
• Do not recalculate the limits every week; constantly refitting them erases the very signals the chart exists to detect.
• Do not react to every wiggle inside the limits — adjusting a stable process adds variation instead of removing it.
• Chart the upstream characteristics that cause defects, not just the final inspection count that reports them.

A control chart never tells you whether the work is good. It tells you whether the process has changed — the question everyone forgets to ask.

Control charts appear early and often in Six Sigma training. Our free White Belt course introduces the concept during its six hours of foundations, and the Green Belt program ($299, about 35 hours) teaches you to build, read, and act on the major chart types — skills tested on its 100-question proctored exam. The advanced variants belong to Black Belt territory, and this glossary covers those separately.