What Does Special Cause Variation Mean in Six Sigma?

Special cause variation is variation produced by a specific, identifiable source that is not part of the process’s normal system — a new material lot, a failing component, a skipped step, an untrained hand. Unlike common cause variation, which is chronic and diffuse, special cause variation is sporadic and assignable: something happened, at a particular time, for a particular reason, and that reason can in principle be found.

How it works

On a control chart, special causes announce themselves as out-of-control signals: a point beyond the control limits, or a non-random pattern such as a long run of points on one side of the center line. The chart’s entire purpose is to separate these signals from routine noise, because the two demand opposite responses. Routine variation is reduced by redesigning the system; a special cause is handled by investigating the specific event, removing the cause, and preventing its return.

Timing matters more here than anywhere else in quality work. Special causes are easiest to diagnose while the evidence is fresh — the shift log still legible, the material lot still on the floor, the operator’s memory still specific. A signal investigated a month later is usually just an anecdote.

A worked example

A commercial bakery charts loaf weight through the day. For weeks the chart shows nothing but calm. Then, mid-morning on a Monday, weights jump and stay high: point after point above the center line, and one beyond the upper limit. Because the signal is timestamped, the investigation is short — the bakery had opened the first bags from a new flour shipment at exactly that hour. The supplier had changed mills, and the flour absorbed water differently. The bakery adjusts hydration, adds an incoming check on flour lots, and the chart settles. The chart did not fix anything; it told the team precisely when and where to look.

• Investigate signals promptly — the evidence behind a special cause goes stale within days.
• Keep an event log beside the chart: material changes, maintenance, staffing, settings. It turns signals into diagnoses.
• Not every special cause is bad news; a favorable shift is knowledge worth capturing and standardizing.
• Removing special causes restores yesterday’s process — it does not improve the underlying system. That takes a project.
• Never widen the control limits to make signals disappear; that is turning off the smoke alarm.

A special cause is the process telling you exactly when to look. The only real mistake is not looking.

Recognizing special cause variation is introduced in our free White Belt course and practiced seriously at Green Belt ($299), where you learn the signal rules on real chart exercises. The proctored exams at every level — 70% to pass, one free retake — test the distinction because so much depends on it.