What Does a c-Chart Measure in Six Sigma?

A c-chart is a control chart that tracks the count of defects found in a constant-sized unit of inspection — flaws per roll of fabric, errors per contract, scratches per finished cabinet. The crucial distinction is between a defect and a defective. A defective is a failed unit; a defect is a single flaw, and one unit can carry several. When what you care about is how much failure each unit carries, the c-chart is the tool.

How it works

The chart’s statistical basis is the Poisson distribution, which describes counts of independent events scattered across a fixed opportunity — a fixed length, area, batch, or document. Each point is simply the number of defects found in one inspection unit, and the control limits are computed from the long-run average count. The one strict requirement is that the area of opportunity stay constant: every roll the same length, every batch the same size. If the inspection unit varies — some rolls long, some short — the count alone is misleading, and the u-chart, which charts defects per unit of opportunity, takes over.

Reading the chart follows the usual rules: counts fluctuating inside the limits are the process’s normal texture; a count beyond the limits, or a persistent run above the average, means a specific cause has entered and is worth hunting.

A worked example

A textile mill inspects every hundred-meter roll of finished fabric and records the number of weaving flaws. Most rolls carry between two and seven, and the chart’s limits enclose that band. One Thursday a roll logs fourteen. The count sits far above the upper limit, so the shift lead walks the line rather than filing the number away — and finds a tension assembly on one loom drifting out of adjustment. Every roll that loom produced that day gets re-inspected, and the assembly goes on the maintenance list with data behind the request.

• Keep the inspection unit genuinely constant — if it varies, switch to a u-chart rather than pretending.
• Count defects, not defective units; if a single pass-fail verdict per unit is what you record, you want a p-chart instead.
• Agree on written defect definitions and categories first, or the chart will measure the inspectors rather than the process.
• Watch the average count: when it is very low, single-unit counts carry little signal, and a larger inspection unit works better.
• Log process events alongside the chart so an out-of-limit count can be matched to its cause quickly.

A p-chart asks how many units failed; a c-chart asks how much failure each unit carries. Different questions, different charts.

The c-chart rounds out the attribute-chart family taught at Green Belt level. Our Green Belt program ($299, about 35 hours, 100-question proctored exam) drills the choice between p, c, and u charts until it is automatic — because choosing the wrong chart quietly invalidates everything drawn on it.