What Is Cp? Definition and Uses

Cp is the potential capability index. It divides the width of the specification by the natural spread of the process — upper limit minus lower limit, over six standard deviations — and answers one narrow question: if this process were perfectly centered, would its output fit inside the specification?

How it works

A Cp of 1.0 means the process spread exactly fills the specification width — no margin at all, so even small drift produces defects. Higher is safer, and 1.33 is a common minimum benchmark in industry, leaving usable margin between the process and its limits. Below 1.0, the process cannot fit the specification even when centered; no adjustment will save it, and the only routes forward are reducing variation or renegotiating the specification.

What Cp deliberately ignores is location. A process can carry a superb Cp while producing constant defects, because its average sits near one limit. That blindness is intentional: Cp isolates the spread question, and its companion index, Cpk, handles centering.

Cp is best understood as a statement about entitlement — the performance the process could reach with nothing more than perfect centering. That framing makes it a planning tool. If Cp is healthy, the improvement conversation is about adjustment and control; if it is poor, the conversation is about variation reduction. Those are different kinds of work with different costs, and honesty about which one you face saves months.

A worked mini-example

An illustrative bottling line fills to a specification of 500 millilitres, plus or minus 10 — limits at 490 and 510, a width of 20. The process is stable with a standard deviation of 2.5 millilitres, so its natural spread is six times that: 15. Cp equals 20 divided by 15, or 1.33. The spread fits the specification with margin to spare. Whether the bottles actually land inside the limits depends on where the average sits — which is exactly the question the next entry’s index, Cpk, exists to answer.

• Cp requires a two-sided specification; with a single limit, use the one-sided capability form instead
• A stable process is a prerequisite; Cp from an unstable process predicts nothing
• Never report Cp alone — pair it with Cpk so centering problems cannot hide
• Cp below 1.0 means reduce variation or renegotiate the spec; centering will not rescue it
• Small samples make Cp jumpy; treat early estimates as provisional

Cp asks whether the process could fit the specification — not whether it does.

Cp belongs to the Green Belt toolkit, taught during capability analysis in the Measure phase of our $299 program. Black Belt revisits it with the sampling theory and non-normal adjustments that serious capability work requires.