Definition

Process capability evaluates how a process’s output variation compares with specified requirements under defined conditions. Capability indices summarize that comparison, but meaningful interpretation requires attention to process stability, measurement quality, study design, and distribution assumptions.

A capability number is not a general certificate of supplier quality. It describes a particular characteristic, process, dataset, and estimation method.

Separate stability from conformance

A stable process can consistently produce output outside specification. An unstable process can temporarily produce acceptable parts while its future behavior remains uncertain.

Control charts help assess process behavior over time. Specification limits define acceptable product requirements; they are not statistical control limits.

VPP recommends reviewing stability and measurement-system suitability before relying on capability claims. A favorable index should not conceal an uncontrolled process or unreliable measurements.

Understand Cp and Cpk

For an appropriate normal-model application, Cp compares specification width with six times the relevant process standard deviation. Cpk also accounts for how close the process mean is to the nearest specification limit.

In plain-text notation:

  • Cp: (USL - LSL) / (6 × sigma)
  • Cpk: The smaller of (USL - mean) / (3 × sigma) and (mean - LSL) / (3 × sigma).

USL and LSL are the upper and lower specification limits. The estimate represented by sigma must be identified and technically appropriate.

Pp and Ppk are related performance indices commonly calculated using overall variation, while Cp and Cpk commonly use a within-process estimate. Confirm the convention used by the method or software instead of comparing labels alone.

A hypothetical capability calculation

Assume a stable process and suitable measurement system, with a specification from 9.900 to 10.100 millimeters. For illustration, the process mean is 10.020 and the relevant standard-deviation estimate is 0.020.

Cp is 0.200 divided by 0.120, or approximately 1.67. Cpk is the smaller of 0.080 divided by 0.060 and 0.120 divided by 0.060, or approximately 1.33.

The lower Cpk reflects the mean’s position closer to the upper limit. It does not, by itself, establish customer approval or a universal pass/fail conclusion.

These assumed values illustrate the arithmetic only. A real study needs adequate representative data, appropriate analysis, and the applicable acceptance requirements.

Inspect the study behind the number

Ask which product revision, machine, tool, cavity, shift, material, and operating conditions the study represents. Combining unlike populations can obscure meaningful differences.

Review the time sequence and subgrouping method. A short, carefully selected run may not reflect the variation experienced during routine production.

If the data are nonnormal, bounded, autocorrelated, or otherwise unsuitable for the assumed method, obtain qualified statistical guidance. Do not force a familiar formula onto every characteristic.

Use capability as part of a readiness decision

Capability evidence belongs alongside capacity, controls, measurement, and customer requirements. An attractive index for one dimension does not establish that a supplier can deliver the required production volume.

Define reassessment triggers for tooling changes, transfers, material changes, and other relevant events. Preserve the study scope and method so later results can be compared appropriately.

Avoid reporting a portfolio average of capability indices as though it describes one process. The practical risk often lies in a specific critical characteristic or operating condition.

Frequently asked questions

  • Is a high Cp enough? No. Centering, stability, measurement quality, and study scope also matter.
  • Are Cp and Pp always calculated identically? No. The variation estimate and method must be confirmed.
  • Does a sample with no defects prove capability? No. Acceptable observations alone do not establish stable long-term performance.

Ask for evidence that supports the decision

VPP helps sourcing teams connect supplier quality claims with operating requirements. Capability studies should be evaluated by qualified technical resources within the applicable customer and industry framework.