Pp and Ppk are process performance indices used to compare process variation with specification limits.

They are related to Cp and Cpk, but they are not identical.

A useful distinction is:

Cp and Cpk are typically used to describe short-term or within-process capability when the process is stable. Pp and Ppk use the overall observed variation in the data.

What Pp measures

Pp compares the total specification width with the overall spread of the process.

Conceptually:

Pp = Specification width ÷ Overall process spread

A larger value means the observed process variation is small relative to the specification range.

Pp does not account for whether the process average is centered between the limits.

What Ppk measures

Ppk considers both variation and centering.

It compares the process average with the nearest specification limit using the overall standard deviation.

If the process is off-center, Ppk will be lower than Pp.

That makes Ppk useful for describing actual observed performance against the specification.

Pp and Ppk vs Cp and Cpk

Process Capability commonly uses Cp and Cpk.

The key difference is the estimate of variation.

Cp and Cpk typically use within-subgroup or short-term variation.

Pp and Ppk use overall variation, which includes longer-term shifts and drift that appear in the full dataset.

If the process is stable and well controlled, the two sets of indices may be reasonably close.

If Ppk is much lower than Cpk, the process may have important long-term instability, shifts, or between-subgroup variation.

Stability comes first

Capability and performance indices can be misleading when the process is unstable.

Before interpreting the numbers, use Control Charts or another appropriate stability method to determine whether the process is behaving predictably.

A single capability index should not replace understanding the process over time.

Centering matters

Imagine a process with narrow variation but an average close to the upper specification limit.

Pp may look strong because the total spread is small.

Ppk will be lower because the process is operating too close to one side of the specification.

That difference is useful.

It tells the team that centering, not only variation reduction, may be required.

Measurement quality matters

Capability analysis is only as reliable as the data.

If the measurement system is poor, the estimated variation can be distorted.

Measurement System Analysis should be considered before trusting capability or performance results for important decisions.

Do not confuse specification with control limits

Specification limits come from customer, design, regulatory, or engineering requirements.

Control limits are calculated from process behavior.

They are not interchangeable.

A process can be statistically stable and still fail specifications.

A process can also produce mostly in-spec output while remaining unstable.

Common mistakes

Calculating indices on unstable data, mixing short-term and long-term interpretations, assuming a high index proves the process is controlled, ignoring centering, and using poor measurement data are common errors.

Practical interpretation

Use Pp and Ppk when the goal is to describe the performance represented by the full dataset.

Use Cp and Cpk when the within-process capability model is appropriate and the process is stable.

Always review:

  • the histogram or distribution;
  • control-chart behavior;
  • sample size;
  • centering;
  • measurement-system quality;
  • actual defect data.

The practical lesson

Pp and Ppk compress useful information into simple indices, but the indices are not the process.

Use them with stability analysis, measurement-system knowledge, and real process understanding.