Process Capability: A Story About What Cpk Told a Plant Manager That His Defect Rate Could Not
This account describes a composite of real situations, with identifying details changed.
The facility was a precision components manufacturer producing machined brackets. Their first-pass yield was 94 percent — strong by the industry's historical standards, and the plant manager was reasonably satisfied. Monthly defect reports were reviewed, the out-of-spec parts were sorted and scrapped, and life went on. Then a Lean Six Sigma Black Belt joined the quality team and introduced two metrics the plant had never tracked: Cp and Cpk.
What Cp and Cpk Actually Measure
Cp (process capability index) measures how much of the specification tolerance is consumed by the natural variation of the process. A Cp of 1.0 means the process variation exactly fills the specification window; a Cp of 1.33 means there is room to spare; a Cp below 1.0 means the process is inherently too variable for the specification. But Cp does not account for where the process is centred. A process could have a Cp of 1.5 and still be producing out-of-specification parts if it is consistently running off-centre. Cpk adjusts for that. Cpk measures capability relative to the nearest specification limit, taking both spread and centring into account. The minimum of the upper and lower capability indices — Cpu and Cpl — is Cpk. A process with a Cpk below 1.33 is generally considered inadequate for precision work.
What the Analysis Found
The plant's main machining centre showed a Cp of 1.41 — adequate process width, well within specification tolerance. But the Cpk was 0.78. That number told a different story: the process was running off-centre, consistently biased toward the upper specification limit.
The 94 percent yield was accurate, but it masked the mechanism. The plant was not producing 6 percent random defects — it was producing a predictable, systematic defect pattern driven by a tool that had drifted from its set-up position over the course of each shift.
The control charts showed the drift clearly in retrospect: the data points were climbing throughout each shift and dropping sharply at the shift change when the machine was reset. No one had been reading the charts as a diagnostic — they were being used as a count of out-of-spec parts, not as a signal about process behaviour.
Once the root cause was identified — inadequate mid-shift tool compensation — a simple procedure change was implemented: a mid-shift dimensional check and tool offset adjustment. Within two weeks, the Cpk had risen to 1.38 and the first-pass yield had reached 98.5 percent.
The Lesson About Defect Rate vs. Process Capability
Defect rate tells you what happened; Cpk tells you why it will keep happening. A 6 percent defect rate is an outcome. A Cpk of 0.78 is a diagnosis. One tells you to sort the bad parts; the other tells you to fix the process.
Cp and Cpk together reveal whether the problem is variation or centring. If Cp is good but Cpk is poor, the process variation is acceptable but the process is not centred on the target. The fix is a setup adjustment, not a reduction in process variation. If both are poor, variation reduction is needed first.
Control charts and capability indices work together. Capability indices summarise a period of process data. Control charts show the process behaviour over time. Using both gives you the diagnostic power neither provides alone.
In 2022, the case for process capability monitoring is stronger than ever. Labour turnover, return-to-office after pandemic remote work, and new equipment qualifying under materials shortages all create risk of process drift. Capability monitoring catches drift before it becomes a defect crisis.
XNM supports public-sector and capital-project clients in applying Lean Six Sigma process improvement methods, including statistical process control and capability analysis. Connect with XNM's strategic advisory team to discuss how process capability monitoring could reduce quality risk in your operations.