Process Capability: Understanding Cp, Cpk, and What They Tell You
Process capability is the ability of a process to consistently produce output within the limits your customer or design specification requires. It sounds simple, but it is one of the most frequently misunderstood concepts in quality management. Organisations measure capability to answer a basic question: can this process reliably give us what we need, or does it require intervention?
Two indices sit at the centre of capability analysis: Cp and Cpk. They are related but distinct, and understanding the difference between them is essential to interpreting capability data correctly.
What Cp Measures: Spread
Cp compares the width of your specification window to the natural spread of your process. The formula is straightforward: Cp equals the difference between the upper and lower specification limits divided by six standard deviations of the process. Six sigma represents essentially all of the natural variation in a stable process (99.73 percent of output, under a normal distribution).
A Cp of 1.0 means the process spread exactly fills the specification window -- if the process is perfectly centred, you will just barely meet spec on both ends. A Cp of 1.33 means the process spread occupies only 75 percent of the specification window, leaving margin on each side. A Cp of 2.0 means the process spread is half the specification width, providing substantial margin for drift and variation.
The limitation of Cp is that it tells you nothing about where the process is centred. A process can have a Cp of 1.5 and still be producing defects if the process mean is shifted toward one of the specification limits. That is where Cpk comes in.
What Cpk Measures: Centring
Cpk accounts for both spread and centring. It measures the distance from the process mean to the nearest specification limit in units of three standard deviations, taking the smaller of the two directions -- the binding constraint. When the process is perfectly centred, Cp and Cpk are equal. Any off-centring causes Cpk to fall below Cp. The gap between the two tells you how much capability potential you are losing to poor centring -- potential you could recover simply by adjusting the process mean.
Interpreting the Numbers
A Cpk of 1.33 is the widely accepted minimum threshold for a capable process. At Cpk 1.33, the process mean is at least four standard deviations from the nearest specification limit, producing approximately 63 defects per million opportunities under ideal conditions. Many automotive and aerospace customers require Cpk of 1.67 or higher for critical characteristics, implying five sigma of separation from the nearest limit. A Cpk below 1.0 means the process is incapable -- it is producing defects under normal operating conditions.
Cpk >= 1.67: Highly capable. Standard for critical-to-safety characteristics in regulated industries.
Cpk 1.33 to 1.67: Capable. Acceptable for most production characteristics.
Cpk 1.0 to 1.33: Marginal. The process meets spec but has little margin for drift.
Cpk < 1.0: Not capable. The process is producing defects under current conditions.
Short-Term vs. Long-Term Capability and the 1.5-Sigma Shift
Capability studies conducted over a short period capture only common-cause variation -- the inherent variability of the process under stable, controlled conditions. Long-term capability studies capture additional sources of variation that accumulate over time: tool wear, raw material batch changes, operator turnover, seasonal environmental effects, and shifts in equipment settings.
The empirical observation behind the 1.5-sigma shift is that long-term performance is typically 1.5 sigma worse than short-term capability predicts. This is why Six Sigma defines a Six Sigma process as producing 3.4 defects per million opportunities -- a process centred at six sigma short-term will drift to roughly 4.5 sigma in the long run. The 1.5-sigma shift is a planning assumption, not a universal law, but a useful reminder that short-term studies paint an optimistic picture.
What to Do When Cpk Is Low
A low Cpk points to one of two problems: too much variation, poor centring, or both. The diagnostic path depends on which is driving the gap.
If Cp is adequate but Cpk is low, the problem is centring. The process has enough inherent capability to meet spec, but the mean has drifted. The corrective action is to adjust the process target -- often a relatively straightforward fix involving a process parameter change or set-point correction. If both Cp and Cpk are low, the problem is excessive variation. Reducing variation requires identifying and eliminating the sources of common-cause variation, often through a DMAIC project that includes a measurement system analysis (to rule out measurement error as the culprit), a process capability study, and designed experiments to identify the key input variables driving the spread.
Capability indices connect directly to customer requirements and design specifications. A specification limit is not arbitrary -- it represents the point at which the product or service fails to meet customer needs. When Cpk is below 1.33, you are telling your customer that your process cannot reliably stay within the boundaries they have defined as acceptable. Capability analysis makes that gap visible and quantifiable, which is the first step to closing it.
XNM Consulting supports organisations in applying statistical process control and process capability analysis to drive measurable quality improvement. Learn more about our strategic advisory services.