Key takeaways
Short answer: Cpk is a process capability index — a statistic that measures whether process output reliably stays within specification limits. OEE is a productivity ratio comparing actual output to ideal. Cpk is a quality statistic; OEE is an effectiveness ratio. They are related (poor Cpk usually drives poor OEE Quality) but answer different questions. Both belong in a manufacturing KPI dashboard. See also Process Validation vs Process Verification.
Cpk (process capability index, k for off-center) is calculated as:
Cpk = min[(USL − mean) / (3σ), (mean − LSL) / (3σ)]
Where USL and LSL are upper and lower specification limits, mean is the process mean, σ is the standard deviation. Cpk answers: how comfortably does the process fit within spec, accounting for both spread and centering?
OEE = Availability x Performance x Quality. It compares actual output to theoretical maximum given the time available.
OEE Quality is the only OEE factor with direct quality content. The other two (Availability, Performance) are time and speed.
When Cpk is low, defects happen more often. Those defects become Quality loss in OEE. So Cpk drives OEE Quality:
The relationship is real but indirect. Cpk does not directly affect Availability or Performance — though chronic Cpk issues often correlate with slow cycles (operators run slowly to avoid scrap) which is Performance loss.
A process can have excellent Cpk and poor OEE if Availability or Performance loss dominate. The process holds spec but the line is down half the time or running at 60% speed.
A process can have great OEE and poor Cpk if defect rates are low overall (good Quality factor) but the process is not capable — the defects happen but the spec is loose, or sampling is too sparse to catch them.
Cpk:
OEE:
Both are KPIs that benefit from continuous measurement, not periodic sampling. Both lose value when calculated incorrectly. Both require honest data inputs to be meaningful.
1. Reporting one without the other. Cpk-only programs miss capacity loss. OEE-only programs miss capability problems hiding behind a Quality factor that looks OK on average.
2. Treating OEE Quality as a substitute for Cpk. OEE Quality is an output ratio; Cpk is a process statistic. They measure different things.
3. Using Cpk on non-normal data. Cpk assumes normal distribution. Non-normal processes need adjusted metrics (Ppk, percentile-based capability).
4. Calculating Cpk over too-long windows. Drift hides capability problems. Roll Cpk over shorter windows for active processes.
OEE platforms typically do not compute Cpk directly — that lives in QC software or LIMS. The integration is what matters: feeding QC results back into OEE Quality, and providing the production context (which line, which SKU, which shift) needed to interpret Cpk movements.
Fabrico's OEE module integrates QC results from external systems into OEE Quality and provides line/SKU/shift context for Cpk analysis in the connected QC system.
See how Fabrico captures this automatically — explore OEE for manufacturing or book a demo.
Cp measures spread relative to spec. Cpk also accounts for how far off-center the process is. Cpk is the stricter metric.
No. Cpk is a quality statistic. OEE has a Quality factor but it is calculated differently.
1.33 is the common minimum. 1.67 or higher for safety-critical or regulated processes.
Often yes. Better Cpk reduces scrap, raising OEE Quality. Both improve together if the cause is process control, not equipment speed.
Statistical process control (SPC) software, LIMS, or quality management systems. OEE platforms typically integrate the result rather than calculating it.
Login
Blog
