Process Capability Cpk: Calculation and Acceptable Values
Process capability Cpk tells you how well a stable process fits inside its tolerance, and whether it is centered. For molded and machined parts, Cpk on critical features shows whether a supplier can hold a dimension lot after lot, not just whether one sample passed. It is widely used in PPAP, supplier approvals and continuous improvement.
Cpk is easy to calculate and easy to misuse. For example, a Cpk of 2.0 from 10 hand-picked parts measured with a poor gauge says very little, while a Cpk of 1.4 from 125 production parts with a validated gauge is meaningful. In this guide, we explain process capability Cpk: what it means, how it differs from Cp, Pp and Ppk, how to calculate it, what values are acceptable, the conditions needed for a valid study, and how to improve capability. For measurement quality, see our guide to gauge R&R.
Process capability Cpk compares the distance from the process mean to the nearest specification limit with three standard deviations of short-term variation: Cpk = min(USL − mean, mean − LSL) ÷ 3σ. Cp compares the tolerance width with 6σ, ignoring centering. Pp and Ppk use overall (long-term) variation. Common targets are Cpk ≥ 1.33 for ongoing production and ≥ 1.67 for critical or new processes. Valid studies need a stable process (in statistical control), enough data (often 25 subgroups or about 100+ values), roughly normal data and a capable measurement system.
In short: Cpk measures centered capability; it is only meaningful for stable processes measured with good gauges.
Key points:
- Cp: spread vs tolerance.
- Cpk: spread and centering.
- Targets: 1.33 typical, 1.67 for critical features.
What Is Process Capability Cpk?
Process capability Cpk is an index that shows how many three-sigma spreads fit between the process mean and the nearest tolerance limit.
Interpretation:
- Cpk = 1.0: process just fits, about 0.27% outside if normal
- Cpk = 1.33: comfortable margin for most features
- Cpk = 1.67: strong margin for critical features
- Cpk < 1.0: process produces out-of-tolerance parts
For example, a bore with tolerance 10.000–10.020 mm, mean 10.012 and σ 0.002 has Cpk = min(0.008, 0.012) ÷ 0.006 ≈ 1.33. Statistical process methods are described in standards from ISO, such as ISO 22514.
Key Takeaway: Above all, Cpk reflects both spread and centering.
What Is the Difference Between Cp, Cpk, Pp and Ppk?
The difference between Cp, Cpk, Pp and Ppk is centering and the type of variation used.
| Index | Variation used | Considers centering |
|---|---|---|
| Cp | Short-term (within subgroup) | No |
| Cpk | Short-term | Yes |
| Pp | Overall (long-term) | No |
| Ppk | Overall | Yes |
For example, a process with Cp 2.0 but Cpk 1.0 has good spread but is off-center; adjusting the offset can raise Cpk close to Cp. Our guide to high volume CNC machined parts explains offset control.
Key Takeaway: In other words, a big gap between Cp and Cpk means “re-center the process”.
How Do You Calculate Process Capability Cpk?
You calculate process capability Cpk from the mean, standard deviation and specification limits.

Steps:
- Collect data in rational subgroups over time.
- Check stability with control charts.
- Estimate σ from within-subgroup variation (for example R̄/d2).
- Calculate the mean.
- Cpu = (USL − mean) ÷ 3σ, Cpl = (mean − LSL) ÷ 3σ.
- Cpk = the smaller of Cpu and Cpl.
For example, for a one-sided limit such as flatness ≤0.05 mm, only Cpu applies. Our guide to the control plan shows where capability is monitored.
Key Takeaway: For this reason, check stability before calculating Cpk.
What Cpk Values Are Acceptable?
Acceptable Cpk values depend on feature criticality and the customer’s requirements.
| Cpk | Typical judgment |
|---|---|
| Below 1.0 | Not capable; 100% inspection or improvement needed |
| 1.0–1.33 | Marginal; improve or control tightly |
| 1.33–1.67 | Capable for most production features |
| 1.67 and above | Capable for critical or new-process features |
For example, an automotive-style program may require Ppk ≥ 1.67 at PPAP and Cpk ≥ 1.33 in production. Our guide to PPAP levels explains submissions.
Key Takeaway: Overall, agree targets per feature before the study.
What Makes a Cpk Study Valid?
A Cpk study is valid when the process is stable, the data are sufficient and the measurement system is capable.
Conditions:
- Statistical control: no special causes on control charts
- Enough data: often 25 subgroups of 4–5, or 100+ individual values
- Representative production: normal settings, materials and operators
- Distribution: roughly normal, or use suitable non-normal methods
- Measurement system: acceptable gauge R&R
- Per cavity or fixture where they differ
For example, Cpk for a 4-cavity mold should be calculated per cavity, because each cavity has its own mean. Our guide to multi-cavity injection molding explains cavity differences.
Key Takeaway: Equally, combine cavities only if their means are similar.
How Do You Improve Process Capability Cpk?
You improve process capability Cpk by centering the process first, then reducing variation.

Centering:
- Adjust tool offsets or mold steel to target nominal
- Use mid-tolerance targets in programs
Reducing variation:
- Stable fixtures and workholding
- Tool life control and planned changes
- Temperature control of machines, molds and parts
- Consistent material lots and drying
- Process windows set by DOE or scientific molding
For example, a molded dimension at the edge of tolerance may be centered by a small steel-safe adjustment, raising Cpk from 0.9 to 1.6. Our guide to scientific molding explains process windows.
Key Takeaway: As a rule, center first; it is usually faster than reducing spread.
When Should Customers Ask for Cpk Data?
Customers should ask for Cpk data on critical features when they need proof of ongoing control.
Typical cases:
- Critical-to-quality (CTQ) dimensions
- New processes or new suppliers
- PPAP or similar approvals
- Features with tight tolerances
What to request:
- Cpk or Ppk per feature and cavity
- Control charts
- Gauge R&R for the measurement used
- Sample size and period
Capability studies and SPC data are supplied on request for agreed features. Our quality control and inspection process records such data.
Key Takeaway: Finally, ask for Cpk together with control charts and gauge R&R.
FAQ
- Q: What is process capability Cpk?
- An index showing how well a stable, centered process fits within its tolerance.
- Q: What is a good Cpk value?
- 1.33 is common for production; 1.67 for critical features.
- Q: What is the difference between Cp and Cpk?
- Cp ignores centering; Cpk accounts for it.
- Q: What is the difference between Cpk and Ppk?
- Cpk uses short-term variation; Ppk uses overall variation.
- Q: How much data is needed for Cpk?
- Often 25 subgroups or about 100+ values from a stable process.
Conclusion
Process capability Cpk shows whether a process can hold a tolerance over time, if the process is stable, the data are sufficient and the gauge is capable. Center first, then reduce variation, and request Cpk with supporting data. To discuss capability on your part, send your 3D model and 2D drawing to our engineers.