If you want a reliable way to measure process capability in Excel, this page gives you everything in one place: Cpk formulas, exact Excel functions, interpretation rules, and a live calculator you can use immediately.
Cpk is the process capability index that tells you how well your process is performing relative to specification limits while also accounting for process centering. In plain terms, Cpk shows how close your process average is to the nearest spec limit after considering variation.
When teams ask, “How do I calculate Cpk in Excel?” they usually want to know if a process is consistently producing output inside tolerance. Cpk gives a single metric to summarize that capability.
Use these standard equations:
Cpu = (USL - Mean) / (3 × Sigma) Cpl = (Mean - LSL) / (3 × Sigma) Cpk = MIN(Cpu, Cpl) Cp = (USL - LSL) / (6 × Sigma)Cp measures potential capability if centered. Cpk measures actual capability with centering included. If Cp is high but Cpk is low, the process is likely off-center.
Put measurement values in cells A2:A101 (or any range).
Set LSL in D2 and USL in E2.
In Excel:
Mean (F2): =AVERAGE(A2:A101) Sigma (G2): =STDEV.S(A2:A101)Use STDEV.S for sample data. Use STDEV.P only if the range is the full population.
If you want one formula without helper cells:
=MIN((USL-AVERAGE(DataRange))/(3*STDEV.S(DataRange)),(AVERAGE(DataRange)-LSL)/(3*STDEV.S(DataRange)))Example with cell references:
=MIN(($E$2-AVERAGE($A$2:$A$101))/(3*STDEV.S($A$2:$A$101)),(AVERAGE($A$2:$A$101)-$D$2)/(3*STDEV.S($A$2:$A$101)))Suppose a shaft diameter process has:
| Metric | Value |
|---|---|
| LSL | 9.95 |
| USL | 10.05 |
| Mean | 10.01 |
| Sigma | 0.012 |
Compute in Excel:
Cpu = (10.05 - 10.01)/(3*0.012) = 1.11 Cpl = (10.01 - 9.95)/(3*0.012) = 1.67 Cpk = MIN(1.11, 1.67) = 1.11Interpretation: the process is limited by the upper side and may need better centering or lower variation to reach stronger capability targets.
| Cpk Range | Interpretation | Typical Action |
|---|---|---|
| < 1.00 | Process not capable | Reduce variation and re-center process |
| 1.00 to 1.32 | Marginal capability | Improve controls and monitor closely |
| ≥ 1.33 | Generally capable | Maintain SPC and periodic review |
| ≥ 1.67 | Highly capable / critical features | Best for tight customer requirements |
| ≥ 2.00 | World-class capability | Sustain with robust control plans |
These thresholds vary by industry, customer, regulatory environment, and characteristic criticality.
Before capability analysis, verify process stability with control charts. Capability metrics are meaningful only when the process is stable.
| Cell | Purpose | Formula |
|---|---|---|
| A2:A101 | Measured values | Raw data |
| D2 | LSL | Manual input |
| E2 | USL | Manual input |
| F2 | Mean | =AVERAGE(A2:A101) |
| G2 | Sigma | =STDEV.S(A2:A101) |
| H2 | Cpu | =(E2-F2)/(3*G2) |
| I2 | Cpl | =(F2-D2)/(3*G2) |
| J2 | Cpk | =MIN(H2,I2) |
| K2 | Cp | =(E2-D2)/(6*G2) |
Cp uses total tolerance width versus process spread and assumes the process is centered. Cpk includes centering by taking the minimum distance to either spec limit.
In standard calculations, Cpk is less than or equal to Cp. If you see Cpk greater than Cp, check formulas, rounding, or cell references.
You can still compute Cpu, Cpl, Cp, and Cpk directly from summary statistics, as long as mean, sigma, LSL, and USL are reliable.
More data is better. Many practitioners start with at least 25 subgroups or around 100 observations for more stable estimates, depending on process behavior.
If your goal is to calculate Cpk in Excel accurately, use a stable dataset, correct specification limits, and the right standard deviation function. Then compute Cpu and Cpl, and take the minimum as Cpk. That simple workflow gives you a practical, decision-ready measure of process capability.
Use the calculator at the top of this page to validate your numbers before building your internal dashboard or quality report.