Cpk = 1.33: The 4σ Target

Is Cpk 1.33 actually acceptable for this characteristic — or is 1.67 the real requirement?

At Cpk 1.33, the nearest specification limit is 3.99 standard deviations from the mean using the stated within-process sigma estimate. The margin on the opposite side depends on centering and Cp.

• Typical PPM: ≈ 33.1 PPM at the nearest limit; ≈ 66.1 PPM only for a perfectly centered, two-sided process
• Sigma equivalent: 3.99σ to the nearest specification limit

Common automotive practice uses 1.33 for established production and 1.67 for some initial or launch studies, but the controlling value is always the applicable customer-specific requirement, drawing, Control Plan, or internal standard. Do not treat these values as universal pass/fail rules.

1. Confirm the source of the 1.33 target: contract, drawing, Control Plan, or internal SOP.
2. Verify the study window matches the reference (short-term vs long-term).
3. Report Cpk with its confidence interval; a lower CL below 1.33 is a failing study for small n.
4. State the σ estimator and normality assumption.

• Applying 1.33 to a special characteristic that actually requires 1.67.
• Reporting Cpk 1.33 without the confidence interval on small samples.
• Using 1.33 as a Ppk threshold; the two are not interchangeable.

• 1.33 corresponds to 3.99σ at the nearest limit — very close to 4σ, not exactly 4σ.
• Cpk 1.33 on a small sample (n = 30) has a wide confidence interval; the true value could be below 1.15.
• Automotive scorecards frequently list 1.33 as "acceptable" while the CSR itself lists 1.67 for special characteristics — always read both.

• Do not use 1.33 for characteristics classified as safety, critical, significant, or regulated.
• Do not use 1.33 as a universal threshold across customers.

• Cpk 1.33 ≈ 3.99σ ≈ 32 PPM at the nearest tail; ≈ 64 PPM two-sided centered.
• To reach 1.67 from 1.33 requires roughly 25% σ reduction at constant centering.

Use the free Process Capability Calculator to compute Cp, Cpk, Pp, Ppk, sigma level, and predicted PPM from your measurement data — no sign-up required.

• Cpk 1.33 is a floor, not a ceiling. Design capacity plans on higher headroom.

• ✓Governing target verified against CSR / drawing, not assumed
• ✓Characteristic class (special vs standard) confirmed
• ✓Sample size adequate for a defensible lower confidence limit
• ✓Study window matches the reference (Cpk vs Ppk)

• Process condition: statistical stability is required.
• Distribution assumption: approximately normal data.
• PPM convention: nearest specification-limit tail.
• Sigma-shift convention: no sigma shift applied.
• Cpk alone identifies distance to the nearest specification limit; it does not determine total two-sided defects unless centering or both tail distances are known.
• Predicted PPM is a model estimate, not a replacement for observed defect data.