Cpk - Process Capability Index (Actual Performance)

Understanding actual process capability - what your process achieves in reality, including centering effects.

Cpk Overview & Fundamentals

What Cpk Measures:

Cpk measures the actual capability of a process, considering both process variation AND centering. It represents the worst-case scenario by taking the minimum distance to either specification limit. Cpk tells you: "Given where my process is actually centered, how capable is it?" This makes it the most practical capability index for production decisions.

Critical Advantages:

• • Reflects actual process performance
• • Accounts for process centering
• • Predicts real defect rates accurately
• • Used for production go/no-go decisions
• • Required for customer capability reports
• • Industry standard for capability assessment

Important Considerations:

• • More complex calculation than Cp
• • Sensitive to process centering changes
• • Requires accurate process mean estimation
• • Can vary with different subgroup strategies
• • Assumes normal distribution
• • May need frequent recalculation

Cpk Statistic & Formula

The Cpk Formula:

Process capability index (actual performance)

USL

= upper specification limit

LSL

= lower specification limit

= process mean (actual center)

= process standard deviation

Cpk ≤ Cp always. Cpk equals Cp only when the process is perfectly centered between the specification limits.

Cpu: (USL − μ) ÷ (3σ)

Cpl: (μ − LSL) ÷ (3σ)

μ: Process mean (actual center)

USL/LSL: Specification limits

Key Assumptions

• • Process output is approximately normally distributed
• • σ is estimated from stable, in-control process data
• • The process mean μ is representative of current, ongoing performance
• • Individual measurements (or subgroup statistics) are statistically independent

Reading the Cpk Capability Histogram

Example Cpk Result & Interpretation

Cpk Performance Standards:

Cpk < 1.0 - Inadequate Performance

Process produces significant defects (>0.27% or 2,700 ppm)

Action: Stop production, improve immediately

Cpk = 1.0-1.33 - Marginal Performance

Acceptable but requires close monitoring (63-2,700 ppm)

Action: Increase inspection, plan improvements

Cpk ≥ 1.33 - Good Performance

Industry standard for capable processes (~63 ppm)

Action: Maintain current performance, optimize

Cpk ≥ 1.67 - Excellent Performance

Superior capability with low defect rates (~0.6 ppm)

Action: Consider tightening specifications

Cpk ≥ 2.0 - World-Class Performance

Six Sigma level performance (~3.4 ppm)

Action: Benchmark for other processes

sixsigmacenter.com Implementation Guide

Example Data Layout for Comprehensive Cpk Analysis

Upload measurement data to calculate both Cp and Cpk with centering analysis. Platform identifies process adjustment needs and provides improvement recommendations.

This is how your spreadsheet or CSV file should be structured before you upload it. Match the column names, keep one row per observation, and leave the analysis to the platform.

Cpk to Defect Rate Conversion

Defect Rate Predictions:

Cpk = 0.67

4.5%

45,000 ppm defects

Cpk = 1.0

0.27%

2,700 ppm defects

Cpk = 1.33

0.006%

63 ppm defects

Cpk = 1.67

0.00006%

0.6 ppm defects

Cpk = 2.0

0.00034%

3.4 ppm defects

Business Impact:

Example: Manufacturing 1 million units annually

• • Cpk = 1.0 → 2,700 defective units per year
• • Cpk = 1.33 → 63 defective units per year
• • Cpk = 1.67 → Less than 1 defective unit per year

Cp vs Cpk Analysis

Understanding the Relationship:

Cp ≈ Cpk (Difference < 0.1)

Process is well-centered, focus on variation reduction

Cp > Cpk (Moderate difference)

Process centering opportunity exists, adjust process aim

Cp >> Cpk (Large difference)

Process badly off-center, immediate centering correction needed

Improvement Strategy Matrix:

Low Cp, Low Cpk

Reduce variation first

Then optimize centering

Good Cp, Low Cpk

Improve centering

Monitor variation

Good Cp, Good Cpk

Maintain performance

Continuous monitoring

Sample Size Requirements

Recommended Sample Sizes for Cpk:

Minimum: 100-125 measurements

Required for reliable mean and standard deviation estimates

Preferred: 200-300 measurements

Good confidence for production decisions

Optimal: 500+ measurements

Excellent precision for critical capability assessments