formula to calculate cpk

What is the formula to calculate Cpk?

The most common process capability index formula is: Cpk = min[(USL − μ)/(3σ), (μ − LSL)/(3σ)]. It tells you how well your process fits within customer specification limits while also accounting for process centering.

In practical terms, Cpk answers this question: How many standard deviations fit between the process mean and the nearest spec limit? The closer your process runs to a limit, the lower your Cpk value becomes.

Meaning of each term in the Cpk equation

• USL = Upper Specification Limit
• LSL = Lower Specification Limit
• μ (mu) = Process mean (average)
• σ (sigma) = Process standard deviation
• Cpu = (USL − μ)/(3σ), capability toward upper side
• Cpl = (μ − LSL)/(3σ), capability toward lower side
• Cpk = minimum of Cpu and Cpl

Because Cpk uses the minimum side, it always reflects the worst-case side of your process. That is why Cpk is typically lower than or equal to Cp.

Cp vs Cpk: quick difference

Cp formula

Cp = (USL − LSL) / (6σ)

Cp measures potential capability assuming the process is perfectly centered.

Cpk formula

Cpk = min[(USL − μ)/(3σ), (μ − LSL)/(3σ)]

Cpk measures actual capability including mean shift. If your mean drifts off target, Cpk drops even if variation stays the same.

Step-by-step: how to calculate Cpk correctly

1) Confirm process stability first

Run control charts and verify only common-cause variation is present. Capability indices are unreliable for unstable processes.

2) Gather enough data

Use a representative sample from normal operating conditions. Avoid cherry-picked runs.

3) Compute mean and standard deviation

Calculate μ and σ from your data (or from validated statistical software).

4) Compute Cpu and Cpl

• Cpu = (USL − μ)/(3σ)
• Cpl = (μ − LSL)/(3σ)

5) Take the minimum value

Cpk = min(Cpu, Cpl). This is your reported index.

Worked example

Suppose a shaft diameter has:

• USL = 10.0
• LSL = 6.0
• Mean (μ) = 8.4
• Standard deviation (σ) = 0.5

Then:

• Cpu = (10.0 − 8.4)/(3 × 0.5) = 1.0667
• Cpl = (8.4 − 6.0)/(3 × 0.5) = 1.6000
• Cpk = min(1.0667, 1.6000) = 1.0667

Interpretation: the process is moderately capable but closer to the upper spec than the lower one.

How to interpret Cpk values

• Cpk < 1.00: Process is not capable of consistently meeting specs.
• Cpk ≈ 1.00: Barely capable; high risk of defects in real production.
• Cpk ≥ 1.33: Common industrial minimum for many processes.
• Cpk ≥ 1.67: Strong capability, often used for critical characteristics.
• Cpk ≥ 2.00: Excellent capability under stable, controlled conditions.

Common mistakes when using the Cpk formula

• Calculating Cpk on an unstable process.
• Mixing multiple process streams into one dataset.
• Using incorrect sigma estimates (short-term vs long-term confusion).
• Ignoring non-normal data distributions.
• Treating a single Cpk snapshot as permanent truth.

How to improve Cpk

Reduce variation

Attack root causes: tool wear, machine vibration, inconsistent materials, operator methods, environmental effects, and measurement error.

Re-center the process mean

If Cp is decent but Cpk is low, your process may be capable but off-center. Re-targeting mean can rapidly increase Cpk.

Control drift over time

Use SPC charts, preventive maintenance, gauge control, and reaction plans to hold gains.

Final takeaway

The core formula to calculate Cpk is simple, but good usage requires discipline: stable process, trustworthy data, and clear interpretation. Use Cpk alongside Cp and control charts for a complete view of process performance.