InsightFab
Knowledge Base/Cpk Below 1.33: Not Your Fault—But You're Accountable
CPK7 min read

Cpk Below 1.33: Not Your Fault—But You're Accountable

When the month-end Cpk report comes in at 1.07 and is rejected by management, it prompts the question: What exactly does Cpk measure? This article clarifies whether the issue is process bias or dispersion, outlining two distinct problems with their respective solutions.

Scenario

The month-end process capability report came out, showing Cpk = 1.07. The customer standard is 1.33. Your supervisor threw the report back: "What happened this month?" You look at the numbers, unable to pinpoint the problem—the machines seem to be running fine, the defect rate is acceptable, but the Cpk just isn't passing.

Plain English Explanation

Cpk is asking one thing: How 'stable' and how 'accurate' is your process?

Think of it like archery:

  • Cp asks: Are your arrows sufficiently grouped together? (Degree of dispersion)
  • Cpk asks: Are your shots grouped together, and are they near the bullseye? (Dispersion + Bias)

Cpk = 1.33 means the specification limits are 4 standard deviations from the process mean, resulting in only 63 defects per million units. This is the industry's "passing line," not excellent, but the minimum requirement.

If Cpk is low, there are only two reasons:

  1. Process dispersion is too large (arrows flying wildly) → Reduce variation; switch to higher-precision equipment or raw materials
  2. Process mean is biased (entire batch shifted to one side) → Adjust the process center; calibrate equipment or parameters

Practical Judgment

Cpk RangeJudgmentAction
< 1.0UnacceptableImmediately identify root cause, cannot ship
1.0–1.33MarginalCustomer might accept, but next batch might fail
1.33–1.67AcceptableGeneral manufacturing standard
≥ 1.67ExcellentAutomotive/aerospace requirement level

To determine bias or dispersion: First, calculate Cp. If Cp is high but Cpk is low, it's a bias problem; simply adjust the center. If both Cp and Cpk are low, it's a dispersion problem; variation must be reduced at the source.

How InsightFab Helps

In Minitab, you have to manually interpret the process capability analysis report to discern the direction of the bias. In InsightFab, after uploading data, it directly gives you green/yellow/red lights and tells you whether 'bias' or 'dispersion' is the main cause, clearly indicating the next step in one sentence.

Golden Quote

"Cpk is not a report card; it's an honest confession of how much room your process has to mess up the next batch."

Want to try it yourself?

Every tool mentioned in this article is available on InsightFab — just upload a CSV to analyze.

Go to Tools

More in this category

The Difference Between Cp and Pp: Many Have Been Using Them Wrong for a Decade

Cp and Pp are both process capability indices whose values typically differ, yet many engineers struggle to differentiate their appropriate applications. Grasping this distinction will ensure your process capability reports are robust and address client inquiries effectively.

Cpk for Non-Normal Processes: What to Do When Data Doesn't Follow a Normal Distribution?

The process capability index Cpk assumes data follows a normal distribution. However, data such as stamping burrs, surface roughness, and leak rates are inherently non-normal. This article explains three calculation methods for non-normal process capability and the logic for their selection.