Sigma-level interpretation and limitations

Learning objectives

• Convert yield / DPMO to a sigma level and back.
• Explain the 1.5σ shift as a modelling convention, not a law.
• Communicate performance without false precision.

What sigma level is

Sigma level standardises defect probability for communication. It is derived from yield or DPMO under a normal-distribution model, often with a 1.5σ long-term shift assumption.

It should always be reported alongside the underlying DPMO or yield so the audience sees the raw performance.

Garment-factory example

A cutting operation at 20,000 DPMO reads roughly 3.6σ short-term or 5.1σ with the 1.5σ shift; the honest reporting shows DPMO plus the shift assumption used.

Method

1. State the calculation method and shift convention.
2. Convert with a free reference table or spreadsheet.
3. Report raw yield/DPMO alongside the sigma value.
4. Do not compare sigma levels across incomparable opportunity definitions.

Common mistakes

• Claiming ‘world-class’ from inflated opportunity counts.
• Comparing measurement-process sigma with inspection-defect sigma.