Process Sigma Calculation converts defect performance into a sigma-level estimate so teams can compare process quality and improvement progress.
Definition
Process Sigma Calculation estimates process performance as a sigma level, often from yield, defects per million opportunities, or defect probability. It provides a common Six Sigma language for comparing defect rates across processes.
The calculation depends on clear definitions of units, defects, opportunities, yield, and whether any long-term shift convention is being used.
History
Process sigma language became widespread through Six Sigma deployments, especially in Motorola and later GE-style programs. It helped leaders translate defect rates into a standardized performance scale, though interpretation requires care.
When to Use
Use Process Sigma Calculation when communicating baseline quality, comparing processes, tracking DMAIC improvement, or translating DPMO into a common metric. Use capability indices or control charts when measurement distribution and stability are the main decision concerns.
Step-by-Step
- Define the unit, defect, and opportunity.
- Collect defect and unit data for a representative period.
- Calculate defects per opportunity or DPMO.
- Convert DPMO or yield to sigma level using the chosen convention.
- Document assumptions, including any shift convention.
- Stratify by product, customer, step, or defect type if needed.
- Use the metric to prioritize improvement, not to hide actual defects.
- Recalculate after verified process changes.
Examples
- Manufacturing: A line converts defects and opportunities into DPMO and sigma level.
- Service: Billing errors per invoice opportunity are translated into a baseline sigma estimate.
- Project tracking: DMAIC teams compare before-and-after sigma levels after corrective actions.
Common Pitfalls
- Inflating opportunity counts to improve the sigma number.
- Mixing defects and defective units.
- Not stating the shift convention.
- Using sigma level as the only customer-risk measure.
- Combining unrelated processes.
- Ignoring process stability and measurement quality.