Sigma Level Calculator (DPMO & Yield)
Convert defects, DPMO, or yield into a process sigma level to benchmark quality on the universal Six Sigma scale. Move between defects per million opportunities, yield and sigma, with the conventional long-term shift applied.
Calculate Sigma Level →What is a Sigma Level?
A sigma level is a standardized measure of process quality that expresses how many standard deviations fit between the process mean and the nearer specification limit. The higher the sigma level, the fewer defects the process produces relative to its tolerance.
The scale is anchored by defects per million opportunities (DPMO). A Six Sigma process produces about 3.4 DPMO, a widely cited benchmark that already includes a 1.5-sigma allowance for long-term process drift. Without that shift, six sigma corresponds to about 2 defects per billion, so the convention matters when interpreting the number.
Because it is unit-free, sigma level lets you compare processes of completely different kinds, a billing process, a machining operation, a call center, on one common quality scale, which is why it is the headline metric of Six Sigma.
In plain terms: Sigma level is a single score for how good a process is at staying inside its limits. More sigma means fewer defects. The famous 3.4 defects per million is the Six Sigma target, and it already bakes in a 1.5-sigma cushion for the fact that processes drift over time.
Three Ways In
From DPMO
Enter defects, units and opportunities per unit. DPMO = defects / (units × opportunities) × 1,000,000, then convert to sigma.
From Yield
Enter the proportion of defect-free output. Yield maps to a defect probability, then to a z-value and sigma level.
Short vs Long Term
Short-term sigma is the raw z from yield. Long-term (the reported figure) subtracts the conventional 1.5-sigma shift to reflect real drift.
Key Formulas
Reading the Sigma Scale
Familiar reference points: about 690,000 DPMO is roughly 2 sigma, 66,800 DPMO about 3 sigma, 6,210 DPMO about 4 sigma, 233 DPMO about 5 sigma, and 3.4 DPMO is 6 sigma. Each step up the scale represents a large drop in defect rate.
Always state whether a sigma figure includes the 1.5-sigma shift. The commonly quoted values (including 3.4 DPMO at six sigma) are long-term figures with the shift applied. Comparing a shifted number to an unshifted one leads to confusion and apparent discrepancies.
Assumptions & Validation
Clear Opportunity Definition
An 'opportunity' for a defect must be defined consistently, or DPMO is not comparable across processes.
If violated: Standardize the opportunity definition before comparing processes.
Approximate Normality
Converting yield to a z-value assumes an approximately normal process for continuous data.
If violated: For attribute or non-normal data, interpret DPMO directly rather than as a normal z.
Consistent Shift Convention
The 1.5-sigma shift is a convention; apply it consistently and state whether figures are short or long term.
If violated: Report both, or clearly label which is shown.
⚠️ Check assumptions first
The biggest source of error in sigma-level reporting is inconsistency in two things: how a defect opportunity is defined, and whether the 1.5-sigma shift has been applied. Two teams can compute very different sigma levels for the same process simply by counting opportunities differently. Fix the opportunity definition and label short-term versus long-term figures before comparing or trending them.
When NOT to Use Sigma Level Calculator
Comparing Group Means
Sigma level benchmarks defect rate, not whether two processes differ statistically. Use a hypothesis test for that.
Unstable Process
Like capability, sigma level projects future performance and assumes stability. Achieve control before benchmarking.
Ill-defined Opportunities
If defect opportunities cannot be defined consistently, DPMO and sigma level are not meaningful; use raw defect counts.
Industry Applications
Baseline & Benchmarking
Establish the current sigma level of a process as a baseline and compare across processes or sites.
Improvement Tracking
Show the sigma-level gain before and after an improvement project as a headline result.
Goal Setting
Translate a target defect rate into a sigma goal that teams across functions can understand.
Transactional Quality
Apply the same scale to service and administrative processes, not just manufacturing.
Frequently Asked Questions
What is DPMO and how does it relate to sigma level?
DPMO is defects per million opportunities, calculated as the number of defects divided by the total number of opportunities, scaled to one million. It is the defect-rate input to sigma level: a lower DPMO corresponds to a higher sigma. A Six Sigma process is defined as about 3.4 DPMO, which is the anchor point of the entire scale.
Why is Six Sigma 3.4 defects per million and not 2 per billion?
A perfectly centered six-sigma process would produce about 2 defects per billion. The commonly quoted 3.4 DPMO figure includes a conventional 1.5-sigma shift that accounts for the tendency of processes to drift off-center over the long term. So 3.4 DPMO is the long-term, shifted value, while 2 per billion is the idealized short-term value.
What is the 1.5-sigma shift?
The 1.5-sigma shift is a convention that adjusts short-term capability to reflect the drift and shifts a process experiences over the long term. Reported sigma levels typically add 1.5 to the short-term z-value, which is why the familiar DPMO benchmarks correspond to shifted figures. When comparing sigma levels, it is essential to know whether the shift has been applied.
How do I convert yield to a sigma level?
First express yield as a proportion of defect-free output, which gives the probability of conformance. The corresponding one-sided z-value from the standard normal distribution is the short-term sigma level, and adding the conventional 1.5-sigma shift gives the reported long-term sigma level. This links yield, defect probability and sigma on one scale.
What counts as an opportunity for a defect?
An opportunity is any point where a defect could occur on a unit, for example each critical dimension on a part or each required field on a form. Because DPMO depends directly on this count, the opportunity definition must be consistent across processes being compared. Inflating opportunities artificially raises the apparent sigma level.
Does sigma level require the process to be normal?
Converting a continuous-data yield into a z-based sigma level assumes approximate normality. For attribute data, where you simply count defective units, DPMO and the resulting sigma level are interpreted directly from the defect proportion. In both cases the process should be stable, since sigma level projects future performance.
Benchmark Your Process on the Sigma Scale
Convert DPMO, yield and defects into a sigma level in one step. Free during Beta.
Calculate Sigma Level →