What is sigma level and how does it relate to DPMO?

Sigma level is a quality score derived from DPMO, defects per million opportunities. You compute DPMO = (Defects / (Units x Opportunities per Unit)) x 1,000,000, then look the value up in a sigma conversion table. In the example, 12,000 DPMO corresponds to about 3.76 sigma. Lower DPMO means higher sigma; 6 sigma is roughly 3.4 DPMO, an extremely capable process.

How do I calculate DPMO and convert it to sigma level?

Multiply units by opportunities per unit for total opportunities, divide defects into that, and multiply by one million. With 6 defects, 100 units, and 5 opportunities each: DPMO = (6 / (100 x 5)) x 1,000,000 = 12,000. Then use a standard sigma table, where 12,000 DPMO lands near 3.76 sigma. The table already bakes in the customary 1.5 sigma long-term shift.

What sigma level is considered good?

6 sigma (about 3.4 DPMO) is world-class; 5 sigma is roughly 233 DPMO; 4 sigma about 6,210; 3 sigma about 66,800. Many processes target 4 sigma or better. The example at 3.76 sigma (12,000 DPMO) is mediocre, acceptable for manual work but poor for high-volume automated production. Moving from 3.76 to 4 sigma means cutting DPMO from 12,000 to about 6,210, roughly halving defects.

Why does my sigma level change when I recount opportunities?

Because opportunities per unit is in the denominator. Raising it from 5 to 10 halves DPMO and inflates sigma without any real improvement. In the example, 5 opportunities gives 12,000 DPMO and 3.76 sigma; using 10 would give 6,000 DPMO and a higher sigma. Define opportunities once, count only genuine defect chances per characteristic, and hold it constant across all measurements for valid comparisons.

How do I count defect opportunities per unit?

Count each independent characteristic that could be defective: one opportunity per critical dimension, solder joint, or spec you actually inspect. If a unit has 5 checked features, opportunities per unit is 5. Do not count cosmetic non-issues or the same feature twice. Be consistent, since the example's denominator of 100 x 5 = 500 depends entirely on that count of 5 staying fixed across every sample.

What is the difference between DPMO and DPU?

DPU is defects per unit: total defects divided by units, or 6 / 100 = 0.06 in the example, ignoring complexity. DPMO normalizes by opportunities too, so DPU is scaled by opportunities per unit and one million: 0.06 / 5 x 1,000,000 = 12,000 DPMO. DPMO lets you compare a simple part against a complex one fairly, which is why sigma tables use DPMO, not DPU.

Quality and Inspection

Sigma Level Formula

Sigma level converts defect rate into a process performance score on the Six Sigma scale. Use it to benchmark process quality, compare across products or lines, and set improvement targets.

Formula

DPMO = (Defects / (Units x Opportunities per Unit)) x 1,000,000

Variables

Defects: Total defects found in the sample
Units: Number of units inspected
Opportunities per Unit: Number of defect opportunities on each unit (each characteristic that could be defective)
DPMO: Defects per million opportunities; used to look up sigma level from a standard conversion table

Understanding the Sigma Level Formula

Sigma level scores process quality on a common scale by first converting defects into DPMO, defects per million opportunities. It normalizes for both volume and complexity: a unit with 5 defect opportunities and one with 50 are compared fairly because you divide by total opportunities, not just units. In the example, 6 defects across 100 units at 5 opportunities each gives 12,000 DPMO, which maps to about 3.76 sigma. Higher sigma means fewer defects per opportunity and a tighter, more predictable process.

Count defects and units from your inspection sample, then decide opportunities per unit deliberately: every characteristic that could fail counts once. Multiply units by opportunities per unit for the denominator (100 x 5 = 500), divide defects into it, and scale by 1,000,000. The result is DPMO, which you look up in a standard sigma conversion table. The table usually includes the 1.5 sigma long-term shift, so 12,000 DPMO reads as roughly 3.76 sigma, not the short-term value.

Interpret sigma against known anchors: 3 sigma is about 66,800 DPMO, 4 sigma about 6,210, 5 sigma about 233, and 6 sigma about 3.4 DPMO. At 3.76 sigma the example process sits between 3 and 4 sigma, common for manual assembly but weak for high-volume automated lines. To move up, reduce the defect count or root-cause the dominant defect mode. Just never inflate the opportunity count to game the number, because that flatters sigma without improving anything.

Worked Example

100 units inspected, 6 defects found, 5 defect opportunities per unit.

DPMO = (6 / (100 x 5)) x 1,000,000
= (6 / 500) x 1,000,000
= 12,000 DPMO
Sigma level lookup: 12,000 DPMO ~ 3.76 sigma

Result: Approximately 3.76 sigma

Common Mistake

Not defining defect opportunities consistently. Changing the opportunity count changes the sigma level without changing the actual process. For valid comparisons, the opportunity count per unit must be defined once and used consistently across all measurements.

Frequently Asked Questions

What is sigma level and how does it relate to DPMO?: Sigma level is a quality score derived from DPMO, defects per million opportunities. You compute DPMO = (Defects / (Units x Opportunities per Unit)) x 1,000,000, then look the value up in a sigma conversion table. In the example, 12,000 DPMO corresponds to about 3.76 sigma. Lower DPMO means higher sigma; 6 sigma is roughly 3.4 DPMO, an extremely capable process.
How do I calculate DPMO and convert it to sigma level?: Multiply units by opportunities per unit for total opportunities, divide defects into that, and multiply by one million. With 6 defects, 100 units, and 5 opportunities each: DPMO = (6 / (100 x 5)) x 1,000,000 = 12,000. Then use a standard sigma table, where 12,000 DPMO lands near 3.76 sigma. The table already bakes in the customary 1.5 sigma long-term shift.
What sigma level is considered good?: 6 sigma (about 3.4 DPMO) is world-class; 5 sigma is roughly 233 DPMO; 4 sigma about 6,210; 3 sigma about 66,800. Many processes target 4 sigma or better. The example at 3.76 sigma (12,000 DPMO) is mediocre, acceptable for manual work but poor for high-volume automated production. Moving from 3.76 to 4 sigma means cutting DPMO from 12,000 to about 6,210, roughly halving defects.
Why does my sigma level change when I recount opportunities?: Because opportunities per unit is in the denominator. Raising it from 5 to 10 halves DPMO and inflates sigma without any real improvement. In the example, 5 opportunities gives 12,000 DPMO and 3.76 sigma; using 10 would give 6,000 DPMO and a higher sigma. Define opportunities once, count only genuine defect chances per characteristic, and hold it constant across all measurements for valid comparisons.
How do I count defect opportunities per unit?: Count each independent characteristic that could be defective: one opportunity per critical dimension, solder joint, or spec you actually inspect. If a unit has 5 checked features, opportunities per unit is 5. Do not count cosmetic non-issues or the same feature twice. Be consistent, since the example's denominator of 100 x 5 = 500 depends entirely on that count of 5 staying fixed across every sample.
What is the difference between DPMO and DPU?: DPU is defects per unit: total defects divided by units, or 6 / 100 = 0.06 in the example, ignoring complexity. DPMO normalizes by opportunities too, so DPU is scaled by opportunities per unit and one million: 0.06 / 5 x 1,000,000 = 12,000 DPMO. DPMO lets you compare a simple part against a complex one fairly, which is why sigma tables use DPMO, not DPU.