AI developers are licensing internal records from established companies to train their models. Process manuals, help desk history, CRM notes, project archives, QA logs and team chat all qualify. It is a license, so you keep your data and keep using it. Interested in selling? See if your business fits the program in under a minute.
Is your business eligible?
You're a match. Your business meets the program's requirements. Start your application to get your data valued.
Start My Application →Unfortunately you do not qualify.
Sigma level, DPMO, defects per unit and yield, all from three simple numbers.
The sigma level is a single number that tells you how capable a process is, from about 1 for a process full of errors to 6 for near perfect output. To get it you need DPMO, the defects per million opportunities, and a conversion that uses the normal distribution. This calculator handles all of it. In the first tab, enter the number of defects found, the units inspected and the opportunities for a defect in each unit. You get DPMO, DPU, yield and the sigma level, plus a plain language rating. The other tabs convert a sigma level into DPMO and yield, or a yield percentage into a sigma level. The standard 1.5 sigma shift is on by default and can be switched off. A reference table covers levels 1 through 6.
| Sigma level | DPMO | Yield | Defect rate |
|---|
An opportunity is any way a unit can fail a customer requirement that you actually check. Agree on the list once and keep it stable. Changing the definition later makes old and new sigma levels impossible to compare.
The same math works for invoices with errors, support tickets reopened, onboarding steps missed or CRM records with wrong data. Any repeatable process with clear pass or fail checks can be measured.
A long history of defects, root causes and fixes shows how your company learns. Those QA records support audits and training, and they are one of the operational data types AI companies license from established businesses.
First find DPMO: defects divided by (units times opportunities per unit), times 1,000,000. Then convert DPMO to a z value using the inverse normal distribution and add 1.5 for the standard shift. The calculator does both steps.
Six sigma means 3.4 defects per million opportunities, or a 99.99966 percent yield, when the 1.5 sigma shift is included. Without the shift, six sigma would be about 0.001 DPMO.
Many business processes run between 3 and 4 sigma, which means roughly 6,000 to 67,000 defects per million opportunities. Safety critical processes, such as in aviation, aim much higher.
DPU is defects per unit and ignores how complex a unit is. DPMO divides by every opportunity for a defect, which lets you compare a simple process with a complex one fairly.
Turn it off if your data already covers a long period with normal variation, or if your organization reports long term sigma. Keep it on to match standard Six Sigma tables and most training material.