2 ways to calculate Six Sigma (for Continuous and Discrete data)

2 ways to calculate Six Sigma (for Continuous and Discrete data)

Six Sigma calculation is the first thing that every six sigma practitioner or student should understand and know how to calculate. Learn the fundamentals here, how to calculate sigma level manually and so easily without the help of any tool like Minitab. I am a veteran as a Black Belt and I can tell you that most of the Black Belts don’t know how to calculate it manually, they tend to depend on some application.

First of all the most basic thing is that sigma level is calculated differently for discrete and continuous type of data.

Quick Recap – Understand your data type

Quick Excel method to calculate Process sigma for discrete data

Defects per million opportunities (DPMO) of a process is the total no. of defects divided by total defect opportunities, multiplied by one million. Synonymous with PPM.

Defects : It is the failure of the process as per specification. It can be in a form of wrong information, wrong opening of call , no closing …. Etc . Denoted by d

Unit (U) –The number of parts, sub-assemblies, assemblies, or systems inspected or tested.

Opportunity (O) –A characteristic you inspect or test. It is the maximum nos. of defects possible in one unit.

Defect (D) –Anything that results in customer dissatisfaction. Anything that results in a non-conformance.

Defects per opportunity – DPO: (D/(U * O))

Defects per million opportunity – DPMO : ( DPO * 1000,000)

1.Number of Units processed

2. Number of Opportunities for error per Unit

3.Total number of Defects

4.Solve for Defects Per Opportunity
DPO = ( D )/ ( U *O )

5.Convert DPO to DPMO
DPMO = DPO * 1,000,000

6.Look up Process Sigma in conversion table

Process Sigma – When my data is discrete

Process Sigma – When my data is Continuous

  • To calculate Sigma for continuous data, we need to calculate Cpk. You can check Process Capability section for more details
  • Cpk compares product specifications relative to centre () of the process
  • Similar to Cp in that it uses the standard deviation of the process, but does not need to have process centered to specification limits.