Check Current Level (Sigma Level)

The data collected can be summarized to get the present level. Present level in the six sigma’s project must be known. Level usually becomes one of parameters (KPI) of projects which quantified and the baseline to determine the improvement direction. The present level shows the capability of our process. The study to check current level of our products/process is Capability Process study, while the output of this study shows the Sigma Level of the process. To check the capability present level, we ensure first the type of our data. Refer to the basic statistic (see basic statistic in Tool Categories); the data can be categorized in two types, Continuous Data and Attribute Data. Each type of data has different way how to calculate the level.

To check the capability process of continuous data, we can use basic metric Cp, Cpk, (Cpm), Pp, Ppk, Z(Bench)-Value (Short-term, Long-term) and Zshift. The output of this metric is Sigma Level. Normal Data is required to conduct this phase (by Normality Test, we can conclude whether data spread in normal distribution or not). While the attribute data, the basic metric are Defect per unit (DPU), Defect per Opportunity (DPO), Defect per Million Opportunity (DPMO). The output of its study is Yield.

Six sigma aims to rise up either sigma level or yield achieving “Six level”. Six in this term shows the Short-term level of sigma.

(See Capability Process practice – Tool Categories –)

Concept of Z-Value and Z(Bench)

As the standardization to justify the process capability, we use Z-value (Z-Value is equal Sigma Level). Z-value measured from the Y’s data by calculating spreading data with its specification. Illustration of measuring Z-value as is shown bellow.

Picture M.8 One Tail Process

Picture M.9 Two Tail Process

While Z(Bench) is Z-value of the process which is calculated by accumulate the probability defect in each tail then transformed to the Z-value.

Concept of ZShort-term (Z_ST), ZLong-term (Z_LT) and Zshift

Refer to concept of Rational Subgrouping , our process have a variation. Picture beside shows that when we take sample group in Time 1 will differ with sample group Time 2, etc. Therefore, the overall of the process may have gap with the target. The gap shows that process have a shifting. To know how much the process shift by its process, we must transform the data to the Z-value first, and then we have Zshift.

Zshift could be calculated by measuring gap Z-value if processes match the target (by equal variance, the process tendency to center) with Z-value the overall.

Z-value that assume match the target often referred as the Z-Short term (Z_ST). Z_ST is Z which is calculated by ignoring external/assignable causes (5M1E) of the process so the position (average) of distribution data is fit with the target. While, Z-value overall is Z calculated from resuming overall groups. This Z often referred as the Z-Long term (Z_LT). So, Zshift is the gap between Z_ST with Z_LT. Because of ZST always higher that ZLT, the formulation of ZShift:

Z_Shift = Z_ST – Z_LT

For Continuous data, we can use that formula to obtain the shifting of process. But, in the attribute Data / discrete data, ZShift are formulated by 1.5 sigma. For a typical process, there is a 1.5 sigma difference between short term and long term process capability on average.