Introduction to Six Sigma

This chapter provides you with a brief overview of Six Sigma.

Six Sigma () is a statistical quality target and, at the same time, a quality management method based on a data-driven business strategy. It requires companies to focus all their processes on promoting the best possible quality for the customer.

Objective

The core element of the method is carrying out data-based improvement projects drawing on proven quality management techniques.

The main objectives of the method are to improve processes, reduce variation and achieve cost savings.

Method

The most frequently used Six Sigma method is the so-called DMAIC control cycle (Define/Measure/Analyze/Improve/Control). The DMAIC approach is used to make existing processes measurable and achieve permanent improvements in those processes.

DMAIC method

s_ppm_sixsigma_dmaic

Basic principle

Many processes that can be recorded stochastically are subject to statistical normal distribution (as defined by C.F. Gauß) and thus the mean (µ) and the standard deviation (sigma) can be uniquely defined for these processes.

Gaussian normal distribution

sixsigma_normal distribution

Normal distribution = Symmetrical distribution (dispersion) of values (variations) around a mean value

Measured values or observation results are plotted on the X-axis and probability values on the Y-axis.

Arithmetic mean (µ) = Average value (in general usage)

Standard deviation (σ) = Measure of deviations in the observed or measured values from the average value (or mean)

Interpretation

In the Six Sigma method, the mean of the normal distribution represents a quality assurance objective, e.g., the planned value of a key measure (measure). The standard deviation σ can be used to determine the relevant range of observations or measurements for quality assurance purposes.

For a normal distribution, the following rules of thumb can be applied:

(µ±σ) includes approximately 70% of observations

(µ±2σ) includes approximately 95% of observations

(µ±3σ) includes more than 99% of observations

As deviations in measure values from a specified planned value are unavoidable, limit values are defined to specify the range in which the measured measure values are still acceptable from a quality management perspective.

sixsigma_limit values

The aim of the Six Sigma method is for almost 100% (more precisely 99.99966%) of all results for a process to lie within a range that stretches from µ-6σ to µ+6σ. In other words, this means that out of a million executions only 3.4 results will have errors.

However, it is normal for every quality feature to have unwanted dispersion (deviations from the mean) in the process results. The mean (or average value) is often not exactly equal to the planned value. (From experience, most companies do not achieve more than 3 to 4 sigma on average.)

Tip

Further information on how Six Sigma works, its methods and tools is available in the book Six Sigma for Dummies by Craig Gygi, Bruce Williams and Neil DeCarlo; Publisher: Wiley-VCH; 1st edition (October 2005).