Six Sigma statistics represents excellence by ensuring that there is no waste (or as little as possible) in the process of production. For the program to be successfully implemented, the production process has to be evaluated and each process that is a step towards the final product dissected to determine where the problem lies. This process at near perfection, with a target of only 3.4 DPMO, or defects per million opportunities of goods produced.

The goal of the Six Sigma Process is near perfection, with a target of only 3.4 DPMO, or defects per million opportunities of goods produced. With this high goal, every process needs to be very thoroughly analyzed to pre-determine faults in the production system before there are flaws like defective or delayed products. This is called statistical capability analysis. A number of statistical methods are used to determine this. One of these methods is the use of graphs.

To understand these Six Sigma statistics, it is important to know that the whole process is based on the notion that if there are six standard deviations between the process mean and the closest specification limit, no item on the production line will fail to meet the standards set by data collection from customers. This is based on the method of calculation used in process capability studies.

Capability studies are done to determine how many standard deviations there are between the process mean and the closest specification limit, indicated in sigma units. With the process standard deviation rising, or the mean of the process moving away from the center of the tolerance, less standard deviations will find a place between the mean and the closest specification limit. This decreases the sigma number and ups the chances of items outside of the specification.  In other words, more products or services will become ‘defective’ or outside the limits of customer satisfaction.

While the above statistical studies help to reduce defective items, it has been established that production processes that are near perfect in the short term do not necessarily remain so in the long term. Other statistical studies need to be carried out to capture this.

If this was the case, the number of sigmas that would have room between the process mean and the closest specification limit would start go down with time when compared to the previous graphs. To explain this, which would in actuality be seen as an increase in process variation, a 1.5 sigma shift, which is empirically-based, is brought into the calculation.

Based on this theory, the assumption is that a process that accommodates 6 sigma between the process mean and the closest specification limit in a short-time will only be able to accommodate 4.5 sigma in the long term. The reason for this will either be that the process means will shift with time or because the long-term typical variation of the process will be more than the shift seen in the short term.  This is why it is highly encouraged that Six Sigma Progress be constantly monitored in the long term.  It is much easier to anticipate and catch a problem before it starts rather than fix a problem that has already occurred.

The reason could also be a combination of both of these factors. This is why the broadly accepted definition of a six sigma process is that it is a process that generates 3.4 defective parts in a million opportunities (DPMO).

There are a number of statistical methods that have to be understood by Six Sigma Professionals who are trained and certified in these methods. The Belts or specialists have to understand these statistics as they advance up the ladder towards the highest Belt which is that of Champion.