One (not Six) Sigma for Process Incapability

Lean Organisations often employ Six Sigma methodology for projects aimed towards reduced variation and almost zero defects. Such organisations are known as “Lean-Sigma” organisations. Statistical Process Control, or SPC, is a technique used within Lean-Sigma companies to ensure that common cause variation (natural variation inherent within the process) is separated from special cause variation such that only special causes are acted upon. This ensures that a process is maintained within statistical control and that it is not being continually and unecessarily adjusted.

Process capability is achieved when the maximum level of natural process variation is well within the specification tolerance limits imposed by the customer. If the process variation follows a normal distribution and the mean measured value is centred on the target value, then the process will always be producing product within specification, as long as the specification limits are beyond the limits of the bell shape distribution curve. A Six Sigma level of capability ensures that the natural process variation is minimised to such an extent that even with a mean value 1.5 standard deviations off-centred around the target value, a process will never be producing more than 3.4 parts per million out of specification.



Fig. 1: The Lean-Sigma Organisation. Sigma Level 6 – 3.4 defects per million opportunities, or 99.9997% efficiency.

In a Fat Manufacturing environment, defects and variation are accepted as part of the norm, given the variation in operator skill level, the lack of standard operating procedures and the high level of error in the measurement techniques employed for quality testing. The aim of a fat organisation is to ensure that the specification range for its products are as wide as possible to accommodate the high process variation. Efforts to strive towards a six sigma defect level are considered as unrealistic, complicated, expensive and therefore pointless.



Fig. 2: The Fat-Sigma Organisation in an ideal world.

Where a customer dictates the specification limits and these are tighter than the process can achieve, “out-of-spec.” product is allowed to pass through the quality gate if the defective attribute is “hardly noticeable” and if “no-one has complained about it before”. If the customer does eventually spot a problem, effort is made to find appropriate in-spec. spot check measurements taken within the organisation that contradict customer claims. Where a defective sample is sent back from the customer for examination, efforts are made to find a measurement technique that produces an in-spec result. In cases where the facts are undisputable, assurances are made to the customer that a quality “toll gate” system has been put in place to ensure 100% inspection of all product leaving the site. The high costs associated with such fire-fighting measures tend to be covered by emergency funding, for example the Six Sigma training and deployment budget.



Fig.3: The Fat-Sigma Organisation in the typical world. A sigma level of 2 – 308,000 defects per million opportunities, or 69% efficiency, would be considered as being pretty good, if it had been measured.

More often than not, Fat-Sigma organisations do not have a system of measurement advanced or robust enough to be able to determine how many of the products they produce are beyond the specification limits communicated to their customers. Similarly a fat organisation will rarely detect whether a supplier’s material is out of specification. Received materials will be assumed to be of good quality and rarely examined. Any quality checks that are employed for key raw materials will tend to use a measurement technique with a larger degree of inherent variation than the level of variation it is attempting to detect, i.e. useless.

Neither will a Fat-Sigma organisation be able to detect whether its process has drifted off-centre, away from the target value. Process drift is guaranteed in fat companies because of the constant manual tweeking that takes place. Reasons for this include:

• A lack of standard operating procedures, instead individual operating procedures that differ from one operator to the next or between shifts. These operating procedures, often simply habits, are usually stored in small personal note books or inside the heads of individual operators.
• A lack of statistical process control or capability analysis. Common cause variation cannot be separated from special cause variation. As a result the process is adjusted even if it is statistically centered around the target value, thus sending the process off-centre. Continued adjustments of this nature send the process out of control. In these situations a Fat-Sigma organisation will not have a clue that it’s a Fat-Sigma organisation.
• Raw material or equipment changes that are implemented without any scientific evaluation of their effect on the process. Paper savings can be made instantly by managers with finance backgrounds. These are subsequently eroded progressively and invisibly by process drift and instability caused by the lack of understanding of the fundamental need to pay attention to process measurement.

Fig.4: Off-centred “One” Sigma Capability, or 69% of product out of spec.


If you are wanting to become a Fat-Sigma organisation, or are already there, then avoid the following reading material: