RE: Ideas about service value

["Vic Forte" <vic@vichara.f9.co.uk>]

> << You can measure some *aspects* of
>  good service, but to rely too much on these measurements can
>  be dangerous, as you may pay too much attention to the "visible"
>  figures, and ignore the far more important "unknown and
>  unknowable". >>
>
>
> (lurker question) So...this response is to emphasize what on the other
> hand...processes?


The important issue here is what is knowable? What is unknown and
unknowable?

When a system is in control with respect a certain variable that is measured
there are two very important things that are knowable.

1) The extent of common cause variation and therefore the degree of
predictability of the process.
2) When a special cause occurs, the cause is knowable.

There is one important thing that is unknowable, and that is the cause of
data points that are within common cause variation.

Since quality is defined as "on target with minimum variation", the really
important issue is the extent of common cause variation (given that the
system is relatively stable and in control).

Common cause variation is reduced by working on the system. You work on the
system by applying the fourteen points, and the system of profound
knowledge.

When a process is in control (there are no special causes) specific data
points within common cause variation are unknowable in the sense that the
variation is caused by the complex network of causal interactions that we
call "the system". These are the really important data that Dr. Deming was
referring to.

What is knowable is the over-all degree of variation of the system and its
predictability and capability.

When working on the system there must perforce be no one-to-one
correspondence between the improvement efforts and individual data points
within common cause variation. Indeed work may go on for some time and
nothing may appear to happen. This does not mean that nothing is happening,
it simply means that the improvements are not yet detectable but are drowned
by the "noise" of common cause variation. But continual improvement has a
cumulative effect and eventually the improvement efforts will have an effect
upon the variation of the system which will become measurable. Nevertheless
it will still be impossible to know exactly what efforts contributed most to
the over all improvement, and in fact it will be impossible to know this.

A phenomenon often observed is a sudden "step" improvement after a long
period of effort. Managers make the mistake of attributing this to a
specific recent action when it is in fact the result of many small changes
that have come together and become finally visible. They are confusing
common cause and special cause variation.

This discontinuous nature of improvement is the main reason why it is know
as "continual improvement" not "continuous improvement".

So continual improvement involves a certain amount of working in the dark.
You need to have faith in the theory and work with it, sometimes for quite a
while, before you get to see measurable results. This process of working on
the unknowable is an inevitable result of the nature of variation as
Shewhart revealed to us, and the nature of knowledge in real-world systems.