Page 243 - Six Sigma Demystified

Part 3 s i x s i g m a to o l s 223

• When should the experiment be scheduled so as to minimize disruptions
to the process?

• Who should specify/execute the experiment?

The project team and key stakeholders must be committed to the experi-
mental design and its plan. Key team personnel, including the black belt, should
be on hand to lead the experiment (although operational personnel will be
responsible for running the process). Through monitoring the course of the
experiment, it is possible the black belt will observe conditions that lead to a
better understanding of process dynamics.
The experimental plan should include projections and allocation of
resources, including costs for personnel, materials, and equipment. Do not
commit all your resources to one design. A good rule of thumb is to spend
no more than 20 percent of your total resources on any one experiment
because each successive experiment will provide information that will be
confirmed or expanded on in subsequent experiments. The summarized
results of each experiment should provide information on what went wrong
and what to do next. Critical factors may not have been included or may not
have been varied sufficiently, so additional trials may be needed to collect
more information.
There are several desirable characteristics of an experimental design:

• It provides distribution of information throughout the region of inter-
est. You may begin the experimental process looking at a wide region
and then narrow your focus to a particular region that looks inter-
esting.

• It includes the necessary conditions to develop a model that predicts the
response, as close as possible to the true response, at all points within the
stated region of interest. This may require three or more levels of particu-
lar factors when nonlinear models are required.
• It allows the analyst to detect a lack of fit in the model.

• It may require blocking to meet the limitations of data collection or when
you wish to add runs to designs (such as folding).
• It allows sequential buildup of design, such as by folding or added axial
points.
• It provides an internal estimate of error variance.