Page 322 - Six Sigma Demystified

302 Six SigMa DemystifieD

Since these objectives conveniently summarized the primary concerns of
their customers, shareholders, and employees, the executive staff decided to use
them as prioritization criteria for the improvement projects.
A score of 1, 3, or 9 is applied to each of the projects for each of the criteria,
with a score of 9 indicating that a given project is highly capable of meeting that
requirement. A score of 1 indicates that the project is not likely to fulfill the
criteria, and a score of 3 indicates that it is likely to meet the requirement.
For example, the first project (“BMP Cell 12 Scrap Reduction”) was consid-
ered to be highly likely to meet the requirements of financial benefits (improve
profitability by 50 percent), process efficiency (95 percent improvement in
cycle time), and process yield/scrap, so the project received a score of 9 for each
of these items. This project was likely to improve the on-time delivery rate to
90 percent and inventory turns (to 10 or higher), so it received a score of 3 for
these items. The sum of these scores is 36, as shown in the “Totals” column.
This score then can be compared with the score for other projects, with the
highest-scored projects receiving the highest implementation priority.
The “Totals” column provides an indication of how well the projects are
geared to meet the criteria the organization has established. A low number
indicates that few projects are likely to meet the criteria.

Multi-Vari Plots

Multi-vari plots, popularized by Dorian Shainin, are used to assign variation to
one of the following:
• Within-piece or within-sample variation. For example, this includes taper or

out-of-round conditions in a manufactured part or temperature or pH
variations in a chemical sample.
• Piece-to-piece variation. This is what we typically think of as within-sub-
group process variation—the variation we see over a short term from one
sample to another.
• Time-to-time variation. This refers to variation changing over a period of time.

When to Use

Analyze Stage

• To categorize variation to eliminate factors
• To investigate interactions among factors