|
Articles
Emergency
Quality Management
Mission Impossible: The Ultimate Facilitation
Challenge
Do You Believe In Magic?
Remembering Root Cause
Analysis
Columns
Conversations
For A Change
by Peter Block
Fox Shows Employees It Has
Heart
by Lynn L. Franzoi
Features
Brief
Cases
Business
News Briefs
Views for a Change
Sites Unseen
The Quality Tool I Never Use
Pageturners
Book Review
|
|
The Quality Tool I Never
Use
Larry Holpp
Wilton, Conn.
Larry Holpp has spent the last 25
years implementing team-based strategies in a wide
variety of organizations to support total quality, worker
involvement, union management cooperation efforts and
manufacturing innovations. His clients include Alberta
Pacific Forest Product, Lowes, Allied-Signal, and General
Electric Capital Corporation. He is the author of
“Managing Teams.” a textbook on team
management published by McGraw-Hill in 1998 and
“Team Turbo Training”, a series of tactical
training modules that will be published this
month.
What is the tool that didn't work
for you?
Scatter Diagrams are used when you need to display what
happens to one variable when another variable changes in
order to test a theory that the two variables are
related. The problem is that there are two ways that data
could be organized to show possible relationships between
variables.
Why didn't it work or why is it useless?
Scatter diagrams demand that we look for examples of
continuous data. Continuous data is data that can be
divided infinitely. If you are in certain service areas
the amount of continuous data might be limited to only
cycle time for example. Scatter diagrams demand
continuous data like time periods, money and percentages.
In businesses dealing in discrete events such as whether
or not a loan closed, the number of meetings that took
place prior to a decision being made or the results of
attitude surveys, scatter diagrams are ineffective.
How would you fix the tool?
For instance, instead of asking whether a deal took too
long, ask instead what the cycle time was from one event
to another. Another way to get to continuous data is to
use percentiles or percentages to reflect the results of
the many surveys we conduct. For example, five percent
answered in such a manner on question number three and 10
percent answered in a similar manner on question eight.
Generally this type of analysis does not provide
meaningful data. A third, and somewhat more technical
approach to data that is not continuous is made by using
logistic regression. This statistical procedure allows us
to show relationships between discrete variables such as
the relationship between common distribution problems and
branch office locations. Logistic regression analyzes the
relationship between different variables that don't have
to be continuous. You can compare one continuous variable
like total sales against regions or individuals. These
become verifiable relationships not just
correlations.
What words of counsel/warning would
you give to someone else before they used the
tool?
Make sure you are using continuous data or have
transformed discrete data to continuous data. And make
certain the client or internal customer considers the
data important.
|
