How to identify the cause(s) of the problem.
|
Understand the
problem symptom(s) and product flow.
Get full knowledge
of the problem by visiting your customer and/or manufacturing plant.
Discuss with them the nature of the problem and ask for its
demonstration. In addition, understand the product flow and how well the
manufacturing process is controlled.
A strategy for
getting valuable information.
Ask anyone for input who was
or is currently involved with the problem.
You
will significantly increase both the quality and quantity of your input
from the information providers if you were to assure them that the
company's objective is to solve the problem and not establish who must be
punished.
If possible, you
should gather all of them in the same room and have a brainstorming
session. Listen carefully not only for what they might think the problem
is, but - more importantly - for their reasoning and any supporting
evidence.
The benefits
from listening to end users and manufacturing personnel.
Our
experience has shown that you should listen very attentively to people
that have direct exposure to the problem on a daily basis. Such people
include equipment operators, inspectors, set-up and maintenance personnel
and quality control technicians.
Don't forget to get
the input of the "grave yard" or third shift, if it exists.
Most plants do their maintenance and other repair work during that
shift.
Get your
hands dirty.
Become
"intimately" involved with the manufacturing process and
quality defects. That means, you may have to sacrifice a meal, a night's
sleep or some weekends away from your family. But at the end, if you work
hard and smart, your efforts will definitely be rewarded by solving the
problem.
Look for
trends.
First,
write down on a large sheet of paper (a flip chart page is recommended)
any trends that begin to emerge.
For example, the
product defect(s) appear to reach a maximum for a particular catalog
number - which is related to a specific product configuration - a
specific shift, season or product produced within a certain time from the
start-up or shut down of equipment or after a weekend or any other two
day holiday (during this time the plant shuts down), an overtime work
period for satisfying above-normal market demands, etc.
Develop
procedures for better characterization of quality defects.
Your
ability to correctly diagnose and fix the problem depends a lot on the
quality of the data you will generate.
Pass/fail testing,
if any, should be replaced with quantitative evaluations. In addition,
you should consider describing defects, environments and failure modes
with greater completeness and clarity than is required merely for
control.
At times, it may be
advisable to put the product through stress conditions which are capable
of magnifying its defects and failure modes.
How many
samples should you test?
All
your procedures need to be both accurate and precise. The number of
samples needed for each test should not exceed N=20. Usually you will get
the same quality of data by testing only as few as five product
samples.
If the product is
very expensive, you should try to come up with some innovative ways of
obtaining almost the same degree of quality data but with non-destructive
testing.
Dissect and
analyze the problem.
Use
proven methodologies for breaking down the problem to its smallest
possible parts. We recommend the use of the Ishikawa "fishbone"
cause-and-effect diagram.
Develop the
fishbone diagram on the same, large sheet that you used earlier to list
emerging trends. Your doing so forces you to be selective and focused.
After all, you have limited space for writing things down, which helps
you to better understand the problem, eliminate weak hypotheses and
visualize potential solutions.
Formulate and
test your theories.
A
culmination of an analysis of trends, existing data and the
cause-and-effect Ishikawa diagram would be the formulation of selectively
few, sound theories. These theories should be tested for their validity
using effective experimental design techniques.
Walk away
from the problem when you approach the "burn out" state.
When
you begin to feel overwhelmed and confused about the problem, it is time
to get re-infused with fresh and creative ideas. So it is highly
recommended that you take a short break to recharge your brain
batteries.
Get temporarily
involved with something else you enjoy and put the problem out of your
mind. Our experience shows that is a very effective and powerful
technique.
.
How to best
implement the derived solution.
|
.
A strategy for a
"win-win" outcome.
Successful
implementation of any program necessitates that the major parties are
engaged in a "win - win" situation. Such an objective might be
difficult but is definitely attainable.
So before you race
ahead for a fast implementation of your action plan, it is highly
advisable that you informally review it with the major players and get
their input. If you do a good enough job at your informal meetings, you
should be able to know before hand what, if any, are the "hot
buttons" of the key people.
How to
develop and implement your plan.
Then,
your objective should be to formulate an action plan which strikes a
balance between the potentially competing needs of certain departments
and those of your project. Your immediate supervisor or sponsor of the
program should be able to steer you clear from any hidden
"mine-fields" and other unpleasant situations.
Your implementation
plan should summarize the history of the problem and provide scientific
evidence for its expected resolution, describe the tangible benefits to
be derived, and include a detailed list of all required tasks with
specified dates of completion and assigned responsibilities.
|
FINAL
CAVEAT: Don't forget to write a "Close Out" report.
|
.
"Those who
do not remember [and document] the past are condemned to relive it"
Very frequently,
companies make the fatal mistake of not permitting the team, who solved
the problem, to document its major experiences and how it managed to
overcome them. Instead management quickly reassigns these experts to put
out other "fires".
This type of
behavior is almost certain to backfire. The expression "those who
do not remember [and document] the past are condemned to relive it",
has been painfully extracted from real life experiences.
The report should
be an objective and accurate summary - but without pointing fingers - of
the problem and its solution, lessons learned, expended resources and
expected benefits. Its typical length should not exceed five pages.
DISCLAIMER:
The
above contents are provided for informational purposes only. These
materials may not be copied or reproduced without express written
permission from TechnoBusiness Solutions. Copyright 1990.
|
