Revolutions In
Management
Shoji Shiba
Alan Graham
David Walden
Productivity Press
Portland, O regon
All rights reserved. No part o f this book may be reproduced or utilized in any
form or by any means, electronic or mechanical, including photocopying,
recording or by any information storage and retrieval system, without permission
in writing from the publisher. Additional copies o f this book are available from
the publisher. Discounts are available for multiple copies by calling the Sales
Department at 8 0 0 -3 9 4 -6 8 6 8 . Address all other inquiries to:
Productivity Press
P.O. Box 13390
Portland, OR 97213-0390
Phone: 503-235-0600
Fax: 503-235-0909
E-mail: service@ppress.com
02 01 00 99 98 97 96 12 11 10 9 8 7 6
Contents
Publisher’s Message xi
Foreword xv
Preface xix
v
VI A NEW AMERICAN TQM
5. Reactive Improvement 73
Identifying the Problem 75
Standard Steps and Tools 85
The 7 Steps: A Case Study 87
The 7 QC Tools 101
Afterword 559
About the Authors 563
Index 565
Publisher's Message
xi
xii A NEW AMERICAN TQM
xv
XVI A NEW AMERICAN TQM
Ray Stata
Chairman, Analog Devices, Inc.
Thomas H. Lee
President, Center for Quality Management
Professor Emeritus and Senior Lecturer
Massachusetts Institute of Technology
Preface
xix
XX A NEW AMERICAN TQM
component technologies all have existed for some time, but only
have a revolutionary effect when joined together.2
We believe that companies cannot succeed in the long run with
out systems and practices that support customer focus, continuous im
provement, total participation, and societal networking. All four
revolutions are necessary; absence of any one is an obvious
major strategic weakness. We formulated the four revolutions
based on Professor Shiba's experiences in Japan, validated them
through observation of Baldrige- (and Deming-) winning Amer
ican companies, and validated them through experiment (suc-
cessfully, so far) through member companies of the Center for
Quality Management implementing TQM based on this view
point. Indeed, the materials of this book are the basis for those
implementations.
pared relevant case studies and helped Shoji Shiba organize and
give the course.
Throughout this book, examples from Japan are cited, espe-
dally from winners of the Deming Prize and from JUSE (Union
of Japanese Scientists and Engineers). Other examples are
winners of the American Baldrige Award and members of the
CQM. The authors have used examples from these sources be-
cause those were the examples they had available to them. In ad
dition, Shoji Shiba drew on his wide experience of bringing TQM
practice to Western countries, and many anecdotes of his are
cited in the book. In its description of TQM, the book does not
take the viewpoint of any particular institution or individual;
rather, it synthesizes successful practices used around the world.
We have focused on issues of general management within
organizations using TQM, rather than on specialized skills such
as those of a quality assurance director, a design assurance di
rector, or a reliability engineer. It is an axiom of TQM that total
quality improvement starts from the top and must have the
commitment of senior management. Therefore, we direct our at
tention to general management concerns first, and treat well-
known quality control and assurance aspects only very lightly.
Many books have been written about the concepts of TQM,
but few exist that describe the practice of it in detail. Here, we
have focused on the practice of TQM methods, with anecdotes
and examples from actual experiences of people using these
methods. The book's purpose is to provide commentary and ex
planation of TQM concepts. It also can be used as the basis for a
course on TQM for senior managers and as reading material for
those studying the topic.
In many cases we focused on what was "a half-step
ahead” 一 not merely on the present, but always on concepts
and practices directly tied to the present. In selecting topics,
thinking a half step ahead meant emphasizing practices that
have been adopted in state-of-the-art companies but are not
yet in common practice worldwide. Chapter 1 discusses this
xxii A NEW AMERICAN TQM
Aggressive Learning
Moving Forward
(2)
Knowing something is quite different both from having ex
posure to information and from possessing actual skill. IAKI
must not become an excuse to tune out information and aban
don the acquisition of skills.
sense that they are widely practiced in Japan and seldom prac
ticed in the United States. Some practices seem foreign to people
because they differ from their habitual styles of analysis and
problem solving. Some practices may seem foreign because the
examples given are from industries, functions, or situations
other than the ones they work in. Many people proclaim their
company’s situation to be unique, but this assertion is seldom
true. Students of TQM must make a commitment either to avoid
NIH or to abandon this attitude if they now hold it.
Prove It to M e (PITM)
NOTES
1. What Is Total Quality Control? The Japanese Way (Englewood
Cliffs, NJ: Prentice-Hall, 1985), Chapter 6.
2. This general pattern was articulated by Graham, with
examples in software development and airplane de
sign ("Softw are Design: Breaking the Bottleneck,”
IEEE Spectrum (March 1982): 43-50).
3. We have derived this model of learning from Matthew J.
Culligan, C. Suzanne Deakins, and Arthur H. Young, Back
to Basics Management (New York: Facts on File, Inc., 1983).
Introduction:
THE EVOLUTION
OF QUALITY
The Evolution of
the Quality Concept
Fitness to standard
Fitness to standard
lower upper
limit limit
lower upper
limit limit
Fitness of cost means high quality and low cost. These are
the two most universal requirements for virtually all customers,
products, and services. To achieve cost reduction while main
taining high quality (with no products outside the bounds), you
must reduce the variability of the production process, so that all
units produced are already within the inspection limits and
none have to be discarded (see Figure 1-2).
• D Delivery
• S Safety
• M Morale
At Motorola, two driving forces are considered to create
customer satisfaction:
1. Cycle time (for example, time from beginning to end of
product development, time to ship a product, or time to
make a product)
2. Normalized process defect rate (errors per million op
portunities to make errors — for example, in insurance
claims handling, customer order taking, semiconductor
fabrication, or '"beeper" manufacturing)
For several organizations (Motorola, Xerox USMG, Yaesu
Book Center, Florida Power & Light), the most important as
pect of quality is customer satisfaction, measured directly by
survey. Some organizations, such as 3M, make product innova-
tion a quality to be improved or maintained, with explicit cor
porate goals.
However, a w eakness remains. Com panies that have
achieved the quality level of fitness of cost are producing highly
reliable, functional products at low cost. But competitors can
create similarly reliable and inexpensive products. Newly in
dustrialized countries can copy their skills of fitness to standard
and fitness to use but have much cheaper labor, yielding low
cost. This happened to Japan in the 1980s. Korea, Hong Kong,
and Taiwan adopted Japanese technology, but had labor costs
that were only one-half or two-thirds the Japanese cost.
For this weakness, the cure that leading Japanese compa
nies began pxirsuing in a standardized way in the early 1980s, is
creating innovative products that will outsell competitors, prod
ucts. Tlus raises product quality to the next level.
Chapter 1: Evolution of the Quality Concept 11
QUALITY CONCEPT 4:
FITNESS TO LATENT REQUIREMENT
NOTE
15
16 INTRODUCTION
ca. 1950s
ca. 1960s
ca. 1970s
ca. 1980s
EVOLUTION OF METHODOLOGY
Company Customer
QUALITY OF CONFORMANCE
VERSUS QUALITY OF DESIGN
Standardization:
C
1 Initial
Statistical
process 、 *
Market Deviations
Adherence to
requirements: Fitness to use
standards Orientation
Market research tools
Reactive
improvement:
7 QC tools,
7 steps
Weakness
Orientation
QFD, 7 management
and planning tools
FUTURE FITNESSES
Region/industry
Organization
Work group
Individual
NOTE
MARKET-IN
Unforeseen
rapid change in
diverse areas
CUSTOMERS
NOTE
MANAGEMENT BY PROCESS
i------------------------------------------------------------------------------- 1
implement them, analyze the result and its causes, and cycle
around again. TQM applies this scientific approach both to im
proving a company's process for anticipating its customers,
needs and also to improving the capabilities of individual staff
members or groups of staff members. Throughout the rest of this
book we will show examples of process as the means of making
learning and improvement more efficient, simplifying communi-
cation, reducing variability of diverse results, building intuition
and creativity, and improving ability to predict the future. TQM
itself can be thought of as a process for helping a company to
learn and improve — to change noncontrollable items into con
trolled or controllable items.
The theme of management by process, and more generally
the process focus, recurs throughout TQM. Managers need such
a viewpoint to understand and improve their own daily work.
And, as this chapter suggests, managers must be able to see the
improvement process as a process in order to provide guidance
and support to subordinates engaging in improvement activity.
The manager’s job is to treat improvement as a problem-solving
process, as described in the rest of this chapter.
Systematic Improvement
Level of
Thought
Level of
Experience
Chapter 4; Improvement as a Problem-Solving Process 49
Choose specific
Sense improvement Plan Standardize
Level of
Thought
Level of
Experience
are many points outside the control limits. Suppose, even if the
worker corrects the process according to the process manual, it
repeatedly produces results that are out of its control limits.
Tliere is obviously something wrong with the process.
In this case, the worker must take data, analyze it, find the
root causes of the problem, and implement appropriate counter
measures. In other words, the worker reacts to a specific prob
lem by using a problem -solving process to make the
improvement — hence the title of this section. For this case,
TQM has a specific standard methodology to follow.
7. Reflect on
process and 6. Standardize
next problem solution
7. Reflect on
process and 6. Standardize
next problems solution
7. Reflect on
process and 6. Standardize
next problem solution
Explore Formulate
2. Collect
r\
3. Analyze Implement
L
5. Evaluate
situation problem and analyze causes solution effects
data
Level of
Experience
Iterative Improvement
higher
quality
you can try an improvement and get real feedback regarding the
direction and distance to targets or goals. It is important to get
improved products or services rapid[y to market or in the hands
of the next process, in order to get this user feedback. In addi-
tion, PDCA is a system for making continuous improvements to
achieve the target or ever-higher performance levels.
The PDC入 cycle is always shown as a circle to indicate the
continuous nature of improvement. All types of improvement
and improvement maintenance require iteration.
PDCA is a method for dealing with all sizes of problems,
including some that may be beyond the scope of the 7 steps. As
will be discussed in more detail in Chapter 15, TQM explicitly
addresses the differing methods needed for incremental im
provements and breakthrough improvements. We call these
PDCAa and PDCA2 (see Figure 4-9).
PDC A1
PROCESS CONTROL
Voice of
the process
Process control
Act to return to
standard or to
initiate reactive
improvement = A
>
Reactive improvement
Proactive improvement
Figure 4-11. Interaction between the SDCA Cycle and the Reactive
Improvement Cycle
C/wpfer 4: Improvement as a Problem-Solving Process 67
CHECK
DO improvement
improvement results
PDCA cycle of
improvement
work ACT to
PLAN standardize or
improvement replan
to rigid
process
Japanese
NOTES
73
74 CONTINUOUS IMPROVEMENT
Level of
Thought
Level of
Experience Data 1 Data 2 Data 3
Control -
Reactive
Proactive
Identify problem
(including weakness orientation)
Run PDCA
▲ Apply standard steps and tools
Management diagnosis
i________________________ _____________________
Develop skill
How does one identify the area of the problem for a reac
tive problem solving effort? The 7 steps do not explicitly include
the part of the WV model at the left side concerning sensing a
problem, exploring the situation, and formulating the problem
to be solved. But these steps 丑re very important. Therefore it
falls to managers to guide a quality team in identifying an ap
propriate problem. (Otherwise, quality teams may take a very
long time, fail to solve the problem, or perhaps spend time solv-
ing an insignificant problem. Most failures of quality teams
occur because managers don't know how to guide the process,
especially step 1 — identifying the problem.
Identification of the problem is the most important aspect
of reactive problem solving. The process of identifying the prob-
lem may be divided into four parts:
76 CONTINUOUS IMPROVEMENT
• weakness orientation
• problem exploration
• careful selection of theme
• clear statement of theme
Weakness Orientation
Problem Exploration
excessive cost
k customer
dissatisfaction
Area of Problem
Human
Equipment Methods Behavior
Yourself 1 1 3
Data collection
or Yourself
potential solution and 2 2 4
implementation others
Others 3 3 4
There are two dimensions in Table 5-1: Can you collect the
data and implement the potential solutions by yourself, or do
others have to do it? and, Is the problem area equipment, meth
ods, or human behavior? The areas marked 1 are easiest for the
beginner to work on. The second and third least difficult areas
are also shown. Working on the human behavior of others is
very difficult, as shown. You can use Table 5-1 to evaluate the
difficulty of the situation in terms of both difficulty of data col
lection and difficulty of potential solution.
Next, having understood the difficulty of the situation, you
might construct a theme selection matrix, which prioritizes prob-
lems for solution in terms of difficulty of the situation (data col
lection and potential implementation), urgency or impact, and
Chapter 5: Reactive Improvement 81
Problem 1
Problem 2
Problem 3
Problem 4
Fill in the matrix in Table 5-2 for each problem being con
sidered. Vote 1 (high), 2 (medium), or 3 (low) in each box. Add
across the rows and put the total in the priority column to pick
the first problem to work on, that is, the problem with the low
est total.
Example 1
Example 2
ume customers."
This is not a theme, however; it is a problem. To convert
this problem to a weakness orientation theme, write:
''Decrease misforecast of 3- to 5-year sales to big-volume cus
tomers
The past may not predict the future, but it is all the data
you have, so it is what you have to use.
Example 4
Example 5
7 QC Steps 7 QC Tools
6. Standardize solution.
Marge Hendriks
David Kneedler
Cheryl McGee
Paula Nowell
Doug Smith
Anne Spagnolo
Julie Vinnacombe
Slide 5-1.
Integrated Circuit
Fabrication
Trim, Assembly
Wafer
-► probe, -► into -► Test
fabrication
separate packages
□ □ □
□ □ □
□ □ □
□ □ □
□ □ □
silicon wafer chips packaged chips
Slide 5-2.
Chapter 5: Reactive Improvement 89
Slide 5-3.
90 CONTINUOUS IMPROVEMENT
Slide 5-4.
Chapter 5: Reactive Improvement 91
Slide 5-5.
; \
Broken and Scratched Pellicle
Data Collection Check Sheet
Date: --------------------------------------
Operator ID# _______________________
Machine # _______________________
Time SL7PA notified ______________________ SL/PA initial
Serial # ___________________________________
Operator inputs: ___________________________________
Location of scratch:
Corrective action:
Slide 5-6.
75%
0%
Pellicle Breakage Cause
Slide 5-7.
3/ 16/1991 — 5/ 14/1991
100%
75%
0 %
Slide 5-8.
94 CONTINUOUS IMPROVEMENT
the mask from the machine (over 70 percent); the next high
est was when the aligner was loading the mask into the ma-
chine (20 percent of occurrences), and the rest of the causes
could not be determined exactly. Thus, our conclusion was
to study why the pellicles broke during unloading.
unloading
pushing mask in
thin plastic ^gl«
lie of mask loader
Man/Womanl
basic _____ l______ \^ | ^ aia
did not ioa(
load
properties
reposition
required worn
mask
holder
Slide 5-9.
C h ap ters:Reactive Improvement 95
12"
10 +
Number of
Pellicles Broken 8
by Machine
During Unloading
6--
2__
0■— White ~ 1
Mask Holder Type
Slide 5-10.
96 CONTINUOUS IMPROVEMENT
the two holders and found that the white holders had
slightly less tolerance than the black holders.
Solution Matrix
Mask
Finger Holder Other
Solutions Alignment Alignment Factors
Tolerance
Slide 5-11.
Chapter 5: Reactive Improvement 97
mask holder
alignment guide
(right side)
c H older
M a sk Lo ad e r A rm (Top View)
Slide 5-12.
Slide 5-13.
Solution Implementation
Dave Kneedler Get first guide made. 5/22/91 Get commitment from
Sonny and George.
Anne Spagnolo/ Inform all shifts of 5/29/91 Anne will inform the
Doug Smith pilot program. other supervisors and
send a PROMIS
message to Photo Mail.
Doug will inform all
Photo technicians.
Anne Spagnolo Designate all black 5/29/91 Julie will inform all shifts
Julie Vinnacombe holders on PE #5 and post a sign on PE’s.
and PE #8.
Slide 5-14.
Before $52,800.00
After $3,200.00
Savings $49,600.00
Estimated annual cycle time lost:
• Before 528days
• After 32days
• Days saved 496days
80%
70%
60%
50%
40%
30%
20%
10%
0%
dling
Slide 5-15.
when, and how it would be done. [See Slide 5-16.] The team
decided to reset its specs for mask holders so that all would
be of the black type in the future, and to install guides on
allremainingalignermachines.
Slide 5-16.
we did not coordinate well with the other two shifts. Third,
someone knowledgeable in mask fabrication should have
been recruited to help from the beginning. For our next
problem, we are considering working on the breakage
caused during loading of the mask. Loading failure is now
our most important pellicle breakage problem. Before we
attack this problem, however, we will examine other prob
lems within wafer fab to see if there is a more important
topic to address with our team.
THE 7 QC TOOLS
Check Sheet
Stratification
Pareto Diagram
Cause-and-Effect Diagram
Graphs
from the center. The example shows that one of the products is
inferior in all dimensions. Of the other two products, one wins
slightly in performance and delivery, and the other wins slightly
in cost and reliability.
Control Chart
Histogram
Scatter Diagram
NOTES
107
108 CONTINUOUS IMPROVEMENT
1 Select theme
3 Analyze causes
5 Evaluate effects
6 Standardize solution
7 Reflect on process
(and next problem)
110 CONTINUOUS IMPROVEMENT
the team and the team members, to identify who undertook this
improvement effort.
Each department says, y/We are different, and we need our
own version of Pareto." However, a common language through-
out the company is needed; TQM is a mass movement that re-
quires discipline.
Weakness Orientation
Schedule Included
frequency
$/unit
Was the data taken so that unusual factors don’t cause mis
leading conclusions? Are an adequate number of samples
taken? Are they taken far enough apart in time to represent in
dependent samples (in the statistical sense)? Was there anything
unusual going on while the data was being taken (e.gv recovery
from a fire)?
Appropriateness of data is a significant issue for improve
ment activities. Be careful about using data collected for process
control purposes (data 3), because it may not be appropriate for
reactive improvement (data 2). Definitions can be skewed, and
errors in collecting data continue for long periods of time if no
118 CONTINUOUS IMPROVEMENT
one has actively tried to make use of it. Data taken before the
improvement activity started is suspect — not necessarily
wrong, just suspect.
4th why
Why? Answer
Understandable Conclusions
In some cases, the team may think that the solution to the
most important root cause is too difficult or beyond its author-
ity to implement. In such cases, it makes sense to choose the
root cause next in importance, gearing the solution toward
that. Still, however, the most important root cause and its po-
tential solution should be reported to the appropriate part of
the organization.
Confirmation of Improvement
\W \ \\ v ❖ VVx
Before After
Rough Model for Printing Costs Rough Model for Improved Printing Costs
Facts of Standardization.
ANALOG
DEVIC ES The Errorbusters QIP Team
Jean Dell'lsola
JGdnne B6rni6r
Pam Mayernik
Nicole Rose
Josephine Hughes
Howard Kane
Jim Johnson
Archie Bryanos
Steve MacKay
Bob Spieth
Al Perry
Ira Moskowitz
Sandy Johnson
Slide 6-1.
□ ANALOG
DEVIC ES
Slide 6-2.
Chapter 6: Management Diagnosis 129
Pareto diagram that of all the areas in the wafer fab, the photo
room had the highest incidence of incorrectly processed
wafers. (See Slide 6-3.) Since this affects yield and thus cost,
the team demonstrated that the photo room was the right
theme to work on.
When the team first began to collect data, it found that the
existing forms used to report mistakes in wafer fab were not
useful. Moreover, operators were reluctant to provide informa
tion out of fear that the data could be used against them. The
forms were traditionally completed by their immediate supervi
sor. As can be seen on this actual form (recopied for clarity on
Slide 6-4), the operator stated that the form is unfair and intimi
dating; the engineer's suggestion was to just "'follow the spec/'
Slide 6-3.
130 CONTINUOUS IMPROVEMENT
Date: 12-12-88
Operator Comments:
From now on Til have the engineer check CB lots. I do not want to sign
this form because the lots could be reworked. If they were scrapped 一
thafs different. Should allow for human error which can be reworked —
I don't think it's fair. This is very intimidating.
Engineering Recommendation:
_______ Follow the spec … please!_____________
Sign Off
Area Supervisor: J .R. Operator:
Area Engineer: P Production
Manager:
Engineering
Manager: ----------
Slide 6-4.
Chapter 6: Management Diagnosis 131
The team decided to collect its own data and make the
form more useful and friendly (Slide 6-5). In addition, it ob-
tained a commitment from management that errors would not
be held against operators unless the error rate was excessive.
V ^
Errorbusters SOS
What Happened:
Used wrong mask, ran lot #F537-l-9843 and then lot #537-1-9840.
The device numbers are dose.
We should have a "mark" on the masks when the device types are almost
the"same"— ► "F537-1" "537-1〃 ________________________________
This rework could have been avoided!!
V _________________________________________________________ J
Slide 6-5.
132 CONTINUOUS IMPROVEMENT
When filling out the new form, an operator on the team sat
down with the other operators involved in the incident, dis
cussed the incident, and jointly wrote comments. The form was
then sent directly and in confidence to the production manager
(that is, around the direct supervisor), so the operators would
not have to fear recrimination for their comments.
With data from their new forms successfully coming in, the
team was able to focus on leading types of problems. One cut of
the data (Slide 6-6) shows the team found that wafers getting on
the wrong track at the coat step and wafers receiving incorrect
alignment at the align step were leading causes of incorrectly
processed wafers in the room. The team moved on to step 3.
Slide 6-6.
Chapter 6: Management Diagnosis 133
Slide 6-7.
3. ROOT CAUSES
\ _________________________________________________ )
Slide 6-8.
134 CONTINUOUS IMPROVEMENT
/ \
4. PLANNING AND IMPLEMENTING SOLUTION
Kanbans
• “KANBAN” ( signal)
- squares with labels (bake, coat, develop) — OR —»
- racks numbered (inspect)
Slide 6-9.
Chapter 6: Management Diagnosis 135
The team recognized that the use of the JIT principles could
decay over time unless the team performed step 6, standardiza
tion. As shown in Slide 6-12, the team instituted JIT meetings
every shift to discuss problems or improvements for the system,
with minutes of these meetings published daily. It also handed
over the system to the supervisors and operators in the room to
run, and agreed to meet only onc^ per month as the Error-
busters to monitor progress and ensure that the gains were
being held.
136 CONTINUOUS IMPROVEMENT
Slide 6-11.
Slide 6-12.
• Review of Process:
Team focused on too many Pareto items at the same time during
Step 2. If team had focused on problems one at a time, the
solution of each one would have been faster.
• Future Work:
_
V___________________ _____________________ J
Slide 6-13.
138 CONTINUOUS IMPROVEMENT
2. Collect and • The decision to collect • The first slides of step 2 are
analyze their own data is the best not numbered.
data point of this Ql story [you • The Pareto diagram on Slide
want to collect your own 6-6 has spaces between its
data and the diagnosing bars and the cumulative
manager should make a distribution curve is not to
point of this]. the same scale as the bars.
• A Pareto diagram is used.
5. Evaluate • A run chart showing improve • The run chart is not convincing.
effects ment is included. • “Before” and “after*’ Pareto
diagrams are not shown.
• A pilot test may not have
been run.
6. Standardize • The team has worked hard to • The solutions are based on
solution standardize the improvement. changes in human behavior
• They have used nice graphics and will be hard to maintain.
to show the various parts of
the solution.
6. Standardize solution.
Act
7. Reflect on process (and next
problem).
NOTES
1. See, for exam ple, G.E.P. Box, J.S. Hunter and W.G.
Hunter, Statistics for Experimenters (New York: John
Wiley & Sons, 1978); and R. Gnanad Esikan, Methods for
Statistical Analysis of Multivariate Observations (New York:
John Wiley & Sons, 1977).
2. The case study was provided by Ira Moskowitz, Produc
tion Manager, Wilmington Wafer-fab, Analog Devices,
Wilmington, Mass.
7
Proactive Improvement
141
142 CONTINUOUS IMPROVEMENT
Level of
Thought
Level of
Experience Data 1 Data 2 Data 3
■Control--------- ►
Reactive- -►
Proactive •
Something is weak
about the process!
Use reactive
improvement to
eliminate weakness.
reality
weakness
plan
Traditional m arket
Principles for collecting type 1 data research m ethod
Intuitive capability: logic may tell you certain Objective processes, e.g.,
data are unimportant, but if intuition says statistical summaries
otherwise, then they are important — human
4 intuition has great capability to find something
new — for instance, something the customer
is doing may be logically irrelevant, but may
actually be the key to something new
2. Stepping-Stone Approach
3. “By Chance”
f
k C C T )-
Act to return to standard
or to initiate reactive
improvement = A
Have
standard = S
Evaluate/
Use do
check
standard = D
effects = C
7. Reflect on
process and 6. Standardize
next problem solution
1. Select 4. Plan
theme solution and...
Level of
Thought
Level of
Experience
6. V ote on th
low -lev el issues and draw conclu
sions. Once the team has reached a
common understanding of the in
dividual facts, their grouping and
hierarchy, and the relatio n sh ip
among the groups, the team votes on the most important low-
level facts (one must think of action in specific, not abstract,
terms). From the im portant low-level facts a conclusion is
drawn. Finally, the team decides what next steps are appropri
ate, given the outcome of the KJ.
7. Reflect on KJ method
process (and
next problem).
MPM
Stage Steps
Warm up
Preparation
Discuss theme
Choose labels
Unconstrained pickup
Count labels chosen
SEMANTICS
Sale spe o ple lie to custom ers to get S alesperson prom ised delivery
orders. without confirming w hether the
[W hat evidence do you h ave of product could be delivered on tim e.
this?]
The translations on the right side of Table 7-5 can still be im
proved (as will be seen shortly), but the first large step has been
taken: translating emotionally charged affective language to the
more objective language of reports. Making a legitimate transla-
tion requires the collaboration of the person making the state-
ment. No mechanical translation will stiffice; a mutual learning
process takes place when one person translates the words of an-
other. Especially when affective language is being removed, the
refinement of language data is like a game of "find the hidden
fact/' A rich body of experiences is revealed, even in the re
stricted domain of a person^ statement. The mutual learning
process involves discovering how to express parts of that expe
rience in a simple and purely informational way.
For example, to design a product to satisfy a customer, first
get the person to verbally describe his or her image of the prod-
uct. This verbal expression is likely to be in emotional or affective
164 CONTINUOUS IMPROVEMENT
language ("I like fast red cars’’). Next, this verbal language must
be converted into measurable parameters that can be used in de
signing and building a tangible product. That is, the customer's
verbal data must be converted into detailed specifications of
components that the customer wants. These specifications are
written in the language of reports.
Similarly, if a machine operator remarks that ''this machine
is awful to operate," or if an executive says, "I'm uncomfortable
with the competitive analysis portion of this year's strategy doc-
ument/7 you would use semantics concepts to convert this emo-
tional language into the language of reports. Only then could
you improve the machine design or operating process. Use se-
mantics concepts to convert affective data into data that precisely
represents reality, and describe reality in the language of reports.
Semantics concepts can also bridge from the language of
reports to clear images in affective language. If the language of
reports says that "we must decrease defects by 68 percent per
year from 1988 to 1992 to catch up with our competitor// then
you can say to the staff in affective language, /7Six Sigma by
1993/,a0 If the detailed design of a new car is expressed in the
language of reports, it can be converted to affective language or
to an image that can be used in marketing the new car (a sleek
red car streaking, top down, across the plain under a clear sky
with a beautiful couple sitting in the front seat, their hair blow-
ing in the wind).
Ladder o f Abstraction
hot/cold scale
hot cold
Chapter 7: Proactive Improvement 169
Fahrenheit degrees
1111 1111 1
1111 1
1111 1111 IIII 1111 1
1111 1
1111 1
1111
0° 10° 20° 30° 40° 50° 60° 70° 80° 90° 100°F
Plan
Lost/ U.S.
Key/ International
Company Unhappy/ Government/ Multidivision TQM Distributor
Target (Non-U.S.)
Good Commercial
A key good commercial yes
Systems
Visit Communi Software
and Advanced
Company Assigned cations Products Computer
Technical
To Division Division
Division
A Levy X X
B Walden X X
C Glabe X X
D Rampe X
E Ide X X X
F Rankin X
G LaVigna X
H Barker X X X X
1 Goldwasser X X X X
174 CONTINUOUS IMPROVEMENT
Do
Check
Back at BBN after the visit, the verbatim notes were tran
scribed onto KJ labels, with one thought per label. Typically this
resulted in 50 to 200 labels. Then a group of people familiar with
the customer (typically the relevant cross-functional customer
support team) used MPM to pick out the most important voices
from the large number of customer voices collected, and orga
nize them into a KJ diagram. This process of picking out the
most important customer voices and structuring them into KJ
diagrams is shown at the left of Figure 7-9.
Two parallel courses of analysis were followed from here:
dealing in the relevant division with specific issues brought up
by customers, and finding the cross-customer, cross-company
issues. Although it is not shown in the figure, dealing with cus-
tomer-specific issues in a specific division typically entailed
using MPM again on the KJ diagram for the customer. This step
Eelllleasolsueos
l ue ueEali
Aoy lPOM
nboJOECnsno Aeyl ue Jepuno >l p
sr ce sds £ sn z£36
-M
-o
oc
TJ
ts
o)
is
.6 —
11
c)
a)
.E
2 A —i s
「
X◄—l
/
M—
CONTINUOUS IMPROVEMENT
e JO
u o p 2 A —2
M-
①p
一
dl/
M
>
o
D
D
5
Q l swu ui
a
c
(
(
2 A l ls
5 Al i s
2
!s>
j
Jo uls sn o
2 2d2
A —
>
\J
}
Q
>l l
.
>
)
SUB E
a
c
2 Alis
I s
/\ I s
5 Al i s
\/ t - i s
T3
}to
S P U B U i9 J e lu o no
5
5
ol u ui
)
)
^sl
to a
a
rt a
T
/
(
176 .LUJO u
寸 1 Jg
10
Chapter 7: Proactive Improvement 177
Act
5. Learn PDCA
and then jumping back out to reflect on what was seen and
heard (see Figure 7-13). These TQM practices define systematic
processes for exploration prior to creation of a hypothesis.
The best method for using the fishbowl approach is obser-
vation: it shows the product in context, it shows you the cus-
tomer’s voice in context, and it increases your sensitivity to the
customer's requirements. If you visit the customer, ask to see
the product in use where the customer really uses it. Often it is
difficult to believe customers' answers, but you must recognize
their validity; you must see and understand what is actually
happening.
NOTES
189
190 CONTINUOUS IMPROVEMENT
Customer
image
Customer
statement
making
customer product and
requirements product process production
-► -► -►
visible and design development
explicit
Concept I
Operationally generation and
defining the ’ 's e le c tio n
/ voice of the
customer
i
Stage 1: Stage 2: Stage 3:
Develop an Convert Operationally
understanding of understanding define
customers* needs into requirements requirements for
and environment development
relations to
company and
products
relation to
trends
market
segments
position current offerings and predict future needs. The BBN cus-
tomer visitation case study presented in Chapter 7 provided an
example of listening to customers from a variety of market seg
ments and in a variety of states of satisfaction. The stripping bas
ket case study presented later in this chapter provides an example
of predicting future needs by listening to lead users.
Eric von Hippel of MIT observes:
Users selected to provide input data to consumer and
industrial market analysis have an important limitation:
Their insights into new product — process, and service —
needs and potential solutions are constrained by their real-
world experience. Users steeped in the present are, thus,
unlikely to generate novel product concepts that conflict
with the familiar.4
high
interaction
with user
phone call
customer visitation
to user customer visitation
with management
. with users
open-ended inquiry
contextual visit to
inquiry homemaker
process o b s e rv a tio n ^ ^ )
gathering
evidence of
human antenna shop user behavior
performance lab 丄 i ethnography
low
interaction CT participant observation
with user
• you still only get what they remember, and you don't
see them in their real work environment
3. Laboratory work (for example, watching users through a
one-way window as they attempt to use a product)
might seem a logical method of observing and dis
cussing the user's work and end results, but
• the work done in such performance laboratories is
often not real user work, and
• the work is not done in the users’ real work environ-
merit
Figure 8-5 indicates some of the environmental and real
work issues, such as interruptions and purpose, that typically
are missing in the laboratory situation. When such environmen
tal and real work issues are not present, companies are unlikely
to understand the needs of the customer completely, especially
the implicit or latent needs.
Interview Script
he been in touch with you about this yet? We would like fif
teen minutes of your time to ensure that our design criteria
are the right ones. Is this a good time for us to talk? (If not,
emphasize that we'd like to have a prototype by May.)
i
r>l i
eEI J E c n s n o >lsef f d ds s .9-8 j3 6
0>
o>
o>
D)
a>
u
.E
<
:
s
•A/ve «
>
(D
M In o a L Il
elpe
J qM®u u
o
a
a
)
)
o
lu
luj
lse
o
ffl
tf
nsuzs M A U IO ld n
^ .
.l s ! i JnoA 〇
一dn )
o
f Bs
a
fJs M I l u!
q o
c
E ou p M MO 95 M
su
enl
cuo
lo
oMO
S
(
X1
^
ra
ocl B i〇
o
e
r
O eld EOJI -u
一
o
JJ
)
U sds u!l§ 8
-
s Ss2
d !
u
vi
(
^r
u
o
*c
a
(a
• 6u!puew
c
>
J 9u>s
.
P S.IB3 ^ s
iuojL
psqpa
pnl
o a
s^
*
U SBO0 1 6p Nv
(
l<D
lq>
l.£
IZ
PIJ
nst
sul
lloJAIluso
X IA IIO 」 >.l8>jseq l0 1 ■punoJB
-
〕
co
TJ
(o x:
I
a)
J
s u Ljllloq 's j
J
.E
A ujlee
i
o
.S
Jno
a
1S lM
•
AM9
fe6
uoPUB 6ulddss *LLU JnoAJ p u n p o
a>
CONTINUOUS IMPROVEMENT
cc
lI
Q s n seo 6u BH〕
I 1 . l> l
15l
l5 l
}
^ 7
— MJ
>l
>
a e usloo
sell % ^
o
.
Lls
yqoleool
paij
nbej ■qs
.
j l o
o>
l-
Lo
9=
J
t3
7
l9
SJn
o u
!a
g .i s s e j dd JspurTY^
co
£
«=
ji
(o
OM6UIXHT
a
'E
slseo o ESO6 BS 厂 x
)
2I;
2>
(
•
ls
!.
ue 8
s8u
oa u
Bnd
Ineo 0 1«e c l o ln E oJo >o
o>
o>
a>
.Ms!j
ca
XI
JZ
as
(D
— d
)
U
OJM
s
)l
l .£ l
i
l o j
a
)
s LU _
*
E_u q
«
^ ^6 gl
*
(
la i
osS
MOL
IS
C .SPU033S
<0
•p 83 M B 9sw e> eM 8pli
IT O I
lt o l
IT 3 l
I j
^slsB O e g o q BLUOJ C O J nlTy^
>
>
D
.6!}J el! qos u!>i
a
s > loo
a
suqA > -
uui二u 【 -
o)
Ape
o 2
a
s
s
e
pu m i ou l J
c
pu oBS MO
£U _
i
i
(
JC
T3
o)
x}
Q)
/ "
..|«3 £ 0 8召 -
<1)
0>
.£ s l
Insa ^ ^
D
o
o
'D
a
ILU®nb^ u o u=J~MS ssg
C
c
o
)
)
a
D£ / ^ o i0 5 uo c i so p a^
a>
a>
b ii
*s
v
c
s l
s o
lseo
sSulleu
al
pl
l
.^
.£
S
i
一^c:ntoud
ntod
doJn
oA E
>
o
usu! UIMOO
I
t
-as l M 0
f
ll
o
jlJ£A >
x
1 eE!
l s
< r
f .
Jal
eM 6
uj
HJO
Msp
jj
q puooas
e l d
e£ u p-e u
>{
0 H=
O a}
a}
i2
a)
.A ULISi^Bq UIMOqs s
cn
JC
u:
售
C
•le
ils
Bqeqi ( 0 1 91U -8 | J
306
31 M8SU0
s co
J
co (
•J addl
4=
o
! u!IA E u| SB l5S
p.B q
q j
.E
ol oo
lsall l « l l
c
cnrodd i u oa noa
la )
s篇 OU2OO 3M q
l.g
ls
? el
luno 3u o o6u!sn o o
/r
l 4c e
>
>
U
J=
(0
u.
l<
.2
a
s
oH E x e£6u e d ol
su< uafMASJeueoi S
Jnl ee
-B
<D
<D
<D
u
!d
d!
Jfsjl
v=
t
0JB0 oj p8J!nb
y
»
ii=l
J=l
(ol
a}l
x :l
J
U0 j ! d e j diu e li
l.g .
l
a
l
l
l
:l
.£
5
~ ^ ! JO ! oi^
> sj sjlE ujll
)
)
)
o
d
x
u
a
l
l
l
l
I
206
uiqsl一 J 9 l e M } l e s li lefM
Z H e n s > 3 O A u e l l M ^ u i u l ol ulo o eu
c-'
o>
a>
a>
o)
ra
(o
jo seueos
Chapter 8: Proactive Improvement in Product Development 207
Here is an example:
Customer requirements:
• basket is positioned at end o f strip p in g m otion
• casting m otion is the same w ith or w ith o u t basket on
Translation:
( —) The basket has Velcro fasteners.
(+ ) The basket fastener can be released w ith one hand.
Translation:
㈠ The basket should be easy to carry when tran siting
to the water.
(+ ) The basket is easy to carry to the water.
Translation:
( - ) The basket is com fortable.
(+ ) The inner edge conform s to the body.
Common Issues
H—
8©>{s q
U i JO
A p o q o L2
-c
l〇u
us M uoeze
oituooMoqeEow
(o
(a
*H
(a
rt
JnoA
tr
I w o ^ js e q lo u o s s o d Lj/6elH0J Aui 0 b m u nM
»*—
HB odE
.9-
»«—
3=
tr
|
o}
T3
.s2
Alwnoowuooun i!
co to
o)
o)
lolu uld d u3« also dszeM BC
I s o JO》 _Als uz«
ull u L{;u c o p > ls q UH uBn d3oq > lu !£ lc o p noA o sLI n o u l
q
*«—
u 5s uo
o)
a)
o)
a}
(D
(a
o[J
<D
CB
o>
OJ
ca
o)
o}
a>
a)
a)
(o
o)
T)
•
.2
'o
'o
A ocl £2SUJ J l
£u o ol >s c
l lq ls c
5E O O zs > l u >elco p
d z J n o A U O i!l
«*—
a>
a>
<l}
a>
TJ
tr
a}
a3
(a
<D
<D
<u
CONTINUOUS IMPROVEMENT
I s m o A s q l s s o d o ui
o eld
} s m B t l q ! s s o d ou
!lI
a3
oOJd-lsm Su0> ls q
C5
to
o)
u{6©/VUM
J p o le a llw J 菅 荃 SBq £ u l i u o Bu c o e u « JP J seM
o B o s=jds
a)
£
«4—
rt
ca
.-tr
a)
(D
u Ja li
<D
a>
o)
MOL!
J luA n u u >fs c B o II
B lco wd o ol 0«
«4—
1 ei〇uzn d UH
(D
a}
T3
<D
ISBoHlun }8>jsBq £ u l S U IB lu al u!l
a>
cd
cd
-'
to
to
L u d 9 l3 = 9 - 2 ^
a}
a)
3ls e ru
uc e ^ c l s ®>l e AI Jn o l >l
s J 1幻 n oxoHOJI u!©>ls q d .u c
>nsn u l
.<2
u ls nu
«*—
a>
o>
a)
(D
JD
(a
a)
<D
is
(d
(a
JO
to
03
.2
a)
a)
(D
(D
(o
(o
父 o slc M O JC l
l Jln u/l Jn u j ou - EJ JOIOu.
u .
l lt i二
q > l co p uEJ su
t5
a]
<D
o>
0)
co
>.
cd
o)
(0
s| Joloo
x:
a}
_r
a)
(D
(o
S J O IO o c 0 9 u B u BJrnBU
eeJ l
UB ® ueu o us ;} u!t
<>1
oi
<D
o)
o)
*v
a>
SI0U H •
(
uil6ue}
d
AlueAewleHM
dlnq s ^ l3
u I • ll p u!T
ounlo us3 0 uzo u nci LUOJ u!l £ l AclaISLjrle》
■D
u d >
ca >t—
JQ
JQ
<l>
is as
a}
■D
<l}
<l}
a>
co
.E
o)
o)
's
.E
ln0ip u
3dJ s®>1 BC I •
Apoq0r l !
-uoNJogSX 1S l>s q 9 £ o s H l« PJO o
.g>-
Q.-
un?c
T3
(D
>.
(])
dm ou s le>lsecl
(o
(o
(o
0
<D
•
o)
03
(D
(D
l s-l se E UBOUUWJ P U B J s B e M S
jo E O S
■c
ca
JI
!
BJ M |s E l E O S u uo EI szleqos
a>
(o a>
a>
a.
ca
tr
(o
(a
szoq is e . u nq c o I
©>1SB C
q lqs s
nfp 0> p u3Elp U B •Isnfpv PJO o
3 S!AJ〇u > LH
»
»-
o>
ffl
.s2
ca
ca
(D
cs
(])
is
a)
(D
a}
a]
(D
lq snfpB
-
op l
e u EJ s o lq lou o o 9J sp n OJd f J
e v>o n aiBs®> S q OL U
■o
OJ
o)
a}
(o
(o
o>
o>
l PJeluuJOO ou
os
to
'o
•o
①E Jln b t n o i a >i 8 o > Jo E olsn o
a>
a5
(D
Ir
UJ
(
d
212 aqr0 90 > -8
>
uouelsil Jiolsno
:
0 2 1
(0
0
1
213
li
M oe
jn > 02 colp c c
ooUJSIoq
CD
JC
o)
(a
H ioM
<
u!u «el Jeu s o p >1
s q
(oQ>
o>
co
sz >{seq
a)
a)
(o
u o uLlP
s o d u zB s
s lo u <
a>
a>
>»
Q}
is
'D
Chapter 8: Proactive Improvement in Product Development
u a s e l o l A S Ss z e u < 9ls l
cc E }SAS u cs s
a)
o)
(D
i5
(o
suospuoo
u!u o
ss /su o
od sp u o oUIAJ > ol^l> epe
s u o s so c u u ejelllp u! U J O M q u o} >ls q l L u c n p ue L
P o d oxs q ① l
s!u os s
B.
o)
o)
05
Q>
o>
a>
a>
co
ca
Jc
(D
(o
a)
s IM @c= s
0 n ocoiloansco o il.{
l
l UB I
cnecll /tAOII lous o l > SB s i s 1 „ > > o s„ o
itr
a}
4i
-o
(D
(o
(o
(o
(o
<D
a>
ca
o)
JO
(D
iAJes/p n O J d es OGIe am ul ICIOJd lu! ld u l q M o u Medxe
oo u o i l u a u lol u > le e MJ O lo o
c-.
<lr
U-
<fl
a>
a>
a>
a)
(B
(o
(a
n o A sn o M s ss J n o A U IO J
'o
's
l >lsq £A qp )s
«!
s al suopunl u llsollue£ JO E o lnoLJS ooIA
a>
03
a>
.®
ns
suolpunr
cd
ol
UILISIreluo npelqjssod
CD
(D
(D
'D
'D
'D
cl lou
u M oBdeAJe
cop
olno l >lsq 6u2oBd 0 e s s
**—
(/)
(D
x:
<D
(o
(o
x:
(D
(a
eMn
a}
e J o is o lA S B
Je
«M o
£o } ussu
4u qM AJJ oolAsee
sl©
>isq 0£ M
«3 cl jouc UIAJ J
Bolo s e
o)
(c
(D
cd
co
<D
<D
a>
ffi
a)
r o d s u e J lc n A S e
J s! UH •
»«—
a}
a}
(D
>is c
l e lu c lA J u£zs u!l L
UII cEz u!l® cluil®c n ul lsAS d 8>l s
co
a}
a)
<D
u o lB ls w e u ll •
o>
<D
a>
o)
ca
a}
c5
o)
a)
(o
is
(D
JO
.g
s e e u oJmn Jn o A s a
l pe lwul s a l r n e e lMo u
H—
c-
co
JC
la lIM
'o
'o
SSSO Jd
u o o E J3eu £ £ uldBs J £ o u s uio o !2ul
3c n
a>
Q}
(D
a)
(c
(5
euilouipsel
u e of u e l e y is B q !
luolclp o ldsl ul p s d d B s I
A IO
uzsfe o u e L 3B E
2solu o
J 一
EOJr Bo
<D
CD
TJ
J=
ln
S IA - U O 》
:!
a)
(o
OJ
to
(D
i〇u p lnolls^y lsBq
's
esn
*B
I
s u o s e o s BJOAO
S Lj n o lcl sl®>ISBC
J£ BI2 e I IS B Z C S 8 0 P SBAUBO
cd
ca
co
to
o)
slq Jnp s q lsnLUI IJ®Blu
6B £IM
ln o o UH• >
! S| xlil uouourv
0}
<fi -D
>*
co
ffi
to
a5
sl
i I A Me®
① Jv l ue ^ s o
ew UH !
Jl suolpunloISBCI uonpo jd o E OJ ou
08
o)
o)
a)
a)
p l i
to
•
214 CONTINUOUS IMPROVEMENT
to the voice and image. You then can im prove on the key item
and custom er requirem ent through ite ra tio n (see Figure 8-8).
visit customer
voices images
customer requirements
MPM
A lte rn a tive ly, the M P M could be done before the voices have
been converted. The firs t m ethod is preferable because it pre
supposes an understanding o f a ll custom er voices before any
are discarded as weak.
U nderstanding the key item in the custom er voice depends
on the context of the statement; sim ilar-appearing voices from
different customers m ay actually mean different things; this is
the reason fo r using the image KJ d u rin g the translation process
and fo r converting the voices in to custom er requirem ents before
e lim ina ting apparently weak voices.
The second m ethod, doing the M PM on raw voices, m ay be
necessary if there are so m any voices that interpretation o f a ll o f
them is im practical. However, it is not necessary to meet a great
m any customers, as was explained in our discussion o f the BBN
case study (see Chapter 7). A fte r exploring w hat was learned in
the in itia l few custom er visits, seek com plem entary in form ation
fro m subsequent custom ers. By d o in g th is successively, you
should be able to gather enough in fo rm a tio n w ith o u t v is itin g
too m any customers or getting too m uch duplicate inform ation.
a>
a>
c>i
oo
co
co
co
ra
ID
a)
il
.E
i2
Hi
■ ECOJj>c
>
D
a
(
O)
0)
(e l d E i s s ze v ls e g )
•5
IP S C P N ^
1
o u u B S IS a l
l IJs
oeJ lpleoj
( B e lt is adjustable)
iJ J Jo o
seli i o Iswxlidy
to
ca
(d
i
5*
v
a
!SI>,iM
c
l l s eE u
S
\ ^ t h e b o d y .^
He wu o ol
a
cd
to
l
l
u s uo
zl
-
)
( conforms to )
=
y ^ 0 s n OMU OS S Slej >e s
(o a}
H
co
ca
SOU SIJep J LP
n
①c u oB \N ! .
g
^ H in e r e d g 2 ^
l
V ol lq MZ > 1 g
r
i >
<l>
i
snoOISlEJ f
<D
a]
<D.
(o
V to wear.
o
uoj}o8d x
0J lucn o ^
D
V e a s y to detach^/ / ^ \
oc
c
o(
i
comfortable P \
Basket is r
n EsolIJop
n
J LIOo
o
lqj
!s!>
Basket is ^ y
c
c
t
t
tf
^LUOOSIPPUB JB9 M 0 1 ASB9
l
l>se
q d s
c>-
.s-
LU lqoJd ull luil Aq uz«l
leu
a>
co
—
o)-
o)
<A
<D
i:
.E
D)
a5
(D
.E
sze Bm
ig
us
ono o
cn noM
AO I Bplnoqsll
»
»
J s si a l ln b a
J J i c n s n o es i III
M
CO
co
fi
220 CONTINUOUS IMPROVEMENT
line is stationary
in basket
the line moves
only when desired line placed in basket stays
there until cast
customer
satisfied
product fully
functional
The five choices in the table are those used in the strip p in g
basket case study. The fo llo w in g alternatives, however, seem to
differentiate better among the responses:
1. I enjoy it that way.
2. It is a basic necessity, or, I expect it that way.
3. I am neutral.
4. I d islike it, b u t I can live w ith it that way.
5. I d islike it, and I can’t accept it.
Based on the responses to the tw o parts o f the question, the
product feature (how fast the w ater drains, in the above exam
ple) can be classified in to one o f six categories: A = attractive, M
= m ust-be, O = one-dim ensional, R = reverse, I = in d iffe re n t,
and Q = questionable. The firs t three categories were defined
above, and these are p rim a rily w hat we are seeking in the Kano
analysis. The other three categories indicate the fo llo w in g situa
224 CONTINUOUS IMPROVEMENT
Customer Dysfunction al
requirements ---------- ► 1. 2. 4.
3. 5.
1 like must-be neutral live with dislike
1. like Q A A A 0
Functional 2. must-be R I I I M
3. neutral R I I I M
4. live with R I I I M
5. dislike R R R R Q
Customer requirement is
A: Attractive O: One-dimensional
M: Must-be Q: Questionable result
R: Reverse I: Indifferent
about "w ater drains quickly," and " I d islike it that w a y" about
"w a ter drains slo w ly," you lo ok at the intersection of the firs t
ro w and fifth colum n and fin d an O, in d ica tin g that the cus
tom er view s speed o f w ater drainage as a one-dim ensional cus
tom er requirem ent.
1 23 4 5l
dislike it.
8b. If the line moves around in the basket, how do like it that way.
you feel? t must be that way.
am neutral.
can live with it that way.
1 23 4 5
dislike it.
9a. If line placed in the basket stays there, how do like it that way.
you feel? t must be that way.
am neutral.
can live with it that way.
1 23 4 5I
dislike it.
9b. If line placed in the basket comes out, how do like it that way.
you feel? t must be that way.
am neutral.
can live with it that way.
1
-
dislike it.
2
*
3
*
10a. If line in the basket is tangle free, how do * like it that way.
4
you feel? 5
-t must be that way.
1
- am neutral.
2
-
- can live with it that way.
3
4
- dislike it.
5
-
10b. If line in the basket is tangled, how do you feel? like it that way.
t must be that way.
am neutral.
can live with it that way.
dislike it.
11a. If line gathers naturally in the bottom of the like it that way.
basket, how do you feel? t must be that way.
am neutral.
can live with it that way.
dislike it.
11b. If line does not gather naturally in the bottom like it that way.
of the basket, how do you feel? t must be that way.
am neutral.
can live with it that way.
dislike it
12a. If the basket causes some drag on the line lik e itth atw ay .
during casts, how do you feel? t must be that way.
am neutral.
can live with it that way.
dislike it.
12b. If line casts from the basket without drag, like it that way.
how do you feel? t must be that way.
am neutral.
can live with it that way.
dislike it.
226 CONTINUOUS IMPROVEMENT
19a. If the basket does not hamper arm or leg 1. I like it that way.
movement, how do you feel? 2. It must be that way.
3. I am neutral.
4. I can live with it that way.
5. I dislike it.
19b. If the basket interferes with arm or leg 1. I like it that way.
movement, how do you feel? 2. It must be that way.
3. I am neutral.
4. I can live with it that way.
5. I dislike it
20b. If the basket is not easily adjusted to different 1. I like it that way.
body positions, how do you feel? 2. It must be that way.
3. I am neutral.
4. I can live with it that way.
5. I dislike it
QO
o>
oc
cn
u:
lo
s lIn s a A o A Jn o u e ^ l o c o u e l n q e l
SAeAJn u o un ei
co
JQ
JS
's
Chapter 8: Proactive Improvement in Product Development
人
cvi CO
A
1
〇>
cc
.ABMI !
cf
①
d
<
is !p
s
ueo
t
ABM llujiM9
>!l
•IB JJno uE B
JC a)
-ABM cnsn uu l
aj
l u o jl o c r ^ s x a
ul > l!
l E o
to
a)
•A B M
td
(D
•A B M llu l VIHSIP
A BAM BLIU ISIM3>l UBO
■IB4n uEe
a}
5. dislike
LX.
to
•ABM lBql q nuul
a)
IB U O ilo u n
I E o
•A B M B M U o
Ei
4. live with
3. neutral
Jlioils n o
a>
a)
2. must-be Functional
的《 〇
|
1 must-be I neutral 1 live with I
一1
2. 3. 4. requirements
Customer
D:
/sfunction al
le> o u >!
le n
l q e iu
0>
lil
(o
.2
228 CONTINUOUS IMPROVEMENT
Customer requirement is
A: Attractive O: One-dimensional
M: Must-be Q: Questionable result
R: Reverse I: Indifferent
one-dim ensional:
water drainage
corrosion-resistance
neutral color
line shifts
natural size loops
ease of adjustment
conformity to body
light weight
outer edge droop
product
interference with
functional movement
su 门 resistance
line fallout
position stability
customer dissatisfied
one-dimensional of cost
(“satisfiers”
) V. to use
must-be
to standard
("dissatisfiers”
)
not
necessary
Quality
variables
A B C D
®
weakly
Customer
A A covered
requirements
〇 〇
not
covered
\
basket can be adjusted
level motions during fishing
to fit the worker
\ \ Quality Measure
Measure how Measure
how well Test if the basket
2d basket fits adjustable how the
accomodates
level basket basket affects
against stripping motion
attachment is walking
body
c
>i
1 s
I ! f |
R equirem ents \ \
i
basket to shoulder
Measure the bend
〇 o©
|1 II 〇>〇
11 艺8
25
\ i 琶
is II
i | I !
1st 2d 3d \
level level level \ I I Is
\
Line placed in
when desired
moves only
the basket
Line is cast
1
W he required
line comes out
without drag
of the basket
Basket naturally
门
LL Bean
Surfcaster i
Other
Target value i |
_ i _ _
® = high O = medium = low (blank) = none
Chapter 8: Proactive Improvement in Product Development 235
236 CONTINUOUS IMPROVEMENT
KJ
D
/ El N G F
C
^ I I ^
=9 E F =9 E F G
i—*—i r 1 i B
A B C D
A B C D E
item in the custom er requirem ent table. Then a tree was con-
structed that classified the low est level groups by com mon p u r
pose. These groups were then grouped b y com m on purpose,
and so on. Finally, the tree was converted to a table.
As described in the exam ple above, after la y in g o u t the
q u a lity table, make an assessment o f the relationship fo r each
custom er requirem ent against each q u a lity m etric. In the strip -
p in g basket case study, an in fo rm a l assessment process was
used to determ ine the strength o f the relationship using a four-
levelscale(seesym bols).
A t this point, the q u a lity table identifies a m inim al, compre-
hensive set of q u a lity m etrics that capture the im portant features
o f customer requirements. The feasib ility row indicates the feasi-
b ility o f m easuring each o f the q u a lity variables. But the transla
tio n m ay s till be am biguous, because d iffe re n t people m ay
understand in d iffe re n t ways how the m etrics are to be mea
sured. You need q u a lity m etrics that always give the same mea-
surements regardless of w ho perform s them. Thus, each q u a lity
m etric requires a w ell-defined measurement plan. The measure-
m ent plan states w h at w ill be done w hen and b y w hom , and
w hat w ill be observed in the process. The m etric's measurement
plans are indicated in the measurement plan row o f the q u a lity
table. A q u a lity m etric w ith a measurement plan is said to be op
erationally defined.
N ote the generality o f the process o f using a tree diagram
to brainstorm m etrics and a m atrix to choose a set that w ill ade
quately measure m u ltip le objectives. The same process can be
Chapter 8: Proactive Improvement in Product Development 237
.E
(0
(Apln 9seola>lseg
a)
r:
(o
238
to
co
oo
l-
q
l JoEOn o -
dll) lea >lueEqouag
dl 6
®
®
S
«_
Customer Benchmark Rating
Competitive
o
Analysis Burchill
<
Product A Product B Product C Product D
1
^
Product
— ①
<
<
Substantial belt 〇
<
〇 0
<
Must-be Tangle-free line
〇
Drag-free casts
〇 <
〇 <
©
®
0
W ater drainage
<
<
C6
Gorrosion resists门
<
<
Line shifts
Demanded 〇 <1 <
<
customer
<1
Loop size
requirements
<
<
〇 〇 〇 〇
Ease of adjustment
<
<
<
One
Outer edge droop
dimensional
CONTINUOUS IMPROVEMENT
<
Movement interference
<
〇 〇
〇 〇
Position stability
〇 <] 〇
Conformity to body
0
0
〇 <
Light weight
1
---------------------------------------------
<
<1
〇 〇 〇 〇 〇 〇 〇 〇 〇 〇 〇 〇 〇 〇 〇
〇
Line fallout
Chapter 8: Proactive Improvement in Product Development 239
Visit customer
notetak
Stage 1
voices images
7 translationX / L .
rules 、 /c h e c k in g
customer requirements
丨MPM
Stage 2
vital few CRs
| KJ
^ Kano analysis
characterization of CRs
士
tree of QMs
Stage 3
quality table
physical
product
bridge
qualitative
C
structure quantitative
and analysis engineering
(stage 2) process
implement
Level product
of
Thought
Level
of -------------------------- ►
Experience try to sell
product
refine
hypothesize product implement
Level of Thought product spec product
il v > oSJd Ml
oeoJdd 3- !
36
*M
o>
oo
co
ui
Q)
a)
IJ.
ui E d ol0> p 0ioj!>u e
(o •s
a}
a)
e 4 S U M 0 pJO
pu p u ,J0uiolsno
co
M-
a)
(D
o) (D
(o
UJ i J !n b e J
o) (0
(d
E
slu Eal l
nbeJ u p ol u UB J p u n lo u u lsj p u n
) a>
co
ito
H=
) (D
) a}
5 a}
3 a)
.E
;5
:6
e
s
s
r
oAUOO
a B
o
A=BUJ
o3d o: 3 e
i
t
o
t
c
c
n
0
UBdo9>9a:
l
e z !J e p 2 e l p
2 e£
opA
s As
l
LoJ oss l MM
-s ca
os
n p o jd J e E o li
ts
1 0 1a4 B eE i J s ll
(
o
J0E2S3
£ !M •OID U Ba b
(o TJ
uog ed E O O
w
s
p n p o jd JU8 00
9J d E
>IJ ELPU 9G I
(0
p oA O JdE I
(D
A I90U 00
S0 53 E !J
SA > 0 M l SE
U
«»—
a}
•bej
.IS
n p o jd
a>
SIACn OMM
ts
o n p o jd
Hield E l /Midi。 NIO d A H la!d
l >> eqr M M
A 9P03P
ca
(D
is
s o l
e>Jep epo9p
el l
Jo E olsn o ro eo o> £6u l G lcojlBJ3d o
a>
>»
(D
:E
244 CONTINUOUS IMPROVEMENT
NOTES
249
250 TOTAL PARTICIPATION
worker manager
k ills hum an creativity. Few people are satisfied doing the same
th in g e ve ry d a y a cco rd in g to s ta n d a rd . T h a t is the jo b o f
m adiines, not hum an beings. TQ M aims instead to develop the
hum an capacity by u n itin g d a ily w o rk and im provem ent w ork.
The TQ M approach is to jo in the functions o f d a ily w o rk (D, in
Figure 9-2) and im provem ent w o rk (I, in the figure) at each level
w ith in the com pany and w ith in each w o rk u n it so that quick
and correct reaction to change m ay occur.
ACT to
Know the
improve the
STANDARD
standard or
SDCA its use
cycle of
daily
work CHECK DO the work
the work according to
against the the standard
standard
CHECK
improvement
results
PDCA cycle of
improvement
work ACT to
standardize or
replan
Know the
STANDARD
Daily
work
DO the work
according to
the standard
standard
S〇DCA
u
E ve ntu ally, how ever, d u rin g the check (C) stage o f the
SDCA cycle you m ay have some vague dissatisfaction w ith the
results. Thus, you act (A ) to in itia te a PDCA im provem ent cycle.
s 〇d c a ------- ,
PDCA
S〇DCA
PDCA
Si
S〇DCA
PDCA
S^CA
u
254 TOTAL PARTICIPATION
Eventually, at the check (C) stage, you are again (or s till)
dissatisfied w ith the result o f the current process (占〇 and again
yo u act (A ) to in itia te an im p ro ve m e n t cycle, w h ic h again
im proves the standard process.
S〇DCA
PDCA
_ _ I
SiDCA
PDCA
___I
—
S〇DCA— ^
PDCA
1
1
s ! d c a -------
;
PDCA
1
1
S2DCA • • • *
PDCA
1
SnDCA
<-------------- --------------►
0% 100%
% of time
Types of Teams
Problem Duration
Set-up Members
to solve of work
Task force ordered selected by work-related duration
by superior management but given of the task
Each o f the three basic types o f teams functions at a typ ica l place
in the com pany hierarchy (see Figure 9-6).
Position in Hierarchy
Cross-functional
c task
c
force
within
QI team
)
functional
area QC circle } permanent
Cross-Functional Teams
Design the w ay the team deals w ith facts to m iium ize con
flic t w ith o u t s tiflin g insight. G ive in tro d u cto ry TQ M and prob-
lem -solving tra in in g (w hich also leads the team to root problem s
and effective conclusions). Leave little tim e fo r unstructured dis
cussion in m eetings. S tick to discovering the facts and using
im provem ent tools — d o n 't leave tim e fo r personal agendas and
com petitive instincts to emerge. Such an in trodu ction means that
a ll team members can speak the same language.
O f course, fo r the team to come to a firm conclusion, mem-
bers need to have the same experiences and stu d y the same
facts. Even if subteams or in d ivid u a ls do th e ir ow n investiga
tio n , they need to re p o rt w h a t they uncover in a factu al and
nonevaluative way.
The focus o f the team should be on process. The team facil
ita to r manages the process and revises it as necessary (if the
team accepts leadership in the process, great progress can be
made). I f the managers w ho created the team ask fo r an in te rim
report, report process, not results. The focus on process e lim i-
nates the p o ssib ility that upper management w ill dictate a solu-
tio n halfw ay through the study effort.
• A nalog Devices
• B olt Beranek and N ew m an
• D ig ita l Equipm ent C orporation
• P olaroid C orporation
參 Teradyne
• Bose
• GE Jet Engine D ivisio n
sos>;ov seoO
(/)
J vo o d Ee l u6« Qs0 0 -6 J
Q.
i»
a>
a)
od
o)
36l
u.
266 TOTAL PARTICIPATION
ods, w h ich its members have to learn and apply to the problem s
they take up. W ith o u t this m ethodology, it w o u ld be d iffic u lt to
im prove the q u a lity o f a product. Second, QC circle activities are
voiuntary. The w orkers decide w hether they w ant to set up a
circle or not.6 If they do organize a circle, the choice o f problem
to focus on is also theirs: they decide on the m anner o f data
gathering, on the p la n n in g o f counterm easures, and on other
m atters. Every step o f the way, it is they w ho decide. In a true
QC circle, there is no such th in g as an order from above.
However, someone in management m ust rem ain responsi
ble fo r a team 's im provem ents. O the rw ise the o rg a n iza tio n
could lose control o f its activities and adversely affect custom er
satisfaction. A system fo r p ro vid in g perm ission from a team’s
management is very im portant, b u t m aintaining control w ith o u t
disem pow erlng a team is a delicate business. Several CQ M com
panies have fo u n d it effective fo r m anagement to suggest the
in itia l problem areas (but not the theme), or to review the theme
together. M anagement also m ay need to review the im plem enta
tio n plan, especially if it affects others. The tric k is to m aintain
awareness o f w hat the team is doing w ith o u t g u id in g the team
too m uch, tru stin g good im provem ent process to produce use
fu l results and then standing behind those results. O f course, if a
team is at risk o f going seriously astray, management m ust pro
vide m ore guidance.
Problem solving in the QC circle is a continuous process.
O ften a group w ill begin by tackling sim ple problem s — such as
cleanliness in th e ir w orkshop o r m iscom m unication between
w orkers — and m ove on to more com plex ones, especially prob
lem s in p ro d u c t q u a lity , p ro d u c tiv ity , and those th a t affe ct
m u ltip le g ro u p s. T h is c o n tin u ity is p o ssib le because th e
com pany supports the group. The com pany gives perm ission
fo r QC circle activities to be held d u rin g w o rk hours and pro
vides conference room s fo r m eetings. This su p p o rt, w ith o u t
w hich the QC circle cannot function, provides the group w ith an
in s titu tio n a l fram ew ork.
Chapter 9: Teamwork Skill 271
Small group
activities
Participation of
every worker
To perform control
and improvement
of the work group to
which they belong
Continuous improvement
and rising levels
of problem solving
no team
self- fulfillment
effectiveness
▲
development is not straight ahead
development
proportional to
time
development is step-by-step
(QC circle, TQM, or whatever)
,
户 development triggered at
appropriate times
time
a a
triggers
Stages of development 1 2 3 4
of the QC Circle
Themes completed 2.5 5.2 7.7 10.2
per year
• im prove q u a lity
• reduce errors
參 inspire team w ork
• create problem -solving capabilities
• im prove com m unications
The graph in Figure 9-13 shows the num ber o f em ployee
circles at the H udson facility.
Participation
100%
Degree of Japanese
Participation start
here
Western
start here
( 2 0- 30%
participation)
enforcement completely
voluntary
Real Results
CHECK
improvement
results
ACT to
standardize
or replan
Standardization
Know the
STANDARD
DO the work
according to
the standard
d ire ctly or to form teams that do. In fact, the extent of (volun
ta ry) QC circle a ctivitie s can be an in d ic a to r o f a m anager's
a b ility to m otivate his or her subordinates.
Diagnosis
The team leader can be a sym bol, as when the CEO leads a
team . B ehavior norm s fa c ilita te team w ork. 7/Speak fro m the
facts” is a good team norm . It is useful to develop a com mon
language, as the CQ M d id w hen its cross-company team m et fo r
a week at the start o f its five-w eek study to learn the language
of TQ M .
Rotation 3
• It reduces defensiveness.
294 TOTAL PARTICIPATION
• It enhances self-esteem.
• It encourages cooperation.
• It identifies problem s and solutions.
Setting up Meetings
Physical Arrangement
________________| 〇
〇〇〇〇〇
〇 〇
〇
0 ^
〇 〇
Degree of
concentration
Time
Chapter 9: Teamwork Skill 299
Time
Yo-One
NOTES
CEO INVOLVEMENT
307
308 TOTAL PARTICIPATION
management
The data supporting the dom ino theory are shown in Fig-
ure 10-3. Shiba measured levels o f prom otion o f QC circles at
various levels w ith in m any companies. He then sorted compa-
nies in to g ro u p s a cc o rd in g to h o w a c tiv e ly the CEO w as
involved.
In Japan, m any companies are subsidiaries o f parent com
panies, and these parent companies p u t pressure on their sub-
sidiaries to im plem ent QC circles. If the CEO is very involved,
the facilitato rs can be h ig h ly effective and other functions (such
as managers' involvem ent and education) can be high. I f CEO
involvem ent is low , the facilitato rs s till m ay try to im plem ent (as
in the bottom curve), b u t q u a lity a ctivitie s w ill rem ain weak.
E m pirically, the CEO's level o f involvem ent is an upper lim it on
t\\e other q u a lity activities o f the company.
W hat does CEO or top management involvem ent mean? In
w orld-class companies the authors have visited, it means hands-
on p a rticip a tio n in TQ M , as is shown in Table 10-1. This table
also characterizes unsuccessful cases; note that sim ply approv
in g and delegating, w h ich is standard m anagem ent practice,
d id n 't w o rk.3
We have provided em pirical evidence o f the necessity fo r
CEO leadership o f TQ M . There are also theoretical bases fo r
CEO leadership. First, as both m ilita ry and business strategists
observe, a tw o -fro n t w ar is m uch m ore d iffic u lt to fig h t than a
one-front war. S p littin g CEO a tte ntion and resources is m ore
d iffic u lt than focusing them . It is more d iffic u lt to lead sim ulta
neous crusades fo r cost reduction, ubiquitous adherence to stan
dards, faster design, and so on, than it is to lead a crusade fo r
quality, w h ich can encompass a ll o f these.
Second, because true charism a is rare, m ost leaders need a
le a d e rsh ip s tyle th a t depends on v is ib le p a rtic ip a tio n and
a rtic u la tio n o f values and strategies m ore than sheer force o f
personality.4 TQ M has evolved m any standardized o p p o rtu n i
ties fo r CEO involvem ent: from in itia l visits to TQ M companies
to involvem ent in planning and p ilo tin g , running the corporate
q u a lity com m ittee, goal setting, ta kin g and then teaching the
Chapter 10: Initiation Strategies 311
l
o £ o
q>
EOo
.E
:« E >
a)
a)
lo>
ulo o
ul
eoUBEJO ®d
co
o
loleAen
l.
aln6
*t:
il
312 TOTAL PARTICIPATION
Do Participate in
improvement activities
to focus on the latent crises that others in the com pany m ay not
yet see. In about h a lf the cases in Japan, TQ M is started because
o f a latent crisis. The CEO m ust be able to visualize the latent
crises and b rin g them to the com pany’s attention.
The example o f Xerox, though often to ld , bears repeating:
"In 1980 we were horrified to learn that the selling
price of the small Japanese machines was our manufactur
ing cost/7 Kearns (the CEO) says. /7We were not tracking the
rate of speed of their improvement. We tended to put a peg
in. By the time we thought we were up or thought we were
close, we found ourselves still off the mark." Kearns began
paying even more attention to Fuji Xerox, the company's
Japanese partner. There, total quality control was the focus
of its New Xerox movement, so the New Xerox movement
became the model for Xerox Corporation’s own focus on
quality.7
Kearns used the rate o f im provem ent inform ation, the cost
inform ation, and one year’s poor p ro fit to id e n tify a latent crisis
and create a p o w e rfu l and lasting source o f change throughout
the organization.8
1. Work Is a Process
3. Teamwork
£dMCflte^;erj/b0dj/.Educatioridevelopsalar\guagefortop-
to -b o tto m and b o tto m -to -to p c o m m u n ic a tio n . E d u c a tio n
enables im provem ent. A t Xerox, Kearn's cascading education
system was seen as the greatest evidence o f his com m itm ent.
M otorola spends $50 m illio n per year on education o f its 100,000
employees. Globe M etallurgical educated its h o u rly w orkers in
statistical process control and gave them calculators.
Chapter 10: Initiation Strategies 321
"In tro d u ctio n " is here used in a narrow sense: as the phase
coining after senior management learning and occurring d u rin g
o r a fte r in itia l p ilo t im p rovem ent projects, goal setting, and
establishm ent o f governance and organ ization . In tro d u c tio n
happens before ro llin g out w idespread education in im prove
m ent skills. F ollow ing is a case study of Teradyne's in trodu ction
strategy.15
U n til 1985 the autom atic test equipm ent in d u stry in w hich
Teradyne w o rks grew w e ll and the com pany prospered. Its
com pany systems were predicated on continuing grow th.
Teradyne had q u a lity program s before. It trained approxi
m ately 2,000 people in Crosby program s, the Teradyne staff has
know ledge o f statistical process control (SPC), and one d ivisio n
held a qua rterly zero defects day. These program s continued up
to 1989. They made good im provem ents from these program s.
H o w e v e r, d u rin g th is p e rio d T e ra d yn e CEO A le x
d 'A rb e lo ff was m ore o f an observer and su p p o rte r than an
active particip ant in Teradyne's q u a lity program s. He had sup
ported q u a lity program s fo r some tim e. Procter and G am ble
ch a irm a n E d w in A rtz t, a lo n g -tim e p ra c titio n e r o f q u a lity
im provem ent, sat on Teradyne's board and was instrum ental in
steering tow ard TQ M . A nd, lik e m any CEOs in the electronics
business, d 'A rb e lo ff had traveled often to Japan, and seen firs t
hand the results o f TQ M . He was also close to Ray Stata, CEO of
A nalog Devices, another Boston-area com pany th a t had been
contem plating im plem entation o f TQM .
As o f 1990 Teradyne faced a crisis. Its in d u stry experienced
excess capacity, d iffic u lt price com petition, and d iffic u lt product
and service com petition. Teradyne itse lf was not grow ing, was
having d iffic u lty sustaining p ro fita b ility, and was dealing w ith a
322 TOTAL PARTICIPATION
4,300
Alex mgt.
commttee UFJper
mgt.
managers
W hen others in the com pany heard that the senior m an-
agers were reading these books, they also w anted to read them.
M any read at least three o f them.
The managers also attended three seminars — a four-day
sem inar b y D em ing, a one-day sem inar at Texas Instrum ents,
and a one-day sem inar from Am erican S upplier Institute, w hich
they brought in-house.
From th e ir study, the Teradyne managers concluded that
they w o u ld adopt a new management m ethod based on three
principles:
2 questions:
• 1 weakness
• 1 thing learned
research in com puters and com m unication, BBN had spun out
several d ivisions to com m ercialize the results o f research. For
m any years, BBN's standard Am erican management system and
its practice of h irin g very capable people had created success
and grow th, b u t in 1989 and 1990 several factors caused a m ajor
business reversal.
O ver the course of nine m onths, BBN’s CEO, chief opera-
tio n s o ffice r, ch ie f fin a n c ia l o ffice r, and d iv is io n presidents
began study o f TQ M and in itia l use o f some o f the TQ M tools.
They attended a six-day course on TQ M fo r senior managers,
offered by the Center fo r Q u a lity M anagement, w hich consoli
dated th e ir th in k in g and gave them a com mon understanding
o f TQ M .
Because custom er focus was a weakness in the corporate
culture, BBN firs t adopted the p rin cip le o f custom er focus and
carried o u t the senior executive custom er v is ita tio n program
described in Chapter 7. A nalysis of the data from the customer
visita tio n program suggested that three com panywide TQ M in i
tiatives should be undertaken to in stitu tio n a lize the fo llo w in g :
<l>
6
o>
a>
-
(D
(o
t
'45
)
o
TOTAL PARTICIPATION
o d ds
nsol rocloao
tr
330
ts
ff
suol ziue o
Chapter 10: Initiation Strategies 331
voice of the
customer
JIT/TQC
cycle time
responsiveness
ADT
NOTES
12. Ib id .
13. Ib id .
14. M a ry W alton, Denting Management at Work (N ew York:
Putnam, 1990).
15. This case study was prepared by O wen Robbins o f Ter-
adyne and presented in December 1990 at the CQ M 's
firs t senior executive courses.
16. N ote that the Teradyne case study matches the m odel,
in w h ic h the c ris is p lu s in fo rm a tio n leads to CEO
involvem ent in TQ M . The crisis plus in form ation about
TQ M p rom pte d Teradyne's CEO to become d ire c tly
in v o lv e d in th e c o m p a n /s q u a lity a c tiv itie s .
17. Teradyne's TQ M efforts and Q IT activities are b u ild in g
on the com pany's previous study and im plem entation
o f Crosby's m ethods. Teradyne is changing vocabulary
o ve r tim e and le ttin g teams evolve h o w th e y w o rk
rather than disbanding them.
18. A g ain , th is matches the m odel in w h ich in fo rm a tio n
plus crisis leads to CEO involvem ent in TQM .
19. BBN's ro llo u t is described b y C liff Scott in "B B N ’s 7
StepsImplementation/’ TfeCewter/brQwflZrtyMflm^e-
merit Journal 1, no. 1. (A utum n 1992): 19-27.
20. These are described in the Coming Total Quality Digest 2
(D irector o f Q uality, C orning, In cv HP-CB-06-4, C orn
ing, N Y 14831).
11
Infrastructure
for Mobilization
337
sol u u E ldEIJ£ alrnomlsej lleuollezce o■T?
a4s
o)
9Jn6
ff
's
o)
o)
iI
TOTAL PARTICIPATION
338
Chapter 11: Infrastructure for Mobilization 339
Benchmarking
Process-oriented
Six sigma
intermediate goals
Cycle time reduction
Benchmarking
ORGANIZATION SETTING
TQM
Steering
Committee
Functional VPs
and
division heads
Functional
divisions
Top management
must be below
top and above
Companywide
direct reports, TQM office
otherwise
little impact
Full-time staffing
(“no job description — must
define job”)
— A d ivision A
Yearly diagnosis ⑦
Department and
All 4 days
section manage
38 H
Department and ment course Specialized
section managers instructor
Mgrs with
Refresher 2 days
5+ yrs.
course 19H
experience
Q C (staff) 4 days
Technical All
course 32 H
Staff
Q C (staff) 3 days
Clerical All
Compulsory course 24 H
training
Supervisor 3 days
Supervisor All
course 24 H
Instructor
Group leader in each
Group leader 3 days
Line and All
course 32 H plant,
circle leader division,
Q C worker 1 day and office
Worker All
course 8H
Part-time
Part-time 1 day
All employee
employees 8H
course
4 mo
1. Basic
5days/mo
course (A)
192 H
3 mo
2. Basic
4 days/mo
course (B)
114H
Specialist Staff Selected
4 days Specialized
training Staff 3. Reliability
39 H instructor
4. Design of 4 days
experiments 36 H
5. Multivariate 4 days
analysis 36 H
6. Others
to teach have already taken the course and have personal expe
rience in the general area b e in g ta u g h t. The course design
emphasizes real cases and group w ork. O nly about a th ird o f the
course is lecture, and m ost o f the lectures are accounts of diverse
examples. CEOs are called on not to teach theory b u t rather to
share personal experiences.
Leaders fo r each day (the CEOs are not called "teachers")
are given videotapes, fu ll classroom materials, and analysis/diag-
nosis from previous offerings to support preparation.
The course fo rm a t require s p re p a ra tio n o f overheads,
w hich ensures a certain am ount of preparation. The overheads
are checked to provide advance w arning o f m ajor problem s of
pace o r om ission. A lw a ys in attendance in each course is an
adm inistrative support person and a course com m ittee member
(i.e., someone fa m ilia r w ith both TQ M and the course), over and
above the people brought in by the leader to facilitate, present
cases, and o th e rw ise su p p le m e n t in s tru c tio n . The process
ensures effective education w ith o u t dictatin g how a CEO goes
about preparing.
A ll o f these strategies are designed to ensure that a ll those
p a rticip ating in a course have in fact learned w hat they need to
learn about TQ M . C ontrast th is w ith the ty p ic a l "'D a rw in ia n
m odel" o f pub lic education, where students w ho do poo rly just
get poor grades.
information
and eliminated the root cause. These were the check and act phases
ofP D C A .
The C om ing senior management and q u a lity officers rec-
ognized after the fact that th e ir people d id need form al instruc
tio n in problem solving, and they q u ic k ly relabeled the in itia l
course as Phase 1 and in s titu te d a Phase 2, w h ich taught the
needed tools and practices.11
The plan phase happens at several levels in the organiza
tion. Senior management derives the design fo r courses directly
from business needs. For example, consider BBN, whose senior
m anagem ent s tu d ie d TQ M and d ecided th a t the com pany
needed to learn m ore about TQ M (call th is "need 1"). Senior
management’s customer visita tio n program , described in Chap-
ter 8, convinced them that the com pany needed to do a better job
o f keeping in touch w ith customers (call this "need 2"), and to
ensure th a t products alw ays w o rked in the fie ld as specified
("need 3"). are sim p lifyin g the internal discussions and leam -
in g th a t led to the educational responses, b u t it is a m atter o f
record that the three hardest pushed and m ost w id e ly diffused
courses at BBN are: (1) a TQ M awareness course, y/TQ M at BBN,7
(m eeting need 1); (2) a course on the procedure fo r customer visi-
tation (m eeting need 2); and (3) a 7 steps course (meeting need
3). As o f this w ritin g , a course is s till being designed fo r the fin a l
item in the in itia l customer visita tio n results, standardization of
product developm ent process ("need 4"), even though the stan-
d a rd iza tio n e ffo rtsa re a lre a d yw e llsta rte d .
P la n n in g d e p lo y m e n t o f courses is done b y the T Q M
office and sometimes b y professional educators, as is design
in g the actual cu rricu lu m m aterials. S im ilarly, the check phase
o f PD CA happens at several levels and tim e horizons. F o llo w
in g the new standard tra in in g practice, participants give d a ily
feedback; some aspects o f courses can be im proved overnight
b y instructors. The use o f course content in im provem ent w o rk
is m onitored b y the TQ M office or corporate education staff on
a three- to six-m o n th tim e scale, and co rre ctive actions are
taken. For exam ple, Teradyne's TQ M office in 1990 ro lle d o u t a
Chapter 11: Iiifrastructure for Mobilization 359
PROMOTIONAL ACTIVITIES
Events
(action)
flyer motto
brochure visual display
flags
symbols
2. Mobilization
3. Deadline Effects
4. Reflection
1. Select theme —
3. Analyze causes —
5. Evaluate effects
6. Standardize solutions
5. Motivation
Excellence in customer
satisfaction
Monitoring QITs
needs improvement
how to improve
NOTES
377
378 TOTAL PARTICIPATION
uT selldsD
o
)
l
,s
o
)
selld o A eaiqi ,1.
o
>
c5
,
Q.
l
-31
al
.n
C
D
i
l
Chapter 12: Phase-In 379
ORIENTATION PHASE
EMPOWERMENT PHASE
ALIGNMENT PHASE
Formal organization
(daily routine function)
problems
Principles of Organization
Elements of
Functions Formal Parallel
Infrastructure
organization organization
Orientation Goal setting Regulation Philosophy
NOTES
391
392 TOTAL PARTICIPATION
Alignment phase
Customer satisfaction;
(getting everyone
Baldrige assessment
to do it)
BENCHMARKING
• Planning phase
1. Id e n tify w hat is to be benchmarked.
2. Id e n tify com parative companies.
3. Determ ine data collection m ethods and collect data.
• A nalysis phase
4. Determ ine current perform ance "gap."
5. Project futu re perform ance levels.
• Integration Phase
6. Com m unicate benchm ark findings and gain accep
tance.
7. Establish functional goals.
• A ctio n Phase
8. Develop action plans.
9. Im plem ent specific actions and m onitor progress.
10. Recalibrate benchmarks.
• Return to step 1
Goal Setting
Means Planning
Customer Focus
Promotion
dem anding of skill — in the sense that beginners are less vul
nerable to making mistakes that significantly distort results 一
than purely proactive approaches for m any functions.
Results-oriented Process-oriented
surface appearance experiences
Goal setting ◎
Means planning ◎
Customer focus 〇
Promotion 〇
Diffusion 〇
Proactive/reactive ◎
Results/process ◎
2a 308,700 ppm
3a 66,810 ppm
4a 6,210 ppm
5a 233 ppm
6a 3.4 ppm
m e m b e rs th a t p e rh a p s 25 p e rce n t of th e v a lu e of a g o o d
com pany's revenue is waste. In one example from Digital, the
actual waste w as 332 minutes and the theoretical w aste w as 57
minutes, showing potential waste of up to 570 percent.
Cycle time reduction is applicable in the factory and else
where:
鑛
Defects seen by
customers ◎
Waste ◎
Overhead costs ◎ ◎
Production costs
◎ ◎
Responsiveness to
customer orders ◎ ◎
R&D productivity ◎
Product innovativeness
and performance ◎
Defect levels and cycle time are drivers not only for the
m ost im portant business goals; these improvement methods are
applicable to nearly any process. We have heard stories of defect
reduction in bread baking, and cycle time reduction in closing
and auditing the accounting books. Like the 5 evils, defects and
cycle time are applicable to a wide range of activities.
NOTES
411
412 TOTAL PARTICIPATION
In the next figure, people and jobs are properly aligned; the
result is a strong force aimed toward com pany goals.
t m i
key m l
90
ak
Chapter 14: Hoshin Management 413
key company
goals
\\\\\
environmental change
The control phase is carried out over the course of the year.
The com pany uses the control by measurement plan to monitor
the results and controls the results and the m eans so that the
planned m eans are in fact carried out and corrective action is
taken when the means and results are not as planned.
A s Figure 14-2 shows, there is overlap am ong the proac
tive, reactive, and control phases of hoshin management. Proac
tive an d re a ctiv e o v erlap at the to p -lev el an n u al h oshins.
Reactive and control overlap where the control by measurement
plan is used to monitor things over the course of the year, and
the control phase is the do part of the reactive phase.
Thus, PD C A cycles exist in hoshin m anagem ent. First,
there is an inner PDCA cycle in the control portion of hoshin
Chapter 14: Hoshin Management 417
The part of the hoshin planning process that deals with the
past will be discussed later in this chapter. The section below
addresses that part that deals with the future, an essential phase
in strategic planning.
Idealized design
planners figure out a path from the ideal com pany to the current
company. By sim ply reversing this path, they have a dem onstra-
ble path from the current com pany to the fu tu re company, and
one lik e ly to be shorter than 10 years.
A ck o ff's m ethods f it n icely in to the hoshin m anagement
m ethods o f TQ M (see Figure 14-6).
division head
y
mid-term plan
department head
Higher-level process
operator method n
higher-level
product
quality
material
For this example w e’ll not include the deadline. It is typ ica l that
the units o f the m etrics change at every level in the deploym ent
and m ay d iffe r from one hoshin to the next at the same level.
D eploym ent w ill go dow n m any levels in the organization, per
haps to the section level.13
Suppose, through analysis o f the past, the environm ent, and
the vision fo r the future, the president concludes that an annual
hoshin should be "reduce costs 20 percent b y decreasing the
length o f business cycles/7 The next step is to deploy relevant
hoshins to each o f the functional divisions below the president
(see Figure 14-9).
------- Manufacturing
President’s Administration/finance
hoshin
y Sales
Reduce costs
20% by
decreasing ------- Marketing
length of
business
cycles
Development
Support
Engineers have
product-out
orientation
Defects not
detected _
until final
inspection
Why are cycle
times long?
Fully qualified
prospects don’t buy
Figure 14-11. Stratifying to Select Problems for the Next Level Hoshin
Reduce costs
20% by Sales: qualified prospects unexpectedly
decreasing decline to purchase at last minute
length of
business
cycles Marketing: each new product requires a new
effort to line up distributors
Sales hoshin
Direct sales: no system for use PDPC to
evaluating and anticipate
addressing objections and
Decrease possible provide counter
average sales ___ objections measures
cycle 30% by
decreasing
qualified
prospects who Sales support: user difficulty in hold user’s hand
don’t buy using product during product
results in trials
unsuccessful
product trials
Manufacturing:
------decrease manufacturing
cycle time 50% by
implementing JIT
Administration/finance:
------decrease key administration
and finance cycles 40% by
flowcharting processes and
eliminating NVA work Direct sales:
President’s
hoshin decrease qualified
Sales: prospects who don’t
— decrease average sales buy 50% by use of
cycle 30% by decreasing PDPC
qualified prospects who
don't buy Sales support:
Reduce costs decrease qualified
20% by prospects who don’t
decreasing buy 50% by improved
length of product trial support
business
Marketing:
cycles
— decrease average time to
release a new product 50% -
by creating strategic alliances
Development:
------decrease average time to
successful product release -
30% by increasing engineers’
understanding of customer
context
------ Support: NA
superior. In this case, the subordinate does not have a hoshin re-
lated to the superior's H oshin A . Focusing on the v ita l few at
every level m ay mean that subordinates do not address a ll o f
the superior’s hoshin goals.
Such alignm ent charts can easily be extended to sum marize
the hoshins o f a ll o f the subordinates o f a manager, as in Figure
14-17. In this example a ll three subordinates have hoshins that
436 TOTAL PARTICIPATION
Superior's Hoshins
Subordinate’s
Hoshin a
0
Hoshins
Hoshin b
0
Hoshin c
0
Figure 14-16. Hoshin Alignment Matrix
Superior’s Hoshins
A B c
a
o
Subordinate 1 b o
c
o
d
0
Subordinate 2 e o
f 0
g 〇
Subordinate 3 h
0
i ©
Figure 14-17. Extended Alignment Matrix
Deploying Metrics
C E ) Manufacturing:
decrease manufacturing
cycle time 50% by 广、
implementing JIT
Q ) Administration/finance:
decrease key administration
and finance cycles 40% by
flowcharting processes and
eliminating NVA work C l ) Direct sales:
President’s
Hoshin decrease qualified
© Sales: prospects who don’t
decrease average sales buy 50% by use of
cycle 30% by decreasing P D P C /? ^
亨 qualified prospects who
don’t buy L @
Reduce costs Sales support:
20% by decrease qualified
decreasing prospects who don't
length of buy 50% by improved
business support
product trial su
cycles C x) Marketing:
decrease average time to
release a new product 50% —
by creating strategic alliances
G ) Development:
decrease average time to
successful product release —
30% by increasing engineers1
understanding of customer
context
1.1 create a
system for ZD
objective of
the daily job
Once the annual Plan has been made (plan) and deployed
over the year or h a lf year (do), it is tim e to check fo r weaknesses
Chapter 14: Hoshin Management 441
Hoshins
objective
of the _
current
daily job
new
daily
job
new
tasks
in the plan or the w ay it was carried out (check) and to act ap
p ro p ria te ly to in fluence the next year's pla n or p ossibly the
long- or m id-term plans (act), (see Figure 14-22).
The check step entails discovering w h y the planned pro
cess d id n 't achieve the desired results. In other w ords, a process
has been created that can now be analyzed using reactive im
provem ent m ethods — you use the 7 QC steps to analyze the
difference between the plan and reality, and determ ine the m ain
causes o f the difference. For instance, using the data taken as
part of the measurement plan, you can discover w hether parts
o f the plan w e ren 't carried out or were carried out b u t d id n 't
produce the planned re su lt (see Figure 14-23). This provides
feedback to the next planning cycle.
442 TOTAL PARTICIPATION
changes
to long- or
mid-term
plan
PRESIDENTS DIAGNOSIS
head
process
N ote that the u nits o f the m etrics stay the same from level
to level o f the deploym ent. This is really a system o f p a rtitio n in g
or dispersion o f objectives rather than deploym ent.
Table 14-4 compares M BO and hoshin management based
on o u r observations o f actual use in U.S. companies.
Sales management often exem plifies the extremes o f MBO.
The cycle show n in the fo llo w in g figure w ill be fa m ilia r to m any
U.S. managers. In the M BO cycle illu stra te d above, the e ffo rt is
to achieve sales perform ance b y c o n tro llin g the investm ent in
salespeople and m o tiva tin g th e ir o u tp u t, w ith little direct effo rt
to understand, teach, and im prove the sales s k ill o f the organi
zation. Managers and salespeople conspire to present sales as an
in d iv id u a l s k ill that depends on the com pany's a b ility to attract
Chapter 14: Hoshin Management 449
Deploy a portion of the top-level target Deploy targets with different metrics to
to each segment at each level (no each segment at each level (change
change of units of measure) of units of measure)
Little monitoring of the means (if it Some targets and metrics are aimed
works, it's OK) at controlling adherence to the means
I
Hire needed people with a purportedly good sales track record
I
Give sales people product training
I
Give sales people quotas and an incentive plan
+
Sales people use their personal sales skills to sell
I
Reward good performers; fire poor performers
I____________
Figure 14-27. An MBO Model for Sales Management
a>
a>
<d
'43
^-
ra
T-
Q)
o)
±:
0)
a)
c)
.E
ul
ll
s n 0一
J > q
oB OIAJI : loA JJo co§ u E
u6ldD
a>
qof uosjAal qof
<l>
(0
aJ
a)
a)
-l I
dE 0 OUBU0HJI voas
CO
(D
E
IT
<•
£:
(D
isbo jo
m—
UBIdxoue UIUJOO
o)
cd
Me>arlsBo0JO
BoeJOrouos>aT
—
CD
cd
±;
s nsaj
aJ
pu u lcuueEeJnse EAqOJUJOO
M
tfl
TOTAL PARTICIPATION
o
D (D
seLId
Hi E A O Ip
)
d
a
0 0 e Bn UBI 2 0 1C0E E8
a>
o)
u lsnSIB0 ss s3q IB n cu Bpu u l
a>
Q.
cd
a> co
o)
a)
.E
J
E01
-PI
Ee> zoB2 p dol > su E
a>
a3
a}
a)
T3
(l)
IT
LUcfl JC
sul soli
B O OJO 牛
e s q d
! nuuB
uUU
CL
o)
AUBdEOo s s u >l 3M«J 9 6 lpu
a>
co
cc
a3
>*
a{
M— to
(l)
EJerPIEyu EUOJ>uecols>
cBId
a)
2p op ou«leLIM o sisaibu
a>
co
SS0JPPB
452
B
le3u u < 一
u e l u eu e s u i l J S O x
)
Q
(
a
(
uluu ldss9uis3g
D
Chapter 14: Hoshin Management 453
outcome
个
means = outcome
means = outcome
step 1 outcome
mea 门
s = outcome
454 TOTAL PARTICIPATION
step 1 — ► outcome
step 1 — ► outcome
s te p 2 、 — ^ 个
step 3 J
Mid-term plan;
environmental
analysis
PDCA2
Hoshin
(breakthrough
Management
improvement)
SDCA
(maintenance)
Daily
Management
PDCA!
(incremental
improvement)
Guidelines
Reduce
costs - Sales: reduce costs 5%
20%
Guidelines
m eans
L > ^ m eans
means means
------Marketing: reduce costs 15%
!—► means
------ Development: reduce costs 15%
!—► means
by implementing JIT
by flowcharting processes
and eliminating NVA work
by increasing engineer’s
understanding of customer context
Production
process building machines jigs manuals
engineering
NOTES
461
462 TOTAL PARTICIPATION
tion, and in 1987 w on the D em ing Prize. This effo rt was led by
the then president, K iyoshi U chim aru, w ho him self has decades
o f engineering management experience.
There are three ways to lo ok at the N IM S story reported in
this book. First, it can be seen as a story o f TQ M phase-in. As de
scrib e d in C h a p te r 12, T Q M pha se-in ty p ic a lly progresses
th ro u g h three phases: o rie n ta tio n (re co g n izin g the need fo r
TQ M and learning its basic principles), em powerm ent (learning
the m ethods o f TQ M and developing s k ill in practicing them ),
and alignm ent (harm onizing the business and TQ M goals and
practices o f the com pany).
The m anagem ent at N IM S began w ith orientation, b u t it
became captivated b y the concept o f hoshin management as a
m ethod o f align ing activities com panywide. Therefore, from the
o rie n ta tio n phase m anagers a tte m p te d to m ove d ire c tly to
hoshin management and the alignm ent phase. Finding, however,
that the N IM S staff d id not have the necessary s k ill to w o rk on
com panyw ide activities, they were forced to go back to the em-
pow erm ent phase to teach the basics of TQ M . W ith a little em
pow erm ent, they trie d alignm ent again and were again forced
to go back fo r m ore em pow ering skills. This cycle repeated sev
eral tim es u n til they had reached tru ly extraordinary levels o f
in d iv id u a l em powerm ent, after w hich they fin a lly succeeded in
alignm ent. In effect, hoshin management became the tool N IM S
used to deploy each year7s im provem ent ideas and to system
atize successful ideas.
Second, the N IM S sto ry is a sto ry o f the a p p lica tio n o f
TQ M in a creative engineering environm ent. We have a ll heard
statements such as //rTQM m ig h t be applicable to m anufactur
ing, b u t it cannot be applied to a creative task such as engineer-
in g // The feeling at N IM S was no different. The N IM S technical
staff cited a ll the usual reasons w h y TQ M could not app ly to
them . H owever, U chim aru makes the p o in t that the histo ry of
science is its e lf a q u a lity im provem ent story: develop a theory
and plan how to test it, do an em pirical experim ent, check the
result to see if the experim ent confirm ed the theory, and act to
Chapter 15: Managerial Development 463
peated over m any years; the tig h t loops represent CAP and the
lo n g arches represent D. The labeled dots along the stra ig h t
lin e segments show successive areas o f im provem ent activity.
The fo llo w in g paragraphs b rie fly describe these successive im
provem ent activities.
Because h o sh in m anagem ent is regarded as a p illa r o f
TQ M in Japan, N IM S im m ediate ly trie d to im plem ent such a
system (see slanting lin e segment nearest to the o rig in of the
fig u re ).2 H ow ever, hoshin m anagem ent d id not im m e d ia te ly
produce results as good as N IM S had sought. In fact, N IM S dis
covered that it was try in g to im plem ent hoshin management (a
tool norm ally used d u rin g the alignm ent phase) before its staff
had gone th ro u g h the em pow erm ent phase (in w h ic h basic
q u a lity capabilities are developed). In other w ords, it attem pted
to skip from the orientation phase o f TQ M phase-in to the align
m ent phase, b u t was forced to go back to the em pow erm ent
phase to b u ild basic capabilities relating to quality.
The next step in N IM S 's im p le m e n ta tio n was to in s ta ll
some q u a lity m etrics. By doing this it was attem pting to make
the developm ent process m ore visible. It discovered firs t that
ge ttin g technical people to agree on a set o f m etrics, or even
agree that there should be m etrics, took m any months. Once the
technical people had agreed on a sm all set o f m etrics, N IM S dis
covered that m etrics alone d id n 't provide m uch im provem ent.
In other w ords, it began to realize the need fo r focusing on pro
cess instead o f results.
N ext N IM S trie d to im prove its developm ent processes by
in itia tin g design reviews. We a ll know about design reviews —
m eetings in w h ic h experienced and s k ille d designers fro m
around the com pany review a proposed design. This was an at-
tem pt to insert the q u a lity fu rth e r upstream , that is, to fin d the
bugs w h ile s till in the design stage rather than im plem enting
poo r designs. This a c tiv ity helped some, b u t not as m uch as
N IM S hoped it w ould.
N IM S then figured out that if fin d in g problem s in the de-
sign fu rth e r upstream was good, then fin d in g problem s in the
466 TOTAL PARTICIPATION
defect defect is
defect should be actually
created detected detected
time
COMPANY STRATEGIES
prevents the desired results and how to fix it. These issues are
discussed later in the chapter.
T hird , there m ust be a system to m obilize m anagerial par
ticip a tio n in im provem ent a c tiv ity in the com pany hierarchy. If
the com pany uses a system o f v o lu n ta ry p a rtic ip a tio n in im
provem ent activities, a manager m ight choose not to participate
一 a decision that can leave those below w ith no leadership or
su p p o rt in q u a lity im p ro ve m e n t. H ow ever, the a c tiv itie s o f
managers d iffe r sig n ifica n tly according to th e ir level in the orga
nization. Therefore, different m obiliza tion strategies are needed
fo r each level of the organization. These strategies are also dis
cussed later in the chapter.
levering
up of
control
process
high
Level of incremental
improvement improvement
levering
up of
control
process
low 1----------------------------------------------------------------------
present T jm e future
Level of incremental
improvement improvement
levering
up of
control
process
low I----------------------------------------------------------------------
present j j me future
The PDCA! cycle is used to im prove the process and its re
sults, as follow s:
P: Pick the problem that is m ost responsible fo r the variatio n
in results, analyze the root causes o f the problem , and plan
countermeasures to fix the root causes.
D: Do the im provem ent.
C: Check that the im provem ent was effective.
A: Standardize it as appropriate, and go to the next im prove
m ent.
because tim es w ill have changed. Even so, you w ant less va ri
ance in the w ay you carry out the plan fo r the next year and the
am ount by w hich you miss objectives. So you ru n the C APD a
cycle. The letters o f CAPD have the fo llo w in g meanings.
CA: Discover w hat was w rong w ith the previous process that
prevented achievement o f the desired results; w hat are the
key things to im prove fo r the next cycle.
P: Determ ine w hat is desired fo r the futu re (e.gv w hat is the
next target).
D : C arry out the plan fo r the year
CA: Check w hether target was achieved, and if not, w h y not
(repeat CAPD).
The purpose o f CAPD is to im prove the process in the face
o f changing goals. Thus, standardization in CAPD differs from
480 TOTAL PARTICIPATION
Control o f Process
baked cake
materials machine
(flour, eggs, etc.) (oven, temperature, etc.)
i-----------------------------------------------------------1
fix the problem . Perhaps the process was not follow ed accurately,
or perhaps the process should be m odified to produce cake w ith
im proved qualities o f moistness. In any case, the focus is on the
actual results, w hat aspect o f the process produced those results,
and how to im prove the process so that better results are ob
tained. In this case, you control the process by m onitoring the de
via tio n o f resets from the standard and im proving the process
as necessary. -
The above exam ple show s th a t there are tw o kin d s o f
m etrics: results m etrics and process m etrics. Few managers un-
derstand the distin ctio n . M ost focus on results w h ile ig n o rin g
process (" I w ant to m otivate the staff," or " I can’t im prove w hat
i can’t m easure"), b u t a focus on results m etrics alone is h ig h ly
u n lik e ly to succeed. As TQ M has spread it has become m ore
com m on fo r m anagers to focus on process m etrics, b u t th is
alone is also u n lik e ly to succeed. The p o in t o f paying attention
to either process or process m etrics is to fin d out w hat is causing
or preventing success. If you don’t look at results m etrics, you
w on’t know w hether the process is w o rkin g or not. If you don’t
lo o k at process m etrics, you w o n 't be able to fig u re o u t w hat
w orked or d id n 't w o rk. Managers m ust understand that process
causes results, and th a t process m etrics te ll you w h y results
m etrics happened. The purpose o f studying both kinds o f m et-
rics is to p in p o in t w h at m ust be done to im prove the process
and thus im prove the results.
There actually are three levels o f control of process in the
concept o f CAPD. Two o f them , control o f process by m onitor
in g deviation o f process and control o f process b y m o n ito rin g
deviation o f results, have already been described.
Consider again the exampie o f cake baking, where the de
sired result is m oist cake.
The process o f m aking m oist cake requires an oven tem per
ature o f 350 degrees. The tem perature can be m onitored, and de-
viations from 350 degrees used to adjust the oven tem perature
u n til it is 350 degrees. This is control o f process by deviation o f
process from standard — a in Figure 15-12.
484 TOTAL PARTICIPATION
Assumption o f Responsibility
o f the section head. Thus, the section head m ust learn about
CAPD to lead the analysis and present its results. The section
head has to give an explanation in fro n t o f these managers. The
d iv is io n head then diagnoses the section head's presentation.
Thus, managers learn a ll the facts of w hat is going on in the d iv i-
sion, as w e ll as CAPD focused on the customer.
August September
N N
Stage 0
Stage 1
Stage 2
Stage 3
Stage 4
Stage 5
Stage 1 Means
Stage 2 Means
says that. Then, w hen she can th in k of nothing more to say, the
schoolmaster says, "yo u are now a member o f this school.” The
g irl goes on to success at the school.
Stage 3 Means
Stage 4 Means
Stage 5 Means
problems
problems
-
〇 〇 〇 〇 〇 〇 〇 〇 〇
integration
of efforts
^ O O O O O O O O O
curiosity
confidence
enthusiasm
success
repeated success
enjoyment
desire to do more
more success
ability
creativity
0 Giving instructions
NOTES
507
508 SOCIETAL NETWORKING
reinvented wheel
I f one co m p a n y d e ve lo p s a m e th o d th a t w o rk s , th a t
m ethod should be diffused so that other companies can use it as
w e ll. For th is reason, Japan has the tra d itio n o f "n o secrets”
about the m ethodology o f q u a lity im provem ents.
There are m any examples in Japanese TQ M o f such d iffu
sion, such as guidelines fo r how QC circles meet and exchange
factory visits. These were not theory. They evolved in QC circles
and w ere a great success, and were then diffused thro ugho ut
Japan. M illik e n ’s "sharing rallies," m entioned in Chapter 11 as
an example o f the d iffu sio n o f q u a lity im provem ent success sto-
ries, are a U.S. in vention that has become popular and has been
diffused to other companies.
One problem in im plem enting TQ M is that fo r some types
o f in fo rm a tio n and s k ill a catch-22 exists that makes in te rn a l
developm ent o f m ethods d iffic u lt, so that societal d iffu s io n be
comes a ll the m ore im p o rta n t. For exam ple, it is d iffic u lt fo r
CEOs to be taught new m ethods o f leadership b y those below
them in the com pany hierarchy; therefore, CEOs m ust get m ost
o f th e ir in fo rm a tio n about TQ M fro m outside th e ir com pany
hierarchy.
A t a ll levels in a corporation, the existence o f success sto
ries and managers experienced in TQ M makes TQ M easier to
practice. B ut h o w does a com pany th a t is ju s t b e g in n in g its
TQ M im plem entation fin d success stories on w h ich to m odel
its ow n activities or experienced managers on w hom to m odel
its behavior? Executives should be lo o kin g outside th e ir ow n
companies or organizations fo r know ledge, trainin g, and exam
ples o f TQ M practice. This chapter is about the stru ctu re o f
Chapter 16: Networking and Societal Diffusion 509
tio n , peo ple w ith s o lid QC backg ro unds are s ittin g on the
boards o f Japan's top corporations. Dr. K o ji Kobayashi, the fo r-
m er president of NEC, is an example of this trend.
The JUSE organization proper is chiefly involved in arrang-
in g logistics fo r its activities rather than itse lf being the source of
q u a lity know ledge. For example, in 1989, the fu ll-tim e staff o f
JUSE was 84, w ith an a d d ition al 31 people em ployed by JUSE
Press. By contrast, about 2,000 people from universities, corpora
tions, and governm ent collaborate to teach JUSE courses and
serve on the committees that create and im prove JUSE courses.
The "lean staff w ith extensive netw orking,/ structure is an in ter
esting innovation, fo r it v irtu a lly guarantees that TQ M kn o w l
edge is in te g ra te d and preserved w ith in c o rp o ra tio n s, n o t
hoarded in JUSE or other external organizations such as consult-
in g companies or universities. Also, members o f JUSE are corpo
rations rather than in dividuals, num bering about 1,850 in 1989.
JUSE's aim is clearly to increase know ledge w ith in corporations,
and its m embership and staffing policies support this aim.
JUSE approaches corporate im plem entation fro m the top
dow n as w e ll as using m iddle-up, m iddle-dow n, and bottom -up
approaches. JUSE's board of directors is made up o f high-rank-
in g industrialists; by tra d itio n , the president o f JUSE is the chair
m an or an ex-chairm an o f the Keidanren, the m ost prom inent
organization o f Japanese CEOs.4 W hen Dem ing taught statistical
q u a lity control in the 1950s, he taught classes w ith presidents of
corporations, not ju st engineers. Joseph Juran's focus in 1954 was
top and m iddle management. As o f i991, nearly 5,000 top execu-
tives had gone through JUSE’s five- day top executive’s course,
and another 9,000 had been particip ants in the JUSE five-day
course fo r executives. Both courses have long w a itin g lists.5
Independent consulting organizations do not play a m ajor
ro le in im p le m e n tin g q u a lity im p ro v e m e n t in Japan, even
though other areas o f lousiness are ric h ly supplied w ith consult-
in g companies, both Japanese and foreign-ow ned.
The Center fo r Q u a lity M anagem ent was designed to be
lik e JUSE in several respects. The m em b ersh ip com prises
514 SOCIETAL NETWORKING
corporations rather than in d ivid u a ls, and the dual focus o f the
CQ M is education and top management. Like JUSE, CQ M uses
in s tru c to rs fro m m em ber co rp o ra tio n s ra th e r than in-house
staff. CQ M was m odeled on JUSE because m any o f JUSE's fea
tures appeared to be w hat the m ember companies m ost needed.
CQ M is described in more detail in the next sections.
In com parison w ith th e ir Japanese counterparts, Am erican
y organizations are fragm ented and num erous. A th o r-
survey o f q u a lity organizations w o u ld therefore be quite
beside the point, at least fo r now, in understanding how a re
gional o r national in fra stru ctu re w orks. How ever, a few b rie f
descriptions should give the fla vo r o f the m em bership, function-
in g ,a n d c u rric u la o fq u a lity o rg a n iz a tio n s in th e U S .
The Am erican Society fo r Q u a lity C ontrol (ASQC) is a pro-
fessional organization fo r q u a lity specialists, and is probably the
largest q u a lity organization in the U nited States.
The Am erican S upplier In stitu te (ASI) was spun o ff from
Ford M o to r’s efforts to im prove the q u a lity of its supplier base.
G O A L/Q P C originated w ith a group o f follow ers of W. Ed-
wards Dem ing, and has evolved in to an independent 仕aining or-
g an izatio n w ith b o th in d iv id u a l and corporate m embers. Its
tools and practices have been adapted from the Japanese ver
sions, observed on study trip s. Instructors are both employees
and free-lancers.
There are m any independent consulting companies, m ost
o f w h ic h have entered the arena o f q u a lity co n su ltin g fro m
other, m ore tra d itio n a l consulting arenas. C onsulting companies
vary enorm ously in focus, ranging from specialty providers o f
e ve ryth in g fro m s ta tistica l softw are to m ission- and v is io n -
setting tra in in g , to "fu ll-sp e ctru m " companies whose services
cover m ore or less the fu ll range o f TQM .
F inally, a unique type o f consu lting com pany should be
m entioned — the TQ M com pany spinoff. In the case o f C om ing
Glass and H o rid a Power & L ig h t, the consulting com pany re
m ains a com pany u n it that provides services both to the parent
com pany and to com m ercial clients. In other cases, a num ber of
Chapter 16: Networking and Societal Diffusion 515
Training
Knowledge Dissemination
Week 1
• Q u a lity Conference fo r Foremen — 92 case studies, and
15 p lant visits and on-site debates
522 SOCIETAL NETWORKING
Week 2
• Q u a lity Conference fo r Managers — 92 case studies, and
9 p lant visits, debates, and panel discussions
• Top Management Conference
• D em ing Prize Ceremony
Week 3
• Conference on TQC in the Service In d u stry
CHANGE AGENTS
• A nalog Devices
• B olt Beranek and Newm an, Inc.
參 Bose
• D ig ita l Equipm ent C orporation
• General Electric
• P olaroid C orporation
• Teradyne, Inc.
C hairm an o f the board was Ray Stata o f A nalog Devices,
president was Professor Lee o f M IT (on a part-tim e, pro bono
basis), and the board o f directors consisted o f the CEOs or other
senior managers o f the founding companies.
524 SOCIETAL NETWORKING
Board of directors
President
Executive Operating
director committee
Staff Standing
committees
------- Seminars
------- Networking
------- Promotion
------- Research, document, and library
------- Finance
1991 Activities
Who
NOTES
533
534 SOCIETAL NETWORKING
Job-specific
capabilities
Skill — engineering,
quality, etc.
Basic education —
reading, writing, history, etc.
change
广 ^ D iffe r e n P ^ X
Job-specific ►( job-specific j
capabilities N^capabilities^/
Phase
Level | | Orientation Empowerment Self-invention
small group
activity ( \
improvement higher and
in one’s — ► wider level of
workplace improvement
For example, the tran sition from com panyw ide 7-steps reactive
problem solving to form al hoshin management is often a transi
tio n from the em pow erm ent phase to an alignm ent phase, as de
scribed in Chapter 12. The term denotes aligning TQ M practices
and systems w ith a ll other practices and systems o f the busi
ness, in a d d itio n to alignm ent o f in d iv id u a ls to com m on p u r-
poses. H o w e v e r, S h o ji S hiba o fte n has c a lle d th is phase
"lo calization," m eaning that the com pany has reinvented TQ M
practices to su it its local situation. This is self-invention at the
com pany level.
Learning b y regions and nations m ay seem less clear-cut
than the other levels we have discussed, perhaps because o f the
relative scarcity o f examples. However, fQ M learning in Japan
and the U nited States seems to fo llo w the orientation/em pow er-
m ent/se lf-in ve n tio n m odel. In Japan, the orientation phase en
com passed the e x p o rt c ris is in the la te 1940s and e a rly to
mid-1950s, and the search fo r foreign experts such as D em ing
and Juran. N ext follow ed an em powerm ent phase in w hich the
lessons o f the 1950s were deployed through a national infrastruc-
ture that included the D em ing Prize process and the intercom -
pany educational processes organized b y JUSE and JSA. The
subsequent history o f Japanese TTQM is the story o f a systematic
search fo r new approaches (self-invention) in response to global
economic and social pressures. The history o f the 1960s through
the present seems analogous to the localization phase fo r organi
zational TQ M learning, and the self-actualization phase fo r team
developm ent.5
T lie U nited States w ent through an orientation phase in the
late 1970s and 1980s, d u rin g w hich people recognized that the
q u a lity practices in m any industries, notably autom obiles and
consumer electronics, were substantially below w orld-class lev-
els. Beginning in the 1980s, m any companies em powered them -
selves b y adopting and practicing the iQ M concepts o f Dem ing
and Juran as practiced in Japan. A t the same tim e, some U.S.
companies had the necessary s k ill to search system atically fo r
im proved approaches. M otorola’s six sigma and Xerox’s bench-
m a rkin g approaches are exam ples. A lth o u g h there are m any
C/wpfer 17: TQM as a Learning System 541
Learning Process
Self
development
stages
Triggers
Team
development
----------------- ►
Time
Learning Process
noble
goals
noble
practices
Learning Process
Learning Process
Outside elements
Region/
change agent 〔 e lf-in v e n tio )
benchmarking
Nation:
Self
development ^mpowerment^
Nonhierarchical elements
stages
to the U n ite d States. S im ila rly, change agents can come fro m
outside a company. D em ing w ent to Japan as a change agent. A
num ber o f Japanese experts have come to the U nited States as
change agents fo r successful q u a lity in itiative s.
In d ivid u a ls, teams, organizations, and countries are m oti-
vated b y outside forces. TQ M system atically takes advantage o f
these outside forces, b u ild in g on th e ir a b ility to create change.
A t every level there are also pulls, or forces that create a de
sire fo r learning. '"Noble goals,/ are one such force. Traditionally,
noble goals reflect loyalty to the (well-being of the) larger unit. That
is, in d iv id u a ls are m otivated b y the desire to serve the team,
company, or country. Sim ilarly, teams are m otivated by the need
fo r the com pany to do w e ll. Ishikaw a often said th a t q u a lity
m ust serve the n a tio n , n o t the com pany. Such lo y a lty to the
h ig h e r u n it reveals an un d e rsta n d in g o f interrelatedness, as
shown in Figure 17-9.
sleooalqozEal
M-
slInd
ts
a l i p cd
I Tzl.
aln6
il
550 SOCIETAL NETWORKING
human goals
Source: How to Operate QC Circle Activities (Tokyo: Union of Japanese Scientists and Engineers,
1985), 4.
-M
iz
a) Eo^sllnd ajn6
i
arn
om J I
UI
co
:IB N
to
u3J
O/v
a u
QC
a>
/UOI63
o)
SOCIETAL NETWORKING
9 Jn n 4S B JIU I
ts
leu
olBNU
B J
O
ra
4=
:u o bzcb 6j o
:i9 1
>
ucJB
D
lI B a n E
(
a
uolpunrenp :IBnp!>!pul
SHJ E ® oll
pJBJ zUON
<l>
co
o}
a)
u JBELIOUoq :U0IIBN
/uo 9y
DJ
o) JE
l
u0 e UBLIO
o)
o)
a)
一 1BZCB6J0
• :uo
suoll udsald 1 0
is
l!s > J 9 is n o :E l
co
a}
U Z 9 > IJB E :IBnp>pul
a>
s8 IPBJd lqoN l
s j
u E0l9 psl
no
(D
a)
0)
ss 30
>
J
.
ul
£e a l
Q
552
C/wpter 17: TQM as a Learning System 553
slopes upw ard, indicating im provem ent in the tra d itio n a l realm
o f in tu itio n and judgm ent when science is applied.
The a d d itio n o f fact-based noble practices m otivates learn-
in g by offering a clear process and appealing to m any types o f
people w ith m any types o f abilities.
nonhierarchical
push,^ noble goals
elements (^infrastructure) Q triggers ^
trained-in c a p a b ilitie s ^
inside routine
elements
^ (〇 < > 0 ) practices
hierarchical practical
elements ^ n frastm ctu re)^ triggers^ goals
NOTES
559
560 A NEW AMERICAN TQM
563
564 A NEW AMERICAN TQM
W V m odel, 47-49,142-144
expanded m odel fo r proac
tive im provem ent, 151-52
and 3 kinds o f data, 145-146
Wallace, John and C.S., 318
W allace Company, 317,530
Weakness orientation, 76-78,
84,112
Corporate Diagnosis
Meeting Global Standards for Excellence
Thomas L. Jackson with Constance E. Dyer
All to o o fte n , strategic planning n eglects an essential first s te p - a n d final
s te p -d ia g n o sis o f th e o rg a n iza tio n ’ s current s ta te . W h a t’s required is a
syste m a tic review o f th e critical fa cto rs in o rg anization al learning and
g ro w th , fa cto rs th at require m o n ito rin g , m e a su re m e n t, a nd m a n a g e m e n t to
ensure th a t y o u r c o m p a n y c o m p e te s su cce ssfully. T h is e xe c u tive w o rk b o o k
pro vid e s a s te p -b y -s te p m e th o d fo r d ia g no sin g an o rg a n iza tio n ’ s strategic
health a nd m easuring its overall c o m p e titive n e s s a g ain st w o rld class s ta n
d a rd s . W ith ch ecklists, ch a rts, a n d detailed e xp la n a tio n s, C orporate
Diagnosis is a practical instruction m a n u a l. T h e pillars o f J a c k s o n ^ d ia g
nostic syste m are stra te g y , stru cture, a n d ca p ab ility. Detailed dia g no stic
q u e stio n s in each area are p ro vid e d as guidelines fo r d e ve lo p in g y o u r o w n
s e lf-a ss e s sm e n t su rve y.
IS B N 1 -5 6 3 2 7-0 8 6 -2 / 10 0 p a g es / $ 6 5 .0 0 / O r d e r C D I A G - B 2 1 8
NOTE: Prices are in U.S. dollars and are subject to change without notice.
llllllllllllllllllllllllllllllllllllllll
X001F5XMS7
New American TQM
Used, Very Good
TvrmTjivi