The+Memory+Jogger+II
Transcript of The+Memory+Jogger+II
TheMemory
Jogger™ IIA Pocket Guide of Tools
for Continuous Improvement& Effective Planning
First Edition
Michael Brassard & Diane Ritter
GOAL/QPC
This handbook is designed to help you and everyperson in your organization to improve daily the
procedures, systems, quality, cost, and yields related toyour job. This continuous improvement process is thefocus of this book—sharing the philosophy and tools
that are fundamental to this effort.
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM1
The Memory Jogger™ was first written, edited, and pro-duced in 1985 by Michael Brassard and Diane Ritter ofGOAL/QPC. They have, in the name of continuous im-provement, worked together again in substantially revis-ing the original tools and writing new material to create amore readable, user-friendly, and comprehensive refer-ence pocket manual. The Memory Jogger™ II now includesthe Seven Quality Control Tools, the Seven Managementand Planning Tools, and a problem-solving case studyexample, which are presented in a format that (we hope)will allow you to find relevant information so easily, you’llalways find space for it in your pocket or pocketbook.
© 1994 by GOAL/QPCAny reproduction of any part of this publication without
the written permission of GOAL/QPC is prohibited.
Edited by Francine OddoCover design by Lori Champney
Layout production by Michele KiersteadGraphics created by Deborah Crovo
GOAL/QPC12B Manor Parkway, Salem, NH 03079-2862
Toll free: 800-643-4316 or 603-893-1944Fax: 603-870-9122
E-mail: [email protected] site: www.goalqpc.com
GOAL/QPC is a nonprofit organization that helps compa-nies continually improve their Quality, Productivity, andCompetitiveness. (Hence, the “QPC” in our name.)
Printed in the United States of AmericaFirst Edition20 19 18
ISBN 1-879364-44-1
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM2
AcknowledgmentsOur sincerest appreciation and thanks go to our dedicated,knowledgeable, and tireless friends who have worked sohard to make this little book a BIG success.
Production & Administrative Staff: Lee Alphen,Steve Boudreau, Paul Brassard, Lori Champney,Deborah Crovo, Brian Hettrick, Michele Kierstead,Fran Kudzia, Larry LeFebre, Francine Oddo,Dorie Overhoff, & Debra Vartanian.
Concept & Content Reviewers: Del Bassett, Casey Collett,Ann McManus, Bill Montgomery, Bob Porter, & Paul Ritter.
Customers & Collaborators: To the customer focus groupsfor their insights and suggestions.
Companies Who Contributed Information for Illustrations:(listed in order of the company name)Bryce Colburne, AT&T TechnologiesRobert Brodeur, Bell CanadaKaren Tate, BGPRip Stauffer, BlueFire PartnersWeston F. Milliken, CUE ConsultingPaula McIntosh, Georgia State Dept. of Human Resources,
Div. of Rehabilitation ServicesDonald Botto, Goodyear Tire & Rubber CompanyTony Borgen & Paul Hearn, Hamilton StandardWalter Tomschin, Hewlett PackardDr. Verel R. Salmon, Millcreek Township School District,
Millcreek Township, PAJack Waddell, Novacor ChemicalsKirk Kochenberger, Parkview Episcopal Medical CenterMarie Sinioris, Rush-Presbyterian-St. Luke’s Medical CenterTim Barnes, SmithKline BeechamDennis Crowley, St. John Fisher CollegeBuzz Stapczynski, Town of Andover, MAKemper Watkins, MSgt., U.S. Air Force, Air Combat CommandCapt. T. C. Splitgerber, U.S. Navy, Naval Dental Center, San Diego
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM3
How to Use The Memory Jogger™ IIThe Memory Jogger™ II is designed for you to use as aconvenient and quick reference guide on the job. Put yourfinger on any tool within seconds!
Know the tool you need? Find it by using the:
Table of Contents. Tools, techniques, the case study—it’sin alphabetical order.Solid tab. This tab runs off the edge of the first page of eachnew section.
Not sure what tool you need? Get an idea by using the:
Tool Selector Chart. This chart organizes the tools bytypical improvement situations, such as working withnumbers, with ideas, or in teams.
What do the different positions of runners mean?
Getting Ready—An important first step is to selectthe right tool for the situation. When you see the“getting ready” position of the runner, expect a briefdescription of the tool’s purpose and a list of benefitsin using the tool.
Cruising—When you see this runner, expect to findconstruction guidelines and interpretation tips. Thisis the action phase that provides you with step-by-step instructions and helpful formulas.
Finishing the Course—When you see this runner,expect to see the tool in its final form. There areexamples from manufacturing, administration/service, and daily life to illustrate the widespreadapplications of each tool.
iv
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM4
ContentsAcknowledgments ........................................................ iiiHow to Use The Memory Jogger™ II ........................... ivTool Selector Chart ....................................................... viIntroduction ..................................................................... 1Activity Network Diagram (AND) .............................. 3
• Gantt Chart .......................................................... 9Affinity Diagram ........................................................... 12Brainstorming ................................................................ 19Cause & Effect/Fishbone Diagram ............................ 23Check Sheet .................................................................... 31Control Chart ................................................................. 36Data Points ..................................................................... 52Flowchart ........................................................................ 56Force Field Analysis ..................................................... 63Histogram ....................................................................... 66Interrelationship Digraph (ID) ................................... 76Matrix Diagram ............................................................. 85Nominal Group Technique (NGT) ............................. 91
• Multivoting ........................................................ 93Pareto Chart ................................................................... 95Prioritization Matrices ............................................... 105Problem-Solving/Process-Improvement Model:
Improvement Storyboard ................................. 115Process Capability ....................................................... 132Radar Chart .................................................................. 137Run Chart ..................................................................... 141Scatter Diagram ........................................................... 145Team Guidelines ......................................................... 150
• Team Stages ..................................................... 150• Conducting Effective Meetings .................... 154
Tree Diagram ............................................................... 156• Process Decision Program Chart (PDPC) .. 160
v
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM5
vi
Tool Selector ChartThis chart organizes the tools by typical improvementsituations, such as working with numbers, with ideas, orin teams.
Workingwith Ideas P
age
#
Gen
erat
ing/
Gro
upin
g
Dec
idin
g
Impl
emen
ting
AND
AffinityBrainstorming
C & E/Fishbone
NGT/Multivoting
FlowchartForce Field
Gantt
IDMatrix
PrioritizationPDPCRadar
Tree
312192356639768591105160137156
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM6
vii
Working withNumbers P
age
#
Cou
ntin
g
Mea
sure
s
Check SheetControl Charts
Data PointsHistogram
Pareto
Process CapabilityRun
Scatter
#
3136526695
132141145
Working in Teams Pag
e #
Impr
ovem
ent R
oadm
ap
Team
Roa
dmap
Storyboard Case Study
Starting TeamsMaintaining Teams
Ending Teams/ProjectsEffective Meetings
115150151153154
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM7
IntroductionIn classrooms to board rooms, on manufacturing
floors and in medical centers, organizations around theworld are using continuous quality improvement (CQI) astheir strategy to bring about dramatic changes in theiroperations. Their purpose is to stay competitive in aworld of instant communication and technologicaladvancement.
These organizations need to meet or exceedcustomer expectations while maintaining a cost-competitive position. Continuous quality improvement(CQI), a systematic, organization-wide approach forcontinually improving all processes that deliver qualityproducts and services, is the strategy many organiza-tions are adopting to meet today’s challenges and toprepare for those down the road.
In pursuing CQI, stick to these four basic principles:
1. Develop a strong customer focus2. Continually improve all processes3. Involve employees4. Mobilize both data and team knowledge to
improve decision making
1. Develop a Strong Customer FocusTotal customer focus includes the needs of bothexternal and internal customers. External custom-ers are the end users, internal customers are yourcoworkers and other departments in the organiza-tion.
2. Continually Improve All Processes• Identify them. A process is a sequence of repeat-
able steps that lead to some desired end oroutput: a typed document, a printed circuitboard, a “home-cooked” meal, arrival at work,and so on.
1
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM9
EquipmentPeople
MaterialsPolicies
Environment
Processstep
Product orservice
Desired endproduct or
service• • •
InputsValue Add Output/Input Final Output
• Improve them. Use the Plan, Do, Check, Act(PDCA) Cycle: PLAN what you want to accom-plish over a period of time and what you mightdo, or need to do to get there. DO what youplanned on doing. Start on a small scale!CHECK the results of what you did to see if theobjective was achieved. ACT on the informa-tion. If you were successful, standardize theplan, otherwise continue in the cycle to plan forfurther improvement.
3. Involve EmployeesEncourage teams—train them—support them—use their work—celebrate their accomplishments!
4. Mobilize Both Data and Team Knowledge toImprove Decision MakingUse the tools to get the most out of your data and theknowledge of your team members.• Graphically display number and word data;
team members can easily uncover patternswithin the data, and immediately focus on themost important targets for improvement.
• Develop team consensus on the root cause(s) ofa problem and on the plan for improvement.
• Provide a safe and efficient outlet for ideas at alllevels.
Use this book and the tools in it to focus on, improve,involve employees in, and direct your path towardcontinuous quality improvement.
2
MJII Frontmatter 06/14/01 10/7/03, 9:33 AM10
©1994 GOAL/QPC Activity Network/Gantt 3
Activity NetworkDiagram (AND)Scheduling sequential& simultaneous tasks
Why use it?To allow a team to find both the most efficient path andrealistic schedule for the completion of any project bygraphically showing total completion time, the neces-sary sequence of tasks, those tasks that can be donesimultaneously, and the critical tasks to monitor.
What does it do?• All team members have a chance to give a realistic
picture of what their piece of the plan requires,based on real experience
• Everyone sees why he or she is critical to the overallsuccess of the project
• Unrealistic implementation timetables are dis-covered and adjusted in the planning stage
• The entire team can think creatively about how toshorten tasks that are bottlenecks
• The entire team can focus its attention and scarceresources on the truly critical tasks
How do I do it?1. Assemble the right team of people with firsthand
knowledge of the subtasks
2. Brainstorm or document all the tasks needed tocomplete a project. Record on Post-its™
Activity Network/Gantt 3
AND 3-11N 10/7/03, 9:17 AM3
©1994 GOAL/QPC4 Activity Network/Gantt
3. Find the first task that must be done, and place thecard on the extreme left of a large work surface
Determinetarget
audience fornew topic
Job/Task card Post-it™ Notes
4. Ask: “Are there any tasks that can be done simul-taneously with task #1?”• If there are simultaneous tasks, place the task card
above or below task #1. If not, go to the next step.
5. Ask, “What is the next task that must be done?Can others be done simultaneously?”• Repeat this questioning process until all the recorded
tasks are placed in sequence and in parallel.
Tip At each step always ask, “Have we forgottenany other needed tasks that could be done simul-taneously?”
Determinetarget
audiencefor
new topic Assesscom-
petitor’sofferings
Reviewfeedback
fromsimilar
courses
AND 3-11N 10/7/03, 9:17 AM4
©1994 GOAL/QPC Activity Network/Gantt 5
6. Number each task, draw the connecting arrows,and agree on a realistic time for the completion ofeach task• Record the completion time on the bottom half of
each card.
1. Determine target
audience for new topic
3. Assess competitor's
offerings
2. Review feedback from
similar courses
T = 14 days
T = 21 days
T = 7 days
Tip Be sure to agree on the standard time unit foreach task, e.g., days, weeks. Elapsed time is easierthan “dedicated” time, e.g., 8 hours of dedicatedtime versus 8 hours over a two-week period(elapsed time).
7. Determine the project’s critical path• Any delay to a task on the critical path will be added
to the project’s completion time, unless anothertask is accelerated or eliminated. Likewise, theproject’s completion time can be reduced by accel-erating any task on the critical path.
• There are two options for calculating the totalcritical path and the tasks included within it.Longest cumulative path. Identify total projectcompletion time. Add up each path of connectedactivities. The longest cumulative path is the
AND 3-11N 10/7/03, 9:17 AM5
©1994 GOAL/QPC6 Activity Network/Gantt
quickest possible implementation time. This isthe project’s critical path.
Calculated slack. Calculate the “slack” in the start-ing and completion times of each task. This iden-tifies which tasks must be completed exactly asscheduled (on the critical path) and those that havesome latitude.
1. Determine target
audience for new topic
3. Assess competitor's
offerings
2. Review feedback from
similar courses
T = 14 days
T = 21 days
T = 7 days
14 35
14 35
Earliest Start (ES)
Earliest Finish (EF)
LatestStart(LS)
LatestFinish(LF)
0 140 14
14 2128 35
Finding the critical path by calculating the slack
ES = The largest EF of any previous connected task
EF = ES + the time to complete that task
LS = LF - the time to complete that task
LF = The smallest LS of any connected following task
When ES = LS AND EF = LF, that task is on the critical path, and therefore there is no schedule flexibility in this task.
AND 3-11N 10/7/03, 9:17 AM6
©1994 GOAL/QPC Activity Network/Gantt 7
Tip Determining the longest cumulative path issimpler than calculating the slack, but canquickly become confusing in larger ANDs.The calculated slack option determines the totalproject and slack times; and therefore the totalproject time and critical path are identified “auto-matically.”
Pro
ject
dur
atio
n: 1
61 d
ays
Dev
elop
ing
a N
ew C
ours
e
Pro
ject
S
tart
1. D
eter
min
e ta
rget
au
dien
ce fo
r ne
w to
pic
2. R
evie
w
feed
back
fr
om s
imila
r co
urse
s
4. D
evel
op
cour
se
obje
ctiv
es
6. D
evel
op
cour
se
form
at
3. A
sses
s co
mpe
titor
's
offe
rings
5. C
hoos
e ge
ogra
phic
lo
catio
n fo
r fin
al c
ours
e
T =
14
days
0 1
40
14
T =
7
days
14
2128
35
T =
21
days
14
3514
35
T =
2
days
35 3
754
56
T =
14
days
42 5
642
56
T =
7
days
35 4
235
42
1
See
ne
xt
page
AND 3-11N 10/7/03, 9:17 AM7
©1994 GOAL/QPC8 Activity Network/Gantt
7. W
rite
draf
t co
nten
t9.
Dev
elop
vi
sual
aid
s11
. Con
duct
pi
lot o
fferin
gs
8. D
evel
op
mar
ketin
g st
rate
gy
T =
30
days
56
8656
86
T =
30
days
86 1
1686
116
T =
21
days
116
137
116
137
13. M
odify
dr
aft m
ater
ial
T =
21
days
137
158
137
158
T =
7da
ys56
63
71 7
8
10. S
ecur
e a
trai
ning
site
T =
14
days
63 7
778
92
12. D
evel
op
and
prin
t br
ochu
reT
= 2
1da
ys77
98
92 1
13
14. D
istr
ibut
e co
urse
br
ochu
reT
= 4
5da
ys 9
8 1
4311
3 15
8
15. C
ondu
ct
trai
ning
T =
3da
ys15
8 1
6115
8 1
61
Pro
ject
F
inis
h1
from
pr
evio
us
page
Dev
elop
ing
a N
ew C
ours
e
AND 3-11N 10/7/03, 9:17 AM8
©1994 GOAL/QPC Activity Network/Gantt 9
VariationsThe constructed example shown in this section is in theActivity on Node (AON) format. For more information onother formats such as Activity on Arrow (AOA) and Prece-dence Diagram (PDM), see The Memory Jogger Plus+®.
Another widely used, schedule-monitoring method is theGantt chart. It is a simple tool that uses horizontal bars toshow which tasks can be done simultaneously over the lifeof the project. Its primary disadvantage is that it cannotshow which tasks are specifically dependent on each other.
Information provided courtesy of BGP
Qtr
4, 1
993
Qtr
1, 1
994
Qtr
2, 1
994
Qtr
3, 1
994
IDN
ame
1 2 3 4 5 6 7 8 9 10 11 12 13 14
ISO
orie
ntat
ion
Pla
n tra
inin
gA
udito
r tra
inin
gP
lan
pre-
audi
t
Pre
-aud
it
Pre
-aud
it pr
ep.
Pre
-aud
it re
port
Cor
rect
ive
actio
n
Gen
erat
e sc
hedu
leP
lan
audi
t
Aud
it pr
epar
atio
n
Aud
it re
port
Oct
. N
ov.
Dec
.Ja
n. F
eb.
Mar
.A
pr.
May
Jun
.Ju
l. A
ug.
Sep
.
1/3
1/21
1/28
ISO
900
0 A
udit
Sch
edul
e
Rel
ease
doc
umen
tatio
n re
ques
t fo
r pr
e-au
dit
All
docu
men
tatio
n fo
r pr
e-au
dit
colle
cted
Hol
d pr
e-au
dit o
rient
atio
n m
eetin
g
*Cor
Act
= C
orre
ctiv
e ac
tion
Key
Dat
es
Cor
Act* m
eetin
gs
Ful
l aud
it
Rel
ease
full
audi
t sch
edul
eC
omm
ence
full
audi
tF
inis
h fu
ll au
dit
Hol
d fu
ll au
dit o
rient
atio
n m
eetin
g
2/25
6/13
7/29
8/12
AND 3-11N 10/7/03, 9:17 AM9
©1994 GOAL/QPC10 Activity Network/Gantt
Activity NetworkPhase I
ISO 9000 Certification Audit Schedule
Ear
liest
S
tart
(E
S)
Ear
liest
F
inis
h (E
F)
Late
stS
tart
(LS
)
Late
stF
inis
h(L
F)
Info
rmat
ion
prov
ided
cou
rtes
y of
BG
P
Pro
ject
S
tart
1. IS
O
orie
ntat
ion
T =
4
wee
ks0
4
0
4
2. P
lan
train
ing
T =
1
wee
k4
5
4
5
4. P
lan
pre-
audi
t
T =
2w
eeks
5
75
7
5. P
re-a
udit
prep
arat
ion
T =
3
wee
ks7
10
7 1
0
6. P
re-a
udit
T =
2
wee
ks10
12
10 1
2
3. A
udito
r tra
inin
g
T =
4
wee
ks4
8
6 1
0
1
See
ne
xt
page
AND 3-11N 10/7/03, 9:17 AM10
©1994 GOAL/QPC Activity Network/Gantt 11
Activity NetworkPhase I
ISO 9000 Certification Audit Schedule (continued)
Info
rmat
ion
prov
ided
cou
rtes
y of
BG
P
Pro
ject
F
inis
h7.
Pre
-aud
it re
port
T =
1
wee
k12
13
12 1
3
9. G
ener
ate
audi
t sc
hedu
leT
= 1
w
eek
13 1
413
14
10. P
lan
audi
t
T =
16
wee
ks14
30
14 3
0
11. A
udit
prep
arat
ion
T =
2
wee
ks30
32
30 3
2
13. A
udit
repo
rt
T =
2
wee
ks39
41
39 4
1
8. C
orre
ctiv
e ac
tion
T =
24
wee
ks1
3
37
15
3
9
1
14. C
orre
ctiv
e ac
tion
mee
tings
T =
4
wee
ks41
45
41 4
5
from
pr
evio
us
page
12. F
ull
audi
t
T =
7
wee
ks32
39
32 3
9
Not
e: T
he A
ND
sho
ws
that
the
cert
ific
atio
n p
roce
ss w
ill t
ake
45 w
eeks
. The
bol
d a
rrow
,in
dic
atin
g th
e cr
itic
al p
ath,
cle
arly
sho
ws
thos
e ta
sks
that
mu
st b
e co
mp
lete
d a
ssc
hed
ule
d. T
he ta
sks
off t
he c
riti
cal p
ath
wil
l als
o re
quir
e ca
refu
l mon
itor
ing
sinc
e th
ere
is o
nly
two
wee
ks o
f sl
ack
tim
e in
the
sche
du
le.
AND 3-11N 10/7/03, 9:17 AM11
©1994 GOAL/QPC12 Affinity
What are the issues involved in planningfun family vacations?
Affinity DiagramGathering &grouping ideas
Why use it?To allow a team to creatively generate a large number ofideas/issues and then organize and summarize naturalgroupings among them to understand the essence of aproblem and breakthrough solutions.
What does it do?• Encourages creativity by everyone on the team at
all phases of the process• Breaks down longstanding communication barriers• Encourages non-traditional connections among
ideas/issues• Allows breakthroughs to emerge naturally, even
on long-standing issues• Encourages “ownership” of results that emerge
because the team creates both the detailed inputand general results
• Overcomes “team paralysis,” which is brought onby an overwhelming array of options and lack ofconsensus
How do I do it?1. Phrase the issue under discussion in a full sentence
12 Affinity
Affinity 12-18N 10/7/03, 9:16 AM12
©1994 GOAL/QPC Affinity 13
Tip From the start, reach consensus on the choiceof words you will use. Neutral statements workwell, but positive, negative, and solution-ori-ented questions also work.
2. Brainstorm at least 20 ideas or issuesa) Follow guidelines for brainstorming.b) Record each idea on a Post-it™ note in bold, large
print to make it visible 4–6 feet away. Use atminimum, a noun and a verb. Avoid using singlewords. Four to seven words work well.
What are the issues involved in planning fun family vacations?
Ask kids for ideas
Find a good range of price alternatives
Consider everyone's
hobbies
Determine total budget
Use a creative
travel agent
Find locations with activities for all ages
Recall good vacations of
the past
Look at family pictures
Combine vacation with business trip
Illustration Note: There are 10 to 40 more ideas in a typical Affinity Diagram
Tip A “typical” Affinity has 40–60 items; it is notunusual to have 100–200 ideas.
3. Without talking: sort ideas simultaneously into5–10 related groupingsa) Move Post-it™ notes where they fit best for you;
don’t ask, simply move any notes that you thinkbelong in another grouping.
Affinity 12-18N 10/7/03, 9:16 AM13
©1994 GOAL/QPC14 Affinity
b) Sorting will slow down or stop when each personfeels sufficiently comfortable with the groupings.
Tip Sort in silence to focus on the meaning behindand connections among all ideas, instead of emo-tions and “history” that often arise in discussions.
Tip As an idea is moved back and forth, try to seethe logical connection that the other person ismaking. If this movement continues beyond a rea-sonable point, agree to create a duplicate Post-it™.
Tip It is okay for some notes to stand alone. These“loners” can be as important as others that fit intogroupings naturally.
4. For each grouping, create summary or headercards using consensusa) Gain a quick team consensus on a word or phrase
that captures the central idea/theme of each
What are the issues involved in planning fun family vacations?
Illustration Note: There are 5 to 10 more groupings of ideas in a typical Affinity Diagram
Ask kids for ideas
Find a good range of price alternatives
Consider everyone's
hobbies
Determine total budget
Use a creative
travel agent
Find locations with activities for all ages
Recall good vacations of
the past
Look at family pictures
Combine vacation with business trip
Affinity 12-18N 10/7/03, 9:16 AM14
©1994 GOAL/QPC Affinity 15
grouping; record it on a Post-it™ note and place itat the top of each grouping. These are draftheader cards.
b) For each grouping, agree on a concise sentencethat combines the grouping’s central idea andwhat all of the specific Post-it™ notes add to thatidea; record it and replace the draft version. Thisis a final header card.
c) Divide large groupings into subgroups as neededand create appropriate subheaders.
d)Draw the final Affinity Diagram connecting allfinalized header cards with their groupings.
Illustration Note: There are 5 to 10 groupings of ideas in a typical Affinity. This is a partial Affinity.
What are the issues involved in planning fun family vacations?
Ask kids for ideas
Find a good range of price alternatives
Consider everyone's
hobbies
Determine total budget
Use a creative
travel agent
Find locations with activities for all ages
Recall good vacations of
the past
Look at family pictures
Combine vacation with business trip
Define an ideal vacation
through family consensus
Determine the most your
budget will buy
Use multiple sources for
vacation research
Affinity 12-18N 10/7/03, 9:16 AM15
©1994 GOAL/QPC16 Affinity
Tip Spend the extra time needed to do solid headercards. Strive to capture the essence of all of the ideasin each grouping. Shortcuts here can greatly reducethe effectiveness of the final Affinity Diagram.
It is possible that a note within a groupingcould become a header card. However, don’tchoose the “closest one” because it’s convenient.The hard work of creating new header cards oftenleads to breakthrough ideas.
VariationsAnother popular form of this tool, called the KJMethod, was developed by the Japanese anthropolo-gist Jiro Kawakita while he was doing fieldwork in the1950s. The KJ Method, identified with Kawakita’s ini-tials, helped the anthropologist and his students gatherand analyze data. The KJ Method differs from theAffinity Diagram described above in that the cards arefact-based and go through a highly structured refine-ment process before the final diagram is created.
Affinity 12-18N 10/7/03, 9:16 AM16
©1994 GOAL/QPC Affinity 17
AffinityIssues Surrounding
Implementation of the Business Plan
The
gro
up c
ould
func
tion
mor
e ef
fect
ivel
y
Fun
ctio
nal g
roup
s no
t tr
ustin
g ea
ch o
ther
Gro
up m
embe
rs n
ot
mak
ing
indi
vidu
al
com
mitm
ent t
o su
cces
s of
the
plan
Rew
ards
do
not
com
pens
ate
team
pl
ayin
g
No
stro
ng c
omm
itmen
t to
the
grou
p
Insu
ffici
ent t
eam
ap
proa
ch to
new
pr
oduc
t dev
elop
men
t an
d in
trod
uctio
n
The
m/u
s pe
rcep
tion
Com
mun
icat
ion
betw
een
func
tiona
l gr
oups
diff
icul
t
Gro
up is
not
foca
l po
int f
or c
onfli
ct
reso
lutio
n
Com
mun
icat
ion
issu
es
with
in th
e gr
oup
Info
rmat
ion
prov
ided
cour
tesy
of G
oody
ear
Rec
onci
liatio
n w
ith
corp
orat
e re
sour
ce
allo
catio
n
Ow
ners
hip
of p
lan
does
n’t c
ross
fu
nctio
nal l
ines
Pla
n no
t int
egra
ted
Poo
r de
finiti
on o
f pr
iorit
izat
ion
for
mar
ket
intr
oduc
tion
Fig
htin
g da
ily
prob
lem
s (a
lliga
tor/
swam
p)
Pla
nnin
g ap
proa
ch
not s
tand
ardi
zed
Mea
ns n
ot c
lear
ly
defin
ed
Unr
ealis
tic g
oals
cr
eate
Blu
e S
ky
attit
ude
Pla
n is
not
link
ed to
un
it fin
anci
al g
oals
Our
bus
ines
s pl
anni
ng
appr
oach
mus
t be
impr
oved
Mor
e gr
oupi
ngs
on n
ext p
age
Affinity 12-18N 10/7/03, 9:16 AM17
©1994 GOAL/QPC18 Affinity
AffinityIssues Surrounding
Implementation of the Business Plan (cont.)
Note: The Affinity helped the team bring focus to the manyopinions on business planning. The headers that surfacedbecame the key issues in the ID example (shown in the ID toolsection).
Lim
ited
reso
urce
s ar
e a
chal
leng
e
Com
plex
ity d
riven
by
cust
omer
dem
ands
re
quire
s ad
ded
inve
stm
ent
Cap
ital a
vaila
bilit
y lim
its o
ppor
tuni
ties
Lack
of t
ime
and
reso
urce
s
Lack
inte
grat
ion
of
supp
ort g
roup
pla
ns
Fas
ter
pace
of
prod
uct
intr
oduc
tions
st
retc
hes
reso
urce
s
Fas
t new
pro
duct
in
trod
uctio
ns s
tret
ch
reso
urce
s
Cap
acity
may
not
m
eet n
eeds
Sal
es fo
reca
st is
no
t acc
urat
e
Pro
duct
ion
capa
city
Pro
duct
ion
capa
bilit
y to
su
ppor
t cha
ngin
g re
quire
men
ts
Ext
erna
l fa
ctor
s im
pact
im
plem
enta
tion
New
gov
ernm
ent
regu
latio
ns
Pos
sibi
lity
of
econ
omic
do
wnt
urn
Info
rmat
ion
prov
ided
cour
tesy
of G
oody
ear
Affinity 12-18N 10/7/03, 9:16 AM18
©1994 GOAL/QPC Brainstorming 19Brainstorming 19
BrainstormingCreating bigger& better ideas
Why use it?To establish a common method for a team to creativelyand efficiently generate a high volume of ideas on anytopic by creating a process that is free of criticism andjudgment.
What does it do?• Encourages open thinking when a team is stuck in
“same old way” thinking• Gets all team members involved and enthusiastic
so that a few people don’t dominate the wholegroup
• Allows team members to build on each other’screativity while staying focused on their jointmission
How do I do it?There are two major methods for brainstorming.
• Structured. A process in which each team mem-ber gives ideas in turn.
• Unstructured. A process in which team membersgive ideas as they come to mind.
Either method can be done silently or aloud.
Brainstorming 19-22N 10/7/03, 9:18 AM19
©1994 GOAL/QPC20 Brainstorming
Structured1. The central brainstorming question is stated,
agreed on, and written down for everyone to seeBe sure that everyone understands the question,issue, or problem. Check this by asking one or twomembers to paraphrase it before recording it on aflipchart or board.
2. Each team member, in turn, gives an idea. Noidea is criticized. Ever!With each rotation around the team, any membercan pass at any time. While this rotation processencourages full participation, it may alsoheighten anxiety for inexperienced or shy teammembers.
3. As ideas are generated, write each one in large,visible letters on a flipchart or other writingsurfaceMake sure every idea is recorded with the samewords of the speaker, don’t interpret or abbrevi-ate. To ensure this, the person writing shouldalways ask the speaker if the idea has beenworded accurately.
4. Ideas are generated in turn until each personpasses, indicating that the ideas (or members)are exhaustedKeep the process moving and relatively short—5 to 20 minutes works well, depending on howcomplex the topic is.
5. Review the written list of ideas for clarity and todiscard any duplicatesDiscard only ideas that are virtually identical. It isoften important to preserve subtle differencesthat are revealed in slightly different wordings.
Brainstorming 19-22N 10/7/03, 9:18 AM20
©1994 GOAL/QPC Brainstorming 21
UnstructuredThe process is the same as in the structured methodexcept that ideas are given by everyone at any time.There is no need to “pass” since ideas are not solicitedin rotation.
VariationsThere are many ways to stimulate creative team think-ing. The common theme among all of them is thestimulation of creativity by taking advantage of thecombined brain power of a team. Here are three ex-amples:
• Visual brainstorming. Individuals (or the team)produce a picture of how they see a situation orproblem.
• Analogies/free-word association. Unusual con-nections are made by comparing the problem toseemingly unrelated objects, creatures, or words.For example: “If the problem was an animal, whatkind would it be?”
• 6-3-5 method. This powerful, silent method isproposed by Helmut Schlicksupp in his bookCreativity Workshop. It is done as follows:
a) Based on a single brainstorming issue, eachperson on the team (usually 6 people) has 5minutes to write down 3 ideas on a sheet of paper.
b) Each person then passes his or her sheet ofpaper to the next person, who has 5 moreminutes to add 3 more ideas that build on thefirst 3 ideas.
Brainstorming 19-22N 10/7/03, 9:18 AM21
©1994 GOAL/QPC22 Brainstorming
c) This rotation is repeated as many times asthere are team members, e.g., 6 team mem-bers = 6 rotations, 6 sheets of paper, 18 ideasper sheet.
This interesting process forces team members toconsciously build on each other’s perspectivesand input.
Brainstorming 19-22N 10/7/03, 9:18 AM22
©1994 GOAL/QPC C & E/Fishbone 23
Cause & Effect/Fishbone DiagramFind & cure causes,NOT symptoms
Why use it?To allow a team to identify, explore, and graphicallydisplay, in increasing detail, all of the possible causesrelated to a problem or condition to discover its rootcause(s).
What does it do?• Enables a team to focus on the content of the prob-
lem, not on the history of the problem or differingpersonal interests of team members
• Creates a snapshot of the collective knowledgeand consensus of a team around a problem. Thisbuilds support for the resulting solutions
• Focuses the team on causes, not symptoms
How do I do it?1. Select the most appropriate cause & effect format.
There are two major formats:• Dispersion Analysis Type is constructed by plac-
ing individual causes within each “major” causecategory and then asking of each individual cause“Why does this cause (dispersion) happen?” Thisquestion is repeated for the next level of detailuntil the team runs out of causes. The graphicexamples shown in Step 3 of this tool section arebased on this format.
C & E/Fishbone 23
C_E 23-30N 10/7/03, 9:21 AM23
©1994 GOAL/QPC24 C & E/Fishbone
• Process Classification Type uses the major stepsof the process in place of the major cause cat-egories. The root cause questioning process is thesame as the Dispersion Analysis Type.
2. Generate the causes needed to build a Cause &Effect Diagram. Choose one method:• Brainstorming without previous preparation• Check Sheets based on data collected by team
members before the meeting
3. Construct the Cause & Effect/Fishbone Diagrama) Place the problem statement in a box on the
righthand side of the writing surface.• Allow plenty of space. Use a flipchart sheet,
butcher paper, or a large white board. A papersurface is preferred since the final Cause & EffectDiagram can be moved.
Late pizza deliveries on
Fridays & Saturdays
Causes“Bones”
(Major cause categories)Effect
Tip Make sure everyone agrees on the problemstatement. Include as much information as possibleon the “what,” “where,” “when,” and “how much”of the problem. Use data to specify the problem.
C_E 23-30N 10/7/03, 9:21 AM24
©1994 GOAL/QPC C & E/Fishbone 25
b) Draw major cause categories or steps in the pro-duction or service process. Connect them to the“backbone” of the fishbone chart.
• Be flexible in the major cause “bones” that areused. In a Production Process the traditionalcategories are: Machines (equipment), Methods(how work is done), Materials (components orraw materials), and People (the human element).In a Service Process the traditional methods are:Policies (higher-level decision rules), Procedures(steps in a task), Plant (equipment and space),and People. In both types of processes, Environ-ment (buildings, logistics, and space), and Mea-surement (calibration and data collection) arealso frequently used. There is no perfect set ornumber of categories. Make them fit the problem.
Late pizza deliveries on
Fridays & Saturdays
Machinery/Equipment People
MaterialsMethods
Illustration Note: In a Process Classification Typeformat, replace the major “bone” categories with:“Order Taking,” “Preparation,” “Cooking,” and “Delivery.”
C_E 23-30N 10/7/03, 9:21 AM25
©1994 GOAL/QPC26 C & E/Fishbone
Late pizza deliveries
on Fridays & Saturdays
Machinery/Equipment People
MaterialsMethods
People don’tshow up
Unreliable cars
Ovenstoo small
Driversget lost
Run out of ingredients
Poor dispatching
Poor handling of
large orders
c) Place the brainstormed or data-based causes inthe appropriate category.
• In brainstorming, possible causes can be placed ina major cause category as each is generated, oronly after the entire list has been created. Eitherworks well but brainstorming the whole list firstmaintains the creative flow of ideas without beingconstrained by the major cause categories orwhere the ideas fit in each “bone.”
• Some causes seem to fit in more than one category.Ideally each cause should be in only one category, butsome of the “people” causes may legitimately belong intwo places. Place them in both categories and see howthey work out in the end.
Tip If ideas are slow in coming, use the major causecategories as catalysts, e.g., “What in ‘materials’ iscausing . . . ?”
d) Ask repeatedly of each cause listed on the“bones,” either:
• “Why does it happen?” For example, under “Runout of ingredients” this question would lead to
C_E 23-30N 10/7/03, 9:21 AM26
©1994 GOAL/QPC C & E/Fishbone 27
more basic causes such as “Inaccurate ordering,”“Poor use of space,” and so on.
Late
piz
za
deliv
erie
s on
F
riday
s &
S
atur
days
Peo
ple
don'
t sho
w u
p
Hig
h tu
rnov
erLo
w p
ay
No
team
wor
k No
trai
ning
Peo
ple
Don
't kn
ow to
wn
Driv
ers
get l
ost
Hig
h tu
rnov
er
Get
wro
ng
info
rmat
ion
Poo
r tr
aini
ng
Rus
hed
Mac
hine
ry/E
quip
men
t
Unr
elia
ble
cars
Low
pay
Kid
s ow
n ju
nks
No
mon
ey fo
r re
pairs
Ove
ns to
o sm
all
Poo
r us
e of
spa
cePoo
r tr
aini
ngH
igh
turn
over
No
capa
city
for
peak
per
iods
Met
hods
Poo
r ha
ndlin
g of
la
rge
orde
rs
Lack
of e
xper
ienc
eP
oor
disp
atch
ing
Man
y ne
w s
tree
ts
Don
't kn
ow
tow
n Hig
h tu
rnov
er
Run
out
of i
ngre
dien
ts
Inac
cura
te
orde
ring
Poo
r us
e of
spa
ce
Lack
of
trai
ning M
ater
ials
Hig
h tu
rnov
er
Hig
h tu
rnov
er
C_E 23-30N 10/7/03, 9:21 AM27
©1994 GOAL/QPC28 C & E/Fishbone
• “What could happen?” For example, under “Runout of ingredients” this question would lead to adeeper understanding of the problem such as“Boxes,” “Prepared dough,” “Toppings,” and so on.
Tip For each deeper cause, continue to push fordeeper understanding, but know when to stop. Arule of thumb is to stop questioning when a causeis controlled by more than one level of manage-ment removed from the group. Otherwise, theprocess could become an exercise in frustration.Use common sense.
e) Interpret or test for root cause(s) by one or more ofthe following:
• Look for causes that appear repeatedly within oracross major cause categories.
• Select through either an unstructured consensusprocess or one that is structured, such as NominalGroup Technique or Multivoting.
• Gather data through Check Sheets or other for-mats to determine the relative frequencies of thedifferent causes.
VariationsTraditionally, Cause & Effect Diagrams have been cre-ated in a meeting setting. The completed “fishbone” isoften reviewed by others and/or confirmed with datacollection. A very effective alternative is CEDAC®, inwhich a large, highly visible, blank fishbone chart isdisplayed prominently in a work area. Everyone postsboth potential causes and solutions on Post-it™ notes ineach of the categories. Causes and solutions are re-viewed, tested, and posted. This technique opens upthe process to the knowledge and creativity of everyperson in the operation.
C_E 23-30N 10/7/03, 9:21 AM28
©1994 GOAL/QPC C & E/Fishbone 29
Cause & Effect/FishboneBed Assignment Delay
Sys
tem
inco
rrec
t
Mac
hine
(P
CIS
)Ti
min
g
Hos
pita
l pro
cedu
res
Com
mun
icat
ion
Pat
ient
wai
ts
for
bed
Not
ent
ered
Not
use
d
No
trus
tN
eed
mor
e tr
aini
ng
Fun
ctio
ns n
ot u
sefu
l
Not
use
dpe
ndin
g di
scha
rge
Dis
char
ged
patie
nt
did
not l
eave
Wai
t for
res
ults
Wai
t for
lunc
hW
ait f
or r
ide
Cal
l hou
seke
epin
g to
o la
te
Wai
t for
MD Cal
l hou
seke
epin
g to
o ea
rly
Thi
nk it
will
take
m
ore
time
Pat
ient
arr
ives
to
o ea
rly
Tran
sfer
too
early
fr
om a
noth
er h
ospi
tal
Cal
l hou
seke
epin
g w
hen
clea
n
Nur
sing
sho
rtag
e
Uni
t cle
rk s
taffi
ng
Uni
t cle
rk tr
aini
ng
Res
ourc
es
Uni
t cle
rk u
naw
are
of d
isch
arge
or
tran
sfer
On
brea
kN
ot to
ldShi
ft ch
ange
Res
erva
tion
unaw
are
Not
ent
eredU
nit s
witc
h be
dA
dmitt
ing
unaw
are
bed
is c
lean
Del
ayed
en
try
San
dbag
Too
busy
Inap
prop
riate
E
R a
dmitt
ance
Man
y tr
ansf
ers
Spe
cial
ty b
eds
Car
diac
mon
itors
Dou
ble
room
sP
hysi
cian
did
no
t writ
e or
der
Med
icin
e ad
mit
quot
a
Phy
sici
an m
isus
e –
inpa
tient
MD
pro
cedu
res
Information provided courtesy ofRush-Presbyterian-St. Luke’s Medical Center
C_E 23-30N 10/7/03, 9:21 AM29
©1994 GOAL/QPC30 C & E/Fishbone
Cause & Effect/FishboneCauses for Bent Pins (Plug-In Side)
Information provided courtesy of AT&T
Ben
t pin
s pl
ug-in
sid
e
Han
dlin
gP
eopl
e
Des
ign
Tool
s
Met
hods
Gen
eral
han
dlin
g
Rep
air
Test
Sto
rage
In/O
ut o
f Box
es
Bac
kpla
ne m
ount
ing
Mou
nt to
fram
eH
eavy
/Aw
kwar
d
Gau
ging
Impr
oper
in
sert
ion
Dam
aged
con
nect
orS
tora
ge
Ret
aine
r in
sert
ion
Pla
cem
ent -
pi
ns o
n be
nch
No/
Don
't Li
ke
Fixt
ures
Myl
ar ta
pe
Scr
ibe
slip
sP
lace
men
t on
edg
esA
ssem
bly
diffi
culty
Larg
e pa
rt #
Com
plex
de
sign
004
prep
VA
Sho
ws
pins
on
benc
h
No
Fixt
ure
Car
eles
snes
s
Lack
of a
ttent
ion
Atti
tude O
n-th
e-jo
b tr
aini
ng
Usi
ng w
rong
tool
s
New
tool
s -
long
lead
tim
eIm
prop
er s
izesBits
Not
eno
ugh
Gau
ges
Scr
ews
com
e lo
ose
Dam
aged
con
nect
ors
Test
fixt
ures
Dam
aged
con
nect
ors
Sto
rage
Lack
of
fixtu
res
Not
Eno
ugh
Uni
vers
al
Cas
ting
ties/
plie
rs
No
Stif
fene
r P
late
sB
ad p
anel
al
ignm
ent –
bar
e
Des
igne
r no
t on
site
Des
igne
r ca
n't r
eact
to
pro
blem
s
Not
des
igne
d fo
r m
anuf
actu
ring
2 ba
ckpl
anes
Larg
e pa
rt #
Ass
embl
y di
fficu
lty
C_E 23-30N 10/7/03, 9:22 AM30
©1994 GOAL/QPC Check Sheet 31
1 2 TotalA |||| ||| 7
B ||| |||| 7
C || | 3
Check SheetCounting &accumulating data
Why use it?To allow a team to systematically record and compiledata from historical sources, or observations as theyhappen, so that patterns and trends can be clearlydetected and shown.
What does it do?• Creates easy-to-understand data that come from
a simple, efficient process that can be applied toany key performance areas
• Builds, with each observation, a clearer picture of “thefacts” as opposed to the opinions of each teammember
• Forces agreement on the definition of each condi-tion or event (every person has to be looking forand recording the same thing)
• Makes patterns in the data become obvious quickly
How do I do it?1. Agree on the definition of the events or conditions
being observed• If you are building a list of events or conditions as
the observations are made, agree on the overalldefinition of the project.Example: If you are looking for reasons for latepayments, agree on the definition of “late.”
Check Sheet 31
Check Sheet 31-35N 10/7/03, 9:23 AM31
©1994 GOAL/QPC32 Check Sheet
• If you are working from a standard list of eventsor conditions, make sure that there is agreementon the meaning and application of each one.Example: If you are tracking sales calls fromvarious regions, make sure everyone knowswhich states are in each region.
2. Decide who will collect the data; over what period;and from what sources• Who collects the data obviously depends on the
project and resources. The data collection periodcan range from hours to months. The data cancome from either a sample or an entire popula-tion.
• Make sure the data collector(s) have both the timeand knowledge they need to collect accurate in-formation.
• Collect the data over a sufficient period to be surethe data represents “typical” results during a“typical” cycle for your business.
• Sometimes there may be important differenceswithin a population that should be reflected bysampling each different subgroup individually.This is called stratification.Example: Collect complaint data from businesstravelers separately from other types of travelers.Collect scrap data from each machine separately.
Tip It must be safe to record and report “badnews,” otherwise the data will be filtered.
Check Sheet 31-35N 10/7/03, 9:23 AM32
©1994 GOAL/QPC Check Sheet 33
j
g
DateTotal
3/10 3/11 3/12 3/13 3/14 3/15 3/16
Lab delays 9 4 6 6 3 12 12 52
No beds available 2 7 2 4 5 8 3 31
Incompletepatient
information7 3 1 2 2 4 5 24
Total 33 28 36 30 25 47 38 237
Project: Admission Delays
Location: Emergency Room
Name: (if applicable)
Dates: 3/10 to 3/16
Shift: All
Reason:
cd
h j
e
g
abf i
3. Design a Check Sheet form that is clear, complete,and easy to use• A complete Check Sheet, illustrated below, in-
cludes the following:Source Information
a Name of project b Location of data collectionc Name of person recording data, if it appliesd Date(s)e Other important identifiers
Content Information f Column with defect/event name
Columns with collection days/datesh Totals for each column i Totals for each row
Grand total for both the columns and rows
Check Sheet 31-35N 10/7/03, 9:23 AM33
©1994 GOAL/QPC34 Check Sheet
4. Collect the data consistently and accurately• Make sure all entries are written clearly.Tip Managers and/or team members can do their
part to help the data collector(s) do their job wellby simply showing an interest in the project. Askthe collector(s) how the project is working out.Show your support—tell the data collector(s) youthink it is important to collect the information.Above all—act on the data as quickly as possible!
VariationsDefect LocationShows the concentration of defects by marking a draw-ing of the product each time a defect is found.
Generic computer monitor
xxx XXxxx
Front
xxxxxx
Back
xx
Project: Monitor casing damageLocation: Assembly
Name: (if applicable) Shift: FirstDate: 6/11
Task ChecklistTasks in producing a product or delivering a service arechecked off as they are done. In complex processes thisis a form of “mistake-proofing.”
Check Sheet 31-35N 10/7/03, 9:23 AM34
©1994 GOAL/QPC Check Sheet 35
Check SheetKeyboard Errors
in Class Assignment
Information provided courtesy of Millcreek Township School District,Millcreek Township, Pennsylvania
Cen
terin
gS
pelli
ngP
unct
uatio
nM
isse
d pa
ragr
aph
Wro
ng n
umbe
rsW
rong
pag
e nu
mbe
rsTa
bles
|| ||||
|||||
| ||
|| |
||||| ||| | |||
|
Mis
take
sM
arch
12
3
||| ||||
||||
||||
| ||
||| |||
|| |||
|
||| ||||
||||
||||
||||
| ||| || ||||
Tota
l
8 23 40 4 10 4 13
3435
33To
tal
102
Check Sheet 31-35N 10/7/03, 9:23 AM35
©1994 GOAL/QPC36 Control Charts
Control ChartsRecognizingsources of variation
Why use it?To monitor, control, and improve process performanceover time by studying variation and its source.
What does it do?• Focuses attention on detecting and monitoring
process variation over time• Distinguishes special from common causes of
variation, as a guide to local or managementaction
• Serves as a tool for ongoing control of a process• Helps improve a process to perform consistently
and predictably for higher quality, lower cost,and higher effective capacity
• Provides a common language for discussing pro-cess performance
How do I do it?There are many types of Control Charts. The ControlChart(s) that your team decides to use will be deter-mined by the type of data you have. Use the TreeDiagram on the next page to determine which ControlChart(s) will best fit your situation. Other types ofControl Charts, which are beyond the scope of thisbook, include the Pre-Control Chart, the MovingAverage & Range Chart, the Cumulative Sum Chart,and Box Plots.
UCL
X
LCL
36 Control Charts
Contrl Chrts 36-51N 10/7/03, 9:24 AM36
©1994 GOAL/QPC Control Charts 37
Based on the type of data and sample size you have,choose the appropriate Control Chart.
Choose Appropriate Control Chart
Variable data:measured & plotted on
a continuous scale, e.g., time, temperature,
cost, figures
Attribute data:counted & plotted as discrete events, e.g., shipping errors, %
waste, absenteeism
Defective** data
Defect* data
Variable sample size, usually ≥ 50
Constant sample size, usually ≥ 50
p Chartnp Chartc Chartu Chart
Sample size is small,
usually 3 to 5
Sample size is large,
usually ≥ 10
Sample size = 1
Sample size is small,
median value
X and Rm
* Defect = Failure to meet one of the acceptance criteria. A defective unit might have multiple defects.
** Defective = An entire unit fails to meet acceptance criteria, regardless of the number of defects on the unit.
X and R–
X and s–
X and R~
Constant sample size, usually c > 5
Variable sample size
Contrl Chrts 36-51N 10/7/03, 9:24 AM37
©1994 GOAL/QPC38 Control Charts
Constructing Control Charts
1. Select the process to be charted
2. Determine sampling method and plan• How large a sample can be drawn? Balance the
time and cost to collect a sample with the amountof information you will gather. See the Tree Dia-gram on the previous page for suggested samplesizes.
• As much as possible, obtain the samples underthe same technical conditions: the same machine,operator, lot, and so on.
• Frequency of sampling will depend on whetheryou are able to discern patterns in the data. Con-sider hourly, daily, shifts, monthly, annually,lots, and so on. Once the process is “in control,”you might consider reducing the frequency withwhich you sample.
• Generally, collect 20–25 groups of samples beforecalculating the statistics and control limits.
• Consider using historical data to set a baseline.Tip Make sure samples are random. To establish
the inherent variation of a process, allow theprocess to run untouched, i.e., according to stan-dard procedures.
3. Initiate data collection• Run the process untouched, and gather sampled
data.• Record data on an appropriate Control Chart
sheet or other graph paper. Include any unusualevents that occur.
Contrl Chrts 36-51N 10/7/03, 9:24 AM38
©1994 GOAL/QPC Control Charts 39
4. Calculate the appropriate statisticsa) If you have attribute data, use the Attribute Data
Table, Central Line column.
Attribute Data Table
Type ControlChart
Samplesize Central Line Control Limits
Fractiondefective Variable,
usually≥50p Chart
Numberdefective Constant,
usually≥50np Chart
Numberof defects Constant
c Chart
Numberof defectsper unit Variable
u Chart
*LCLp = p – 3 p(1 – p) n
*UCLp = p + 3 p(1 - p) n
LCLnp = np – 3 np(1 – p)
UCLnp = np + 3 np(1 – p)
LCLc = c – 3 c
UCLc = c + 3 c
*UCLu = u + 3 u n
*LCLu = u – 3 u n
For each subgroup:p = np/n
For all subgroups:p = Σnp/Σn
For each subgroup:np = # defective unitsFor all subgroups:
np = Σnp/k
For each subgroup:c = # defects
For all subgroups:c = Σc/k
For each subgroup:u = c/n
For all subgroups:u = Σc/Σn
np =c =n =
k =
# defective units# of defectssample size withineach subgroup# of subgroups
* This formula creates changing control limits. To avoid this, use average sample sizes n for those samples that are within ±20% of the average sample size. Calculate individual limits for the samples exceeding ±20%.
Contrl Chrts 36-51N 10/7/03, 9:24 AM39
©1994 GOAL/QPC40 Control Charts
b) If you have variable data, use the Variable DataTable, Central Line column.
Variable Data Table
LCLX = X – A3s
Type Control Chart
Samplesize n Control Limits
Average & Range <10, but
usually 3 to 5
Central Line*
–R = (R1+R2+…Rk)
k
Average & Standard Deviation Usually
≥10 –s = (s1+s2+…sk)k
X = (X1+X2+…Xk)=
k
– – –
Median & Range <10, but
usually 3 or 5
–R = (R1+R2+…Rk)
k
X = (X1+X2 +…Xk)k
~ ~ ~~
Individuals& Moving
Range
X and Rm
1
Rm = (R1 + R2 +…Rk–1)k – 1
= (X1+X2 +…Xk)X–
k
–
X = (X1+X2+…Xk)=
k
– – –
Rm = |(Xi + 1 – Xi)|
–
–LCLx = X – A2R–
–LCLR = D3R
–UCLx = X + A2R–
–UCLR = D4R
–UCLs = B4s–LCLs = B3s
–UCLX = X + A2R
~–~
–UCLX = X + E2Rm
–LCLR = D3R
–UCLR = D4R
–LCLX = X – A2R
~~
–LCLX = X – E2Rm
––
–UCLRm = D4Rm
–LCLRm = D3Rm
UCLX = X + A3s––
––
–
––––
X and R–
X and R~
X and s–
~k = # of subgroups, X = median value within each subgroup
*X = ∑Xin
=
=
~
~
Contrl Chrts 36-51N 10/7/03, 9:24 AM40
©1994 GOAL/QPC Control Charts 41
5. Calculate the control limitsa) If you have attribute data, use the Attribute Data
Table, Control Limits column.b) If you have variable data, use the Variable Data
Table, Control Limits column for the correct for-mula to use.
• Use the Table of Constants to match the numericvalues to the constants in the formulas shown inthe Control Limits column of the Variable DataTable. The values you will need to look up willdepend on the type of Variable Control Chart youchoose and on the size of the sample you havedrawn.
Tip If the Lower Control Limit (LCL) of an At-tribute Data Control Chart is a negative number,set the LCL to zero.
Tip The p and u formulas create changing controllimits if the sample sizes vary subgroup to sub-group. To avoid this, use the average sample size,n, for those samples that are within ±20% of theaverage sample size. Calculate individual limitsfor the samples exceeding ±20%.
6. Construct the Control Chart(s)• For Attribute Data Control Charts, construct one
chart, plotting each subgroup’s proportion ornumber defective, number of defects, or defectsper unit.
• For Variable Data Control Charts, construct twocharts: on the top chart plot each subgroup’smean, median, or individuals, and on the bottomchart plot each subgroup’s range or standarddeviation.
Contrl Chrts 36-51N 10/7/03, 9:24 AM41
©1994 GOAL/QPC42 Control Charts
Table of Constants
Sample size
n
X and R Chart X and s Chart
A2 D3 D4 A3 B3 B4 c4*2 1.880 0 3.267 2.659 0 3.267 .7979
3 1.023 0 2.574 1.954 0 2.568 .8862
4 0.729 0 2.282 1.628 0 2.266 .9213
5 0.577 0 2.114 1.427 0 2.089 .9400
6 0.483 0 2.004 1.287 0.030 1.970 .9515
7 0.419 0.076 1.924 1.182 0.118 1.882 .9594
8 0.373 0.136 1.864 1.099 0.185 1.815 .9650
9 0.337 0.184 1.816 1.032 0.239 1.761 .9693
10 0.308 0.223 1.777 0.975 0.284 1.716 .9727
Sample Size
n
X and R Chart X and Rm Chart
A2 D3 D4 E2 D3 D4 d2*2 - - - - 0 3.267 2.659 0 3.267 1.128
3 1.187 0 2.574 1.772 0 2.574 1.693
4 - - - - 0 2.282 1.457 0 2.282 2.059
5 0.691 0 2.114 1.290 0 2.114 2.326
6 - - - - 0 2.004 1.184 0 2.004 2.534
7 0.509 0.076 1.924 1.109 0.076 1.924 2.704
8 - - - - 0.136 1.864 1.054 0.136 1.864 2.847
9 0.412 0.184 1.816 1.010 0.184 1.816 2.970
10 - - - - 0.223 1.777 0.975 0.223 1.777 3.078
~
~
* Useful in estimating the process standard deviation σ.Note: The minimum sample size shown in this chart is 2because variation in the form of a range can only be calculatedin samples greater than 1. The X and Rm Chart creates theseminimum samples by combining and then calculating thedifference between sequential, individual measurements.
Contrl Chrts 36-51N 10/7/03, 9:24 AM42
©1994 GOAL/QPC Control Charts 43
• Draw a solid horizontal line on each chart. Thisline corresponds to the process average.
• Draw dashed lines for the upper and lower con-trol limits.
Interpreting Control Charts• Attribute Data Control Charts are based on one
chart. The charts for fraction or number defective,number of defects, or number of defects per unit,measure variation between samples. Variable DataControl Charts are based on two charts: the oneon top, for averages, medians, and individuals,measures variation between subgroups over time;the chart below, for ranges and standard devia-tions, measures variation within subgroups overtime.
• Determine if the process mean (center line) iswhere it should be relative to your customerspecifications or your internal business needs orobjectives. If not, then it is an indication thatsomething has changed in the process, or thecustomer requirements or objectives havechanged.
• Analyze the data relative to the control limits;distinguishing between common causes and spe-cial causes. The fluctuation of the points withinthe limits results from variation inherent in theprocess. This variation results from commoncauses within the system, e.g., design, choice ofmachine, preventive maintenance, and can onlybe affected by changing that system. However,points outside of the limits or patterns within thelimits, come from a special cause, e.g., humanerrors, unplanned events, freak occurrences, thatis not part of the way the process normallyoperates, or is present because of an unlikelycombination of process steps. Special causes must
Contrl Chrts 36-51N 10/7/03, 9:24 AM43
©1994 GOAL/QPC44 Control Charts
be eliminated before the Control Chart can be usedas a monitoring tool. Once this is done, the pro-cess will be “in control” and samples can be takenat regular intervals to make sure that the processdoesn’t fundamentally change. See“Determining if Your Process is Out of Control.”
• Your process is in “statistical control” if the pro-cess is not being affected by special causes, theinfluence of an individual or machine. All thepoints must fall within the control limits and theymust be randomly dispersed about the averageline for an in-control system.
Tip “Control” doesn’t necessarily mean that theproduct or service will meet your needs. It onlymeans that the process is consistent. Don’t confusecontrol limits with specification limits—specifica-tion limits are related to customer requirements,not process variation.
Tip Any points outside the control limits, onceidentified with a cause (or causes), should beremoved and the calculations and charts redone.Points within the control limits, but showing in-dications of trends, shifts, or instability, are alsospecial causes.
Tip When a Control Chart has been initiated andall special causes removed, continue to plot newdata on a new chart, but DO NOT recalculate thecontrol limits. As long as the process does notchange, the limits should not be changed. Controllimits should be recalculated only when a perma-nent, desired change has occurred in the process,and only using data after the change occurred.
Contrl Chrts 36-51N 10/7/03, 9:24 AM44
©1994 GOAL/QPC Control Charts 45
Tip Nothing will change just because you chartedit! You need to do something. Form a team toinvestigate. See “Common Questions for Investi-gating an Out-of-Control Process.”
Determining if Your Process is“Out of Control”
A process is said to be “out of control” if either one ofthese is true:
1. One or more points fall outside of the control limits
2. When the Control Chart is divided into zones, asshown below, any of the following points are true:
Zone AZone BZone CZone CZone BZone A
Upper Control Limit(UCL)
Average
Lower Control Limit(LCL)
a) Two points, out of three consecutive points, are onthe same side of the average in Zone A or beyond.
b) Four points, out of five consecutive points, are onthe same side of the average in Zone B or beyond.
c) Nine consecutive points are on one side of theaverage.
d) There are six consecutive points, increasing ordecreasing.
e) There are fourteen consecutive points that alter-nate up and down.
f) There are fifteen consecutive points within ZoneC (above and below the average).
Contrl Chrts 36-51N 10/7/03, 9:24 AM45
©1994 GOAL/QPC46 Control Charts
A B C C B A
UC
L
LC
LX=
12a
2b2c
2d2e
2f
12a
2b2c
2d2e
2f
Test
s fo
r C
on
tro
l
Sou
rce:
Llo
yd S
. Nel
son,
Dir
ecto
r of
Sta
tist
ical
Met
hod
s, N
ashu
a C
orp
orat
ion,
New
Ham
psh
ire
Contrl Chrts 36-51N 10/7/03, 9:24 AM46
©1994 GOAL/QPC Control Charts 47
Common Questions for Investigatingan Out-of-Control Process
❐ Yes ❐ No Are there differences in the meas-urement accuracy of instruments/methods used?
❐ Yes ❐ No Are there differences in the methodsused by different personnel?
❐ Yes ❐ No Is the process affected by the environ-ment, e.g., temperature, humidity?
❐ Yes ❐ No Has there been a significant changein the environment?
❐ Yes ❐ No Is the process affected by predictableconditions? Example: tool wear.
❐ Yes ❐ No Were any untrained personnel in-volved in the process at the time?
❐ Yes ❐ No Has there been a change in the sourcefor input to the process? Example:raw materials, information.
❐ Yes ❐ No Is the process affected by employeefatigue?
❐ Yes ❐ No Has there been a change in policiesor procedures? Example: mainten-ance procedures.
❐ Yes ❐ No Is the process adjusted frequently?❐ Yes ❐ No Did the samples come from different
parts of the process? Shifts? Indi-viduals?
❐ Yes ❐ No Are employees afraid to report “badnews”?
A team should address each “Yes” answer as a potentialsource of a special cause.
Contrl Chrts 36-51N 10/7/03, 9:24 AM47
©1994 GOAL/QPC48 Control Charts
Individuals & MovingRange Chart
IV Lines Connection Time
Process/Operation: IV Lines Connection Open Heart Admissions
Characteristic:Time in seconds
Sample Size:One
Sample Frequency:Each patient
Individuals:Ranges:
Department:Intensive CareBy:EW
Date:6/10
720
630
540
450
360
270
180
90
600
480
540
240
420
450
480
690
240
360
450
300
480
120
240
210
210
180
240
300
300
130
120
300
180
210
— 120
60 300
180
30 30 210
450
120
90 150
180
360
120
30 0 30 60 60 0 120
60 80 120
30
EW
EW
EW
EW
EW
EW
EW
EW
MA
EW
EW
EW
EW
EW
EW
EW
MA
EW
EW
EW
EW
EW
EW
EW
EW
EW
4/9
11 12 13 25 30 5/2 3 4 5 9 13 14 14 16 17 19 20 22 23 23 24 30 30 6/6 6
X
Rm
Who?
When?
420360300 240 180 120 60 0
UCL
X
LCL
Rm
k = 26n = 2
∑X = 8470∑R = 2990
UCL = 645UCL = 392
LCL = 7LCL = 0
LCL
UCL
X = 325.77R = 119.6
––
0
Sec
onds
Information provided courtesy ofParkview Episcopal Medical Center
Note: Something in the process changed, and now it takes lesstime to make IV connections for patients being admitted foropen heart surgery.
Contrl Chrts 36-51N 10/7/03, 9:24 AM48
©1994 GOAL/QPC Control Charts 49
60
50
40
30
20
10
% Failed
Month
Year
UCL
LCL
Historical Statistics:p = 39 UCL = 47 LCL = 31
p
Regular Hours Flex Time
1992 1993
40 36 36 42 42 40 20 26 25 19 20 18 16 10 12 12
Jul
Aug
Sep Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep Oct
Information provided courtesy of U.S. Navy,Naval Dental Center, San Diego
Note: Providing flex time for patients resulted in fewer appoint-ments missed.
p ChartGeneral Dentistry: Percent of
Patients Who Failed to Keep Appointments
Contrl Chrts 36-51N 10/7/03, 9:24 AM49
©1994 GOAL/QPC50 Control Charts
u ChartShop Process Check
Solder Defects
Information provided courtesy of AT&T
10000 –
9000 –
8000 –
7000 –
6000 –
5000 –
4000 –
3000 –
2000 –
1000 –
0 –
Jan
2 –
Jan
3 –
Jan
4 –
Jan
5 –
Jan
6 –
Jan
7 –
Jan
8 –
Jan
9 –
Jan
10
–
Jan
11
–
UCL
UDef
ect
Rat
e (P
PM
)
LCL
Checks (3 per day)
Historical Ave.: 2974 ppmHistorical UCL: 8758 ppmHistorical LCL: 0 ppm
Contrl Chrts 36-51N 10/7/03, 9:24 AM50
©1994 GOAL/QPC Control Charts 51
X & R Chart
Note: Hours 1, 16 and 22 should be reviewed to understandwhy these sample averages are outside the control limts
3.0
n = 10 parts randomly sampled each hour
2.01.81.61.41.21.0.8.6.4.2
LCL
UCL
4.0
4.5
5.0
3.5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Ave
.
3.76
4.21
4.29
4.36
4.13
3.77
4.17
4.21
4.22
4.00
4.30
4.20
4.32
4.18
4.02
3.71
4.08
4.23
3.98
4.46
3.96
3.63
4.48
4.30
4.29
4.13
1.01
1.27
0.48
1.32
1.52
1.03
1.15
1.07
0.70
2.05
0.95
0.99
1.06
1.21
1.33
0.78
1.21
1.23
1.08
1.64
1.20
0.98
0.91
1.19
1.03
1.14
LCL
UCL
X=
Hou
rly
#X
R
Dia
met
er i
n M
illi
met
ers
R
Information provided courtesy of BlueFire Partners, Inc.and Hamilton Standard
Contrl Chrts 36-51N 10/7/03, 9:24 AM51
©1994 GOAL/QPC52 Data Points
Data PointsTurning datainto information
What Type of Data Do You Have?• Words?• Numbers?
– Attribute data? Attribute data can be countedand plotted as discrete events. It includes thecount of the numbers or percentages of good orbad, right or wrong, pass or fail, yes or no.Example: Number of correct answers on a test,number of mistakes per typed page, percentdefective product per shift.
– Variable data? Variable data can be measuredand plotted on a continuous scale.Example: Length, time, volume, weight.
Do You Need to Collect Data?• If you need to know the performance of an entire
population, the more economical and less timeconsuming method is to draw a sample from apopulation. With a sample, you can make infer-ences about, or predict, the performance of apopulation. Basic sampling methods:– Random. Each and every observation or data
measure has an equally likely chance of beingselected. Use a random number table or ran-dom number generator to select the samples.
– Sequential. Every nth sample is selected.– Stratified. A sample is taken from stratified
data groups.
36
78
51
307
43
93
6/8
?
52 Data Points
Data Points 52-55N 10/7/03, 9:26 AM52
©1994 GOAL/QPC Data Points 53
Can You Categorize Your Data Into Subgroups?• When you stratify data, you break it down into
meaningful subcategories or classifications, andfrom this point you can focus your problem solving.
Example: Data often comes from many sourcesbut is treated as if coming from one. Data on minorinjuries for a plant may be recorded as a singlefigure, but that number is actually the sum total ofinjuries by 1) type (cuts, burns, scrapes), 2) location(eyes, hands, feet), and 3) department (mainte-nance, shipping, production). Below is an exampleof how data has been stratified by plant department.
What Patterns are Important in Your Data?Predictable patterns or distributions can be describedwith statistics.
• Measures of location– Mean (or average). Represented by X (or X-bar),
the mean is the sum of the values of the sample(X1, X2, X3 . . . Xn) divided by the total number(n) of sampled data.Example: For the sample: (3, 5, 4, 7, 5)
(3 + 5 + 4 + 7 + 5)
5X = = 4.8
35-30-25-20-15-10-
5-0-
35-30-25-20-15-10-
5-0-
# In
jurie
s/M
onth
Dept. A and B Combined
∆ = Dept. Ao = Dept. B
J F M A M J
# In
jurie
s/M
onth
J F M A M J
∆∆
∆ ∆ ∆∆
o o oo o o
(Appears fairly stable over time.) (Dept. A is running higher and may be increasing over time.)
Data Points 52-55N 10/7/03, 9:26 AM53
©1994 GOAL/QPC54 Data Points
– Median. When sampled data are rank ordered,lowest to highest, the median is the middlenumber.Example: For the sample: (3, 5, 4, 7, 5)Median of (3, 4, 5, 5, 7) = 5When there are an even number of values, themedian is the average of the middle two values.Example: For the sample: (2, 5, 7, 4, 5, 3)Median of (2, 3, 4, 5, 5, 7) = 4.5
– Mode. The most frequently occurring value(s)in a sample.Example: For the sample: (3, 5, 4, 7, 5)Mode = 5
• Measures of variation– Range. Represented by R, the range is the dif-
ference between the highest data value (Xmax)and the lowest data value (Xmin).Example: For the sample: (3, 5, 4, 7, 5)R = 7 – 3 = 4
– Standard Deviation. Represented by s, the stan-dard deviation of a sample measures the varia–tion of the data around the mean. The lessvariation there is of the data values about themean, X, the closer s will be to zero (0).
Data Points 52-55N 10/7/03, 9:27 AM54
©1994 GOAL/QPC Data Points 55
Example: For the sample: (3, 5, 4, 7, 5) X = 4.8
s = [(3 – 4.8)2 + (5 – 4.8)2 + (4 – 4.8)2 + (7 – 4.8)2 + (5 – 4.8)2 ]
5 – 1
= [3.24 + .04 + .64 + 4.84 + .04]
4
= 8.8 4
= 2.2
= 1.48
The square of the standard deviation, s, is referred to asthe variance. Variance is not discussed in this book.
Data Points 52-55N 10/7/03, 9:27 AM55
©1994 GOAL/QPC56 Flowchart
FlowchartPicturingthe process
Why use it?To allow a team to identify the actual flow or sequenceof events in a process that any product or servicefollows. Flowcharts can be applied to anything from thetravels of an invoice or the flow of materials, to the stepsin making a sale or servicing a product.
What does it do?• Shows unexpected complexity, problem areas,
redundancy, unnecessary loops, and where sim-plification and standardization may be possible
• Compares and contrasts the actual versus theideal flow of a process to identify improvementopportunities
• Allows a team to come to agreement on the stepsof the process and to examine which activitiesmay impact the process performance
• Identifies locations where additional data can becollected and investigated
• Serves as a training aid to understand the com-plete process
How do I do it?1. Determine the frame or boundaries of the process
• Clearly define where the process under studystarts (input) and ends (final output).
• Team members should agree to the level of detailthey must show on the Flowchart to clearly un-derstand the process and identify problem areas.
56 Flowchart
Flow Chart 56-62N 10/7/03, 9:27 AM56
©1994 GOAL/QPC Flowchart 57
• The Flowchart can be a simple macro-flowchartshowing only sufficient information to under-stand the general process flow or it might bedetailed to show every finite action and decisionpoint. The team might start out with a macro-flowchart and then add in detail later or onlywhere it is needed.
2. Determine the steps in the process• Brainstorm a list of all major activities, inputs,
outputs, and decisions on a flipchart sheet fromthe beginning of the process to the end.
3. Sequence the steps• Arrange the steps in the order they are carried
out. Use Post-it™ notes so you can move themaround. Don’t draw in the arrows yet.
Tip Unless you are flowcharting a new process,sequence what is, not what should be or the ideal.This may be difficult at first but is necessary to seewhere the probable causes of the problems are inthe process.
4. Draw the Flowchart using the appropriate symbolsAn oval is used to show the materials, informa-tion or action (inputs) to start the process or toshow the results at the end (output) of theprocess.
A box or rectangle is used to show a task oractivity performed in the process. Although mul-tiple arrows may come into each box, usuallyonly one output or arrow leaves each activity box.
Flow Chart 56-62N 10/7/03, 9:27 AM57
©1994 GOAL/QPC58 Flowchart
A diamond shows those points in the processwhere a yes/no question is being asked or adecision is required.
A circle with either a letter or a number identifiesa break in the Flowchart and is continued else-where on the same page or another page.
Arrows show the direction or flow of the process.
• Keep the Flowchart simple using the basic sym-bols listed above. As your experience grows, useother, more graphic symbols to represent thesteps. Other symbols sometimes used include:– A half or torn sheet of paper for a report com-
pleted and/or filed– A can or computer tape wheel for data entry
into a computer database– A large “D” or half circle to identify places in the
process where there is a delay or wait for furtheraction
• Be consistent in the level of detail shown.– A macro-level flowchart will show key action
steps but no decision boxes.– An intermediate-level flowchart will show ac-
tion and decision points.– A micro-level flowchart will show minute detail.
• Label each process step using words that areunderstandable to everyone.
• Add arrows to show the direction of the flow ofsteps in the process. Although not a rule, if youshow all “yes” choices branching down and “no”choices branching to the left, it is easier to follow
➝
A
Flow Chart 56-62N 10/7/03, 9:27 AM58
©1994 GOAL/QPC Flowchart 59
the process. Preferences and space will later dic-tate direction.
• Don’t forget to identify your work. Include thetitle of your process, the date the diagram wasmade, and the names of the team members.
5. Test the Flowchart for completeness• Are the symbols used correctly?• Are the process steps (inputs, outputs, actions,
decisions, waits/delays) identified clearly?• Make sure every feedback loop is closed, i.e.,
every path takes you either back to or ahead toanother step.
• Check that every continuation point has a corre-sponding point elsewhere in the Flowchart or onanother page of the Flowchart.
• There is usually only one output arrow out of anactivity box. If there is more than one arrow, youmay need a decision diamond.
• Validate the Flowchart with people who are noton the team and who carry out the process ac-tions. Highlight additions or deletions they rec-ommend. Bring these back to the team to discussand incorporate into the final Flowchart.
6. Finalize the Flowchart• Is this process being run the way it should be?• Are people following the process as charted?• Are there obvious complexities or redundancies
that can be reduced or eliminated?• How different is the current process from an ideal
one? Draw an ideal Flowchart. Compare the two(current versus ideal) to identify discrepanciesand opportunities for improvements.
Flow Chart 56-62N 10/7/03, 9:27 AM59
©1994 GOAL/QPC60 Flowchart
1.0 Determine party size
2.0Find
location
3.0Invite guests
1.1 Decide on budget
1.2 Decide on guest list
2.1 Decide theme
2.2 Select location
3.1 Complete invitations
3.2 Send invitations
• • •
Planning a Party
VariationsThe type of Flowchart just described is sometimes re-ferred to as a “detailed” flowchart because it includes,in detail, the inputs, activities, decision points, andoutputs of any process. Four other forms, describedbelow, are also useful.
Macro FlowchartRefer to the third bulleted item in Step 1 of this sectionfor a description. For a graphic example, see Step 2 ofthe Improvement Storyboard in the Problem-Solving/Process Improvement Model section.
Top-down FlowchartThis chart is a picture of the major steps in a workprocess. It minimizes the detail to focus only on thosesteps essential to the process. It usually does not in-clude inspection, rework, and other steps that result inquality problems. Teams sometimes study the top-down flowchart to look for ways to simplify or reducethe number of steps to make the process more efficientand effective.
Flow Chart 56-62N 10/7/03, 9:27 AM60
©1994 GOAL/QPC Flowchart 61
Deployment FlowchartThis chart shows the people or departments respon-sible and the flow of the process steps or tasks they areassigned. It is useful to clarify roles and track account-ability as well as to indicate dependencies in the se-quence of events.
Workflow FlowchartThis type of chart is used to show the flow of people,materials, paperwork, etc., within a work setting. Whenredundancies, duplications, and unnecessary com-plexities are identified in a path, people can take actionto reduce or eliminate these problems.
Chris Karin Lauren
No
Yes
Is there time to do graphics?
Draws graphics
Writes ad
Plans ad
Sends adout
Ad completed
Flow Chart 56-62N 10/7/03, 9:27 AM61
©1994 GOAL/QPC62 Flowchart
FlowchartProposed Patient
Appointment Procedure
Appointment book marked “failure”
Preparation of appointment book
Opening of appointment book
Appoint-ment
shore or fleet?
Appointment issued (PT reminded to
confirm 24 hours prior to
appointment)
Refer to ships corpsman, inform PT they can call at 1500
to make their own appointments for the
next working day
Failure report submitted from front desk to fleet liaison
Did patient call in to
confirm 24 hours prior to
appoint-ment?
Appointment book marked
“patient showed”
Appointment canceled and slot refilled by
new patient
Standby patient placed in
appointment slot
Patient given confirmation number
Fleet liaison sends failure notices to
commands
Yes
Yes No
Fleet
Does patient show for appoint-
ment?
No
Shore
Information provided courtesy ofU.S. Navy, Naval Dental Center, San Diego
Flow Chart 56-62N 10/7/03, 9:27 AM62
©1994 GOAL/QPC Force Field 63
Force FieldAnalysisPositives &negatives of change
Why use it?To identify the forces and factors in place that supportor work against the solution of an issue or problem sothat the positives can be reinforced and/or the nega-tives eliminated or reduced.
What does it do?• Presents the “positives” and “negatives” of a situ-
ation so they are easily compared• Forces people to think together about all the aspects
of making the desired change a permanent one• Encourages people to agree about the relative
priority of factors on each side of the “balance sheet”• Encourages honest reflection on the real underly-
ing roots of a problem and its solution
How do I do it?1. Draw a large letter “T” on a flipchart
a) At the top of the T, write the issue or problem thatyou plan to analyze.
• To the far right of the top of the T, write a descrip-tion of the ideal situation you would like toachieve.
b) Brainstorm the forces that are driving you to-wards the ideal situation. These forces may beinternal or external. List them on the left side.
+
Force Field 63
Force Field 63-65N 10/7/03, 9:28 AM63
©1994 GOAL/QPC64 Force Field
c) Brainstorm the forces that are restraining move-ment toward the ideal state. List them on the rightside.
2. Prioritize the driving forces that can be strength-ened or identify restraining forces that wouldallow the most movement toward the ideal state ifthey were removed• Achieve consensus through discussion or by us-
ing ranking methods such as Nominal GroupTechnique and Multivoting.
Tip When choosing a target for change, rememberthat simply pushing the positive factors for achange can have the opposite effect. It is oftenmore helpful to remove barriers. This tends tobreak the “change bottleneck” rather than justpushing on all the good reasons to change.
Force Field 63-65N 10/7/03, 9:28 AM64
©1994 GOAL/QPC Force Field 65
Force FieldFear of Public Speaking
Not
e: T
he F
orce
Fie
ld c
an h
elp
ind
ivid
ual
s an
d te
ams
sele
ct ta
rget
s fo
r cha
nge.
Gen
eral
ly,
whe
n yo
u f
ocu
s on
res
trai
ning
for
ces,
not
dri
ving
for
ces,
thi
s w
orks
bes
t. Fo
r ex
amp
le,
usi
ng in
dex
car
ds
for
key
poi
nts
may
red
uce
the
fear
“M
ay f
orge
t wha
t to
say.
”
Incr
ease
s se
lf-es
teem
Hel
ps c
aree
r
Com
mun
icat
es id
eas
Con
trib
utes
to a
pla
n/so
lutio
n
Enc
oura
ges
othe
rs to
spe
ak
Hel
ps o
ther
s to
cha
nge
Incr
ease
s en
ergy
of g
roup
Hel
ps c
larif
y sp
eake
r’s
idea
s by
get
ting
feed
back
from
oth
ers
Ham
s ca
n be
ham
s (r
ecog
nitio
n fr
om o
ther
s)
Hel
ps o
ther
s to
see
new
per
spec
tive
Pas
t em
barr
assm
ents
Afr
aid
to m
ake
mis
take
s
Lack
of k
now
ledg
e on
the
topi
c
Afr
aid
peop
le w
ill b
e in
diffe
rent
Afr
aid
peop
le w
ill la
ugh
May
forg
et w
hat t
o sa
y
Too
reve
alin
g of
per
sona
l tho
ught
s
Afr
aid
of o
ffend
ing
grou
p
Fea
r th
at n
ervo
usne
ss w
ill s
how
Lack
of c
onfid
ence
in p
erso
nal a
ppea
ranc
e
+ D
rivi
ng
Fo
rces
Res
trai
nin
g F
orc
es –
Idea
l sta
te: T
o sp
eak
conf
iden
tly,
clea
rly, a
nd c
onci
sely
in a
ny s
ituat
ion.
Force Field 63-65N 10/7/03, 9:28 AM65
©1994 GOAL/QPC66 Histogram
HistogramProcess centering,spread, and shape
Why use it?To summarize data from a process that has been col-lected over a period of time, and graphically present itsfrequency distribution in bar form.
What does it do?• Displays large amounts of data that are difficult
to interpret in tabular form• Shows the relative frequency of occurrence of the
various data values• Reveals the centering, variation, and shape of the
data• Illustrates quickly the underlying distribution of
the data• Provides useful information for predicting future
performance of the process• Helps to indicate if there has been a change in the
process• Helps answer the question “Is the process capable
of meeting my customer requirements?”
How do I do it?1. Decide on the process measure
• The data should be variable data, i.e., measuredon a continuous scale. For example: temperature,time, dimensions, weight, speed.
66 Histogram
Histogram66-75N 10/7/03, 9:29 AM66
©1994 GOAL/QPC Histogram 67
2. Gather data• Collect at least 50 to 100 data points if you plan on
looking for patterns and calculating thedistribution’s centering (mean), spread (varia-tion), and shape. You might also consider collect-ing data for a specified period of time: hour, shift,day, week, etc.
• Use historical data to find patterns or to use as abaseline measure of past performance.
3. Prepare a frequency table from the dataa) Count the number of data points, n, in the sample.
9.99.89.7
10.29.99.39.0
10.09.59.6
10.39.59.99.9
9.39.89.4
10.110.710.2
9.59.79.79.49.69.5
10.410.2
10.210.1
9.69.89.39.29.79.4
10.610.1
9.79.89.39.8
9.49.9
10.010.110.3
9.99.79.89.59.59.79.99.69.3
10.19.79.8
10.39.99.79.89.4
10.110.110.1
9.210.2
9.6
9.69.89.9
10.09.89.99.89.6
10.010.2
9.810.0
9.79.5
9.99.9
10.110.210.3
9.89.3
10.09.89.89.7
10.09.79.6
10.110.010.4
9.89.59.59.6
10.310.1
9.510.0
9.79.7
10.7
9.89.6
10.010.7
9.99.49.79.89.69.3
10.09.7
10.7
In this example, there are 125 data points, n = 125.b) Determine the range, R, for the entire sample.
The range is the smallest value in the set of datasubtracted from the largest value. For ourexample:
R = Xmax – Xmin = 10.7 – 9.0 = 1.7
c) Determine the number of class intervals, k, needed.
Histogram66-75N 10/7/03, 9:29 AM67
©1994 GOAL/QPC68 Histogram
For our example, 125 data points would be di-vided into 7–12 class intervals.
Tip These two methods are general rules of thumbfor determining class intervals. In both methods,consider using k = 10 class intervals for ease of“mental” calculation.
Tip The number of intervals can influence the pat-tern of the sample. Too few intervals will pro-duce a tight, high pattern. Too many intervalswill produce a spread out, flat pattern.
d)Determine the class width, H.• The formula for this is:
Number of Data Points
Under 5050 – 100
100 – 250Over 250
Number of Classes (k)
5 – 76 – 107 – 12
10 – 20
H = R = 1.7 = .17 k 10
• Round your number to the nearest value with thesame decimal numbers as the original sample. Inour example, we would round up to .20. It is usefulto have intervals defined to one more decimalplace than the data collected.
e) Determine the class boundaries, or end points.
• Method 1: Take the square root of the total number ofdata points and round to the nearest whole number.
k = 125 = 11.18 = 11 intervals
• Method 2: Use the table below to provide a guide-line for dividing your sample into a reasonablenumber of classes.
Histogram66-75N 10/7/03, 9:29 AM68
©1994 GOAL/QPC Histogram 69
• Use the smallest individual measurement in thesample, or round to the next appropriate lowestround number. This will be the lower end pointfor the first class interval. In our example thiswould be 9.0.
• Add the class width, H, to the lower end point.This will be the lower end point for the next classinterval. For our example:
9.0 + H = 9.0 + .20 = 9.20
Thus, the first class interval would be 9.00 andeverything up to, but not including 9.20, that is,9.00 through 9.19. The second class intervalwould begin at 9.20 and be everything up to, butnot including 9.40.
Tip Each class interval must be mutually exclusive,that is, every data point will fit into one, and onlyone class interval.
• Consecutively add the class width to the lowestclass boundary until the k class intervals and/orthe range of all the numbers are obtained.
f) Construct the frequency table based on the valuesyou computed in item “e.”A frequency table based on the data from ourexample is shown below.
1 9.00-9.19 9.1 l 1 2 9.20-9.39 9.3 l l l l l l l l 9 3 9.40-9.59 9.5 l l l l l l l l l l l l l 16 4 9.60-9.79 9.7 l l l l l l l l l l l l l l l l l l l l l l 27 5 9.80-9.99 9.9 l l l l l l l l l l l l l l l l l l l l l l l l l 31 6 10.00-10.19 10.1 l l l l l l l l l l l l l l l l l l 22 7 10.20-10.39 10.3 l l l l l l l l l l 12 8 10.40-10.59 10.5 l l 2 9 10.60-10.79 10.7 l l l l 5 10 10.80-10.99 10.9 0
Class#
Class Boundaries
Mid-Point Frequency Total
Histogram66-75N 10/7/03, 9:29 AM69
©1994 GOAL/QPC70 Histogram
4. Draw a Histogram from the frequency table• On the vertical line, (y axis), draw the frequency
(count) scale to cover class interval with the high-est frequency count.
• On the horizontal line, (x axis), draw the scalerelated to the variable you are measuring.
• For each class interval, draw a bar with the heightequal to the frequency tally of that class.
40-
30-
20-
10-
0-
Thickness
Specifications: 9 ± 1.5 USL
Spec.Target
9.2 9.4 9.6 9.8 10.0 10.2 10.4 10.6 10.89.0
Fre
quen
cy
Histogram66-75N 10/7/03, 9:29 AM70
©1994 GOAL/QPC Histogram 71
5. Interpret the Histograma) Centering. Where is the distribution centered? Is
the process running too high? Too low?
Processcentered
CustomerRequirement
Processtoo high
Processtoo low
b) Variation. What is the variation or spread of thedata? Is it too variable?
Processwithin
requirements
CustomerRequirements
Processtoo variable
Histogram66-75N 10/7/03, 9:29 AM71
©1994 GOAL/QPC72 Histogram
c) Shape. What is the shape? Does it look like anormal, bell-shaped distribution? Is it positivelyor negatively skewed, that is, more data values tothe left or to the right? Are there twin (bi-modal)or multiple peaks?
Tip Some processes are naturally skewed; don’texpect every distribution to follow a bell-shapedcurve.
Tip Always look for twin or multiple peaks indi-cating that the data is coming from two or moredifferent sources, e.g., shifts, machines, people,suppliers. If this is evident, stratify the data.
d)Process Capability. Compare the results of yourHistogram to your customer requirements orspecifications. Is your process capable of meetingthe requirements, i.e., is the Histogram centeredon the target and within the specification limits?
Normal Distribution
Positively Skewed Negatively Skewed
Bi-Modal Distribution Multi-Modal Distribution
Histogram66-75N 10/7/03, 9:29 AM72
©1994 GOAL/QPC Histogram 73
Tip Get suspicious of the accuracy of the data if theHistogram suddenly stops at one point (such as aspecification limit) without some previous de-cline in the data. It could indicate that defectiveproduct is being sorted out and is not included inthe sample.
(a) Centered and well withincustomer limits.
Action: Maintain present state.
LowerSpecification
Limit
UpperSpecification
Limit
Centering and Spread Compared toCustomer Target and Limits
Target
(b) No margin for error.Action: Reduce variation.
(c) Process running low.Defective product/service.
Action: Bring averagecloser to target.
(d) Process too variable.Defective product/service.Action: Reduce variation.
(e) Process off center andtoo variable. Defective
product/service.Action: Center better and
reduce variation.
Histogram66-75N 10/7/03, 9:29 AM73
©1994 GOAL/QPC74 Histogram
Tip The Histogram is related to the Control Chart.Like a Control Chart, a normally distributed His-togram will have almost all its values within ±3standard deviations of the mean. See Process Capa-bility for an illustration of this.
VariationsStem & Leaf PlotThis plot is a cross between a frequency distributionand Histogram. It exhibits the shape of a Histogram,but preserves the original data values—one of its keybenefits! Data is easily recorded by writing the trailingdigits in the appropriate row of leading digits.
In this example, the smallest value is .057 and thelargest value is .164. Using such a plot, it is easy to findthe median and range of the data.
• Median = middle data value (or average of thetwo middle values) when the data is ranked fromsmallest to largest.
For this example, there are 52 data points. Therefore,the average of the 26th and 27th value will give themedian value.
Median = (.113 + .116)/2 = .1145
• Range = Highest value – lowest value = .164 – .057 = .107
.05 7
.06 4
.07 5 7
.08 1 9 3 9
.09 7 4 8 2 6 9 4
.10 7 2 0 4 3 5 9
.1 1 3 1 9 3 7 3 8 6 6
.12 2 4 8 0 8 9 0 5
.13 2 5 2 7 7 6
.14 0 3 6 9
.15 4 7
.16 4
.05 7
.06 4
.07 5 7
.08 1 3 9 9
.09 2 4 4 6 7 8 9
.10 0 2 3 4 5 7 9
.1 1 1 3 3 3 6 6 7 8 9
.12 0 0 2 4 5 8 8 9
.13 2 2 5 6 7 7
.14 0 3 6 9
.15 4 7
.16 4
Data ascollected
Data rank-ordered
Histogram66-75N 10/7/03, 9:29 AM74
©1994 GOAL/QPC Histogram 75
HistogramTime Distribution of Calls
0
10
20
30
40
50
60
8:00
8:30
9:00
9:30
10:0
0
10:3
0
11:0
0
11:3
0
12:0
0
12:3
0
13:0
0
13:3
0
14:0
0
14:3
0
15:0
0
15:3
0
16:0
0
16:3
0
17:0
0
>17
:00
HOTrep data May 22 to August 4
Information provided courtesy ofSmithKline Beecham
Note: The Histogram identified three peak calling periods atthe beginning of the workday and before and after the tradi-tional lunch hour. This can help the HOTreps synchronizestaffing with their customer needs.
Histogram66-75N 10/7/03, 9:29 AM75
©1994 GOAL/QPC76 Interrelationship
What are the issues relatedto reducing litter?
• If using an original statement, (it didn’t comefrom a previous tool or discussion), create a com-
InterrelationshipDigraph (ID)Looking for drivers& outcomes
Why use it?To allow a team to systematically identify, analyze, andclassify the cause and effect relationships that existamong all critical issues so that key drivers or outcomescan become the heart of an effective solution.
What does it do?• Encourages team members to think in multiple
directions rather than linearly• Explores the cause and effect relationships
among all the issues, including the most contro-versial
• Allows the key issues to emerge naturally ratherthan allowing the issues to be forced by a domi-nant or powerful team member
• Systematically surfaces the basic assumptions andreasons for disagreements among team members
• Allows a team to identify root cause(s) even whencredible data doesn’t exist
How do I do it?1. Agree on the issue/problem statement
76 Interrelationship
ID 76-84N 10/7/03, 9:30 AM76
©1994 GOAL/QPC Interrelationship 77
plete sentence that is clearly understood andagreed on by team members.
• If using input from other tools, such as an AffinityDiagram, make sure that the goal under discussionis still the same and clearly understood.
2. Assemble the right team• The ID requires more intimate knowledge of the
subject under discussion than is needed for theAffinity. This is important if the final cause andeffect patterns are to be credible.
• The ideal team size is generally 4–6 people. How-ever, this number can be increased as long as theissues are still visible and the meeting is well facili-tated to encourage participation and maintain focus.
3. Lay out all of the ideas/issue cards that have eitherbeen brought from other tools or brainstormed• Arrange 5–25 cards or notes in a large circular
pattern, leaving as much space as possible fordrawing arrows. Use large, bold printing, includ-ing a large number or letter on each idea for quickreference later in the process.
Lack of respect for
othersLack of
awareness of impact
Inadequate penalties
Notenough
receptacles
Lack of parental
examples
Unneces-sary
packaging
A
D
E
F
C
B
ID 76-84N 10/7/03, 9:30 AM77
©1994 GOAL/QPC78 Interrelationship
4. Look for cause/influence relationships betweenall of the ideas and draw relationship arrows• Choose any of the ideas as a starting point. If all of the
ideas are numbered or lettered, work through themin sequence.
• An outgoing arrow from an idea indicates that itis the stronger cause or influence.
Continued next page
1
3
2
4
A
B
C
D
E
F
Decision: “B” causes or influences “A”
Or?
A
B
C
D
E
F
Decision:No relationship
Or?
A
B
C
D
E
F
Decision:No relationship
Or?
A
B
C
D
E
F
Decision: “E” causes or influences “A”
Or?
Ask of each combination:1) Is there a cause/influence relationship?
2) If yes, which direction of cause/influence is stronger?
ID 76-84N 10/7/03, 9:30 AM78
©1994 GOAL/QPC Interrelationship 79
5 6A
B
C
D
E
F
Decision:No relationship.
"A" is completed.
Or?A
B
C
D
E
F
Decision: “B” causes or influences “C.”
Now begin with “B” and repeat thequestions for all remaining
combinations.
Or?
Tip Draw only one-way relationship arrows in thedirection of the stronger cause or influence. Makea decision on the stronger direction. Do not drawtwo-headed arrows.
5. Review and revise the first round ID• Get additional input from people who are not on
the team to confirm or modify the team’s work.Either bring the paper version to others or repro-duce it using available software. Use a differentsize print or a color marker to make additions ordeletions.
6. Tally the number of outgoing and incoming arrowsand select key items for further planning• Record and clearly mark next to each issue the
number of arrows going in and out of it.• Find the item(s) with the highest number of outgo-
ing arrows and the item(s) with the highest num-ber of incoming arrows.
• Outgoing arrows. A high number of outgoing arrowsindicates an item that is a root cause or driver. This isgenerally the issue that teams tackle first.
ID 76-84N 10/7/03, 9:30 AM79
©1994 GOAL/QPC80 Interrelationship
Lack of respect for
others
Lack of awareness of impact
Inadequate penalties
Not enough receptacles
Lack of parental
examples
Unnecessary packaging
A
B
C
D
E
F
In=1 Out=1
In=4 Out=1
In=1 Out=1.5
In=1.5 Out=1
In=0 Out=5
In=2 Out=0
Driver
Outcome
• Incoming arrows. A high number of incoming ar-rows indicates an item that is a key outcome. Thiscan become a focus for planning either as a mean-ingful measure of overall success or as a redefini-tion of the original issue under discussion.
Tip Use common sense when you select the mostcritical issues to focus on. Issues with very closetallies must be reviewed carefully but in the end,it is a judgment call, not science.
7. Draw the final ID• Identify visually both the key drivers (greatest
number of outgoing arrows) and the key outcomes(greatest number of incoming arrows). Typicalmethods are double boxes or bold boxes.
What are the isues relatedto reducing litter?
Informationprovided courtesy
of CUE Consulting
ID 76-84N 10/7/03, 9:30 AM80
©1994 GOAL/QPC Interrelationship 81
VariationsWhen it is necessary to create a more orderly display ofall of the relationships, a matrix format is very effective.The vertical (up) arrow is a driving cause and thehorizontal (side) arrow is an effect. The example belowhas added symbols indicating the strength of the rela-tionships.The ”total“ column is the sum of all of the ”relationshipstrengths“ in each row. This shows that you are workingon those items that have the strongest effect on thegreatest number of issues.
Cau
se/
Dri
ver
Res
ult
/R
ider
Tota
l3 1 16
0 4 24
2 2 18
1 3 22
4 0 24
LogisticSupport
CustomerSatisfaction
Education& Training
PersonnelIncentives
Leadership
Lo
gis
tic
Su
pp
ort
Cu
sto
mer
Sat
isfa
ctio
n
Ed
uca
tio
n&
Tra
inin
g
Per
son
nel
Ince
nti
ves
Lea
der
ship
ID – Matrix Format
= 9 Significant= 3 Medium= 1 Weak
Relationship Strength
Information provided courtesy of U.S. Air Force, Air Combat Command
ID 76-84N 10/7/03, 9:30 AM81
©1994 GOAL/QPC82 Interrelationship
Interrelationship DigraphIssues Surrounding
Implementation of the Business Plan
Communica-tion issues within the
group
External factors impact
implemen-tation
Means not clearly defined
Plan not integrated
Fast new product
introductions stretch
resources
Lack of time and resources
No strong commitment to the group
In = 3 Out = 2
In = 2 Out = 4
In = 1 Out = 2
In = 0 Out = 2
In = 2 Out = 0
In = 1 Out = 3
Driver
Driver
Planning approach not standardized
In = 0 Out = 5
Outcome
Outcome
In = 5 Out = 0
Capacity may not
meet needs
In = 5 Out = 1
Information providedcourtesy of Goodyear
Note: “The drivers” from this ID will be used as the goal inthe Tree example shown at the end of the Tree Diagram/PDPCsection.
ID 76-84N 10/7/03, 9:30 AM82
©1994 GOAL/QPC Interrelationship 83
Interrelationship DigraphA Vision of Andoverin the 21st Century
Information provided courtesyof Town of Andover, MA
See next page for close up
Controlgrowth
In=5 Out=2
Preserve andpromote thenatural stateof Andover
In=9 Out=1
Encourage avariety ofmeans of
transportation
Total communitycommitment to
the task oflearning
Preservelegacy ofAndover
In=4 Out=7
Citizens treatedas valued
activecustomers
In=0 Out=11
All people’slives valued
and enhanced
In=6 Out=4
Thoughtful andresponsiblefinancing of
valued services
In=4 Out=7
The physicalsystems in town
workingsmoothly
In=4 Out=5
Maintain adiversity in
housing
In=3 Out=5
People shophere for their
needs bychoice
In=6 Out=2
Sense ofcommunity
through townactivities and
facilities
In=4 Out=3
In=4 Out=3
In=5 Out=3
ID 76-84N 10/7/03, 9:30 AM83
©1994 GOAL/QPC84 Interrelationship
Interrelationship DigraphA Vision of Andoverin the 21st Century
Close up
1 This is the driver. If the focus on the citizen as acustomer becomes the core of the town’s vision theneverything else will be advanced.
2 This is the primary outcome. It puts the preserva-tion of nature in the town as a key indicator of thevision working.
Information provided courtesyof Town of Andover, MA
Preserve andpromote thenatural stateof Andover
In=9 Out=1
Preservelegacy ofAndover
In=4 Out=7
Citizens treatedas valued
activecustomers
In=0 Out=11
1
2
Driver
Outcome
ID 76-84N 10/7/03, 9:30 AM84
©1994 GOAL/QPC Matrix 85Matrix 85
Matrix DiagramFinding relationships
Why use it?To allow a team or individual to systematically iden-tify, analyze, and rate the presence and strength ofrelationships between two or more sets of information.
What does it do?• Makes patterns of responsibilities visible and
clear so that there is an even and appropriatedistribution of tasks
• Helps a team get consensus on small decisions, en-hancing the quality and support for the final decision
• Improves a team’s discipline in systematicallytaking a hard look at a large number of importantdecision factors
Types of MatricesMost Common
• L-shaped matrix. Two sets of items directly comparedto each other or a single set compared to itself.
Tourfacility
Reviewpersonnel& safety policies
Review business values
Introduceto team
members
HumanresourcesDivisionmanager
Supervisor
Associates
Tasks
Resources
Orienting New Employees
Conclusion: Supervisors and associates have taken on theorientation role rather than the traditional human resource function.
Primaryresponsibility
Teammembers
Resources
a b c123
Matrix 85-90N 10/7/03, 9:31 AM85
©1994 GOAL/QPC86 Matrix
• T-shaped matrix. Two sets of items compared to acommon third set.
Uncommon• Y-shaped matrix. Three sets of items compared to each
other. It “bends” a T-shaped matrix to allow com-parisons between items that are on the vertical axes.
Rarely Used• X-shaped matrix. Four sets of items compared to
each other. It is essentially two T-shaped matri-ces placed back to back.
Communicate organization
spirit
Communicate purpose of
organization
Resolve practical concerns
Reduce anxiety
GoalsTour
facility
Reviewpersonnel& safetypolicies
Review business values
Introduceto team
members
Human resources
Division manager
Supervisor
Associates
ResourcesTasks
Conclusion: The most important purpose of orientation is to reduce anxiety, and the most effective tasks focus on the personal issues.
Orienting New Employees
ResponsibilityPrimaryTeam membersResources
ImpactHighMediumLow
Matrix 85-90N 10/7/03, 9:31 AM86
©1994 GOAL/QPC Matrix 87
• C-shaped matrix. Shows the intersection of threesets of data simultaneously. It is a three-dimen-sional graphic.
• You can find more complete information on the Y-,X-, and C-shaped matrix in The Memory JoggerPlus+®.
How do I do it?1. Select the key factors affecting successful imple-
mentation• The most important step is to choose the issues or
factors to be compared. The format is secondary.Begin with the right issues and the best formatwill define itself. The most common use is thedistribution of responsibilities within an L-shaped or T-shaped matrix.
2. Assemble the right team• Select individuals that have the influence/power
to realistically assess the chosen factors.Tip When distributing responsibilities, include
those people who will likely be involved in theassigned tasks or who can at least be part of areview team to confirm small group results.
3. Select an appropriate matrix format• Base your choice of format on the number of sets of
items and types of comparisons you need to make.
4. Choose and define relationship symbols• The most common symbols in matrix analysis are
, , . Generally they are used to indicate:= High = 9= Medium = 3= Low = 1
Matrix 85-90N 10/7/03, 9:31 AM87
©1994 GOAL/QPC88 Matrix
• The possible meanings of the symbols are almostendless. The only requirement is that the teamcomes to a clear understanding and creates anequally clear legend with the matrix.
5. Complete the matrix• If distributing responsibilities, use only one “pri-
mary responsibility” symbol to show ultimateaccountability. All other core team members canbe given secondary responsibilities.
Tip Focus the quality of the decision in each matrixcell. Do not try to “stack the deck” by consciouslybuilding a pattern of decisions. Let these patternsemerge naturally.
Tip Interpret the matrix using total numerical val-ues only when it adds value. Often the visualpattern is sufficient to interpret the overall results.
VariationsThe matrix is one of the most versatile tools available.The important skill to master is “matrix thinking.” Thisapproach allows a team to focus its discussion on re-lated factors that are explored thoroughly. The separateconclusions are then brought together to create high-quality decisions. Use your creativity in determiningwhich factors affect each other, and in choosing thematrix format that will help focus the discussion to-ward the ultimate decision.
Matrix 85-90N 10/7/03, 9:31 AM88
©1994 GOAL/QPC Matrix 89
MatrixLogistics Annual Plan
LQC
Obj
ectiv
es(M
atrix
)
Mea
sure
s
Sch
edul
es(A
ND
)
Res
earc
hcu
stom
er n
eeds
via
QF
D
Cap
ture
cus
tom
erco
mm
ents
Reducecustomer cost
Sur
vey
cust
omer
satis
fact
ion
Continueimplementation of total quality
Continue upgradingtech., prof., &
managerial skills ofemployeesPromote
environmentalresponsibility inour operations
Del
ight
our
cust
omer
s
%sa
tisfa
ctio
nvi
a su
rvey
List
of
cust
omer
need
s by
key
proc
esse
s#
ofco
mm
ents
or
# of
com
plai
nts
2003
2
004
TQ
Impl
emen
tatio
n(T
ree)
1 See
next
page
Con
tinue
toim
plem
ent
tota
l qua
lity
Goa
l
=
9 S
tron
g in
fluen
ce/r
elat
ions
hip
=
3 S
ome
influ
ence
/rel
atio
nshi
p
= 1
Wea
k in
fluen
ce/r
elat
ions
hip
Bla
nk =
No
influ
ence
/rel
atio
nshi
p
Qua
rter
Qua
rter
1
2
3
4
1
2
3
4
Info
rmat
ion
prov
ided
cou
rtes
y of
Bel
l Can
ada
Matrix 85-90N 10/7/03, 9:31 AM89
©1994 GOAL/QPC90 Matrix
MatrixLogistics Annual Plan
Goa
ls(A
ND
)C
o-R
espo
nsib
ility
(Mat
rix)
Cos
t/Ben
efit
Ana
lysi
s
Status*
Sur
vey
cust
omer
satis
fact
ion
Res
earc
hcu
stom
er n
eeds
via
QF
D
Cap
ture
cus
tom
erco
mm
ents
2003
2
004
LQC
Departments
Boards ofmanagement
Logiqual
Otherstakeholders
Resourcesrequired ($)
Tangiblebenefits ($)
75%
cust
omer
satis
fact
ion
80%
cust
omer
satis
fact
ion
Fie
ld tr
ial
100%
cust
omer
need
sga
ther
ed
1%tra
nsac
tions
1.9%
03
0
4
03
0
4
25K
25K
TQ
Impl
emen
tatio
n(T
ree)
1fr
ompr
evio
uspa
ge
= 9
Prim
e re
spon
sibi
lity
= 3
Sec
onda
ry r
espo
nsib
ility
= 1
Kep
t inf
orm
ed
25K
25K
Cau
tion
Sto
pped
On
targ
et*S
tatu
s:
Del
ight
our
cust
omer
s
Con
tinue
toim
plem
ent
tota
l qua
lity
Goa
l
Info
rmat
ion
prov
ided
cou
rtes
y of
Bel
l Can
ada
Matrix 85-90N 10/7/03, 9:31 AM90
©1994 GOAL/QPC NGT/Multivoting 91
Nominal GroupTechnique (NGT)Ranking for consensus
Why use it?Allows a team to quickly come to a consensus on therelative importance of issues, problems, or solutions bycompleting individual importance rankings into ateam’s final priorities.
What does it do?• Builds commitment to the team’s choice through
equal participation in the process• Allows every team member to rank issues with-
out being pressured by others• Puts quiet team members on an equal footing
with more dominant members• Makes a team’s consensus (or lack of it) visible;
the major causes of disagreement can be dis-cussed
How do I do it?1. Generate the list of issues, problems, or solutions
to be prioritized• In a new team with members who are not accus-
tomed to team participation, it may feel safer todo written, silent brainstorming, especially whendealing with sensitive topics.
2. Write statements on a flipchart or board
A.B.C.D.
2+21+13+44+3
A.B.C.D.
2+21+13+44+3
NGT/Multivoting 91
Nominal Group 91-94N 10/7/03, 9:34 AM91
©1994 GOAL/QPC92 NGT/Multivoting
3. Eliminate duplicates and/or clarify meanings ofany of the statements• As a leader, always ask for the team’s permission
and guidance when changing statements.
4. Record the final list of statements on a flipchartor boardExample: Why does the department have inconsis-tent output?
A Lack of trainingB No documented processC Unclear quality standardsD Lack of cooperation with other
departmentsE High turnover
• Use letters rather than numbers to identify eachstatement so that team members do not get con-fused by the ranking process that follows.
5. Each team member records the corresponding letterson a piece of paper and rank orders the statementsExample: Larry’s sheet of paper looks like this:
A 4B 5C 3D 1E 2
• This example uses “5” as the most importantranking and “1” as the least important. Sinceindividual rankings will later be combined, this“reverse order” minimizes the effect of team
Nominal Group 91-94N 10/7/03, 9:34 AM92
©1994 GOAL/QPC NGT/Multivoting 93
members leaving some statements blank. There-fore, a blank (value = 0) would not, in effect,increase its importance.
6. Combine the rankings of all team members
Larry Nina Norm Paige Si TotalA 4 5 2 2 1 = 14B 5 4 5 3 5 = 22C 3 1 3 4 4 = 15D 1 2 1 5 2 = 11E 2 3 4 1 3 = 13
“No documented process,” B, would be the highestpriority. The team would work on this first and thenmove through the rest of the list as needed.
VariationsOne Half Plus OneWhen dealing with a large number of choices it may benecessary to limit the number of items ranked. The “onehalf plus one” approach would rank only a portion ofthe total. For example, if 20 ideas were generated, thenteam members would rank only the top 11 choices. Ifneeded, this process could be repeated with the remain-ing 9 items, ranking the top 5 or 6 items, (half of 9 = 4.5+ 1 = 5.5), until a manageable number are identified.
Weighted MultivotingEach team member rates, not ranks, the relative impor-tance of choices by distributing a value, e.g., 100 points,across the options. Each team member can distributethis value among as many or as few choices as desired.
Nominal Group 91-94N 10/7/03, 9:34 AM93
©1994 GOAL/QPC94 NGT/Multivoting
Example:
Larry Nina Norm Paige Si TotalA 20 10 = 30B 40 80 50 100 45 = 315C 30 5 10 25 = 70D 5 10 20 = 35E 10 10 20 10 = 50
With large numbers of choices, or when the voting forthe top choices is very close, this process can be re-peated for an agreed upon number of items. Stop whenthe choice is clear.
Nominal Group 91-94N 10/7/03, 9:34 AM94
©1994 GOAL/QPC Pareto 95
Pareto ChartFocus on key problems
Why use it?To focus efforts on the problems that offer the greatestpotential for improvement by showing their relativefrequency or size in a descending bar graph.
What does it do?• Helps a team to focus on those causes that will
have the greatest impact if solved• Based on the proven Pareto principle: 20% of the
sources cause 80% of any problem• Displays the relative importance of problems in a
simple, quickly interpreted, visual format• Helps prevent “shifting the problem” where the “so-
lution” removes some causes but worsens others• Progress is measured in a highly visible format
that provides incentive to push on for more im-provement
How do I do it?1. Decide which problem you want to know more
aboutExample: Consider the case of HOTrep, an inter-nal computer network help line: Why do peoplecall the HOTrep help line; what problems arepeople having?
2. Choose the causes or problems that will be moni-tored, compared, and rank ordered by brainstorm-ing or with existing dataa) Brainstorming
Pareto 95
Pareto 95-104N 10/7/03, 9:34 AM95
©1994 GOAL/QPC96 Pareto
Example: What are typical problems that usersask about on the HOTrep help line?
b) Based on existing dataExample: What problems in the last month haveusers called in to the HOTrep help line?
3. Choose the most meaningful unit of measure-ment such as frequency or cost• Sometimes you don’t know before the study
which unit of measurement is best. Be preparedto do both frequency and cost.Example: For the HOTrep data the most impor-tant measure is frequency because the projectteam can use the information to simplify soft-ware, improve documentation or training, orsolve bigger system problems.
4. Choose the time period for the study• Choose a time period that is long enough to
represent the situation. Longer studies don’t al-ways translate to better information. Look first atvolume and variety within the data.
• Make sure the scheduled time is typical in orderto take into account seasonality or even differentpatterns within a given day or week.Example: Review HOTrep help line calls for 10weeks (May 22–August 4).
5. Gather the necessary data on each problem categoryeither by “real time” or reviewing historical data• Whether data is gathered in “real time” or his-
torically, check sheets are the easiest method forcollecting data.Example: Gathered HOTrep help line calls databased on the review of incident reports (historical).
Pareto 95-104N 10/7/03, 9:34 AM96
©1994 GOAL/QPC Pareto 97
7. List the problem categories on the horizontal lineand frequencies on the vertical line• List the categories in descending order from left
to right on the horizontal line with bars aboveeach problem category to indicate its frequency orcost. List the unit of measure on the vertical line.
8. (Optional) Draw the cumulative percentage lineshowing the portion of the total that each problemcategory representsa) On the vertical line, (opposite the raw data, #, $,
etc.), record 100% opposite the total number and50% at the halfway point. Fill in the remainingpercentages drawn to scale.
Tip Always include with the source data and thefinal chart, the identifiers that indicate the source,location, and time period covered.
6. Compare the relative frequency or cost of eachproblem category
Example:
Problem Category Frequency Percent (%)Bad configuration 3 1Boot problems 68 33File problems 8 4Lat. connection 20 10Print problems 16 8Reflection hang 24 12Reflection sys. integrity 11 5Reflections misc. 6 3System configuration 16 8System integrity 19 9Others 15 7
Total 206
Pareto 95-104N 10/7/03, 9:34 AM97
©1994 GOAL/QPC98 Pareto
b) Starting with the highest problem category, drawa dot or mark an x at the upper righthand cornerof the bar.
• Add the total of the next problem category to thefirst and draw a dot above that bar showing boththe cumulative number and percentage. Connectthe dots and record the remaining cumulativetotals until 100% is reached.
Information provided courtesy of SmithKline Beecham
HOTrep Problem Data220
200
180
160
140
120
100
80
60
40
20
Boo
t pro
blem
s (6
8)
Ref
lect
ion
hang
(24
)
Lat.
conn
ectio
n (2
0)
Sys
tem
inte
grity
(19
)
Sys
tem
con
fig. (
16)
Prin
t pro
blem
s (1
6)
Ref
lect
ion
syst
. int
egrit
y (1
1)
File
pro
blem
s (8
)
Ref
lect
ions
mis
c. (
6)
Bad
con
figur
atio
n (3
)
Oth
ers
(15)
100%
90
80
70
60
50
40
30
20
10
33%
12%10% 9% 8% 8% 5% 4% 3% 1% 7%
45%
55%
64%
72%
80%
85%
89%92%
93%
# of
Rep
orte
d O
ccur
renc
es
Pareto 95-104N 10/7/03, 9:34 AM98
©1994 GOAL/QPC Pareto 99
9. Interpret the results• Generally, the tallest bars indicate the biggest con-
tributors to the overall problem. Dealing withthese problem categories first therefore makescommon sense. But, the most frequent or expen-sive is not always the most important. Alwaysask: What has the most impact on the goals of ourbusiness and customers?
VariationsThe Pareto Chart is one of the most widely and cre-atively used improvement tools. The variations usedmost frequently are:
A. Major Cause Breakdowns in which the “tallestbar” is broken into subcauses in a second, linkedPareto.
B. Before and After in which the “new Pareto” barsare drawn side by side with the original Pareto,showing the effect of a change. It can be drawn asone chart or two separate charts.
C. Change the Source of Data in which data iscollected on the same problem but from differentdepartments, locations, equipment, and so on,and shown in side-by-side Pareto Charts.
D. Change Measurement Scale in which the samecategories are used but measured differently.Typically “cost” and “frequency” are alternated.
Pareto 95-104N 10/7/03, 9:34 AM99
©1994 GOAL/QPC100 Pareto
Information provided courtesy of Goodyear
50
40
30
20
10
0
Reduced Payment Freight BillsContract Rate Disputes
120110100
908070605040302010
0
Cla
ssReduced Payment Freight Bills
Total Bills (329)
Adv
ance
d ca
rrie
r
Load
ed to
full
visi
ble
capa
city
Del
iver
y to
min
e si
tes
Min
. chr
g.
NY
C a
rb.
Mis
c.
SW
C
NO
A
% o
f con
trac
t dis
pute
s#
of b
ills
Con
trac
t rat
edi
sput
es
Can
ada
Orig
inal
dest
inat
ion
Mis
c.
Wei
ght
Rec
onci
led
Deb
t
Byn
d.
ParetoA. Major Cause Breakdowns
Pareto 95-104N 10/7/03, 9:34 AM100
©1994 GOAL/QPC Pareto 101
120
110
100 90 80 70 60 50 40 30 20 10 0
Janu
ary
bills
(32
9)
Red
uced
Pay
men
t Fre
ight
Bill
s–
Afte
r S
tand
ardi
zatio
n –
June
bill
s (5
6)
# of bills
Contract rate disputes
Class
Canada
Original destination
Misc.
Weight
Reconciled
Debt
Bynd.
Info
rmat
ion
prov
ided
cou
rtes
y of
Goo
dyea
r
ParetoB. Before and After
Pareto 95-104N 10/7/03, 9:34 AM101
©1994 GOAL/QPC102 Pareto
Information provided courtesy ofU.S. Navy, Naval Dental Center, San Diego
0
5
10
15
20
25
30
35
For
got
Wor
kloa
d
Per
sona
lB
usin
ess
Leav
e
Mis
c.
Tran
sfer
red
Vehi
cle
% 31 25 21 8 8 4 2
% o
f fai
led
appo
intm
ents
Reason for Failed AppointmentsSource of Data is: Shore Commands
ParetoC. Change the Source of Data
Pareto 95-104N 10/7/03, 9:34 AM102
©1994 GOAL/QPC Pareto 103
Reason for Failed AppointmentsSource of Data is: Fleet Commands
ParetoC. Change the Source of Data
0
5
10
15
20
25
30
Cor
psm
an
Und
erw
ay
For
got
Per
sona
lB
usin
ess
Vehi
cle
TAD
/Tr
ansf
er
Sch
ool
% 27 25 23 9 7 5 4
% o
f fai
led
appo
intm
ents
Information provided courtesy ofU.S. Navy, Naval Dental Center, San Diego
Pareto 95-104N 10/7/03, 9:34 AM103
©1994 GOAL/QPC104 Pareto
ParetoD. Change Measurement Scale
25
20
15
10
5
0
Shi
ppin
g
Inst
alla
tion
Del
iver
y
Cle
rical
Mis
c.
42% of all complaints
Field Service Customer Complaints
# of
com
plai
nts
20,000
15,000
10,000
5,000
0
Inst
alla
tion
Cle
rica
l
Ship
pin
g
Deliv
ery
Mis
c.
13% of total cost
Cost to Rectify Field Service Complaints
Dolla
rs (
$)
Pareto 95-104N 10/7/03, 9:34 AM104
©1994 GOAL/QPC Prioritization 105
PrioritizationMatricesWeighing your options
Why use it?To narrow down options through a systematic ap-proach of comparing choices by selecting, weighting,and applying criteria.
What does it do?• Quickly surfaces basic disagreements so they
may be resolved up front• Forces a team to focus on the best thing(s) to do,
and not everything they could do, dramaticallyincreasing the chances for implementation success
• Limits “hidden agendas” by surfacing the criteriaas a necessary part of the process
• Increases the chance of follow-through becauseconsensus is sought at each step in the process(from criteria to conclusions)
• Reduces the chances of selecting someone’s “petproject”
How do I do it?There are three methods for constructing PrioritizationMatrices. The outline that follows indicates typicalsituations for using each method. Only the “Full Ana-lytical Criteria Method” is discussed here. The othersare covered fully in The Memory Jogger Plus+®.
a b c
ParetoPrioritization 105
Prioritization 105-114N 10/7/03, 9:35 AM105
©1994 GOAL/QPC106 Prioritization
Full Analytical Criteria Method
Typically use when:• Smaller teams are involved (3–8 people)• Options are few (5–10 choices)• There are relatively few criteria (3–6 items)• Complete consensus is needed• The stakes are high if the plan fails
Consensus Criteria MethodThis method follows the same steps as in the FullAnalytical Criteria Method except the ConsensusCriteria Method uses a combination of weightedvoting, and ranking is used instead of paired com-parisons.
Typically use when:• Larger teams are involved (8 or more people)• Options are many (10–20 choices)• There are a significant number of criteria (6–15
items)• Quick consensus is needed to proceed
Combination ID/Matrix MethodThis method is different from the other two methodsbecause it is based on cause and effect, rather thancriteria.
Typically use when:• Interrelationships among options are high and
finding the option with the greatest impact iscritical
Prioritization 105-114N 10/7/03, 9:35 AM106
©1994 GOAL/QPC Prioritization 107
Full Analytical Criteria Method
1. Agree on the ultimate goal to be achieved in aclear, concise sentence• If no other tools are used as input, produce a clear
goal statement through consensus. This state-ment strongly affects which criteria are used.
Choose the most enjoyable vacation for the whole family
2. Create the list of criteria• Brainstorm the list of criteria or review previous
documents or guidelines that are available, e.g.,corporate goals, budget-related guidelines.
• Cost• Educational value• Diverse activity• Escape reality
Tip The team must reach consensus on the finalcriteria and their meanings or the process is likelyto fail!
3. Using an L-shaped matrix, weight each criterionagainst each other• Reading across from the vertical axis, compare
each criterion to those on the horizontal axis.• Each time a weight (e.g., 1, 5, 10) is recorded in a
row cell, its reciprocal value (e.g., 1/5, 1/10) must berecorded in the corresponding column cell.
• Total each horizontal row and convert to a relativedecimal value known as the “criteria weighting.”
Prioritization 105-114N 10/7/03, 9:35 AM107
©1994 GOAL/QPC108 Prioritization
4. Compare ALL options relative to each weightedcriterion• For each criterion, create an L-shaped matrix
with all of the options on both the vertical andhorizontal axis and the criteria listed in thelefthand corner of the matrix. There will be asmany options matrices as there are criteria tobe applied.
• Use the same rating scale (1, 5, 10) as in Step 3,BUT customize the wording for each criterion.
• The relative decimal value is the “option rating.”
CostEduc.value
Diverseactivity
Escapereality
RowTotal
RelativeDecimal
Value
Cost 5
5
10 5
5
5
5.3
10.2
20
.60
.15
.28
.55
.02
Criteria
Criteria
Grand Total 36.1
Row TotalRating scores addedGrand TotalRow totals addedRelative Decimal ValueEach row total ÷ by the grand total
Educ.value
Diverseactivity
Escapereality
110
15
15
15
15
15
= Equally important= More important= Much more important= Less Important= Much less important
15
101/5
1/10
Criterion vs. Criterion
Prioritization 105-114N 10/7/03, 9:35 AM108
©1994 GOAL/QPC Prioritization 109
Tip The whole number (1, 5, 10) must always rep-resent a desirable rating. In some cases this maymean “less,” e.g., cost, in others this may mean“more,” e.g., tasty.
110
Disney World
Gettys-burg
New York City
Uncle Henry's
Row Total
RelativeDecimal
Value
5
10 5
5.3
15.2
.40
25
.12
.33
.01
.54
Grand Total 45.9
Cost
Disney World
Gettys-burg
New York City
Uncle Henry's
5
10
10
Continue Step 4 through three more Options/Criterion matrices, like this:
15
110
110
15
15
Options vs. Each Criterion (Cost Criterion)
= Equal cost= Less expensive= Much less expensive= More expensive= Much more expensive
15
101/5
1/10
Escape reality
Options
Opt
ions
Crt .
Diverse activity
Options
Opt
ions
Crt .
Educational value
Options
Opt
ions
Crt .
Prioritization 105-114N 10/7/03, 9:35 AM109
©1994 GOAL/QPC110 Prioritization
5. Using an L-shaped summary matrix, compareeach option based on all criteria combined• List all criteria on the horizontal axis and all
options on the vertical axis.• In each matrix cell multiply the “criteria weight-
ing” of each criterion (decimal value from Step 3)by the “option rating” (decimal value from Step4). This creates an “option score.”
• Add each option score across all criteria for a rowtotal. Divide each row total by the grand total andconvert to the final decimal value. Compare thesedecimal values to help you decide which optionto pursue.
.54 x .15(.08)
Cost(.15)
Educa-tionalvalue(.28)
Diverse activity(.55)
Escape reality(.02)
Row Total
RelativeDecimal
Value(RT ÷ GT)
Criteria
Disney World
Gettys-burg
New York City
Uncle Henry's
Optns.
.12 x .15(.02)
.33 x .15(.05)
.01 x .15(0)
.24 x .28(.07)
.37 x .28(.10)
.01 x .28(0)
.37 x .28(.10)
.40 x .55(.22)
.10 x .55(.06)
.49 x .55(.27)
.01 x .55(.01)
.65 x .02(.01)
.22 x .02(0)
.12 x .02(0)
.01 x .02(0)
.32
.22
.37
.09
.32
.22
.38
.09
.54 x .15(from Step 4 matrix) (from Step 3 matrix)
(.08)Option score
1.00Grand Total
Summary MatrixOptions vs. All Criteria
Prioritization 105-114N 10/7/03, 9:35 AM110
©1994 GOAL/QPC Prioritization 111
6. Choose the best option(s) across all criteriaTip While this is more systematic than traditional
decision making, it is not a science. Use commonsense and judgment when options are rated veryclosely, but be open to non-traditional conclusions.
VariationsSee The Memory Jogger Plus+® for full explanations ofboth the Consensus Criteria Method and the Combina-tion ID/Matrix Method. The Full Analytical CriteriaMethod, illustrated in this book, is recommended be-cause it encourages full discussion and consensus oncritical issues. The Full Analytical Criteria Method is asimplified adaptation of an even more rigorous modelknown as the Analytical Hierarchy Process. It is basedon the work of Thomas Saaty, which he describes in hisbook Decision Making for Leaders. In any case, use com-mon sense to know when a situation is importantenough to warrant such thorough processes.
Prioritization 105-114N 10/7/03, 9:35 AM111
©1994 GOAL/QPC112 Prioritization
PrioritizationChoosing a Standard
Corporate Spreadsheet ProgramB
est u
seof
har
dwar
e
Eas
e of
use
Max
imum
func
tiona
lity
Bes
tpe
rfor
man
ce
Tota
l(1
4 cr
iteria
)
Rel
ativ
eD
ecim
al V
alue
Best useof hardware .20 .10 3.7 .01
Ease of use 5.0 35.4 .08
Maximumfunctionality 10.0 69.0 .17
Bestperformance 45.2 .11
Grand Total(14 criteria) 418.1
5.0
5.05.0
5.0
.20
.20
.20
.20
Criteria
Information provided courtesy of Novacor Chemicals
Note: This constructed example, illustrated on three pages,represents only a portion of the prioritization process and onlya portion of Novacor’s spreadsheet evaluation process.Novacor Chemicals assembled a 16-person team, comprisedmainly of system users and some information systems staff.The team developed and weighted 14 standard criteria andthen applied them to choices in word processing, spreadsheet,and presentation graphics programs.
1 Weighting criteria (described in Step 3)This is a portion of a full matrix with 14 criteria in total.
This example continued next page
Prioritization 105-114N 10/7/03, 9:35 AM112
©1994 GOAL/QPC Prioritization 113
PrioritizationChoosing a Standard
Corporate Spreadsheet Program (cont.)
Best integration–internal
Total
1.00 2.00 .33
1.00
1.00
6.00
1.00
1.00
1.00 2.00
2.00
.33
.33
Relative Decimal
Value
ProgramA
ProgramB
Program
C
ProgramA
ProgramB
ProgramC
Grand Total
This example continued next page
Lowest ongoing
costTotal
.10 .20 .30 .02
10.00 5.00 15.00 .73
5.00 .20 5.20 .25
20.50
ProgramA
ProgramB
ProgramC
ProgramA
ProgramB
ProgramC
Relative Decimal
Value
Grand Total
Information provided courtesy of Novacor Chemicals
2 Comparing options (described in Step 4)These are just 2 of 14 matrices.
Prioritization 105-114N 10/7/03, 9:35 AM113
©1994 GOAL/QPC114 Prioritization
PrioritizationChoosing a Standard
Corporate Spreadsheet Program (cont.)
3 Summarize Option Ratings Across All Criteria (described in Step 5)
This is a portion of a full matrix with 14 criteria in total.
Criteria Total(across 14
criteria)
.16 .18
.48 (.04) .30 .33
.49
.90
ProgramA
ProgramB
ProgramC
Easy to use (.08)
Best integration int. (.09)
Lowest ongoing
cost (.08)
.44
Relative Decimal
Value
.03 (.01)
.48 (.04)
.33 (.03)
.33 (.03)
.33 (.03)
.02 (0)
.73 (.06)
.25 (.02)
Grand Total
Options
Result: Program C was chosen. Even though 14 out of the 16team members were not currently using this program, theprioritization process changed their minds, and preventedthem from biasing the final decision.
Information provided courtesy of Novacor Chemicals
Prioritization 105-114N 10/7/03, 9:35 AM114
©1994 GOAL/QPC PS/PI Model 115
Problem-Solving/Process-Improvement ModelImprovementStoryboard
There are many standard models for making improve-ments. They all attempt to provide a repeatable set ofsteps that a team or individual can learn and follow. TheImprovement Storyboard is only one of many modelsthat include typical steps using typical tools. Followthis model or any other model that creates a commonlanguage for continuous improvement within yourorganization.
Plan
1. Select the problem/process that will be ad-dressed first (or next) and describe the improve-ment opportunity.
2. Describe the current process surrounding theimprovement opportunity.
3. Describe all of the possible causes of the prob-lem and agree on the root cause(s).
4. Develop an effective and workable solution andaction plan, including targets for improvement.
Do
5. Implement the solution or process change.
Check
6. Review and evaluate the result of the change.
Act
7. Reflect and act on learnings.
P
D
C
A
PS/PI Model 115
PS/PI 115-131N 10/7/03, 9:37 AM115
©1994 GOAL/QPC116 PS/PI Model
Depending on your formal process structure, Step 1may be done by a steering committee, managementteam, or improvement team. If you are an improvementteam leader or member, be prepared to start with Step 1 orStep 2.
1. Select the problem/process that will be addressedfirst (or next) and describe the improvement op-portunity.• Look for changes in important business indicators• Assemble and support the right team• Review customer data• Narrow down project focus. Develop project pur-
pose statement
Typical toolsBrainstorming, Affinity Diagram, Check Sheet, ControlChart, Histogram, Interrelationship Digraph, ParetoChart, Prioritization Matrices, Process Capability, Ra-dar Chart, Run Chart
SituationStop ’N Go Pizza* is a small but recently growing pizzadelivery business with six shops. After a period of rapidgrowth, Stop ’N Go Pizza experienced a six-month declinein volume. Customers were leaving. Top managementformed a mixed team of store managers, kitchen staff,and delivery personnel to find out why, and to generatean implementation plan to correct the situation. Theteam used both the Run Chart and Pareto Chart.
* The name Stop ’N Go Pizza, and the data associated with thiscase study are fictional. Any similiarity to an actual companyby this name is purely coincidental.
Plan
PS/PI 115-131N 10/7/03, 9:37 AM116
©1994 GOAL/QPC PS/PI Model 117
Run ChartAverage monthly volume of deliveries
(per shop)2400
2100
1800
1500
1200
900
600
300
0Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Un
it V
olu
me
1,951deliveries
Pareto ChartTypes of customer complaintsTotal=2520 October-December
(across 6 shops)2500
2000
1500
1000
500
0
Late
del
iver
ies
Wro
ng o
rder
Col
d fo
od
Tast
e
Oth
er
100%
75
50
25
(1890)
(220) (206) (117) (87)Tota
l # o
f Cus
tom
er C
ompl
aint
s
Illustration note: Delivery time was defined by the total time from when the order was placed to when the customer received it.
PS/PI 115-131N 10/7/03, 9:37 AM117
©1994 GOAL/QPC118 PS/PI Model
Pareto ChartLate delivery complaints
Total=1890 October-December(across 6 shops)
# of
Lat
e D
eliv
ery
Com
plai
nts
400350300250200150100500
C A B F D E
Shops
= Friday
= Saturday
= Other
(391)(358)
(313)(295)
(275) (258)
DecisionLate deliveries, (late from the time of order), were by farthe most frequent customer complaint across all loca-tions, especially on Fridays and Saturdays.
Team purpose statementReduce late deliveries on Fridays and Saturdays.
2. Describe the current process surrounding the im-provement opportunity.• Select the relevant process or process segment to
define the scope of the project• Describe the process under study
Typical toolsBrainstorming, Macro, Top-down, and DeploymentFlowcharts, Tree Diagram
SituationThe team began to understand the overall process forproducing and delivering their product and problems
PS/PI 115-131N 10/7/03, 9:37 AM118
©1994 GOAL/QPC PS/PI Model 119
that contributed to the project focus. The team used aMacro Flowchart.
Receive order
Prepareingredients
Bakepizza
Assembleorder
Deliver orderReceivepaymentReturn to shop
Process for producing and deliveringStop ’N Go Pizza
DecisionIt became obvious that the “late deliveries” went farbeyond the physical delivery process. Everything inthe Macro Flowchart affected the “order-to-eating”time. This total process improvement became theteam’s focus.
3. Describe all of the possible causes of the problemand agree on the root cause(s).• Identify and gather helpful facts and opinions on
the cause(s) of the problem• Confirm opinions on root cause(s) with data
whenever possible
Typical toolsAffinity Diagram, Brainstorming, C & E/Fishbone Dia-gram, Check Sheet, Force Field Analysis, Interrelation-ship Digraph, Multivoting, Nominal Group Tech-nique, Pareto Chart, Run Chart, Scatter Diagram
PS/PI 115-131N 10/7/03, 9:37 AM119
©1994 GOAL/QPC120 PS/PI Model
SituationThe team brainstormed all of the possible causes for“late deliveries” and then continued to ask “Why?” sothat possible cause patterns could emerge. The teamused a C & E/Fishbone Diagram and Run Charts.
C &
E/F
ish
bo
ne
Dia
gra
mR
easo
ns fo
r la
te p
izza
del
iver
ies
Late
piz
za
deliv
erie
s on
F
riday
s &
S
atur
days
Peo
ple
don'
t sho
w u
p
Hig
h tu
rnov
erLo
w p
ay
No
team
wor
k No
trai
ning
Peo
ple
Don
't kn
ow to
wn
Driv
ers
get l
ost
Hig
h tu
rnov
er
Get
wro
ng
info
rmat
ion
Poo
r tr
aini
ng
Rus
hed
Mac
hine
ry/E
quip
men
t
Unr
elia
ble
cars
Low
pay
Kid
s ow
n ju
nks
No
mon
ey fo
r re
pairs
Ove
ns to
o sm
all
Poo
r us
e of
spa
cePoo
r tr
aini
ngH
igh
turn
over
No
capa
city
for
peak
per
iods
Met
hods
Poo
r ha
ndlin
g of
la
rge
orde
rs
Lack
of e
xper
ienc
eP
oor
disp
atch
ing
Man
y ne
w s
tree
ts
Don
't kn
ow
tow
n Hig
h tu
rnov
er
Run
out
of i
ngre
dien
ts
Inac
cura
te
orde
ring
Poo
r us
e of
spa
ce
Lack
of
trai
ning M
ater
ials
Hig
h tu
rnov
er
Hig
h tu
rnov
er
PS/PI 115-131N 10/7/03, 9:37 AM120
©1994 GOAL/QPC PS/PI Model 121
Run Charts
Average turnover rate of employees(company-wide)
70
60
50
40
30
20
10
0
43.25%
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
%
Average training hours of new employees
14
12
10
8
6
4
2
0
8 hours
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Ave
rage
# o
f Hou
rs
PS/PI 115-131N 10/7/03, 9:37 AM121
©1994 GOAL/QPC122 PS/PI Model
DecisionThe C & E/Fishbone Diagram repeatedly pointed to“turnover” and “lack of training” as root causes. Thisapplied to ALL areas of the operation, NOT just in theactual delivery portion of the process. Subsequent datashowed that as the business rapidly grew, less time wasput into training all employees. With this lack of adequatetraining, many employees felt a great deal of pressure atthe busiest times. They also felt that they were unable to dotheir jobs well. This combination of work pressure andlack of self-confidence often caused employees to quit.
4. Develop an effective and workable solution andaction plan, including targets for improvement.• Define and rank solutions• Plan the change process: What? Who? When?• Do contingency planning when dealing with new
and risky plans• Set targets for improvement and establish
monitoring methods
Typical toolsActivity Network Diagram, Brainstorming, Flowchart,Gantt Chart, Multivoting, Nominal Group Technique,PDPC, Prioritization Matrices, Matrix Diagram, TreeDiagram
SituationThe team used the combination of the Tree Diagram,Prioritization Matrices, Responsibility Matrix, andGantt Chart to create a workable plan that attacked theheart of the problem.
DecisionThe team focused on the most effective, efficient way todeliver the new training. They developed a four-month
PS/PI 115-131N 10/7/03, 9:37 AM122
©1994 GOAL/QPC PS/PI Model 123
Tree DiagramReduce turnover rate of employees
(company-wide)
implementation plan that featured the creative use ofvideotapes, role plays, peer ratings, and so on. The teamset the following targets based on past performance:
• Reduce turnover rate from 62 to 30 percent• Reduce average time of order-to-delivery from 40
to 25 minutes• Reduce customer complaints of late deliveries by
50 percent, without increasing other complaintcategories
• Increase average monthly volume to 2400 unitsper shop from the current 1891 units
Reduce turnover through
improved training in all facets of
operation
Delivertraining
conveniently
Delivertraining
effectively
At convenient
times
At convenientlocations
Before shift
During shift
After shift
At shop
At home
Hold trainee interest
Improve performance
Develop videos
Use texts
Use role play
Evaluate each other
Pre/post tests
Set clear performance
standards
PS/PI 115-131N 10/7/03, 9:37 AM123
©1994 GOAL/QPC124 PS/PI Model
Prioritization MatrixSelecting the best training program components
The total = the sum of [rating values x criteria weighting]For example, to find the total of the “Train before shift” row,do the following:[ (9) x .60] + [ (9) x .19] + [ (3) x .19] + [ (3) x .01] = 7.70Note: Weighting values of each criterion come from a criteria
matrix not shown.Task options come from the most detailed level of theTree Diagram on the previous page.
Effe
ctiv
enes
s
Fea
sibi
lity
Tim
e
Cos
t
Tota
l
Train before shift 7.70
Train during shift 3.78
Train after shift 1.77
Train at the shop
Train at home 5.29
Develop videos 8.83
Use texts 4.16
Use role play 8.89
Evaluate each other
Pre/post test 8.83
Set clear perfor-mance standards 8.89
Criteria & Weighting
Tasks & Options
7.70
7.70
(.60) (.19) (.19) (.01)
= 9 Excellent = 3 Fair = 1 Poor
PS/PI 115-131N 10/7/03, 9:37 AM124
©1994 GOAL/QPC PS/PI Model 125
Matrix & Gantt Chart CombinedNew training program timeline
Managers
Employees
Human resources
President
Janu
ary
Feb
ruar
yM
arch
Apr
il
Trai
n at
the
shop
be
fore
the
shift
Dev
elop
vid
eos
Use
rol
e pl
ay
Eva
luat
e ea
ch o
ther
Use
pre
/pos
t tes
t
Set
cle
ar p
erfo
rman
ce
stan
dard
s
Res
pons
ibili
ty
Task
s* = P
rimar
y re
spon
sibi
lity
= S
econ
dary
/team
mem
ber
= N
eed
info
rmat
ion
to/fr
om
*The
se w
ere
the
high
est r
ated
task
s fr
om th
e
Prio
ritiz
atio
n M
atrix
on
the
prev
ious
pag
e.
PS/PI 115-131N 10/7/03, 9:37 AM125
©1994 GOAL/QPC126 PS/PI Model
5. Implement the solution or process change.• It is often recommended to try the solution on a
small scale first• Follow the plan and monitor the milestones and
measures
Typical toolsActivity Network Diagram, Flowchart, Gantt Chart,Matrix Diagram, and other project management meth-ods, as well as gathering ongoing data with Run Charts,Check Sheets, Histograms, Process Capability, andControl Charts
SituationThe team used the Responsibility Matrix and GanttChart to guide the training rollout. The original planwas to start in only two shops. There were some initialproblems with employees not being paid for the train-ing time as well as some managers not getting coverageduring training.
DecisionThe solution was to pay the employees for one-half thetraining time and to set up overlapping shifts for bettercoverage for managers. Once these adjustments weremade, the training went so well and had so manyunintended benefits that it was expanded to all sixshops.
Do
PS/PI 115-131N 10/7/03, 9:37 AM126
©1994 GOAL/QPC PS/PI Model 127
6. Review and evaluate the result of the change.• Confirm or establish the means of monitoring the
solution. Are the measures valid?• Is the solution having the intended effect? Any
unintended consequences?
Typical toolsCheck Sheet, Control Chart, Flowchart, Pareto Chart,Run Chart
SituationThe training plan was developed in January and Febru-ary and rolled out in March and April. The team wentback to the original Pareto Chart data to compare it tothe current customer complaint data. In addition, theycollected current turnover and delivery time measuresto test the original connection with improved training.
Check
The Pareto Chart on the next page shows the types ofcustomer complaints after the training plan was rolled outin March and April.
Pareto ChartBefore employee training
0
500
1000
1500
2000
2500
La
ted
eliv
-e
rie
s
Tast
e
Wro
ng
o
rde
r
Co
ld
foo
d
Oth
er
75%
9% 8% 5% 3%
Tota
l # o
f Cust
om
er C
om
pla
ints
Total=2510October–December
(across 6 shops)
PS/PI 115-131N 10/7/03, 9:37 AM127
©1994 GOAL/QPC128 PS/PI Model
Pareto ChartAfter employee training
DecisionThe team reviewed the original targets, which were setin Step 4.
Turnover: While not at the 30 percent average, ithad decreased for six consecutive months from ahigh of 62 percent to 44 percent last month.Delivery Time: Steadily declined from a high of 40minutes to its most recent level of 28 minutes.Customer Complaints: Overall, complaints werereduced by 52 percent and within “late deliveries”by 63 percentSales Volume: The average volume last month wasat 2250 units, up for the third straight month.
Unexpected resultFor the first time, customers complained about the lackof a good selection on the menu. While Stop ’N Go Pizzawas working toward increasing its speed of pizza deliv-eries and standardizing its processes, the variety of themenu was perhaps too simplified.
Total=1310May–July
(across 6 shops)
Tota
l # o
f Cus
tom
er C
ompl
aint
s
53%
20%14%
5% 5% 3%
0
200
400
600
800
1000
Late
deliv
-er
ies
Sel
ec-
tion
Tast
e
Wro
ng
orde
r
Col
d fo
od
Oth
er
PS/PI 115-131N 10/7/03, 9:37 AM128
©1994 GOAL/QPC PS/PI Model 129
Run Charts
Average turnover rate of employees(company-wide)
40
35
30
25
20
15
10
5
0Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
Ave
rage
Tim
e (in
min
utes
)
28minutes
Average delivery time(company-wide)
80
70
60
50
40
30
20
10
0Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
% 44%
PS/PI 115-131N 10/7/03, 9:37 AM129
©1994 GOAL/QPC130 PS/PI Model
7. Reflect and act on learnings.• Assess the results and problem-solving process
and recommend changes• Continue the improvement process where
needed; standardization where possible• Celebrate success
Typical toolsAffinity Diagram, Brainstorming, ImprovementStoryboard, Radar Chart
SituationSix months after the new training was started, the team metto evaluate its results and process. Team members used aRadar Chart to illustrate their assessment of the team.
DecisionThe Radar Chart showed strong agreement amongteam members on the performance of “Results,” “Useof tools,” and “Impact on customers.” The performanceand consensus among team members were both lowerin “Standardization” and “Teamwork.” When the teampresented its storyboard to top management, the majorresult was a complete overhaul of basic training contentand delivery, as well as the new performance measuresthat would continue to be monitored. The final celebra-tion was . . . what else . . . an all-expense-paid Stop ’N GoPizza party!
Future PossibilitiesMore efficient mapping, routing, and dispatching ofpizza deliveries, as well as more staff cross-training.
Act
PS/PI 115-131N 10/7/03, 9:37 AM130
©1994 GOAL/QPC PS/PI Model 131
Radar ChartTeam evaluation of itself after new training
Note: The “x” mark indicates the team’s average performancerating while the shaded area indicates the range of ratingswithin the team.
5
4
3
2
1
Use of Tools
Teamwork
Results
Standardization
Impact on Customers
X
X
X
X
X
PS/PI 115-131N 10/7/03, 9:37 AM131
©1994 GOAL/QPC132 Process Capability
LSL USLProcess CapabilityMeasuring conformanceto customer requirements
Why use it?To determine whether a process, given its natural varia-tion, is capable of meeting established customer re-quirements or specifications.
What does it do?• Helps a team answer the question, “Is the process
capable?”• Helps to determine if there has been a change in
the process• Helps to determine percent of product or service
not meeting customer requirements
How do I do it?1. Determine the process grand average, X, and the
average range, R• Use a stable Control Chart, which means the
process is stable and normally distributed.
2. Determine the Upper Specification Limit (USL)and the Lower Specification Limit (LSL)• The USL and LSL are based on customer require-
ments. Recognize that these specification limitsare based solely on customer requirements and donot reflect the capacity of the process.
3. Calculate the process standard deviation• Process cabability is based on individual points
from a process under study. Information from aControl Chart can be used to estimate the process’
132 Process Capability
Proc Capability 132-136N 10/7/03, 9:36 AM132
©1994 GOAL/QPC Process Capability 133
where R and s are the averages of the subgroupranges and standard deviation, and d2 and c4 arethe associated constant values based on the sub-group sample sizes. See the Table of Constants inControl Charts.
• The process average is estimated simply by X, X,X.
4. Calculate the process capability• To measure the degree to which a process is or is
not capable of meeting customer requirements,capability indices have been developed to com-pare the distribution of your process in relation tothe specification limits.
• A stable process can be represented by a measureof its variation—six standard deviations. Com-paring six standard deviations of the processvariation to the customer specifications providesa measure of capability. Some measures of capabil-ity include Cp and its inverse Cr , Cpl , Cpu, and Cpk.
average and variation (standard deviation, s).• σ is a measure of the process (population) stan-
dard deviation and can be estimated from infor-mation on the Control Chart by
R sd2 c4
σ = or σ =
USL – LSL6 σ
Cp =
Cp (simple process capability)
Tip While Cp relates the spread of the process rela-tive to the specification width, it DOES NOT lookat how well the process average is centered to thetarget value.
˜
Proc Capability 132-136N 10/7/03, 9:36 AM133
©1994 GOAL/QPC134 Process Capability
The process variation exceeds specification. Defectives are being made.
Cp<1
Cp=1
Cp>1
USLLSL
The process is just meeting specification. A minimum of .3% defectives will be made, more if the process is not centered.
The process variation is less than specification, however, defectives might be made if the process is not centered on the target value.
Cpl , Cpu, and Cpk (process capability indices)
The indices Cpl and Cpu (for single-sidedspecificiation limits) and Cpk (for two-sided speci-fication limits) measure not only the processvariation with respect to the allowable specifica-tion, they also take into account the location of theprocess average. Cpk is considered a measure ofthe process capability and is taken as the smallerof either Cpl or Cpu
X – LSL USL – X3 σ 3 σ
Cpl = Cpu = Cpk = min {Cpl, Cpu}
Tip If the process is near normal and in statisticalcontrol, Cpk can be used to estimate the expectedpercent of defective material. Estimating the per-centages of defective material is beyond the scopeof this book and can be found in statistical books.
Proc Capability 132-136N 10/7/03, 9:36 AM134
©1994 GOAL/QPC Process Capability 135
LSL USLTarget
XCpl Cpu
X – LSL USL – X
Actual Spread, 3 σ Actual Spread, 3 σ
• If the process is not capable, form a team toidentify and correct the common causes of thevariation in the process.– Process capability, based on individual data of
the process population, is used to determine ifa process is capable of meeting customer re-quirements or specifications. It represents a“snapshot” of the process for some specificperiod of time.
– Control Charts use small sample sizes overtime and look at the averages. The controllimits are natural limits of variation of theaverages within the sample. These limits are not tobe confused with specification limits, whichare for individual data points in the population.
VariationsThe construction steps described in this section are basedon process capability of a Variable Data Control Chart.The process capability of an Attribute Data Control Chartis represented by the process averages p, np, c, and u.
Proc Capability 132-136N 10/7/03, 9:36 AM135
©1994 GOAL/QPC136 Process Capability
Process CapabilityDie Cutting Process
207 208 209 210 211 212 213 214 215
USLLSL
X
Cp = USL – LSL = 213 – 207 = 6 = 1.938 6σ 6(.516) 3.096
Cpl = X – LSL = 212.5 – 207 = 5.5 = 3.553 3σ 3(.516) 1.548
Cpu = USL – X = 213 – 212.5 = 0.5 = 0.323 3σ 3(.516) 1.548
Cpk = min {Cpl, Cpu} = 0.323
Since Cpk<1, defective material is being made.
R 1.2d2 2.326
σ = = = .516
A Control Chart was maintained, producing the following statistics:X = 212.5 R = 1.2 n = 5Spec. = 210 ± 3 USL = 213 LSL = 207
Proc Capability 132-136N 10/7/03, 9:36 AM136
©1994 GOAL/QPC Radar 137
Radar ChartRating organizationperformance
Why use it?To visually show in one graphic the size of the gapsamong a number of both current organization perfor-mance areas and ideal performance areas.
What does it do?• Makes concentrations of strengths and weak-
nesses visible• Clearly displays the important categories of per-
formance• If done well, clearly defines full performance in
each category• Captures the different perceptions of all the team
members about organization performance
How do I do it?1. Assemble the right team/raters
Tip It is critical to get varied perspectives to avoidorganization “blind spots.”
2. Select and define the rating categories• The chart can handle a wide number of categories,
with 5–10 categories as an average.• Brainstorm or bring headers from an Affinity
Diagram to create the categories.• Define both non-performance and full perfor-
mance within each category so ratings are doneconsistently.
Radar 137
Radar Chts 137-140N 10/7/03, 9:38 AM137
©1994 GOAL/QPC138 Radar
3. Construct the chart• Draw a large wheel on a flipchart with as many
spokes as there are rating categories.• Write down each rating category at the end of
each spoke around the perimeter of the wheel.• Mark each spoke on a zero to “n” scale with “0”
at the center equal to “no performance” and thehighest number on the scale at the outer ringequal to “full performance.” Performance can bemeasured either objectively or subjectively.
= Individual rating= Team consensus rating
A
B
C
D
E
F
G
H
4. Rate all performance categoriesa) Individual: Each person rates in silence, using
multicolored markers or adhesive labels directlyon the flipchart.
b) Team: Through consensus or an average of individualscores, get a team rating. Take into account both theclustering and the spread of the individual ratings.
A
B
C
D
E
F
G
H
5
45
646
72
Radar Chts 137-140N 10/7/03, 9:38 AM138
©1994 GOAL/QPC Radar 139
Tip Make the team rating highly visible on thechart. Be sure to differentiate the team ratingsfrom individual ratings on the chart by color ortype of mark.
5. Connect the team ratings for each category andhighlight as needed
Tip A gap score can be added to each category bysubtracting the team rating score from the highestnumber on the rating scale, e.g., on a scale of “10,”a team rating of “4” produces a gap score of “6” incategories B and E.
6. Interpret and use the results• The overall ratings identify gaps within each cat-
egory but not the relative importance of the cat-egories themselves. Work on the biggest gap inthe most critical category.
• Post the resulting Radar Chart in a prominentplace, review progress regularly, and update thechart accordingly. It is a great visual “reportcard.”
A (5)
B (6)
C (5)
D (4)
E (6)
(4) F
(3) G
(8) H5
4
5
646
72
Gap scores are in parentheses.
Radar Chts 137-140N 10/7/03, 9:38 AM139
©1994 GOAL/QPC140 Radar
RadarTQC Review Scores
4
3
2
1 4
3
2
1
4
3
2
1
4
3
2
1
4
3
2
1
Goal
World leader
Improvement Cycle
CustomerFocus
Planning Process
Total Participation
Process Management
Company's goal: to have 80% of all entities (34) achieve an overall score of >3.5.
To compute overall score:Sum of average
scores from each category# of categories
12.525
= 2.5 (maximum is 5)=
Range of ratings within the team
Average
Information provided courtesy of Hewlett Packard
Radar Chts 137-140N 10/7/03, 9:38 AM140
©1994 GOAL/QPC Run 141
Run ChartTracking trends
Why use it?To allow a team to study observed data (a performancemeasure of a process) for trends or patterns over aspecified period of time.
What does it do?• Monitors the performance of one or more pro-
cesses over time to detect trends, shifts, or cycles• Allows a team to compare a performance mea-
sure before and after implementation of a solu-tion to measure its impact
• Focuses attention on truly vital changes in the process• Tracks useful information for predicting trends
How do I do it?1. Decide on the process performance measure
2. Gather data• Generally, collect 20-25 data points to detect
meaningful patterns.
3. Create a graph with a vertical line (y axis) and ahorizontal line (x axis)• On the vertical line (y axis), draw the scale related
to the variable you are measuring.– Arrange the y axis to cover the full range of the
measurements and then some, e.g., 11/2 timesthe range of data.
Run 141
Run Chts 141-144N 10/7/03, 9:38 AM141
©1994 GOAL/QPC142 Run
• On the horizontal line (x axis), draw the time orsequence scale.
4. Plot the data• Look at the data collected. If there are no obvious
trends, calculate the average or arithmetic mean.The average is the sum of the measured valuesdivided by the number of data points. The me-dian value can also be used but the mean is themost frequently used measure of the “centering”of the sample. (See Data Points for more informa-tion on averages.) Draw a horizontal line at theaverage value.
Tip Do not redraw this average line every timenew data is added. Only when there has been asignificant change in the process or prevailingconditions should the average be recalculatedand redrawn, and then only using the data pointsafter the verified change.
(x axis)Time or sequence
(y a
xis)
Mea
sure
men
t
Average
Run Chts 141-144N 10/7/03, 9:38 AM142
©1994 GOAL/QPC Run 143
5. Interpret the Chart• Note the position of the average line. Is it where it
should be relative to a customer need or specifica-tion? Is it where you want it relative to yourbusiness objective?
Tip A danger in using a Run Chart is the tendencyto see every variation in data as being important.The Run Chart should be used to focus on trulyvital changes in the process. Simple tests can beused to look for meaningful trends and patterns.These tests are found in Control Charts in the“Determining if Your Process is Out of Control”section. Remember that for more sophisticateduses, a Control Chart is invaluable since it issimply a Run Chart with statistically-based limits.
Run Chts 141-144N 10/7/03, 9:38 AM143
©1994 GOAL/QPC144 Run
RunAverage Number of Days
for Determining Eligibility for Services
Ave . #
400
350
300
250
200
150
100 50 0
Dec '92
Jan
'93
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
'94
Pro
cess
Ave
rage
Info
rmat
ion
prov
ided
cou
rtes
y of
Geo
rgia
Sta
te D
epar
tmen
tof
Hum
an R
esou
rces
, Div
isio
n of
Reh
abili
tatio
n S
ervi
ces
Note: Eligibility requirements changed in May, making itmuch simpler for the department staff to make determina-tions. The trend is statistically significant because there are sixor more consecutive points declining.
Run Chts 141-144N 10/7/03, 9:38 AM144
©1994 GOAL/QPC Scatter 145Scatter 145
Scatter DiagramMeasuring relationshipsbetween variables
Why use it?To study and identify the possible relationship betweenthe changes observed in two different sets of variables.
What does it do?• Supplies the data to confirm a hypothesis that two
variables are related• Provides both a visual and statistical means to test
the strength of a potential relationship• Provides a good follow-up to a Cause & Effect
Diagram to find out if there is more than just aconsensus connection between causes and the effect
How do I do it?1. Collect 50–100 paired samples of data that you
think may be related and construct a data sheet
Course
123•••
40
AverageSession Rating(on a 1–5 scale)
4.23.74.3•••
3.9
AverageExperience of
Training Team (days)
220270270
•••
625
Theory: There is a possible relationship between the number of days of experience the training team has received and the ratings of course sessions.
Scatter 145-149N 10/7/03, 9:39 AM145
©1994 GOAL/QPC146 Scatter
2. Draw the horizontal (x axis) and vertical (y axis)lines of the diagram• The measurement scales generally increase as
you move up the vertical axis and to the right onthe horizontal axis.
4.54.44.34.24.14.03.93.83.73.63.5
150 200 250 300 350 400 450 500 550 600 650
Dependent variable (“effect”)
Independent variable (“cause”)
Looking for relationships,not cause & effect
4.54.44.34.24.14.03.93.83.73.63.5
150 200 250 300 350 400 450 500 550 600 650
Ove
rall
Rat
ing
of th
e S
essi
on
Average Experience of Training Team (in days)
4.0 y axis430 x axis
3. Plot the data on the diagram• If values are repeated, circle that point as many
times as appropriate.
Information provided courtesy of Hamilton Standard
Scatter 145-149N 10/7/03, 9:39 AM146
©1994 GOAL/QPC Scatter 147
4. Interpret the data• There are many levels of analysis that can be
applied to Scatter Diagram data. Any basic statis-tical process control text, like Kaoru Ishikawa’sGuide to Quality Control, describes additional cor-relation tests. It is important to note that all of theexamples in this chapter are based on straight-line correlations. There are a number of non-linear patterns that can be routinely encountered,e.g., y = ex, y = x2). These types of analyses arebeyond the scope of this book.
• The following five illustrations show the variouspatterns and meanings that Scatter Diagrams canhave. The example used is the training sessionassessment previously shown. The patterns havebeen altered for illustrative purposes. Pattern #3is the actual sample.
Tip The Scatter Diagram does not predict cause andeffect relationships. It only shows the strength ofthe relationship between two variables. Thestronger the relationship, the greater the likeli-hood that change in one variable will affectchange in another variable.
Scatter 145-149N 10/7/03, 9:39 AM147
©1994 GOAL/QPC148 Scatter
1. Positive Correlation. An in-crease in y may depend on anincrease in x. Session ratingsare likely to increase as trainerexperience increases.
2. Possible Positive Correla-tion. If x is increased, y mayincrease somewhat. Other vari-ables may be involved in thelevel of rating in addition totrainer experience.
Trainer Experience
4.5
4.0
3.5
150 400 650
Rat
ing
y
x
Trainer Experience
4.5
4.0
3.5
150 400 650
Rat
ing
y
x
Trainer Experience
4.5
4.0
3.5
150 400 650
Rat
ing
y
x
Trainer Experience
4.5
4.0
3.5
150 400 650
Rat
ing
y
x
Trainer Experience
4.5
4.0
3.5
150 400 650
Rat
ing
y
x
3. No Correlation. There is nodemonstrated connection be-tween trainer experience andsession ratings.
4. Possible Negative Correla-tion. As x is increased, y maydecrease somewhat. Othervariables, besides trainer ex-perience, may also be affect-ing ratings.
5. Negative Correlation. A de-crease in y may depend on anincrease in x. Session ratingsare likely to fall as trainer ex-perience increases.
Scatter 145-149N 10/7/03, 9:39 AM148
©1994 GOAL/QPC Scatter 149
ScatterCapacitance vs. Line Width
2.1
2.0
1.9
1.8
1.7
1.6
40 45 50 55 60 65 70
Cap
acita
nce
(pic
ofar
ads)
Line Width/Spacing (in microns)
Note: This Scatter Diagram shows that there is a strong positiverelationship between these two variables in producing micro-electronic circuits. Since capacitance measures a critical perfor-mance of a circuit, anything that affects it positively or nega-tively is also critical. The diagram shows that line width/spacing is something to watch closely, perhaps using a ControlChart or another type of statistical process control (SPC) tool.
Information provided courtesy of AT&T
Scatter 145-149N 10/7/03, 9:39 AM149
©1994 GOAL/QPC150 Team Guidelines
TeamGuidelinesFrom “me” to “we”
Starting TeamsThe most critical task for any new team is to establish itspurpose, process, and measures of team progress. Oncethe team has developed the following guidelines andcharters specific to its purpose, they should be recordedon a flipchart and posted at each team meeting forreference.
• Develop a Team Behavior Charter– Groundrules. Develop consensus groundrules of
acceptable and unacceptable individual andteam behavior.
– Decision making. Determine whether decisionswill be made by consensus, majority rule, oranarchy! Discuss whether there are, or shouldbe, exceptions to when the team should notfollow its usual process.
– Communication. Recognize the value of listen-ing and constructive feedback, and make theeffort, every day, to communicate constructively!
– Roles and participation. Discuss how the teamwill choose a leader, and generally how theteam process will be led. The individuals andteam must take responsibility to encourageequal participation.
– Values. Acknowledge and accept the uniqueinsight of each member of the team.
• Develop a Purpose Charter– Establish the answer to why the team exists.
150 Team Guidelines
Team Guides 150-155N 10/7/03, 9:40 AM150
©1994 GOAL/QPC Team Guidelines 151
– Bring together the individuals who wouldwork well together as a team. Determinewhether each person has the knowledge, skills,and influence required to participate effec-tively on the team.
– The team should discuss who its customers are.If the team has multiple customers, decidewhich customers have the highest priority, orat least how their needs will be balanced.
• Develop Measures of Team Progress– Discuss and agree on the desired signals, which
the team can assess both objectively and subjec-tively, that will indicate the team is making progress.
– Discuss and agree on the types of measures andoutcomes that will indicate the team hasreached success or failure.
– Estimate the date when the project should becompleted.
Maintaining MomentumMany teams enjoy terrific starts and then soon fizzle.The real challenge is to keep a team focused on itspurpose and not the histories of its members and theirrelationships to one another.
• Agree on the Improvement Model to Use– Standard steps. Use your organization’s stan-
dard step-by-step improvement process orchoose from the many published options. (Seethe Improvement Storyboard in the ProblemSolving/Process Improvement Model sectionfor one such standard process.)
Team Guides 150-155N 10/7/03, 9:40 AM151
©1994 GOAL/QPC152 Team Guidelines
– Data. Gather relevant data to analyze the cur-rent situation. Define what you know, andwhat you need to know, but know when tostop. Learn, as a team, to say when your workis good enough to proceed to the next step inthe process.
– Develop a plan. Use your organization’s stan-dard improvement model to provide the over-all structure of a project plan. Estimate timesfor each step and for the overall project. Moni-tor and revise the plans as needed.
• Use Proven Methods Based on Both Data andKnowledge– Data-based methods. Use tools in this booklet,
e.g., Run Chart, Pareto Chart, that reveal pat-terns within data. These tools often take theemotion out of discussions and keep the pro-cess moving.
– Knowledge-based methods. Many of the methodsin this booklet, e.g., Affinity Diagram, Interre-lationship Digraph, help to generate and ana-lyze ideas to reveal the important informationwithin. They help create consensus, which isthe ideal energy source for a team.
• Manage Team Dynamics– Use facilitators. A facilitator is someone who
monitors and helps team members to keeptheir interactions positive and productive. Thisis the stage when a facilitator can help the teamstay focused on its purpose while improving itsworking relationship.
Team Guides 150-155N 10/7/03, 9:40 AM152
©1994 GOAL/QPC Team Guidelines 153
– Manage conflict. As teams grow, so do conflicts.This is a natural process as communicationbecomes more open. The entire team can learntechniques for conflict resolution and use thefacilitator as a resource.
– Recognize agreement. Managing agreement isoften as much of an effort as managing dis-agreement. Test for agreement often and writedown the points of agreement as they occur.
– Encourage fair participation. Each team membermust eventually take responsibility for partici-pating consistently in all discussions. Like-wise, the entire team should be constantlyworking to “pull back” the dominant membersand draw out the quieter members.
Ending Teams/ProjectsMost teams and all projects must eventually end. Bothoften end in unsatisfactory ways or don’t “officiallyend” at all. Before ending, the team should review thefollowing checklist:
❐ We checked our results against our originalgoals and customer needs.
❐ We identified any remaining tasks to be done.
❐ We established responsibility for monitoringthe change over time.
❐ We documented and trained people, when nec-essary, in the new process.
Team Guides 150-155N 10/7/03, 9:40 AM153
©1994 GOAL/QPC154 Team Guidelines
❐ We communicated the changes to everyoneaffected by them.
❐ We reviewed our own team’s accomplish-ments for areas of improvement.
❐ We celebrated the efforts of the team with alunch, newsletter article, special presentation tothe company, or other expression of celebration.
❐ We feel proud of our contribution and ac-complishments, our new capabilities, and ournewly defined relationships with coworkers.
Conducting Effective MeetingsPreparation:
• Decide on the purpose of the meeting• Develop a meeting plan (who, what, where,
when, how many)• Identify the meeting leader• Prepare and distribute the agenda• Set up the meeting area
Beginning:• Start on time• Introduce the meeting leader• Allow team members to introduce themselves• Ask for a volunteer timekeeper• Ask for a volunteer recorder• Review, change, order the agenda• Establish time limits• Review prior meeting action items
Team Guides 150-155N 10/7/03, 9:40 AM154
©1994 GOAL/QPC Team Guidelines 155
Meeting Etiquette:• Raise your hand and be recognized before
speaking• Be brief and to the point• Make your point calmly• Keep an open mind• Listen without bias• Understand what is said• Avoid side conversations• Respect other opinions• Avoid personal agendas• Come prepared to do what’s good for
the company• Have fun
Ending:• Develop action items (who, what, when, how)• Summarize the meeting with the group• Establish the date and time for a follow-up meeting• Evaluate the meeting• End on time• Clean the meeting area
Next Steps:• Prepare and distribute the meeting activity report• Follow up on action items• Go to “Preparation”
Team Guides 150-155N 10/7/03, 9:40 AM155
©1994 GOAL/QPC156 Tree
Tree DiagramMapping the tasksfor implementation
Why use it?To break any broad goal, graphically, into increasinglevels of detailed actions that must or could be done toachieve the stated goals.
What does it do?• Encourages team members to expand their think-
ing when creating solutions. Simultaneously, thistool keeps everyone linked to the overall goalsand subgoals of a task
• Allows all participants, (and reviewers outsidethe team), to check all of the logical links andcompleteness at every level of plan detail
• Moves the planning team from theory to the realworld
• Reveals the real level of complexity involved inthe achievement of any goal, making potentiallyoverwhelming projects manageable, as well asuncovering unknown complexity
How do I do it?1. Choose the Tree Diagram goal statement
Goal: Increase workplace suggestions
156 Tree
Tree 156-164N 10/7/03, 9:40 AM156
©1994 GOAL/QPC Tree 157
• Typical sources:– The root cause/driver identified in an Interre-
lationship Digraph (ID)– An Affinity Diagram with the headers as major
subgoals– Any assignment given to an individual or team
• When used in conjunction with other manage-ment and planning tools, the most typical sourceis the root cause/driver identified in the ID.
Tip Regardless of the source, work hard to cre-ate—through consensus—a clear, action-ori-ented statement.
2. Assemble the right team• The team should consist of action planners with
detailed knowledge of the goal topic. The teamshould take the Tree only to the level of detail thatthe team’s knowledge will allow. Be prepared tohand further details to others.
• 4-6 people is the ideal group size but the TreeDiagram is appropriate for larger groups as longas the ideas are visible and the session is wellfacilitated.
3. Generate the major Tree headings, which are themajor subgoals to pursue• The simplest method for creating the highest, or
first level of detail, is to brainstorm the major taskareas. These are the major “means” by which thegoal statement will be achieved.
Tree 156-164N 10/7/03, 9:40 AM157
©1994 GOAL/QPC158 Tree
• To encourage creativity, it is often helpful to do an“Action Affinity” on the goal statement. Brain-storm action statements and sort into groupings,but spend less time than usual refining the headercards. Use the header cards as the Tree’s first-level subgoals.
Increase workplace suggestions
Goal Means
Create a workable process
Create capability
Measure results
Provide recognition
Tip Use Post-it™ notes to create the levels of detail.Draw lines only when the Tree is finished. Thisallows it to stay flexible until the process is fin-ished. The Tree can be oriented from left to right,right to left, or top down.
Tip Keep the first level of detail broad, and avoidjumping to the lowest level of task. Remember: “Ifyou start with what you already know, you’llend up where you’ve already been.”
Tree 156-164N 10/7/03, 9:40 AM158
©1994 GOAL/QPC Tree 159
4. Break each major heading into greater detail• Working from the goal statement and first-level
detail, placed either to the extreme left, right ortop of the work surface, ask of each first-levelitem:“What needs to be addressed to achieve the goalstatement?”Repeat this question for each successive level ofdetail.
• Stop the breakdown of each level when there areassignable tasks or the team reaches the limit to itsown expertise. Most Trees are broken out to thethird level of detail (not counting the overall goalstatement as a level). However, some subgoals arejust simpler than others and don’t require asmuch breakdown.
Increase workplace
suggestions
Goal
MeansCreate a workable process
Create capability
Create simple input system
Streamline evaluation
Create quick implementation
Provide information
Supply cost data
Provide single-page
documentation
Go online
Make evaluation a key manager responsibility
Do local evaluation
Monitor approved suggestions monthly
Provide a projected implementation date at
time of approval
Why?
How?
Tree 156-164N 10/7/03, 9:40 AM159
©1994 GOAL/QPC160 Tree
5. Review the completed Tree Diagram for logicalflow and completeness• At each level of detail, ask “Is there something
obvious that we have forgotten?”• As the Tree breaks down into greater detail (from
general to specific) ask, “If I want to accomplishthese results, do I really need to do these tasks?”
• As the Tree builds into broader goals (from thespecific to the general) ask, “Will these actionsactually lead to these results?”
• Draw the lines connecting the tasks.Tip The Tree Diagram is a great communication
tool. It can be used to get input from those outsidethe team. The team’s final task is to considerproposed changes, additions or deletions, and tomodify the Tree as appropriate.
VariationsThe Process Decision Program Chart (PDPC) is a valu-able tool for improving implementation through con-tingency planning. The PDPC, based on the Tree Dia-gram, involves a few simple steps.
1. Assemble a team closest to the implementation
2. Determine proposed implementation steps• List 4–10 broad steps and place them in sequence
in the first Tree level.
3. Branch likely problems off each step• Ask “What could go wrong?”
4. Branch possible and reasonable responses offeach likely problem
Tree 156-164N 10/7/03, 9:40 AM160
©1994 GOAL/QPC Tree 161
5. Choose the most effective countermeasures andbuild them into a revised plan
Goal
Step #1 Step #2 Step #3
Likely problems
Reasonable countermeasures
X = Difficult/impossibleO = Selected
X O X O O X
Tree 156-164N 10/7/03, 9:40 AM161
©1994 GOAL/QPC162 Tree
Awarding Unrestricted Financial Aid
Establish total financial aid budget
Decide who receives financial aid†
Decide how financial aid is awarded†
Determine amount for returning
students
Determine amount for new students†
Not sure about retention
Not sure about tuition rate of
increase
Not sure about outside sources
of aid
• Increase enrollment research – existing data• Additional quantitative data• Additional qualitative data
• Seek quicker decision on tuition rate increase• Plan for variety on rates of increase
• Survey students – expected resources• Participate in state & federal financial aid organizations
† Further information exists but is not shown
Information provided courtesy of St. John Fisher College
PDPC (Tree Variation)Awarding Unrestricted Financial Aid
Note: The PDPC surfaced a lack of accurate information as amajor problem. By anticipating this and filling the most criticalinformation gaps, the budget can be more accurate.
Tree 156-164N 10/7/03, 9:41 AM162
©1994 GOAL/QPC Tree 163
TreeImprove
Business Planning Interaction
Information providedcourtesy of Goodyear
Improve interactionamong functional
areas represented inthe group, in the
creation andimplementation of an
effective business plan
Developprocedures toassure teameffectiveness
Understandgoals &roles of
other areas
Interactiontechniques
Make groupmeetings more
effective
Provide system tocommunicate
progress
Write & circulatemeeting minutes
Interaction mustoccur with
regularfrequency
Publish & adhereto agenda, with
team input
Distributetracking charts
of teamperformance
Establish methodof communicating
progress orproblems
Require eachfunction toperiodicallyreport status
Show functionalinterdependencies
in plandevelopment
Use consensus-building techniquesin plan development& implementation
Identifyrelationships anddependencies in
project plans
Each functionshows its plan tofulfill overall plan
Participate in jointtraining of planning
methods
Use facilitatorapproach at
meetings
Use appropriatetools to makedecisions andsolve problems
1Seenextpage
Tree 156-164N 10/7/03, 9:41 AM163
©1994 GOAL/QPC164 Tree
TreeImprove
Business Planning Interaction(continued)
Improve interaction
. . . effectivebusiness plan
Develop procedures toassure team effectiveness
Understand goals & roles of
other areas
Understand regional & functional objectives
Improve understanding
of other functional
areas
Management decisions
affecting group must be
communicated ASAP
Define objectives &
role of the group
Each functionalarea to explain/discuss division
objectives
All functional areas must be represented &
contribute
Develop matrix of evaluation & rewards by
functional areas
Have “field trips” to different
functional areas
Interaction techniques
Rotate associates among functions
1from
previous page
Information providedcourtesy of Goodyear
Tree 156-164N 10/7/03, 9:41 AM164
IMPROVING THE WAY ORGANIZATIONS RUN