06-Lssbb-Pg1a Vol 1 (Mod. 1-13) (Ingles)

LEAN SIX SIGMA
BLACK BELT*
(Volume 1)
Participant
ParticipantGuide
Guide
Texts: Knowledge Based Management

Lean Six Sigma: A Tools Guide
Basic Statistics: Tools for Continuous Improvement
Understanding Industrial Designed Experiments
Knowledge Notebook DMAIC
Software: SPC XL
DOE KISS
DOE PRO
SimWare PRO
Copyright © 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
06-LSSBB-PG1A
1650 Telstar Drive, Ste 110

Colorado Springs, CO 80920
Phone: 719-531-0777
Fax: 719-531-0778
email: aapa@airacad.com
web-site: www.airacad.com
All rights reserved. Do not reproduce.

USER COPYRIGHT ACKNOWLEDGEMENT
I hereby acknowledge that this and all other Air Academy

Associates training materials are proprietary and confidential
and that copyright for the materials belongs to Air Academy
Associates. I agree to exercise the highest degree of care to
protect Air Academy Associates training materials and any
associated confidential information from unauthorized use and
disclosure. I understand that reproduction or foreign language
translation of Air Academy Associates’ materials without
specific written permission is a violation of copyright. Since
trust is the cornerstone of any strong relationship, I understand
that any violation of this agreement may irreparably harm my
standing with Air Academy Associates, and in the event of any
such violation, Air Academy Associates shall be entitled to
seek relief for any damages/harm which it may sustain as a
result thereof.
ANONYMOUS SURVEY
1. Rate your level of Lean Six Sigma knowledge on a scale of

0 to 10.
(0 → no knowledge; 10 → perfect knowledge)
2. What does Lean Six Sigma mean to you? (25 words or

less)
3. What is your commitment level on a scale of 0 to 10 to the

Lean Six Sigma initiative?
(0 → absolutely no commitment; 10 → “I couldn’t be more
fired up!”)
Copyright Air Academy Associates

LEAN SIX SIGMA
INFORMATION SHEET
NAME:
Job:
Your Phone / E-mail:
NAME OF SPONSOR:
Phone / E-mail of Champion/Sponsor:
NAME OF PROJECT:
Brief Project Description (25 Words or Less):
Expected Yearly ROI from Project:

How long at your company?:
Background in Lean, Six Sigma, or Statistics:
What are 3 important pieces of information you’d like to share about

yourself?
1.
2.
3.

CONTENTS
Goals
Outcomes
Course Guidelines
Introductions
WEEK 1 Module 1 Lean Six Sigma...The Journey Begins
Module 2 Lean Six Sigma Fundamentals
Module 3 Defining the Project and Managing Change
Module 4 Understanding the Voice of the Customer

(VOC) and Defining a Process
Module 5 Measure…Making Sense out of Data Using

Graphical and Measurement Tools
Module 6 Measurement System Analysis
WEEK 2 Module 7 Analyzing the Causes of Poor Performance
Module 8 Techniques for Narrowing the Focus
Module 9 Drawing Conclusions from Sampled Data
Module 10 Improving the Process and Work Flow
Module 11 Controlling Process Performance / Realizing

and Holding the Gains
Module 12 DMAIC Summary – Putting it All Together
Module 13 References, Glossary of Terms, Course

Evaluations
Copyright Air Academy Associates i

CONTENTS (cont.)
WEEK 3 Module 14 Review of DMAIC Tools and Techniques
Module 15 Additional Lean Tools to Define and Measure
Module 16 Analyze Data Using Probability Distributions
Module 17 Regression Modeling to Analyze and
Improve
Module 18 Introduction to the Use of Design of
Experiments (DOE) to Analyze and Improve
Module 19 Two-Level Designs
Module 20 Three-Level Designs
Module 21 Rules of Thumb and DOE Design Selection
WEEK 4 Module 22 Quick Review of DMAIC and DOE

Module 23 Additional Lean Improvement Tools
Module 24 Additional Hypothesis Tests to Analyze and
Improve
Module 25 DOE Diagnostics and Assessments
Module 26 Advanced DOE Analysis Techniques and
Designs
Module 27 Historical Data Analysis
Module 28 Lean Six Sigma Mini-Project
Module 29 Advanced Black Belt Topics and Training
Module 30 References, Glossary of Terms, and Course
Evaluation Forms
Copyright Air Academy Associates ii

GOALS (Weeks 1 & 2)
- To understand the nature of a Lean Six Sigma culture and

to be able to contribute substantially to its development
- To know the basic principles of Lean Six Sigma
- To understand and apply Lean Six Sigma tools and
concepts, including the DMAIC roadmap, on your project
for the purpose of improving the bottom line
- To understand the need for a well-defined project and that
good project management practices, along with teamwork,
change management practices, and voice of the customer
are all needed for project success
- To understand the concepts of value and value stream
- Develop an appreciation for the need and use of statistics
in process improvement
- Be able to describe a process graphically and numerically
- Be able to use and apply the basic graphical tools and
charting techniques
- Understand methods and tools for prioritizing and focusing
attention
- Be able to apply the concepts of flow and pull to both
manufacturing and transactional processes
- Understand the relationship between variation and the
cost of poor quality
- Be able to apply the tools for removing non-value added
activities and reducing the Cost of Waste
Copyright Air Academy Associates iii

GOALS (Weeks 1 & 2) (cont.)
- Be able to conduct a Measurement System Analysis (MSA)

to determine the integrity of the data we are getting
- Be able to reduce the time for changeovers, both in
production and transactions
- Be able to apply 5S and Visual Controls to the workplace
- To understand the difference between a batch and queue
philosophy and one-piece flow
- To know what Takt time and Cycle Time are, and how to
calculate them
- Be able to construct and interpret a confidence interval
- Know what a 2-sample hypothesis test is and be able to use
statistical tests and Rules of Thumb to detect shifts in
average and standard deviation
- Be able to select the appropriate control chart to use, know
how to construct it, and be able to recognize out-of-control
symptoms
- Know the difference between process control and process
capability, and be able to compute process capability
- Understand the need for and use of control plans for
sustaining the gains
- Measure the results from your Lean Six Sigma project and
understand strategies for holding the gains
Copyright Air Academy Associates iv

GOALS (Weeks 3 & 4)
- Gain confidence applying the Lean Six Sigma tools and

methodology to increase product and process knowledge,
to reduce the Cost of Waste and enhance business
profitability and customer value
- Become familiar with three of the most commonly
occurring distributions (normal, binomial, and Poisson)
and be able to apply them to your project
- To determine normality of a distribution using goodness of
fit
- Be able to transform non-normal data to normal data
- Be able to use software to create value stream maps and
improve a process
- Be able to apply a 1-sample hypothesis test, paired t-test,
and ANOVA in the correct situations
- Know what regression analysis is and be able to interpret
regression analysis computer output
- Understand the need for modern methods of
experimentation
- Know the steps required for successful experimentation
- Be able to set up designs and conduct simple analysis
- Know the Rules of Thumb associated with DOE
- Be familiar with all of the best design techniques to include
the objectives, similarities/differences, and
advantages/disadvantages of each
- Understand the advantages and disadvantages of different
variance reduction strategies
Copyright Air Academy Associates v

GOALS (Weeks 3 & 4) (cont.)
- Know how to recover from experiments that do not confirm

- Be able to interpret output from non-orthogonal designs and
historical data
- Know how to clean up data from non-DOE experiments
- Be able to convert coded models into non-coded models
- Know how to optimize processes having multiple responses
- Be able to use diagnostics to assess a regression model
- Be able to use Kanban and inventory calculations to
improve processes
- Be able to apply the correct tools to a comprehensive mini-
project
- Be able to set up and interpret robust designs
- Understand the need for randomization
- Be able to determine and correct Lack of Fit problems with
regression models
- Understand how production sequencing can improve a
process
- Be able to apply Little’s Law
- Be able to calculate Overall Equipment Effectiveness
- Understand the concept of Total Productive Maintenance
- Become familiar with the principles and process of Design
for Six Sigma (DFSS)
- Be familiar with the advanced topics of High Throughput
Testing and discrete-event simulation
Copyright Air Academy Associates vi

COURSE OUTCOMES
This training will provide each participant with the ability to:
1. Identify and Define opportunities within their business unit

processes to which Lean Six Sigma tools and techniques can
and will be applied.
2. Measure and Analyze a process using the Lean Six Sigma

methodology.
3. Improve a process using Lean Six Sigma tools and techniques.
4. Control a process by using Lean Six Sigma tools and Realize

return on investment from this training to improve the bottom
line.
5. Repeatedly apply the Define – Measure – Analyze – Improve –

Control (DMAIC) Strategy to produce success stories and
improve the corporate scorecard.
6. Communicate at all levels of an organization.
7. Change the culture of an organization.
Copyright Air Academy Associates vii

COURSE GUIDELINES
• Start and stop times
• Breaks
• You are responsible for learning
• Active participation
• Questions are good
• Importance of applying right away
INTRODUCTIONS
• Name
• Organization / Your Role
• Background in Lean and/or Six Sigma
• Expectations
Copyright Air Academy Associates viii

MODULE 1
LEAN SIX SIGMA...

THE JOURNEY BEGINS
LEAN SIX SIGMA...
THE JOURNEY BEGINS
• IN THIS MODULE, WE WILL DISCUSS:
– THE WHAT AND WHY OF LEAN SIX SIGMA

• Definitions
• Goals
• Benefits
• Impact Zones
– KEY ELEMENTS AND INFRASTRUCTURE
– ROLES AND RESPONSIBILITIES
– LEAN SIX SIGMA PROJECT MASTER STRATEGY
– IMPORTANCE OF KNOWLEDGE
• Role of Questions
– COURSE PHILOSOPHY AND INSTRUCTIONAL
APPROACH
Copyright Air Academy Associates 1-1

WHY ARE WE HERE?
WORLD CLASS QUALITY

BUSINESS PROFITABILITY
CUSTOMER VALUE
Providing a
BETTER product or service,
FASTER, and
at a LOWER COST
than our competition
VARIATION and WASTE … the enemies of

customer satisfaction
“Always know the language of the enemy.”
Knowledge

LEAN SIX SIGMA
GOALS AND BENEFITS
• BUSINESS RESULTS
– Generate business growth and improve profitability
– Achieve total customer satisfaction resulting in
increased market share and improved profit margin
• Remove wasteful/non-value added activities
• Decrease defects, cycle time, and inventory;
increase first pass yields and improve quality
• Expand knowledge of products and processes
through characterization and optimization
• BREAKTHROUGH TECHNOLOGIES
– Develop leaders in breakthrough technologies to
meet stretch goals associated with Better, Faster,
Lower Cost Products and Services
– Provide a common set of tools and techniques (a
common methodology)
• WORLD CLASS CULTURE

– Develop a world class culture for competitive
advantage
– Improve communication and teamwork

OLD PHILOSOPHY OF QUALITY
– Quality is based on conformance to specifications

(“Goalpost View”)
– Test an inspection used to ensure we satisfy

customer requirements
– in spec = “good”
– out of spec = “bad”
Loss due to Loss due to

scrap and scrap and
rework No Loss rework
Lower Upper
Spec Spec
Limit Limit
(LSL) (USL)
- What are some of the

problems with this
quality philosophy?

EXERCISE: WHY NOT JUST
TEST / INSPECT DEFECTS OUT?
You cannot rely on test and inspection to find all

defects, because no test or inspection is 100%
effective! If you don’t believe this, then try this
exercise.
We will simulate the process of visual inspection.

Let’s assume that the letter “e” represents a defect.
The instructor will give you one minute to inspect this
page and count the total number of “e’s” that you find.
Write down your number on a slip of paper and give it
to the instructor.
Hopefully you will find all of them. But, remember that

this is only one page. What if your job was to inspect
all of the pages in this manual?

DEFINITIONS
Traditional Lean…..
– Views time as a strategic competitive weapon
– Focuses on the removal of wasted efforts, that is, any
activity that does not add value or is not necessary to
produce a product or provide a service
– Concentrates on improving flow to meet customer
requirements with minimum inventories in the least
amount of time
Traditional Six Sigma….

– Focuses on business profitability and growth
– Concentrates on driving out defects and process
variation using a structured methodology
– Is data-driven
Lean Six Sigma…

– Combines the proven and powerful elements of both
Lean and Six Sigma
– Focuses on customer value and business profitability
– Uses a structured methodology for reducing non-value
added activities and variation
– Empowers teams with the best tools to leverage talent
and deliver accelerated business results
S ix
Six
ean
LLean S igm
Sigma a

LEAN SIX SIGMA EVOLUTION
Standard
Standard Model
Model Jidoka
Jidoka
Interchangeable
Interchangeable Costing
Costing Variety
Variety S.
S.Toyoda
Toyoda
Parts
Parts
Eli
EliWhitney
Whitney
Mass Just
Just––In
In––Time
Mass/ /Batch
Batch Time
K. Toyoda
K. Toyoda
Alfred
AlfredP.
P.Sloan
Sloan
Assembly
Assembly
Line
Line Toyota
Mass Toyota
MassProduction
Production Production
Waste Production
Waste System
Henry System
Elimination
Elimination HenryFord
Ford T.
T.Ohno
Ohno
System
System
Supermarket
Supermarket Lean
Synchronization Systems
Systems Employee
Synchronization Employee Six Sigma
Partnership
Partnership
Total
TotalQuality
Quality Drucker
Drucker
Time E.
E.Deming,
Time&&Motion
Motion Deming,
Division et
etal
DivisionofofLabor
Labor al
F.
F.Taylor
Taylor Six
SixSigma
Sigma
SQC
SQC Motorola
Motorola
Shewhart
Shewhart
Western
WesternElectric
Electric
DOE
DOE
Taguchi
Taguchiet
etal
al

LEAN SIX SIGMA PRINCIPLES
1. Specify value in the eyes of the customer
2. Identify the value stream and eliminate

waste / variation
3. Make value flow at the pull of the customer
4. Involve, align and empower employees
5. Continuously improve knowledge in pursuit

of perfection

LEAN SIX SIGMA IMPACT ZONES
Customer
Better, Faster, Lower Cost
Capability Culture
Tools and Values,
Methodology Leadership,
The Way We Work

LEAN SIX SIGMA –
“SAME OLD THING” or SOMETHING
DIFFERENT?
Zero
Defects
TQM
CPI
Quality
Circle
Re-
Engi-
neering
• How many quality initiative funerals have you

witnessed?
• How many more can our people handle?
• How do we prevent the Lean Six Sigma initiative

from the same fate as all the others?
Copyright Air Academy Associates 1 - 10
KEY ELEMENTS OF LEAN SIX SIGMA
… AS A BUSINESS STRATEGY
• Executive Ownership
• Leadership Alignment and Communication
• Infrastructure (Roles and Responsibilities)

– Champions (Sponsors)
– Master Black Belts, Black Belts, Green Belts
• Selecting the Right Projects
• Training (the Right Tools)
• Motivation
• Financial Accountability
• Information Technology Involvement
• Sales and Marketing
• Supply Chain

LEAN SIX SIGMA INFRASTRUCTURE
• Lean Six Sigma Leadership Team
(focused on business results)

• Champions (Sponsors, Mentors)
PROJECTS
• Master Black Belts (Masters)
• Black Belts (Experts)
• Green Belts (Specialists)
• Team Members

ROLES AND RESPONSIBILITIES
Lean Six Sigma Leadership Team:
• Trained in Lean Six Sigma
• Develop a Lean Six Sigma Master Plan to accelerate

Lean Six Sigma performance
• Work with management sponsors to identify Lean Six

Sigma Black Belt candidates
• Develop a focused curriculum
• Schedule training
• Serve as mentors for Lean Six Sigma Black Belt

candidates
• Identify Champions (Sponsors, Mentors)
• Determine certification requirements
• Work with management sponsors on reward system and

propagating success stories
• Determine annual re-certification requirements and

duration for Black Belt / Green Belt status
• Develop a Lean Six Sigma Network to enhance

communications
• Review and improve the Lean Six Sigma Process
• Monitor project selection and progress

COMPARISON OF ROLES
Green Belt Black Belt Master Black Belt Champion
(Specialist) (Expert) (Master) (Sponsor, Mentor)
• well respected • well respected • technically excellent in • senior manager
• proficient in basic and • master of basic and the knowledge and • respected leader and
advanced tools advanced tools application of Lean and mentor of business
• able to turn data into • able to turn data into statistical tools issues
PROFILE information information • excellent communicator • strong proponent of Lean

• respected Lean Six Six Sigma who asks the
Sigma role model at all right questions
levels of the company • serves as a business
mentor for Black Belts
• leads important • leads strategic, high • technical mentor for Lean • provide resources and
process improve- impact process Six Sigma Black Belts strong leadership for
ment teams improvement projects and Green Belts projects
• leads, trains and • change agent • internal consultant and • inspires a shared vision
ROLE coaches on tools • teaches and mentors trainer of Lean Six Sigma • establishes plan and
and analysis cross-functional team • generates breakthrough creates infrastructure
• assists Black Belts members and Green Belts thinking for improving • develops metrics
• typically part-time on • full-time project leader Lean Six Sigma process • converts gains into $
a project • converts gains into $ • may serve on Lean Six • may serve on Lean Six
Sigma Leadership Team Sigma Leadership Team
• two 1-week sessions • five or six 1-week • 1-2 weeks of Train the • 3 days of Champion
with one month in sessions with three to Trainer / plus additional training
between to apply six weeks in between to Master Black Belt • Lean Six Sigma
TRAINING
• project review in 2nd apply training beyond Black development and
session • project review in every Belt training implementation plan
session • at least one for every 5
• approximately 10% of the Black Belts
NUMBERS • at least 5 for each • at least 2% of number of
Black Belt number of Black Belts
1 - 14
(ROT) employees
CHARACTERISTICS OF
LEAN SIX SIGMA PROFESSIONALS
(Champions, Masters, Black Belts, and Green Belts)
• Highly respected by superiors, peers and subordinates

• Understands the “big picture” of the business
• Emphasizes results and bottom line
• Speaks the language of management (money, time, organizational
dynamics, etc.)
• Committed to do whatever it takes to excel
• Sponsored by a V.P., Director of Business Unit Manager
• Expert in specific field
• Excellent written and verbal communicator
• Inspires others to excel
• Challenges others to be creative
• Capable of consulting, mentoring, and coaching
• Leads change by challenging conventional wisdom, developing and
applying new methodologies, and creating innovative strategies
• Creative, critical, out-of-the box thinker
• Allows room for failures and mistakes with plan to recover
• Accepts responsibility for choices
• Responds well to criticism
• Encourages commitment, dedication and teamwork
• Unites a team to a core purpose
• Able to communicate all sides of an issue
• Solicits diverse ideas and viewpoints
• Empathizes
• Works on win-win solutions
• Disagrees tactfully, does not overreact
• Acts decisively under pressure
• Anticipates and confronts problems early and corrects causes
• Effectively identifies priorities from a business standpoint
• Manages limited resources in a highly efficient and effective manner
• Careful not to overtask any team member
• Understands people limitations
• Displays a genuine concern about others
• More concerned about business success than personal gain
• Does not lord his/her expertise over others
• Recognized by his/her results instead of title
LEAN SIX SIGMA PROJECTS
• Vital Role
– Learn and accelerate use of tools and methods
– Focus on knowledge
– Linked to business goals, bottom-line improvement,
and customer satisfaction
• Project Selection Considerations

1. What is the Cost of Poor Quality (COPQ)?
2. What is the Cost of Doing Nothing (CODN)?
3. What is the timeliness/scope/likelihood of success
of the project?
4. Can the project be tracked from a finance point of
view?
5. Can the entire organization be excited about
undertaking this project?
6. How does this project fit within the big picture and
trends in the business?
7. Will the project help us increase market share
and/or increase profits and/or generate savings?
8. Are there clear accountability links to management?
9. Can the project be defined so that sub optimization
is not a problem?
10. Can project success be propagated throughout the
organization to accelerate other successes?

LEAN SIX SIGMA PROJECT MASTER
STRATEGY (DMAIC)
… the Roadmap
What is important to the
customer and the
business? Clearly
define the problem,
project, and team.
Define
Pr
oj thi
ec s
t s wa
uc y
ce
ss
…
Measure
Define the current

(“as-is”) process and
gather baseline data
to assess
performance.
Analyze
What needs
improvement?
Identify the root
cause(s) of problems.
I mprove
Make the changes

needed to improve the
process based on data
and analysis.
Control
Measure success. Plan for
holding the gains. Monitor
critical outputs to prevent
repeat problems.

DMAIC ROADMAP
Define Project
Define
Define Team, Stakeholders,
Voice of the Customer
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Analyze Narrow the Focus
Tollgate Review
Improve the Process;

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Monitor Performance;
Control
Measure Success
Plan for Holding the Gains
No
OK
Yes
Celebrate
DEFINE
Define
Identify opportunities and select project

based on business needs
Define project charter; problem statement, Measure

goals, scope, measures, sponsor
Identify cross-functional team; key

resources; stakeholders Analyze
Understand the voice of the customer and

ensure measurable CTC’s
Improve
Create a high level process map of
the process being studied
Control
Build Knowledge /Project Notebook
Review with project sponsor, team,

etc. and revise project charter if
necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go
Yes
MEASURE
Close Down Project

MEASURE
Define
Understand the process and create
detailed process maps
Measure
Plan for data collection; Establish or
refine metrics if necessary
Validate the measurement system and

Analyze
improve it as needed
Collect and present baseline data;

Measure current performance Improve
Review with project sponsor,

team, etc. and make adjustments
as necessary Control
Update Knowledge/Project
Notebook
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
ANALYZE Close Down Project

ANALYZE
Define
Determine the gap between current

performance and goal
Measure
MEASUREMENT METHOD MACHINE
Carburetor Transmission
Gage for
Ma
Ride Brakes Tire Rotation

Au
nu
Tire Press.
Tire Press.
to
al
Estimate of
Full Tank Maint. Sched. Tune-Ups Weight Tire Type
Ra
Re
Identify potential causes of variation

Front or
All
dia
gu
WX
Speedometer Rear Drive

lar
l
Engine Warm-
Type Fast Start Up Time Radio
Air Conditioner
GAS
MILEAGE
Driver Gas Type Highway or City
Training
A
Weather Conditions
# of Ice
Oil Type
Passengers Snow
Terrain
Pr
A
Rain
Hu
ec
Te
ipi
mi
s
Wi
mp
ain
tat
dit
e
t
Fla
nd
tud
era
ion
y
unt
Driver
tur
Alti
Mo
Fuel Additive
PEOPLE MATERIALS ENVIRONMENT
Acctng
Identify the non-value added activities,
HR
Legal
sources of waste, and defects and/or Analyze
Functional
Area
Benefits
problems
Frequency or
Financial Measure
Collect data and screen the list of

potential causes
Brown Red Yellow Tan Green
Improve
Find and prioritize the root

causes of the problem(s)
Control

etc. and make adjustments as
necessary
Update Knowledge/Project Notebook
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
IMPROVE Close Down Project
IMPROVE
Define
Identify improvement opportunities
Test and screen solutions Measure
170
Pilot solutions, measure results, and

160
150
140
130
120
110
100
90
80
validate improvement
1
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
Month Number
Analyze
Mistake proof the process; Update

process maps and develop/document
SOPs
Improve
necessary
Control
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
CONTROL Close Down Project

CONTROL
Define
UCL
CENTERLINE
Monitor performance and measure
success
LCL
Measure
Performance No
confirmed and Gap Analysis
project goals
achieved? Analyze
Yes
Determine control plan for

holding the gains Improve
Summarize and communicate key

learnings; recommend any future
actions or follow-up; validate business
impact Control

team, etc.
Toll Revise control plan or

Gate Review No Go
Charter follow-up project if
Go/No Go improvement not sufficient
?
Done
Celebrate

THE IMPORTANCE OF KNOWLEDGE
ZERO
DEFECTS ISO 9000
TQM
DEMING SPC DOE
BALDRIGE
CRITERIA LEAN
QFD
The Quality Improvement Puzzle:

“A Set of Disjointed Pieces?” or “Do the Pieces Fit Together?”
Which steps
have you Job Security
taken?
Long Term Success
Return On Investment
(Improved Bottom Line)
Quality Improvement
(Better, Faster, Lower Cost)
Knowledge
“Knowledge has become the key economic resource

and the dominant, if not the only, source
of comparative advantage.”
Peter Drucker

PROFOUND KNOWLEDGE
"Hard work and best efforts put forth without guidance of

profound knowledge may well be the root of our ruination.
There is no substitute for knowledge ....”
Deming
PROFOUND KNOWLEDGE
Ability to Ask and Answer

Profound Questions

14 QUESTIONS WE NEED TO ANSWER!
1. Which value stream are you supporting and who is the recipient of the value, i.e.,
who is the customer? Who is the value stream owner and who are the players or
DEFINE
team members? How well does the team work together?

2. Within the value stream, which process or processes have the highest priority for
improvement? Show me the data that led to this conclusion.
3. How is the process performed? How does the value flow? What activity is value
added and what is non-value added?
4. What are the process performance measures, i.e., how will we gauge if a
MEASURE
process is improving? Why did we choose those? How accurate and precise is
the measurement system? Show me the data.
5. What are the customer-driven requirements or specifications for all of the
performance measures? Are the process performance measures in control and
how capable is the process? Show me the data. What are the improvement
goals for the value stream or process performance measures?
6. What kinds of waste and cost of poor quality exist in the value stream or process
and what is the financial and/or customer impact? Show me the data.
7. What are all the sources of variability in the value stream or process and which of
those do we control? How do we control them and what is our method of
documenting and maintaining this control? Show me the data.
ANALYZE
8. Are any sources of waste or variability supplier-dependent? If so, what are they,
who are the suppliers, and how are we working together to eliminate waste and
variability? Show me the data.
9. What are the key input variables that affect the average and standard deviation of
the measures of performance? How do you know this? Show me the data.
10. What are the relationships between the measures of performance and the key
input variables? Do any of the key input variables interact? How do you know
for sure? Show me the data.
IMPROVE
11. What settings or values for the key input variables will optimize the measures of
performance? How do you know this? Show me the data.
12. For the optimal settings of the key input variables, what kind of variability still
exists in the performance measures? How do you know? Show me the data.
13. Have we implemented a process flow and control system to sustain the gains
CONTROL
and continuously improve the process? Show me the data.

14. How much improvement has the value stream or process shown in the past six
months? How much time and/or money have our efforts saved the company?
Show me the data.

PROJECT SPECIFIC QUESTIONS
… begin a Knowledge Notebook and add questions

(and answers) as you move forward with your
project.

COURSE PHILOSOPHY
Which steps
have you Job Security
taken? Long Term Success
Return On Investment
(Improved Bottom Line)
Quality Improvement
(Better, Faster, Lower Cost)
Knowledge
Questions Questions
Managers Managers
Need to Need to
Knowledge Answer Ask
KBM
Based
Management
(KBM)
Tools and Techniques
to Answer Questions
and Improve the
Scorecard
Knowledge Gained
KISS
LOW HIGH
Statistical Complexity

LEAN SIX SIGMA TRAINING:
INSTRUCTIONAL APPROACH
A Keep-It-Simple-Statistically (KISS) approach is used,
with the intention to avoid statistical complexity. Statistics is not
presented as an “end”, but rather the means to gaining
knowledge for making good decisions which are critical for
success. There are a variety of Lean Six Sigma tools and
techniques, and we will use the "Present/Practice/Apply/Review"
instructional strategy. That is, we will present a tool or method,
give you a chance to practice that tool in class, then have you
apply that tool to your project, and finally have you review the
results of the application to your project. A final report will be
written to document your success story and its impact to the
company’s bottom line.

MODULE 1 KEY TAKEAWAYS

MODULE 2
LEAN SIX SIGMA FUNDAMENTALS

LEAN SIX SIGMA FUNDAMENTALS
• IN THIS MODULE, WE WILL DISCUSS
– LEAN SIX SIGMA AND ITS KEY PRINCIPLES
– THE CONCEPTS OF VALUE AND VALUE

STREAM
– DEFINING PROCESSES USING IPO DIAGRAMS
– KEY TERMINOLOGY
• TYPES OF DATA
• DISTRIBUTION, MEAN, STANDARD DEVIATION
• MEASURES OF QUALITY (Cp, Cpk, FPY, RTY, sigma
capability)
– PF/CE/CNX/SOP
• THE FIRST LINE OF DEFENSE AGAINST VARIATION
– MEASURING AND UNDERSTANDING THE

COST OF POOR QUALITY / COST OF WASTE

WHAT IS LEAN SIX SIGMA?
X Metric Based on the Standard Deviation
X Vision
X Benchmark
X Philosophy
X Methodology and Set of Tools
X Aggressive (Stretch) Goals
X Value Stream Enabler and Enhancer

RECALL: THE FIVE LEAN SIX SIGMA
PRINCIPLES
1. Specify value _________________________
2. Identify the value stream and eliminate

_____________________
3. Make value flow _______________________
4. Involve, align and _____________________
5. Continuously __________________________

VALUE
..
value |’val-(,)yu , n
1. : relative worth, utility, or importance; degree of excellence
2. : the monetary worth of something : marketable price
- Webster’s New Collegiate Dictionary
• Value is always specified in the eyes of the customer
• Goal is to provide the customer defect-free products and

services that are
– what they want
– when they want it
– at the desired price

WHAT IS A VALUE STREAM?
• The VALUE STREAM is the entire set of activities

performed to transform the products and services into
what is required by the customer.
• It includes all of the actions (both value added and non-

value added) required to bring a product or service from
concept to raw materials all the way to the customer.
• It looks at both product flow and information flow.
• It is used to identify key areas for improvement – all Lean

Six Sigma projects will target some value stream or some
process or processes within a value stream.
The VALUE STREAM
Suppliers
Suppliers Design
Design Procure
Procure Make
Make Sell
Sell Customers
Customers
A Primary Focus is TIME,
Product and / or Service Flow
Information Flow: Quickly In All Directions

VALUE STREAM DRILL DOWN
Total Value
Stream
External External Internal Your Ware- Distributor Retailer Customer
Supplier Supplier Supplier Facility houses
Value Stream in Facility
Receiving Testing Component Final Final Packaging Shipping

Assembly Assembly Testing
Final Assembly Process Step
X X
Eliminate Non-value Added Steps
Stabilize Remaining Steps by

Reducing Variability
Typical Process Improvement Efforts Begin on a Single Process or

Process Step Rather Than The Larger Value Stream

DEFINITION OF A PROCESS
INPUTS OUTPUTS
People
Material Perform a service
Equipment
PROCESS
Policies Produce a product

A blending of
Procedures
inputs to
Methods achieve the
desired outputs Complete a task
Environment
VALUE STREAM
(also known as a SIPOC model)

EXAMPLE OF A TRANSACTIONAL PROCESS
(within a financial services value stream)
Customer
om
e ry
Inc H isto
dit
Cr e
U/W accuracy
Broker/Correspondent
Application forms
r
pe -mail Fax
Pa E
Criteria
Procedures Underwriting
Regulations Process Cycle Time
Policies
Software
Underwriter
e /
i enc ing e w ny
n
er rai Tim mpa
p T
Ex co Cost / per
Property application
e n
ue
Ag tio
l
a
Va
c
Lo
Application
Under-
writing Approval
Fund
Service

EXAMPLE OF A TRANSACTIONAL PROCESS
(within a software value stream)
Coding SOPs
Cost per Line of Code
Programming Language
Experience of Programmer
Software Time to Code (Lines per day)
Coding Process
Design Documentation Errors Detected in
Internal Testing
Module Length
Time to Fix Errors
Module Complexity
Str of P
Pro cisio
Da Va
Errors Detected in the Field

(#
(Deount)
I/O
uc ath
ta riab
C
c ed n
tur s)
(# les
ur a
al
of
l
Analysis
Design
Phase
Phase Coding
Phase Testing
Phase Operational
Phase

THE ANATOMY OF A PROCESS
(Constructing an IPO Diagram)
PROCESS
Describe the
process Keep it
short, using 2-10
words as a rule of
thumb. A blending of
inputs to
achieve the
desired outputs
Some examples of processes are:
Purchasing Payroll
Accounts Receivable Help Desk
Accounts Payable Shipping
Hiring Scheduling
Training Web Site Creating
Documentation Management of Lean Six Sigma
Advertising Database Management
Drug Testing Personnel Placement

OUTPUTS FROM THE PROCESS
(Constructing an IPO Diagram) cont.
OUTPUTS
Perform a service
Describe the PROCESS
measurable outcomes
of the process. These
should be measures
related to process Produce a product
performance, speed, A blending of
cost, etc. inputs to
achieve the
desired outputs Complete a task
The outputs should be measurable outcomes from the process

that will help us determine (i.e., measure) if a process is
improving. These are: "Critical to Customer (CTC)" or "Critical
to Quality (CTQ)" outcomes that are measurable.
Some examples of outputs are:
# items misfiled Loans past due

% on time shipments Orientation cost
Time to return calls Stock price
Errors per document $ saved
Processing time Wait time
Calls per day % on time arrival
Cost per employee % Market share
Customer wait time # of returns

INPUTS TO THE PROCESS
(Constructing an IPO Diagram) cont.
INPUTS
People
Material PROCESS
Describe the process
inputs. These are the
Equipment sources of variation
that impact the output.
Policies A blending of
inputs to
Procedures achieve the
desired outputs
Methods
The inputs to a process are items that affect the outcome of the
process. You may or may not have control over these items.
Some examples of inputs are:
# of People Advertising time slot

Payment method Hours training
# days training Years experience
# of phone calls Market size
Type network Monthly revenue
Day of week Appointments per hour
Time of day Size of monitor

DIFFERENT LEVELS OF IPOs
Satellite View
People
Market Share
Technology
Corporate Profit
Policies
Mgmt. Methods Stock Price
Aircraft View
People Yield / FPY
Suppliers Company Profit

Business Unit
Division or Inventory
Materials
Directorate
Turnover in Personnel
Methods
In the Trenches
People
Methods Cost
Materials
Process Variables Performance
My Process
Environment
Time
Measurement

Pricing Policy
Sales Volume Alignment of Marketing
Manpower & Product Development
# of Defects Actual Market Share
Training Product Performance
# of Contract Errors
Incentive System Advertising
# of Lost Sales Ti $
m Sp Ki
nd % of Products that Attain
in s
Competitor Intelligence # of Sales Oriented g en
tO Expected Market Share
Sales n Marketing
Customer Complaints
Distribution Process Process
Contract Completion
Cycle Time Targeted Market
Follow-Up Policy G Ed In Ag Time Until Maximum
en u co e
Profit Margin de ca m Market Share is Reached
Std. Oper. Sales Proc. r tio e
n
Market Share Le

ve
Product Line Intelligence l
Sales Person Uniformity ROI from Marketing
Customer Relationship Sales Staff Endeavor
Sales Identification of # Tr
of ai
Payment Policy Customer Needs ni
(Sales & Marketing) ng (Sales & Marketing)
Errors in Hiring
# of people
Recruitment Method
# of Errors
Payment method
Time to Hire
Job Descriptions Hiring
Payroll
Tax Tables Process Deposit Dates Process
Cost per Hire
Benefit Plans
Hourly Rate
Tax Amount
Orientation Cost
Vacation Policy Background Checks
(HR) (HR)
2 - 14
Purchase Order (PO)
Cycle Time Type of Raw Material
Payment Policy
Mold Temperature
Sole Source vs. Bid FPY Thickness of Molded Part
Holding Pressure
Bid Evaluation Purchasing PO Cost Manufacturing
Process Holding Time Injection
Molded Parts % Shrinkage from Mold Size
Negotiations Time Until Delivery Process
Gate Size
JIT vs. Larger Quantities # of Errors Screw Speed

Number of Defective Parts
Percent Regrind
Shipping Method Inventory Levels
(Procurement) Moisture Content

(Manufacturing)
Coding SOPs
Method Programming Language Cost per Line of Code

Experience of Time to Code (Lines per
Programmer day)
Sequencing of Steps Time Software
Assembly Coding Errors Detected in
Design Documentation
Process Process Internal Testing
Module Length
Tool Type Time to Fix Errors
# of Defects
Module Complexity
Pr D Errors Detected
at (# Stru
oc
ed a of ct in the Field
Team Structure (# Pa ura
ur of
al I th l
(D /O s)
ec Va
is
io riab
n le (Software)
(Manufacturing) C s)
ou
nt
)
2 - 15
Monitor Size
Mouse Type
Feed Rate # of Sorted Errors # Pages Per Day
Type Keyboard
Automatic Desktop
Mail Sorter Years Experience
Speed Cycle Time Publishing
Process Process Defects Per Page
(R&D)
Computer Speed
Proportion of Amount of RAM

Gate Height Damaged Mail
Defects Per Document
Software Used
Type Computer
(Design)
Type notification # of Wire Turns

# of days
Automated or
Manual Type of Wire Chip Inductor
Accounts
Design
Receivable % defaulted Inductance
Process
Wire Insulation Level
Net Policy
$ outstanding Wire Gauge

Methods of payment
accepted
(Accounting) (Design)
2 - 16
Type of Training
Instructional Materials # of Calls
# of Students Ratio of Abandoned
Like (Course Evaluation) Type of Call Calls
Site of Training
# of Operators
Accountability of Training
Student & Trainer Process Learn (Pre vs Post Test) Help Desk Time per Call
Operator Call Center
Student Background Kn Tr Ex
o w
ai
ni pe
Amount of Time for le ng rie
Use (# of Applications) dg nc
Training e e Cost per Call
of
Pr
Kind of Trainers od

uc
Individual vs Team t Time Until Final
ROI ($) Resolution
Has Management Been Information on Failures
Trained?
(Training or H.R.)
(Service)
Method for Obtaining

Bill Information Time to Complete Type Product Time to Deliver
a Bill
Data Entry Method

Billing Receiving
Process # of Errors per Bill PO Included Process
Type of Bill Holding Time
Billing Clerk Skill

Level Total Cost to Prepare and
Send a Bill
Value of Delivery % Errors Delivering
# of Billing Clerks
(Finance) (Shipping)
2 - 17
EXERCISE
Develop your own IPO diagram for some process that

is or will be an integral part of your Lean Six Sigma
project. Use the template below to identify at least 4
input variables and at least 2 output variables. More
arrows can be added if needed.

THE FUNDAMENTAL RELATIONSHIP
Xs Ys
Y=f(X)
• Pronounced “Y is a function of X”
• Ys are the outputs and Xs are the inputs – the process

outputs are a function of the process inputs – they
depend on the inputs
• One of the goals in Lean Six Sigma is to study the

relationship between the Xs (inputs) and the Ys
(outputs), identify which Xs are most critical, and
determine how to change or control those critical Xs to
achieve the desired output and satisfy the customer

VARIABLES AND ATTRIBUTE DATA
• Variables Data
– Also called continuous or measurement data
– Actual measured values are recorded
– Data can assume a range of values on a continuous
scale
– Examples:
• Dimension in thousandths of an inch
• Distance in feet
• Operating temperature in degrees Fahrenheit or Celsius
• Attribute Data
– No direct measurement is made
– The presence or absence of something is recorded
– We classify or count a feature or characteristic of
the product or process
– Examples:
• Number of nonconforming articles
• Invoice errors per week
• Test results (pass/fail, good/bad)
• Presence or Absence of a date code stamp
• We will discuss methods to deal with both types of data,

because we will likely experience both on Lean Six
Sigma projects
• Given the choice, which type of data do you think is

preferred and why?

VARIABLES AND ATTRIBUTE DATA
(Exercise)
For each of the following, identify if the data is variable (V) or

attribute (A):
____ 1. The shipping time is 8 hours

____ 2. There are bubbles in the paint on the hood of the car
____ 3. The payment was received on time
____ 4. The call was returned in 6.5 minutes
____ 5. There are 10 errors on the form
____ 6. The wait time for a customer service rep is 3 minutes
____ 7. The windshield has 2 cracks
____ 8. The weight of the ball is 9 ounces
____ 9. The fan speed is 45 rpm
____ 10. The fan is off

GRAPHICAL MEANING OF Y, S, and σ
(variables data)
y (measure of performance)
130 140 150 160 170
∑y i
Center of Process: y= i=1
n
∑ ( y − y)
2
i
Process Variance: s =
2 i =1
n −1
Process Standard Deviation: s = s2 ≈ σ

GRAPHICAL VIEW OF VARIATION AND
“SIX SIGMA” PERFORMANCE
Normal Distribution
Each unit of measure is a numerical value on a continuous scale.
VARIATION: COMMON AND SPECIAL CAUSES

PIECES VARY FROM EACH OTHER:
SIZE SIZE SIZE SIZE
BUT THEY FORM A PATTERN THAT, IF STABLE, IS CALLED A DISTRIBUTION:
SIZE SIZE SIZE

DISTRIBUTIONS CAN DIFFER IN:
LOCATION SPREAD SHAPE
SIZE SIZE SIZE

… OR ANY COMBINATION OF THESE.
-6σ -5σ -4σ -3σ -2σ -1σ 0 +1σ +2σ +3σ +4σ +5σ +6σ
68.27%
95.45%
99.73%
99.9937%
99.999943%
99.9999998%
Typical Areas Under the Normal Curve

SPECIFICATION LIMITS
Limits that are imposed by the customer (internal and

external). For example, an accounts payable activity might
set a 40 day maximum on the length of time to pay a bill for
the purpose of taking advantages of any discounts.
However, management (the internal customer) may impose
a lower specification limit of 30 days in order not to pay the
bill too early and thus impose cash flow problems on the
system.
The picture
of the
process
LSL = 30 days USL = 40 days
(internal customer) (external customer)
20 22 24 26 28 30 32 34 36 38 40 42 44
Any bills being paid more than 40 days out or less than 30
days would be classified as an error or defect in the billing
paying process.
Question: If the accounts payable distribution were as

shown above, with the LSL and USL as shown, what would
be the sigma level of this process performance measure?

GRAPHICAL VIEW OF VARIATION AND
DEFINITION OF SIGMA LEVEL
The Sigma Level of a process performance measure
compares the Voice of the Process with the Voice of the
Customer, and it is defined as follows:
Sigma Level is the number of standard deviations (or σ’s) that

fit between the center ( y ) of the process and the nearest
specification (or requirement) limit.
3σ Process Centered
• Process is WIDER
than the
Example 1: σlevel = 3
specifications,
causing waste and
LSL USL cost of poor quality
σ σ σ
Determined by Determined by
the customer the customer
-6σ -5σ -4σ -3σ -2σ -1σ 0 +1σ +2σ +3σ +4σ +5σ +6σ
WASTE WASTE
Example 2: σlevel = 6
6σ Process Centered
LSL USL
σσσσ σ σ • Process FITS well
within the
specifications, so
even if the process
shifts, the values fall
well within tolerances
-6σ -5σ -4σ -3σ -2σ -1σ 0 +1σ +2σ+3σ +4σ+5σ+6σ

A “SIX SIGMA” CAPABLE PROCESS
Sigma Capability is a measure of process capability. It is related
to the defect rate and the complexity of the process or product.
The most commonly used definition of a Six Sigma Capable

Process is:
3.4 defects per million (dpmo)
This handy reference table can be used to convert between

defects per million opportunities (dpmo) and the sigma capability
of a process
σ Capability DPMO Yield

2 308,537 69.1%
3 66,807 93.3%
4 6,210 99.4%
5 233 99.97%
6 3.4 99.99966%
Sigma Defects per Million First Pass

Capability Opportunities (rolled throughput) Yield
Notes:
Yield is the probability that whatever we are producing (manufactured part, PO,
paperwork, etc.) will pass through the entire process without rework and without
defects.
Opportunities are a measure of complexity of the product or process (for example,

the # of value-added features, steps, etc.)

PUTTING QUALITY LEVELS IN
PERSPECTIVE
• 99% is pretty high quality, right? Why then is “6 sigma”,

or 99.9996% yield the stretch goal?
• Think about it this way:
At 99% good At 99.9996% good

(3.8 sigma capability) (6 sigma capability)
Unsafe drinking water for Only one unsafe minute
almost 15 minutes each day every 7 months
5,000 incorrect surgical Only 1.7 incorrect

operations per week operations per week
200,000 wrong drug Only 68 wrong prescriptions

prescriptions each year per year
No electricity for 7 hours each One hour without electricity

month every 34 years
88 hours of computer system 2 minutes of computer

downtime per year system downtime per year

OTHER MEASURES OF CAPABILITY
(Variables Data)
Cp: Process Capability Potential
Cp = USL − LSL
6σ
Cp PPM
15.0
7.0
4.5 << 1PPB
3.0
2.0 .0018
1.67 < 1PPM
1.33 55
1.0 2,700
.9 6,900
.8 16,400
.67 45,500
.5 133,600
LCL LSL TARGET USL UCL
Cpk: Process Capability
⎛ USL − y y − LSL ⎞⎟
Cpk = min ⎜⎜ or
⎝ 3σ 3σ ⎟⎠
Formal Definition of a Six Sigma Capable Process

Cp = 2.0
Cpk = 1.5
dpm = 3.4
EXERCISE 1: RAPID-FIRE STATAPULT®
LAUNCHING
Objective
To fire the Statapult® and record the distance in inches for each of the launches. The
measured distance will be from the back base of the launcher to the point where the ball
hits the floor. Use the following grid to record the distances in the order in which they were
obtained (to the nearest inch). This will establish a process baseline.
Launch Number Distance (inches)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Rules of Engagement
(1) Every shot will be launched from a pull-back angle of 177°, with the peg position
in the second hole from the top. Each person on the team will perform an
equitable number of launches (or as close thereto as possible). “Launching”
means pulling back and releasing.
(2) There will be a time limit of 15 seconds between successive firings. Each group
will be required to police themselves. There will be a penalty associated with late
firings. Practice shots are not allowed.
(3) Plot the distances recorded above as points on the run chart form on the next
page. Connect adjacent points with straight lines, and record the longest distance
(Max) and the shortest distance (Min) and compute the Range = Max - Min.
Distance (inches)
180
176
172
168
164
160
156

152
148
144
140
136
132
128
124
120
116
112
108
104
100
96
92
88
84
80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Launch Number
2 - 30
PRIOR TO COLLECTING ANY DATA
1. ESTABLISH AN ENVIRONMENT SUITABLE FOR

ACTION.
– Management Responsiveness
– Employee Training and Motivation
– Plan for Documentation
2. DEFINE THE PROCESS.

– General Flow
– Specific Flow
3. DETERMINE CUSTOMER REQUIREMENTS.

– Survey
– Interview
– Cross Functional Teams
– QFD
4. DETERMINE THE MEASURES OF PERFORMANCE TO

BE MANAGED.
– The Customer’s Needs
– Current and Potential Problem Areas
– Correlation Between Characteristics
5. DETERMINE COPQ AND OPPORTUNITIES FOR

IMPROVEMENT.
6. MINIMIZE EXTRANEOUS VARIATION.

HOW TO REDUCE EXTRANEOUS VARIATION
AND TURN AN ART INTO A SCIENCE
(1) PROCESS FLOW (PF) or PROCESS MAP

(2) CUSTOMER DRIVEN CAUSE AND EFFECT (CE)
Output(s)
and How What Who
- - -
Specs - - -
- - -
- - -
- - -
- - -
- - -
C= Controlled (Constant) SOPs - - -
- - -
- - -
N= Noise - - -
X= Experimental
2 - 32
PROCESS FLOW DIAGRAM
A process flow diagram (sometimes called a flowchart) is a

visual representation of all the major steps in a process. It
helps us understand the process better, identify critical or
problem areas, and identify improvements. Non-value added
steps should be identified and removed. Each step should
be assessed with regard to both input and output variables.
In building a process flow diagram, we will use five standard
symbols.
This Symbol… Represents… Some Examples are…
Receive Trouble Report

Start/Stop
Machine Operable
Approve/Disapprove
Accept/Reject
Decision Point
Yes/No
Pass/Fail
Drop Off Travel Voucher

Activity
Open Access Panel
A
Connector (to
another page or B
part of the diagram)
Represents
direction of flow
Process Flow Diagram Symbols (KISS)

PROCESS FLOW DIAGRAM EXAMPLE
Patientcalls
Patient calls No Patienttold
Patient toldtoto
hospital IsIsan
anopen
open
holdfor
for1st
1st
hospital lineavailable?
available? hold
mainnumber
number line availableoperator
operator
main available
Yes
Patientdescribes
Patient describes
typeofofcare
type care
required
required
Operator
Operator
transferspatient
transfers patient
totoappropriate
appropriateoffice
office
Emergency
Emergency Family
Family AcuteMinor
Acute Minor
Room
Room Optometry
Optometry
Practice
Practice IllnessClinic
Illness Clinic
(E.R.)
(E.R.)
E.R.personnel
E.R. personnel Appointment
Appointment Appointment
Appointment Appointment
Appointment
consultpatient
consult patient made
made made
made made
made
Flow Chart: Hospital Appointment Process Flow

WEB SITE CHANGE FLOW CHART
Requested Change
Technical Return for more

Fail
Review information
Pass
Example Web
Page Creation
Management
Fail Rework
Review
Pass
Sample Integration with

Existing Web Site
Q.C. Fail Rework
Pass
Final Integration with

Existing Web Site
Q.C. Fail Rework
Pass
Complete

STAFFING PROCESS
Process Description: Process Customer: Customer Valid Requirements: Performance Measure:
STAFFING Hiring Manager Timely Requisition/Approval Staffing Q1: % of Late Filled & Open Positions
Position Higher Hiring Manager Human Resources HR Support Candidate

Step/Time Authority
SOURCING APPROVAL
Identify the Need
Create and Approve Job Requisition and Description
Post Job
Develop Hiring Strategy
Conduct Ext. No
Recruiting?
Yes
Submit
Conduct Ext. Recruiting Collect Candidate Responses Resume
Screen Paperwork
No Any Potential
Candidates?
Yes
Forward to Hiring Mgr.
Screen Paperwork
No Any Potential
Interviewees?
Yes
INTERVIEWING
Conduct and Document Interviews
Evaluate and Rank Interviewees
No Found Final
Candidate?
Yes
Set Terms & Conditions of Offer No

OFFER / START
Accept
Extend Offer Offer?
Yes
Initiate New Hire Process
Send Offer Package No

Issues
Re-
solved
No ?
Contact Candidate Successful?
Yes
Yes Candidate
Reports

VALUE
VALUE = f (performance, cost, time)
Value added steps are defined as steps that:

1. Improve performance, function, etc.
2. Decrease cost (lower price)
3. Deliver the product or service faster (lower cycle time)
VALUE ADDED NON-VALUE ADDED

CAUSE-AND-EFFECT DIAGRAM
CAUSES EFFECT
(Sources of Variation or Variables) (Response or Output Variable)
Procedures/Policies Environments Material
RESPONSE
Other Equipment People
Partition the Variables:

C=
N=
X=

PARTITIONING THE VARIABLES
C = Constant
• To hold a variable as constant as possible requires
controlling the variable via Mistake Proofing and SOPs
to eliminate errors and reduce variation.
• Controlling a variable or holding it as constant as
possible doesn't just happen. It must be engineered into
the process.
• Mistake Proofing: The process of eliminating conditions
(errors) that lead to variation in the CTCs and ultimately
cause defects.
N = Noise
• Noise variables are those that are not being controlled or
held as constant as possible.
• Mistake Proofing is needed to change an "N" variable to
a "C" variable.
X = Experimental
• These are key variables that can be controlled and held
constant at different levels or settings for the purpose of
determining the effect of this variable on the CTC.

MEASUREMENT
MEASUREMENT METHOD
METHOD MACHINE
MACHINE
M
Gage for Au an
Ride Brakes Tire Rotation t o
ua
Tire Press. l
Tire Press.

Estimate of
R R
ad Al eg
Front or ia l W ul
l X ar
Speedometer Engine Warm- Rear Drive
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type Pr
Passengers H ec Snow
Terrain
A
Te
B
m
um ip Rain
W id ita
s
in pe it y tio
d
e
ra
at
n
a in
tu
Fl
ud
nt
Driver re
u
l tit
A
Mo
Fuel Additive
PEOPLE
PEOPLE MATERIALS
MATERIALS ENVIRONMENT
ENVIRONMENT
2 - 40
Convenience Service Level of Appreciation
Did not Difficult to use Too many Keep asking
consolidate mistakes Don’t for the same
Discourteous
accounts understand information
or impersonal
Delays my needs

Locations not
convenient Not helpful Don’t provide
Not enough Don’t value me
Don’t value differential
choice of
me / Know me service
access
Poor hours
Low Customer
Range is Don’t trust Exploits Retention Rate
limited customers
Fees changed
Manner of
Too big Salesman gave
denial was
Rates not false expectations
Products not Not as good demeaning
competitive
competitive as competitor
Out of touch with Competitor
Took too long to
Product terms community approved
communicate
and conditions quickly
denial /
Fees and fines
acceptance
too high
Product Image Denied Credit
* As viewed by customer
2 - 41
LINKING PF/CE/CNX AND IPO TOGETHER
Process Flow
Step 1
(PF)
Step 2
No
Step 3 Step 4
Yes
Step 5
Step 6
Generating
Variables Associated with this Process
can be Done Using a Cause-and-Effect Diagram with
Different Sets of Variable Categories.
Two examples are shown here.
CE/CNX CE/CNX
Using Traditional 6 M's Using Process Steps
Materials Method Machine Step 1 Step 2 Step 3
X1 X3 N1 X1
C1 C2 N2
X2 C1
y y
N1 C2 C3
N3 C3 N3
N2 X2 X3
Manpower Measurement Mother Nature Step 4 Step 5 Step 6
C1 C2 C3
IPO
x1
x2 y
x3

N1 N2 N3
EXERCISE 2: RAPID-FIRE STATAPULT®
LAUNCHING
(after PF/CE/CNX/SOP)
Objective: Reduce Variation
1. Create an “as-is” process flow for “shooting the statapult®”

2. Complete a cause-and-effect diagram
3. Label inputs as C or N
4. Write simple, one-line SOPs for each C
5. Revise the process flow as necessary
6. Repeat the statapult® launches using the same Rules of Engagement from the
first exercise (pull back angle = 177°, time limit of 15 seconds between shots,
etc.)
7. Record the distances below and plot the points on the run chart form on the
next page. Connect adjacent points with straight lines, and record the longest
distance (Max) and the shortest distance (Min) and compute the Range = Max -
Min.
Launch Number Distance (inches)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

Distance (inches)
180
176
172
168
164
160
156

152
148
144
140
136
132
128
124
120
116
112
108
104
100
96
92
88
84
80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Launch Number
2 - 44
LEAN SIX SIGMA PHILOSOPHY:
WORLD CLASS QUALITY

BUSINESS PROFITABILITY
CUSTOMER VALUE
Providing a BETTER product or service,

FASTER, and
at a LOWER COST than our competition.
Variation and Waste … the enemies of customer satisfaction

“Always know the language of the enemy.”
Lean Six Sigma (kept simple):

“the relentless and ruthless pursuit of variation and
waste reduction”

WHY DO WE NEED LEAN SIX SIGMA?
Total Product or Service Price to Customers
Profit Profit Budget Constraints and

Profit Profit
Competition Drive a
Waste Lowered Price
Waste
(COPQ)
(COPQ)
Total
Total
Cost Profit Profit
Cost Profit Profit
to Theoretical
to Theoretical
Produce Costs
Produce Costs
i.e., Cost
or i.e., Cost
of Doing Waste COPQ
or of Doing Waste COPQ
Provide the Right (COPQ)
Provide the Right (COPQ)
Things* Theoretical
Things* Theoretical
Right Theoretical COSTS
Right Theoretical COSTS
the First COSTS
the First COSTS
Time
Time
**
0 **
a. b. c. d. e.
“The price of gaining knowledge is

nothing compared to the cost of
ignorance.”
Anonymous

THE COST OF POOR QUALITY
Where Does the Cost of Poor Quality

(Waste or Muda) Come From? …
Internal Failure Costs
External Failure Costs
Appraisal Costs
Prevention Costs
Lost Opportunity Costs

WHY “SIX” IN LEAN SIX SIGMA?
• There is a correlation between a company’s

COPQ and the σ rating of its key processes
COPQ (% of Sales)
40
35
30
25
20
15
10
5
2 3 4 5 6
Sigma Rating

COMPUTING COST OF POOR QUALITY
Instructions
Refer to pages 2-30 and 2-44 and convert the Run charts
shown on these pages to Histograms, using the 4-inch intervals as
class intervals. The Histogram bars can be extended to the right side
of the sheet, as the instructor will demonstrate. The student may
then choose any 12-inch range (3 consecutive 4-inch intervals) to be
the specification range, but choosing the 12-inch range that contains
the most points is preferable. Draw those spec limits on the
Histogram and complete the following table:
Before After
PF/CE PF/CE
(from p. 2-30) (from p. 2-44)
N= total number of observations =
D (Defectives) = number of observations outside

of spec =
R (Rework) = number of observations outside of

spec but still within 4 inches of the
spec limit =
S (Scrap) = number of observations outside of 4

inches beyond the spec limits =
COST COMPUTATION
Cost of Inspection (Time + Manpower)
= N • $5.00 =
Cost of Defect Disposition (Time + Manpower)
= D • $10.00 =
Cost of Rework (Time + Manpower + Material)
= R • $50.00 =
Cost of Scrap (Time + Manpower + Material
+ Removal)
= S • $100.00 =
Total Cost

MEASURES OF QUALITY
(Attribute Data)
• The yield of a particular process or step is often

used as a metric for assessing quality
• Many definitions of “yield”
• What’s the yield of the process below?
5 units are test/in- 1 unit acceptable

started spection (i.e., "pass")
4 units
rework
test/in- 3 units acceptable

spection (i.e., "pass")
1 unit
scrap

FIRST PASS YIELD (FPY) AND
ROLLED THROUGHPUT YIELD (RTY)
The number of units completely defect free

• First pass yield = (i.e., no rework or repair)
(FPY) The number of units started
• Why is first pass yield important?
• Rolled throughput yield (RTY) can be thought of

as the first pass yield of an entire process across
multiple process steps
– It is the proportion of units completing an
entire process (sequence of steps) without
ever having to be reworked or repaired
– RTY = (FPYstep1) • (FPYstep2) • (FPYstep3) • …
• Example:
step 1 step 2 step 3

RTY = ???
FPY = .94 FPY = .90 FPY = .95

HOW PROCESS CAPABILITY IMPACTS CYCLE TIME
AND RESOURCE ALLOCATION
Step X Step Y
LS US LS US
Desired
Desired
... End State
EndState

Start No Defect No Defect
Defect Defect Defect Defect
Test
Test Test
Test
Analyze
Analyze Analyze
Analyze
* *
Fix
Fix Fix
Fix
Every Time a Defect is Created During a Process (Step), it Takes Additional Cycle Time to Test,
Analyze, and Fix.
2 - 52
* These Non-Value Added Activities Typically Require Additional Floor Space, Capital Equipment,
Material, and People.
EXERCISE: ILLUSTRATING FIRST PASS YIELD
(Rolled Throughput Yield) (Round 1)
Directions:
1. Each team will be given a deck of cards (52).
2. Each team will have 2 operators, each of whom will drop one
card at a time onto a target area.
Start Target Target Finished
Product to
Supplier Area # 1 Area # 2 Customer
Station 1 Station 2
3. The method of drop will be to hold the card at arms length over
the target using two fingers on edge (as demonstrated by the
instructor).
4. After release, the card will fall and either be completely in the
target area or not. Only those cards that fall completely within
the target area may move to the next station.
5. Objective: to deliver 20 completed products or units to
the customer.
6. Team Personnel: 2 Operators
2 Inspectors
1 Material Supplier
1 Timer
1 Foreman
7. Data To Track:
A. Total number of good units moved on to the next station (or
customer). Inspector at each station should record this.
B. Total number of drops at each station. Inspector should
keep track of this, or a separate auditor could perform this
function.
C. Total time from first drop to completion of customer order
(20 good units).
D. Total number of cards of the 52 used in the exercise (52 -
whatever supplier has left over).

8. Compute:
A = total # of good units B = total # of drops at
moved on each step
A1 = B1 =
A2 = B2 =
Stepwise yield: Y1 = A1/B1 =
Y2 = A2/B2 =
First Pass Yield (Rolled Throughput Yield) = Y1 • Y2 =
Total Cost = ($5 • D) + ($2 • [B1 + B2]) =
Total Cost
Average Cost Per Unit produced = =
20
C
Average Cycle Time = =
20

Start Target Target Finished

Product to
Supplier Area # 1 Area # 2 Customer
Station 1 Station 2
Objective: Improve the process and deliver 20 completed

products or units to the customer.
Directions:
1. As a team, brainstorm opportunities for improvement.
2. Have a tollgate review with your customer (instructor) to discuss
your improvement ideas and get approval.
3. Train the team on the new process and repeat the exercise at
the instructor’s signal.
4. Track the same data as in round 1 and report your first pass
yield, average cycle time per unit produced, and average cost
per unit produced.
Constraints:
1. Cards must still be dropped one at a time.
2. You cannot eliminate one of the steps in the process (i.e., must
still use two target areas)
3. You may not damage the cards. (no bending, tearing, etc.)
4. You must still hold the card at arms length over the target.

8. Compute:
A = total # of good units B = total # of drops at
moved on each step
A1 = B1 =
A2 = B2 =
Stepwise yield: Y1 = A1/B1 =
Y2 = A2/B2 =
First Pass Yield (Rolled Throughput Yield) = Y1 • Y2 =
Total Cost = ($5 • D) + ($2 • [B1 + B2]) =
Total Cost
Average Cost Per Unit produced = =
20
C
Average Cycle Time = =
20

RELATIONSHIP OF Cp, Cpk, Yield*,
Complexity, PPM AND σcapability
Standard Deviations from Mean

OVERALL YIELD vs SIGMA
100K
(Distribution Shifted 1.5σ) Cp = 1
Cpk = 0.5
# of Parts
(Steps) ±3σ ±4σ ±5σ ±6σ Centered
±1.5 Sigma Shift
10K
1 93.32% 99.379% 99.9767% 99.99966% Cp = 1.33
Cpk = 0.83
7 61.63 95.733 99.839 99.9976 Cp = Cpk = 1
10 50.08 93.96 99.768 99.9966 1K
20 25.08 88.29 99.536 99.9932
40 6.29 77.94 99.074 99.9864
60 1.58 68.81 98.614 99.9796
100 Cp = 1.67
80 0.40 60.75 98.156 99.9728 PPM Per Part (Step)
Cpk = 1.17
100 0.10 53.64 97.70 99.966
150 --- 39.38 96.61 99.949 Cp = Cpk = 1.33
200 --- 28.77 95.45 99.932 10

300 --- 15.43 93.26 99.898
400 --- 8.28 91.11 99.864 Cp = 2
Cpk = 1.5
500 --- 4.44 89.02 99.830 1
600 --- 2.38 86.97 99.796
Cp = Cpk = 1.67
700 --- 1.28 84.97 99.762
800 --- 0.69 83.02 99.729
900 --- 0.37 81.11 99.695 0.1
1000 --- 0.20 79.24 99.661
1200 --- 0.06 75.88 99.593
3000 --- --- 50.15 98.985 0.01
17000 --- --- 1.91 94.384
38000 --- --- 0.01 87.880
Cp = Cpk = 2
70000 ---USE FOR --- --- 78.820
BENCHMARKING 0.001
150000 --- --- --- 60.000
2 3 4 5 6 7
Design Specification Width (Sigma) = 3 * Cp
Source: SIX SIGMA RESEARCH INSTITUTE

Motorola University
Motorola Inc.
*
Yield is defined here as the percentage of all products, services or transactions
in which all of the parts are defect-free or all steps are performed perfectly, that
is, without error.
WHY LEAN “SIX” SIGMA?
• Worldwide benchmarking (1986-87) found the “best

in class” to have 6σ quality
• A 4σ company will spend >10% of revenue on

internal and external repair
• A 6σ company will spend <1% of revenue on internal

and external repair
Motorola
Defects / M
100K
DEFECT REDUCTION
66810
10K
Average 6210
Company
1K
233
100 >> 100x
10
“Best in Class” 3.4
1
2 3 4 5 6 7
Sigma Capability
Note: The ± 1.5 sigma shift is included.

HOW DO WE ACHIEVE
LEAN SIX SIGMA?
• Use the DMAIC Roadmap for gaining knowledge

needed to make the process better, faster, and/or
lower cost
• Keep focused on the 5 key Lean Six Sigma

principles:
1. Specify value in the eyes of the customer
2. Identify the value stream and eliminate waste /

variation
3. Make value flow at the pull of the customer
4. Involve, align and empower employees
5. Continuously improve knowledge in pursuit of

perfection

SUMMARY OF TOOLS FOR QUALITY
IMPROVEMENT
Process Flow Diagram (Sec. 2.2.1) Cause & Effect Diagram with CNX
(Sec. 2.2.2)
Input
Input
Manpower
Manpower Materials
Materials
No
Decision?
Decision? Causes or
Sources of
Variation Effect or
Effect or
Yes Response
Response
ProcessStep
Process Step
Output
Output
Methods
Methods Machines
Machines
• a graphic representation of the sequence of

steps in a process
• a structured approach to identify root
• used to define and understand the complete causes of a given effect
process
• used with brainstorming to identify
• helps identify non-value added steps and variables which cause a particular
locate problem areas and bottlenecks outcome or response to vary
• suggests data collection points in the process • identifies variables to be held constant
(C), noise variables (N), and critical
• identifies loops or repeated sequences of
variables for possible experimentation (X)
steps
Pareto Diagram (Sec. 2.2.4) Histogram (Sec. 2.2.5)

Frequency
Frequency
LSL USL
y
2 4 6 8 10 12 14 16 18
• a bar chart for non-numerical categorical • a bar chart for numerical categories; bars
descriptors whose bars are in descending need not be in descending order
order
• shape provides form of distribution of data
• used to identify and separate the most
• central tendency and variability are easily
frequently occurring categories from the
estimated
less important categories
• spec limits can be superimposed to
• may be used in a stratified or nested
estimate process capability
fashion to zero in on precise problem area

IMPROVEMENT (cont.)
Failure Mode and Effect Analysis Run Chart (Sec. 2.2.7)

(FMEA) (Sec. 2.3)
Component Failure Mode Failure Effect S Cause O D R "Fool Proof" Plan
Absenteeism
S = Severity of Failure
1: Low 3: Moderate 5: High
O = Probability of Occurrence
D = Probability of Escaped Detection Time
R = S•O•D = Risk estimate
• used to identify and evaluate potential • plot of data over time
failure modes in order to prioritize and • used to record and display trends in data
“fool proof” the critical ones over time
• performed on both products and • detects meaningful changes in a process
processes
Control Chart (Sec. 2.2.8, Chap. 9) Scatter Diagram (Sec. 2.2.9)

mpg
UCL
CENTERLINE
LCL
Vehicle Weight
• used to identify when special causes are • shows the relationship between two
affecting the process average or variables (positive, negative, none)
variation
• provides visual estimate of correlation
• separates natural variation from special coefficient (strength of a linear
causes of variation relationship)
• monitors process performance over time • allows “eyeball” fit of the data
• helps reduce process variability • identifies outlying points (outliers)
• helps determine what kind of action
should be taken

IMPROVEMENT (cont.)
Regression Modeling (Chap. 7) Design of Experiments (Chap. 8)
Run A B AB y1 y2 y3 ... y s
y
1 - - +
2 - + -
20 ŷ = 18 − .65x
3 + - -
15 4 + + +
10
ΔA Δ Δ
ŷ = y + A + B B + AB A ⋅ B
5 2 2 2
δA δ δ
5 10 15 20
x ŝ = s +A + B B + AB A ⋅ B
2 2 2
• systematizes an efficient way to collect
data
• develops a model relating a response
variable to input variable(s) • used to identify important factors affecting
the process
• allows prediction of responses at levels
other than where data was collected • identifies important interactions between
process variables
• quantifies the strength of the model
• can be used to optimize a process
• identifies outliers
• used to build a model relating how inputs
affect the outputs
Gage Capability Study (Sec. 9.13) Quality Function Deployment (Chap. 11)
UCL
LCL
• determines proportion of total variability in • identifies customer requirements

the process attributed to the measurement
• uses customer requirements and
system
benchmark data to identify critical
• estimates precision and accuracy of the parameters
measuring system
• weighting scheme rates the importance of
• repeatability and reproducibility are critical factors
studied

IN SUMMARY… LEAN SIX SIGMA MEANS
KEEPING THE RIGHT SCORE
Be Wiser
Products Services Timelines Budgets

Shipped Rendered
An Incomplete Scorecard
Performance Schedule Cost

(Better) (Faster) (Lower Cost)
– Operational Cost
– Advertising Cost
– Stock price – Milestones
– Materials Cost
Organization – Market share – Timelines
– Taxes
– Organizational Climate – Reaction Time
– Budgets
– Profits
– Development Cost
– Meeting Customer Needs – Design Cost
– Development Time
– Customer Complaints – Product Price
Product – Design Time
– Reliability – Lifecycle Cost
– Design Changes
– Availability – Cost of Poor Quality
– Profit Margin
– Defect Rate – Process Development Time

– Manufacturing Cost
Process – Cpk – Down Time
– Cost of Poor Quality
– Customer Complaints – Daily Production Rate
– Turnover Rate
– Product Rate
– Training Cost
– Defect Rate – Training Time
– Legal Costs
People – Safety Violation Rate – Rehabilitation Time
– Absentee Costs
– Absenteeism – Learning Curve Time
– Ergonomic Costs
– Suggestion Rate
– Benefit Costs
A Balanced Scorecard
“If you are not keeping score,

you are just practicing.”
Vince Lombardi
SUMMARIZING THE POWER OF
LEAN SIX SIGMA
X Provides common measurement and common

goals
X Encourages a common vision shared by all
X Promotes teamwork and esprit de corps
X Combines aggressive goals with a method and a set

of tools
X Requires the application of tools throughout entire

lifecycle of a product or service
X Correlates improved cycle time, reduced defects,

and lower cost
Better products and services

created faster
at lower cost
=
Return on Investment


MODULE 3
DEFINING THE PROJECT AND

MANAGING CHANGE
DEFINE PHASE
– THE DEFINE PHASE OF DMAIC
– SELECTING AND DEFINING A PROJECT
– ELEMENTS OF A PROJECT CHARTER
– IMPORTANCE OF GOOD PROJECT

MANAGEMENT AND TEAMWORK
– UNDERSTANDING CHANGE AND

INGREDIENTS FOR SUCCESSFUL
CHANGE

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
DEFINE
Define




Improve
Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go
Yes
MEASURE
Close Down Project

IDENTIFY OPPORTUNITIES AND
SELECT A PROJECT
• OPPORTUNITIES … WHERE TO LOOK

– Business needs; strategic goals
– Customer needs, issues, complaints
– Bottlenecks; problem areas
– Processes with long cycle times, high rework and/or
defect rates
– Previous project work
– Employee suggestions
– Other sources?
• CHARACTERISTICS OF GOOD PROJECTS:

– No known solution in advance
– Good data available (or able to be gathered)
– Aimed at making a process better, faster, and/or
lower cost
– Drives business profitability and/or enhances
customer value
– Clear, precise definition of the problem, project
goals and scope of the project
– Resources and support available for working on the
project

PROJECT SELECTION MATRIX
• A prioritization matrix can help when there are

many project ideas or it is difficult to decide
• Steps:
– List project opportunities
– Determine project selection criteria and importance
rating (on a scale from 1(low) to 10 (high))
– Rate the relationship between each project and
each criteria (0 = none, 1 = low, 3 = med, 9 = high)
– Calculate weighted sum by cross multiplication
• Example:
Example Project Selection Criteria
(Importance)
Doable in High Apply
2-3 Likelihood across
High ROI months of Success business
(9) ( 10 ) (8) (7)
Project Ideas Weighted Sum
Project #1 9 1 3 1 122
Project #2 9 9 3 3 216
Project #3 3 9 9 1 196
Project #4 1 3 1 9 110
Total: 644
Pareto Chart
250
200
Weighted Sum
150
100
50
0
Project #2 Project #3 Project #1 Project #4
Copyright Air Academy Associates Project

3-5
EXERCISE
• The SigmaZone Company has 3 candidate Lean Six
Sigma projects under consideration.
– Saturn
– Neptune
– Jupiter
• Given the following information, determine which project

you recommend pursuing by completing a project
prioritization matrix.
– To help meet business goals, it is desirable for

projects to have high ROI (over $200,000), a high
likelihood of success, strong linkage to at least 2 of
the 5 key strategic business initiatives, the ability to
get results quickly (complete in 3 months or less),
and resource availability.
– The most critical requirements are high ROI and

high likelihood of success.
– Linkage to business initiatives and ability to get

results quickly are moderately important
– Resource availability is of lower importance,

because the company feels it can effectively
address any issues with people or time.

EXERCISE (cont.)
• Results from discussion of each project being
considered:
– Saturn:
• moderate ROI (about $100,000)
• moderate likelihood of success (about 60%)
• moderate linkage to strategic business goals (1 of 5)
• quick to complete (< 3 months)
• lots of available resources
– Neptune:
• high ROI ($250,000)
• moderate likelihood of success (about 60%)
• high linkage to strategic business goals (3 of 5)
• moderately quick to complete (about 3-4 months)
• low availability of resources
– Jupiter:
• low ROI (<$50,000)
• high likelihood of success (>95%)
• no linkage to key, strategic business initiatives
• long to complete (>6 months)
• lots of available resources
• Complete the Project Selection matrix on the next page,

based on the supplied information

EXERCISE (Completed Matrix)
• What project would you recommend?
• Any comments or other observations?
Project Selection Criteria

(Importance)
Project Ideas Weighted Sum
Total:

DEFINING THE PROJECT
• Project details should be discussed and agreed to

with the Project Sponsor/Champion and ultimately
the team
• Project details should be summarized in a project

charter
• Key Elements of a Project Charter

– Description of product or process
– Problem statement
– Opportunity / Business Impact
– Project goals
– Project measures
– Project sponsor
Lean Six Sigma Project Charter
Project Name: Project Start Date:

Your name: Target Completion Date:
Division/Location:
Project Sponsor:
– Team 1. Product / Process

Describe the product or
process involved
Describe the current pain.

2. Problem Statement Why are you working on this
project?
Describe the importance to

3. Business Benefit the business. What is cost of
(Impact) doing nothing or the estimated
return?
List the specific goals of this

4. Project Goal(s)
project.
Primary Measure(s) Secondary Measure(s)
What are the primary and

secondary metrics? Note
5. Project Measure(s)
baseline data for each (if
available) and goal.
Who are the customers and

6. Customers and suppliers? What are their
Suppliers critical requirements or
concerns?
List your team members and

7. Project Team
their department/function/role.
For planning purposes,

8. Percent of Time
indicate what % of time you
Dedicated to Project
will devote to this project

PROBLEM STATEMENT
• Everyone on the team must agree on the problem
• Problem Statements should:

– Define what is wrong
– State the magnitude and/or impact of the
problem (the current pain)
– Address the “so what?” (what is the cost of
doing nothing or the business impact)
– Not specify a solution
• Example:
– “We have huge differences between our
forecasted and actual demands for software,
resulting in writeoffs of $1 million per quarter
in software worldwide. This is causing us to
miss our revenue projections for the quarter
and impacts profitability.”

PROBLEM STATEMENT (EXERCISE)
Consider the following problems statements. Identify
each as a good or poor example. If poor, explain why
and what is needed to improve it.
1. “Our customers are not happy with us.”
2. “Customers are complaining because our response

time is slow so we need to install the ABC software
package to help us speed up response time.”
3. “Customers are complaining because health care

claims are taking an average of 48 days to process.
We’re incurring $10,000 in overtime expenses each
month to keep up with the claims and our customer
satisfaction rating has slipped by 30% over the past
year.”

PROJECT GOALS
• Project Goals should:

– Be derived from the problem statement
– Define the desired improvement
– Make use of a verb, noun, number, and date
– Not be too narrowly or too widely focused
– Not specify a solution
• Example of Project Goals:

– Poor: Improve claims processing
– Poor: Install ABC software to speed up processing
of claims
– Better: Reduce the average cycle time for
processing claims by 50% by August 30th, in order
to improve customer satisfaction and reduce
overtime expenses.

PROJECT GOALS
Remember to make project goals:
Specific
Measurable
Achievable
Relevant
Timely

PROJECT MEASURES
• Used as indicators to measure success
• Directly related to project goals
• Green Belt and Black Belt projects may have both

primary and secondary measures
• Primary measures are the main focus of the

project and should be limited to one or at most two
• Secondary measures are additional things you

will measure and track as you work on your
project, although not your primary focus. Why?
Give an example.

PROJECT GOALS AND MEASURES
(Quick Check)
Consider the following scenario:
Gary Greenbelt has been assigned to improve a

shipping process. In particular, the big problem has
been with damaged shipments arriving at the customer
site. His department is receiving a lot of customer
complaints and are having to pay huge premium freight
charges to get replacement shipments back to the
customers quickly. Gary’s project goals, agreed upon
with his Sponsor are to:
• Reduce the percentage of damaged shipments by
60% by August 30th.
• Reduce the premium freight charges by at least
50% by August 30th.
What are Gary’s primary project measures?
What are possible secondary project measures?

PROJECT SCOPING
• Critical activity for project success. Must be

discussed with your Sponsor when defining the
project and goals.
• What’s the problem with a scope that is too

broad?
• What’s the problem with a scope that is too

narrow?
• When reviewing your project with your Sponsor,

here are some items to discuss and agree on:
– What resources are available to the team?
– What (if anything) is considered “out of bounds” for
the team?
– What (if any) constraints must the team work under?
– What percent of time will you have available to work
on the project?
“in scope”
“out of scope”

QUANTIFYING PROJECT BENEFITS
(Business Impact)
• Projects will result in hard benefits (direct impact

to the bottom line) and/or soft benefits
• Examples
– Cost Takeout / Cost Avoidance
• reduction in scrap
• reduction in raw materials
• reduction in labor or planned headcount
• freight savings
• interest expense reduction
• advertising and agency fee reduction
• warehouse space reduction
– Cash Flow
• reduction in inventory levels
• inventory carrying cost reduction
• reduction of past due receivables
– Growth
• increase in capacity/throughput
• increase in market share / sales revenue
– Reduction of Non-Value Added Activity

• reduced rework, waiting, administration, etc.
Q#
2,14
The financial impact of your project must be

verified with Finance

QUANTIFYING PROJECT BENEFITS
(Exercise)
• Individually (5 minutes)
– How will you measure the financial impact of your
project? Document some ideas that you have right
now for potential hard and/or soft savings.
• In small groups (15 minutes)

– Take turns sharing your ideas for capturing the
financial impact of your project and give each other
feedback.

THE PROJECT CHARTER
• The project charter summarizes the project
• Example charter format:

Lean Six Sigma Project Charter
Project Name: Project Start Date:

Your name: Target Completion Date:
Division/Location:
Project Sponsor:
Describe the product or

1. Product / Process
process involved
Describe the current pain.

2. Problem Statement Why are you working on this
project?
Describe the importance to

3. Business Benefit the business. What is cost of
(Impact) doing nothing or the estimated
return?
List the specific goals of this

4. Project Goal(s)
project.
Primary Measure(s) Secondary Measure(s)

What are the primary and
secondary metrics? Note
5. Project Measure(s)
baseline data for each (if
available) and goal.
Who are the customers and

6. Customers and suppliers? What are their
Suppliers critical requirements or
concerns?
List your team members and

7. Project Team
their department/function/role.
For planning purposes,

8. Percent of Time
indicate what % of time you
Dedicated to Project
will devote to this project

EXERCISE / HOMEWORK
• Individually (10 minutes)

– Consider your project
– Review (or create) your problem statement/purpose
and project goals. Apply the S.M.A.R.T. criteria to
your project goals and ensure that there is not a
solution specified in advance.
• In groups of two or three (15 minutes)

– Share your problem statement/purpose and project
goals with each other
– Give feedback to each other

TEAMWORK
• Successful projects require a team effort
• Core team should be made up of no more

than 4-8 members
• Effective Lean Six Sigma project teams
– are cross-functional and adjust

membership as needed
– recognize the value of listening, being

open-minded, and basing decisions on
facts and data
– involve everyone and have clearly defined

roles and expectations
– conduct effective meetings

TEAM ROLES
• Team Sponsor / Champion

• Ensures that the team has the resources and
support it needs to be successful
• Directly responsible and accountable for team results
• Team Leader
• Facilitates team meetings
• Keeps Sponsor informed on team progress and
resource needs
• Core Team Members

• Active participation in (weekly) meetings
• Provide technical input and assist with data collection
and analysis
• Complete action items
• Extended Team Members

• Active participation in sub-team meetings.
Provide technical input and assist with data collection and
analysis
Complete action items
• Review core team decisions and provide
recommendation and approval on final output of
team
• Examples: Finance, Customer Service, Sales, IT,
etc.

KICKOFF TEAM MEETING
• Purpose
• Typical Activities
– Validate the project charter
– Validate the problem statement
– Validate the project goals
– Establish team ground rules and code of
conduct

EXAMPLE CODE OF CONDUCT
(Ground Rules)
• All information will be communicated to all team

members through the meeting minutes and
emailed to all team members.
• All issues will be discussed and conflicts resolved

by consensus.
• One person talks at a time. No side

conversations.
• Check your job title at the door. There is no

inequality in the team
• Decisions are based on what’s right, not who’s

right.
• Contribute with flexibility. No fixed positions.
• Decisions are based on data and analysis, not

opinions.

DEALING WITH SOME COMMON TEAM
PROBLEMS (EXERCISE)
In small groups, discuss your assigned rows in the table
below, and identify some strategies or techniques you
might use for dealing with the specific problem or situation.
Problem (symptom) Strategy for Dealing With

Disagreements (emotions run high;
there are many legitimate differences
of opinions)
Power and Authority (managers or

others discourage discussion; team
members don’t feel free to say what
they think)
Overbearing “experts” (experts

discourage discussion about their
areas of expertise; technical jargon
and complex principles referred to
without explanation)
Lack of Focus (several topics
discussed at the same time; lots of
distractions; little progress is made)
Too much agreement (no one

disagrees; no alternatives are offered;
different perspectives are quickly
dismissed)
Uneven participation (some team

members talk too much, while others
talk too little)
Lack of follow-through (tasks don’t

get done on time; people don’t do
their assignments between meetings)

HELPFUL HINTS AND “BEST
PRACTICES” FOR TEAMS
• Start each meeting with “PACER”

– Purpose
– Agenda
– Code of Conduct (ground rules)
– Expectations
– Roles and Responsibilities
• Utilize a “parking lot” for capturing ideas not directly

related to the topic under discussion or those issues
which you want to temporarily set aside
• Do an evaluation at the end of each meeting

– Plus/Delta
– Short Survey
• Capture action items and summarize/review them at the

conclusion of the meeting
• Use data to support beliefs and opinions
• Encourage questions and always check for

5 min Check In
understanding 3 min PACER
10 min Actions from
Last Mtg
15 min Subgroup Reports
• Involve everyone; emphasize participation 10 min
5 min
Plans
Review Actions
2 min Mtg. Evaluation

STAKEHOLDER ANALYSIS
(Mobilizing Commitment for the Project)
• Identify the key stakeholders. Stakeholders are
individuals or groups who are affected by or who can
affect your team’s work.
• Understand who is “for” and “against” the project
• Identify sources of resistance and anticipate issues and

concerns
– Technical (cost, skills, etc.)
– Political (authority, power, etc.)
– Cultural
• Develop strategies for gaining needed support
Sample Project Stakeholder Analysis / Planning Worksheet
Level of Acceptance
Key Issues /
Dept. 1 2 3 Strategy
Stakeholder Concerns
blocker neutral supporter

STAKEHOLDER ANALYSIS (cont’d)
• For each stakeholder, consider
– What do you need from this stakeholder?
– What are their interests?
– How can this stakeholder benefit from the project?
– How might this stakeholder be affected or

inconvenienced by the project?
– What possible resistance might I see from this

stakeholder? (technical, political, cultural)
– What are the best strategies for communicating with

and influencing this stakeholder?
Focus on Win/Win Strategies

STAKEHOLDER ANALYSIS
(Exercise / Homework)
• In teams
• Select one project from among the team members

to use for this exercise and select a facilitator
• Complete two or three lines of the stakeholder

analysis. (at least one of the lines should deal
with a blocker). Focus your discussion on the
issues/concerns and strategies you can use for
dealing with each stakeholder. (avoid using
names wherever possible)
• Be prepared to report out
Dept./ Level of Acceptance

Issues /
1 2 3 Strategy
Stakeholder Concerns
blocker neutral supporter

MANAGING YOUR PROJECT
• Project management is key to a successful Lean Six

Sigma project
• Make sure you and your team have a clear

understanding of:
– project goals (what will be a successful termination of the
project)
– start and end dates
– primary and secondary measures
– key players (team members, stakeholders, support
resources, Finance rep, etc.)
– resources (monetary, manpower, equipment, etc.)
• Have a Project Plan and Schedule

– Organize work to ensure that necessary tasks are
completed on time
– Utilize project milestones, Gantt charts, etc.
– Know your key deliverables. Use the DMAIC roadmap as a
guideline.

PROJECT PLAN AND SCHEDULE
• Helpful to use a work breakdown structure (WBS) to identify
tasks to be accomplished
– Keep it current. If there are changes to the project then
update the WBS
• For each task, estimate the time duration and assign resources
• Example:
Team Resources
ID TASK Duration Larry Chris Jane Fred
Assess the
1.0 measurement A
system capability
30
1.1 Brief the team I A I I
minutes
Select items to be 30
1.2 R A D
included in the study minutes
Determine the 15
1.3 I A
sample size minutes
Take data
1.4 2 days R I A
(measurements)
1.5 Analyze the data 1.5 days A I, R
Determine actions
1.6 0.5 days I A, S
to be taken
S = sign-off
R = review
D = data
I = input
A = accountable

GANTT CHARTS
• A chart using bars, lines, and other symbols to depict
multiple time-based activities on a horizontal time scale.
• The most common chart used in project management.
• Named after Henri Gantt, the inventor
DAY #
ID TASK 1 2 3 4 5 6 7
Assess the measurement

1.0
system capability
1.1 Brief the team
Select items to be
1.2
included in the study
Determine the sample
1.3
size
Take data
1.4
(measurements)
1.5 Analyze the data
Determine actions to be
1.6
taken
. .
. .
. .
Collect Data on Recent

5.2
Customer Complaints

GANTT CHARTS
• Can become overwhelming quickly due to the size
• Project management software, such as Microsoft

Project, is helpful
– Allows for easy manipulation of the chart, “rolling up” of
tasks, etc.
• Even if software is not available, simple Gantt charts

which are hand-sketched or done in Powerpoint can still
be useful for helping schedule and organize your project

CHANGE
• LEAN SIX SIGMA INVOLVES CHANGE
• LEAN SIX SIGMA CHALLENGES US TO:

– Think differently
– Work differently
– Ask questions and challenge the status quo
– Make decisions with facts and data
– Use new tools and methodologies
“Culture Eats Change for Breakfast”

-Jim Hunt*
*Jim Hunt is a change management consultant

and President of Change2, Inc.

TEAM EXERCISE
• On the next page, you will find several quotes

related to the topic of “change”
• As a team,
– Review the quotes and see if you can match

the quote with the famous person who said it.
The names of the authors, in no particular
order, are located on the page following the
quotes. Write the letter of your answer on the
lines before each quote.
– Select one quote that is particularly

meaningful to your team. Summarize it on a
flipchart by identifying a few key points related
to the quote. What does it have to do with
Lean Six Sigma? Why is it important? Be
prepared to discuss your thoughts.

QUOTES RELATED TO “CHANGE”
1. “Humans are allergic to change. They love to say, ‘We’ve

always done it this way’. I try to fight that. That’s why I have
a clock on my wall that runs counter clockwise.”
2. Any change, even a change for the better, is always

accompanied by drawbacks and discomforts.”
3. “The world we have created is a product of our thinking; it

cannot be changed without changing our thinking.”
4. “It is not the strongest of the species that survive, nor the
most intelligent, but the ones most responsive to change.”
5. “Progress is impossible without change, and those who

cannot change their minds cannot change anything.”
6. “Know what’s weird? Day by day, nothing seems to change,

but pretty soon…everything’s different.”
7. “Progress is a nice word. But change is its motivator, and

change has its enemies.”
8. “To improve is to change; to be perfect is to change often.”
9. “Failure is the opportunity to begin again more intelligently.”
10. “I get by with a little help from my friends.”
11. “But of the good leader, when his work is done, his aim
fulfilled, the people will say “We did this ourselves.”

FAMOUS AUTHORS OF
QUOTES RELATED TO “CHANGE”
A. Lao-Tzu
• Chinese philosopher
B. Henry Ford
• U.S. automobile industrialist
C. Arnold Bennett
• Novelist
D. Albert Einstein
• Scientist and one of the greatest thinkers of the 20th century
E. George Bernard Shaw

• Playwright
• Awarded Nobel Prize for Literature in 1925
F. Calvin
• From the Calvin and Hobbes cartoon strip, by Bill Watterson
G. Grace Hopper
• Pioneer computer scientist
• First woman to receive the prestigious National Medal of Technology in
1991.
H. Robert F. Kennedy
• United States Senator
• Brother of John F. Kennedy; assassinated in 1968
I. The Beatles
• Famous British Rock and Roll band
J. Charles Darwin
• British naturalist
• Known for his theory of evolutionary selection
K. Sir Winston Churchill

• Famous British politician and statesman
• Awarded the Nobel Prize for Literature in 1953

REACTIONS TO CHANGE
• What are some normal reactions to change?

– Denial - Anger - Bargaining - Depression -
Exploration - Acceptance
Emotional Level
(Intensity)
Transition Curve
High
• Acceptance
• Denial
• Anger
• Exploration
• Bargaining
• Depression
Low
Time
• Why is change difficult? What makes some resist

change?

REACTIONS TO CHANGE (cont’d)
• While change is often difficult, what makes some

openly embrace change?

READINESS FOR CHANGE
• The “Frontier” Model

KEY INGREDIENTS FOR
SUCCESSFUL CHANGE
DD + VF + FS > R *
DD = ___________________________
VF = ___________________________
FS = ___________________________
R = ___________________________
* Change Model used with permission from Kelly and Frank Petrock.

6 MAJOR POINTS A CHANGE AGENT
MUST REMEMBER
1. WIIFM?
2. Organization’s Status?
a. Company in Crisis
b. Market Leader
c. Middle-of-the-Pack
3. Organization’s Tradition / Experience with Change
4. Incremental vs Breakthrough Change
5. Show Me the Data!

- COPQ
- Trend Data
- Benchmarking
- Benefits of Change
6. Trumpet Our Successes (as a result of change)


MODULE 4
UNDERSTANDING THE
VOICE OF THE CUSTOMER (VOC) AND
DEFINING A PROCESS
goals
business
value
customer
UNDERSTANDING THE VOICE OF THE
CUSTOMER (VOC) AND DEFINING A
PROCESS
– THE DEFINE PHASE OF DMAIC (cont.)
– UNDERSTANDING THE VOICE OF

CUSTOMER AND INTRODUCTION TO
QUALITY FUNCTION DEPLOYMENT (QFD)
– DESCRIBING A PROCESS USING A SIPOC

DIAGRAM

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
DEFINE
Define




Improve
Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go
Yes
MEASURE
Close Down Project

CUSTOMER EXPERIENCE
• Think about a recent experience you had as a

consumer
• What comes to mind about this experience?
+ -

MANAGING THE EXPERIENCE
(Outcome vs. Process)
Customer Retention Grid*
Value
Added
OUTCOME
Satis-
factory
Fails
Dissatisfied Satisfied Dazzled
PROCESS
*Copyright Dr. Tom Connellan

THE VOICE OF THE CUSTOMER (VOC)
Suddenly, a heated exchange took place

between the king and the moat contractor.
Source: The Far Side

The Far Side Millennium Off-the-Wall Calendar 2000
Far Works, Inc.

DETERMINING WANTS AND NEEDS
Kano’s Model
CUSTOMER SATISFACTION
Very Satisfied
Exciting Quality
(Delighters) One-Dimensional
Quality
Did Not Fully DEGREE OF

Achieve Achieved ACHIEVEMENT
Expected Quality
(Must Be)
Very Unsatisfied
Expected Quality
• Not specifically requested, but assumed to be present
NEEDS • If present, customer neither satisfied nor dissatisfied
• If absent, customer very dissatisfied
One-dimensional Quality
• Specifically requested items; specifications
• Items from a typical survey (stated wants)
WANTS • If present, customer satisfied
• If absent, customer dissatisfied
Exciting Quality
• Unknown to the customer
• Most difficult to define and develop
DELIGHTERS
• If present, customer very satisfied
• If absent, customer neither satisfied or dissatisfied

SOME OF THE WAYS WE HEAR
THE VOICE OF THE CUSTOMER
• Market studies
• Customer research
• Sales calls
• Unplanned contacts with the customer
• Customer phone conversations (calls for

assistance)
• Customer observation
• Customer complaints
Some of the Tools we use to Understand the Customer:

- Questionnaires - Focus Groups
- Interviews - Historical Data
- Expert Studies - Affinity Diagrams
- House of Quality (aka HOQ and QFD)

UNDERSTANDING THE
VOICE OF THE CUSTOMER (VOC)
HOW ARE WE GOING

WHAT’S IMPORTANT
TO MEASURE OUR
TO THE CUSTOMER
ABILITY TO SATISFY
(in their own language)
WHAT THE
CUSTOMER WANTS?
(VOICE OF THE CUSTOMER) (MEASURABLE CHARACTERISTICS)
• “I don’t like to wait on hold a • hold time (goal: less than 1 minute)
Customer Service
long time.”
• “I hate getting bounced • % of customers routed to the correct
around to multiple people.” person the first time (goal = 100%)
• “I want to have good • % of customers surveyed replying
customer service.” “satisfied” or better (goal 99%)
• “The reps should be • % of customers greeted by name; %
friendly.” of customers not interrupted when
talking (goal 100%)
• “I hate filling out this • form completion time (goal: less

Loan Application
form – it takes too long.” than 10 minutes to complete)
• “I don’t understand what • % loan applicants responding form

they’re asking for.” is “easy” to complete (goal: at
least 99%)
• % of the form users who

understand the instructions on the
form without having to ask for
clarification (goal: at least 90%)

QUALITY FUNCTION DEPLOYMENT
(QFD)
• A formal approach to capturing the voice of the customer
and integrating it into every aspect of the design and
delivery of products and services (See chapter 11 in the
Basic Statistics text for more detailed information and
examples)
• Uses a series of carefully constructed matrices, called

Houses of Quality (HOQ)
Functional Req.
(CTC's)
Customer Req.
Customer
Expectations
House of
Quality
Prod. Design
#1 Characteristics
Functional Req.
House of
(CTC's)
Quality
Mfg. Process
#2 Characteristics
Performance
Characteristics
Prod. Design
CTC's House of
Quality
Mfg. Process
#3 Control
Product
Characteristics
Mfg. Process
Design CTC's House of

Quality
#4
Process
Design CTC's
Mfg. Process
Control

HOUSE OF QUALITY #1
(KISS-Style)
HOW
What can we measure or control to ensure
customer requirements are met? (These are
the CTCs, or critical to customer performance
measures)
Relationships
IMPORTANCE
WHAT
RATING
How do our measures relate to our

customer requirements?
What does
the 9 = strong relationship
customer 3 = medium relationship
want? 1 = weak relationship
0 = no relationship
1 = less
5 = more
PRIORITY / IMPACT
What do we need to focus on?
This is a weighted sum.
Example: Internet Service Provider

City Coverage
Transfer Rate
Logon Time
Ease of Use
Importance
Setup Time
Web Filter
Customer Requirements
A: Ability to block sites 1 3 9 3 1
B: Quick Logon 3 9 1 3
C: Fast Downloads 4 1 1 9 3 1 3
D: My kids can use it 5 3 1 9 1
E: Have access on the road 2 1 1 1 3
F: Short setup Time 3 1 3 9
Rank-Ordered Sum 9 48 56 62 28 45

HOUSE OF QUALITY #1 (cont.)
• Facilitates group discussion and decision making
• Helps us focus on the customer and their requirements
• Useful thought process for:

• determining what to measure (CTCs) in our project
• identifying highest priority requirements
• identifying process outputs on the IPO diagram
• identifying potential secondary measures for our
project
• Exercise (Optional) – Practice completing a House of

Quality for a pizza delivery process. The customer
requirements and importance ratings (from surveys and
talking with customers) are given. Brainstorm at least 4
CTCs (“hows”), assess their relationship with the
customer requirements (using 1, 3, 9 or blank) and then
calculate the weighted sum to determine the highest
priority measures.
Importance
Customer Requirements
A: Fast Delivery 5
B: Pizza hot when it arrives 4
C: Correct order 5
D: Food Tastes good 3
Rank-Ordered Sum

CREATING A HIGH LEVEL PROCESS MAP
(Moving From IPO to SIPOC)
• Before leaving the Define Phase, the Lean Six

Sigma project team must be able to define,
understand, and communicate the process being
studied
• The IPO diagram, covered previously, is a good

first start at describing the process at a high level
• Why is it critical to describe the process in terms

of inputs and outputs?
• Where do the outputs come from?
# of Calls
Ratio of Abandoned Calls

Type of Call
# of Operators
Help Desk Time per Call
Operator
Call Center
Ex
Tr
Kn
ai
pe
ow
ni
Cost per Call

rie
ng
le
nc
dg
e
e
of
Pr
Time Until Final

od
uc
Resolution
t
Information on Failures
(Service)

CREATING A HIGH LEVEL PROCESS MAP
(From IPO to SIPOC) (cont.)
• A SIPOC diagram is an extension of a simple IPO

diagram
• It includes information about suppliers and

customers
• To create a SIPOC diagram:

– name the process
– identify the outputs and customers
– identify the inputs and suppliers
S I P O C
Suppliers Inputs Process Outputs Customers
Materials, Any activity Measurable
methods, that takes outcomes of
people, etc. inputs and the process
required for creates
the process outputs
Who provides Who receives
the inputs to the outputs
your process? of the
process?

SIPOC DIAGRAM (cont.)
• Things to keep in mind when reviewing your

IPO/SIPOC diagram
– Are all of the critical to customer outputs

included?
– Have you ensured that the outputs are

measurable?
– Have you considered measures related to

better, faster, and lower cost?
– Have you indicated your primary project

measure(s) on the IPO/SIPOC diagram

SIPOC DIAGRAM EXAMPLE / EXERCISE
• Complete the SIPOC diagram below for the process of

ordering a pizza.
4. 3. 2. 1.
Suppliers: Inputs: Process: Outputs: Customers:
Telephone
Ordering for
Take-Out
Pizza
(from receipt of call to
delivery of order to
cook)
1. Who is/are the customers?
2. What are their requirements, translated into measurable

outcomes?
3. What are some of the inputs are needed to achieve the

desired output?
4. For each input, who is/are the suppliers?


MODULE 5
MEASURE … MAKING SENSE

OUT OF DATA USING GRAPHICAL AND
MEASUREMENT TOOLS
MAKING SENSE OUT OF DATA USING
GRAPHICAL AND MEASUREMENT
TOOLS
– DETAILED PROCESS MAPPING
– GRAPHICAL ANALYSIS OF DATA

Pareto charts
Histograms
Box plots
Run charts
Scatter diagrams
– NUMERICAL MEASURES
Measures of Location (mean, median)
Measures of Dispersion (range, standard deviation,
variance)
Measures of Quality for Variables Data
– Cp
– Cpk
– Dpm
– Sigma level
– Sigma capability
Measures of Quality for Attribute Data
– FPY
– dpu
– dpmo
– Sigma capability
Measuring Correlation
– USING SPC XL SOFTWARE AND

INTERPRETING OUTPUT

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
MEASURE
Define
Measure

Analyze


Notebook
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

CREATING DETAILED PROCESS MAPS
• Goal of the Measure phase is to capture the

current, “as-is”, process
• Different types of maps

– process flow diagram
– detailed process map / value stream map
(VSM)
– physical flow map
– time value map

PROCESS FLOW (PF) DIAGRAM
• Visual representation of all the major steps in a process,
whose primary purpose is to promote understanding
• Helps identify critical or problem areas, data collection

points, value and non-value added steps, and potential
areas for improvements. Non-value added steps should
be identified and removed.
• There are many different symbols, although we generally

use five standard (basic) symbols:
This Symbol… Represents… Some Examples are…
Receive Trouble Report

Start/Stop
Machine Operable
Approve/Disapprove
Accept/Reject
Decision Point
Yes/No
Pass/Fail
Drop Off Travel Voucher

Activity
Open Access Panel
A
Connector (to
another page or B
part of the diagram)
Represents
direction of flow
Process Flow Diagram Symbols (KISS)

PROCESS FLOW (PF) DIAGRAM (cont.)
• Can be accomplished at different levels
• To keep things simple, start at a high level, with 5-

10 steps, and then break it down further as
needed
Helpful Hints When Constructing a PF:

• Ensure that all steps, even if not part of the formal
process, are included.
• Walk the process! You can learn a lot by just

watching.
• When developing the flow chart, write each of the

steps onto sticky notes and post them onto a large
sheet of paper. As the team reviews and discusses
the process, the sequence of steps may change and
omitted steps may need to be added. Once it is
agreed that the flowchart looks right, the arrows can
be added to complete it.
• Create a “parking lot” for ideas as you are creating

the process flow. It is human nature to jump to
conclusions about possible solutions. Team
members will offer ideas, so take advantage of this
by capturing the ideas on a parking lot for later use.

PROCESS FLOW (PF) EXAMPLES
• High Level
Go to the Get desired Check Out Put Unload and

grocery items and place Drive
and Pay for items in put items
store in shopping cart Home
groceries car away
• More Detailed Level
Check prices, Get desired

Go to the Get out Locate
discounts, coupons
Have all
grocery desired items and items on
grocery for competing brands
store items place in list?
list to find the best deal shopping cart
no
yes
Unload Determine
and Scan Bag Method of
yes groceries Groceries Payment
Proceed Want to use
to Check self-service Put
Pay for items in
Out lane? Proceed Unload
Determine groceries car
no to Check items; wait
Method of
Out Lane for clerk to
Payment
scan/bag
Drive Unload and put

Home items away

PROCESS FLOW (PF) EXAMPLES (cont.)
Requested Change
Technical Return for more

Fail
Review information
Pass
Example Web
Page Creation
Management
Fail Rework
Review
Pass
Sample Integration with

Existing Web Site
Q.C. Fail Rework
Pass
Final Integration with

Existing Web Site
Q.C. Fail Rework
Pass
WEB SITE CHANGE FLOW CHART

Complete

STAFFING PROCESS
Process Description: Process Customer: Customer Valid Requirements: Performance Measure:
STAFFING Hiring Manager Timely Requisition/Approval Staffing Q1: % of Late Filled & Open Positions
Position Higher Hiring Manager Human Resources HR Support Candidate

Step/Time Authority
SOURCING APPROVAL
Identify the Need
Create and Approve Job Requisition and Description
Post Job
Develop Hiring Strategy
Conduct Ext. No
Recruiting?
Yes
Submit
Conduct Ext. Recruiting Collect Candidate Responses Resume
Screen Paperwork
No Any Potential
Candidates?
Yes
Forward to Hiring Mgr.
Screen Paperwork
No Any Potential
Interviewees?
Yes
INTERVIEWING
Conduct and Document Interviews
Evaluate and Rank Interviewees
No Found Final
Candidate?
Yes
Set Terms & Conditions of Offer No
Accept
OFFER / START
Extend Offer Offer?

Yes
Initiate New Hire Process
Send Offer Package No

Issues
Re-
solved
No ?
Contact Candidate Successful?
Yes
Yes Candidate
Reports

VALUE
VALUE = f (performance, cost, time)
Value added steps are defined as steps that:

1. Improve performance, function, etc.
2. Decrease cost (lower price)
3. Deliver the product or service faster (lower cycle time)
VALUE ADDED NON-VALUE ADDED

PROCESS FLOW (PF) SUMMARY
• What are some things to look for? Considerations

when building a process flow?

VARIATION OF THE PROCESS FLOW
• Enhancements to the process flow include turning
it into a “detailed process map” or simplified
“value stream map” by adding information such
as:
– Results or desired outcome at each step
(What is needed?)
– Inputs needed to achieve desired outcome
– Measurements taken at each step
– Data collection results, such as:

• Cycle time
• Number of people
• WIP, Inventory levels
• Defects, rework
• Yield
• Changeover times
• Distance Traveled
• etc.

DETAILED PROCESS MAP EXAMPLE
• Statapult® Launching (with outcomes and data

collection boxes for each step)
Ball doesn’t fall out Ball

Proper Angle Ball doesn’t stick in cup launches to Able to measure
Ready to Go Proper Tension Ready to Fire target Check if spec met
Set up Statapult® Pull Back Load Ball Release Measure and

and Get Ball the Arm in Cup arm Retrieve
cycle time = 35 sec. cycle time = 15 sec. cycle time = 5 sec. cycle time = 3 sec. cycle time = 13 sec.
material availability dpm = 5,000 dpm = 10,000 yield = 99% 15% out of spec
= 85% accident rate =
throughput = 100/hr 1/1,000

OTHER TYPES OF PROCESS MAPS
(Physical Process Flow Map)
DOCK
Warehouse
Physical Process Map

2
3 1
Eng. Work
Supplier Adm Ass’t Engineering
Mgr. Queue
14 4 18
Administration Building
Engineering Building
10 11
13
Room
Mail
WIP Accounting Controller

Queue Admin.
9
16 17
In-Bound Receiving Out-Bound

Work Queue Dock Work Queue
15
Purchasing
Operations
Department
12 6 5
8
Purchasing WIP Purchasing WIP
Admin. Queue Director Queue
7
• Plot the path on a layout

• Identify each step on the layout
• Connect the steps
• Measure / estimate the distance traveled

OTHER TYPES OF PROCESS MAPS
(Time Value Map) Value Added
Non-value Added
Value added: 30%
Non-value added: 70%
Total active
time: 10%
Time Value Map

Start End
Day Day
1 235
Value
Adding
Non-Value
Adding Queue
Queue / Feedback Loop Queue / Waiting Rework Loop

Waiting
Total cycle time = 235 days
118 minutes
value add
a measure of
process efficiency is
(value added time) / (total
lead time)
Activity
Time
• Place steps, queue’s along the time line as they occur

–Place value adding steps above the line,
–Place non-value adding steps below the line
• Create step segments proportional to the task times.
• Add in queue / non-activity times between steps.
• Draw in feedback loops & label yield percentages.
• Determine total cycle time.
• Categorize each non-value adding step into one of seven categories of waste.
• Sum activity / non-activity times.
• Sum value / non-value adding times by category of waste.

LEAN SIX SIGMA JOURNEY REQUIRES
DATA COLLECTION
The DMAIC ROADMAP
customer and the
business? Clearly
define the problem,
project, and team.
Define
Pr
oj thi
ec s
t s wa
uc y
ce
ss
…
Measure
Define the current

to assess
performance.
Analyze
What needs
improvement?
Identify the root
I mprove
Make the changes

and analysis.
Control
repeat problems.

WHY COLLECT DATA?
• To gather the facts for good decision making
• Because perception and intuition are not always reality
• To identify/verify problem areas and bottlenecks
• To characterize our processes
• To baseline the process
• As an indicator of process performance
“If you cannot measure it, you cannot improve it.”
“When you can measure what you are speaking

about, and express it in numbers, you know
something about it; but when you cannot measure it,
when you cannot express it in numbers, your
knowledge is of a meager and unsatisfactory kind; it
may be the beginning of knowledge, but you have
scarcely, in your thoughts, advanced to the stage of
science.”
Lord Kelvin
(1821 – 1907)

ROLE OF QUESTIONS IN DETERMINING
WHAT DATA IS NEEDED
• Questions help guide you through your project and

give you a starting point (think about what data
you will need in order to answer these questions?)
– The 14 Questions
– DMAIC Phase Tollgate Questions
• Other sources for identifying needed data

– The DMAIC roadmap / activities
– Process mapping activities
– Questions from project sponsor, team,
stakeholders, etc.
– Others?
Data Information Knowledge
Statistical Questions;
Tools Critical
Thinking

PLANNING FOR DATA COLLECTION
• To help get organized, a planning worksheet such as the

one below, can help you as you identify the data you and
your team needs to collect.
Sampling plan
How?
Metric Formula Who (time period, Accuracy
Definition (collection
(Data) (if appl.) Responsible? sample size, of data?
method)
source, etc.)
Start with first

From the time a
quarter data for
package enters
part #’s xyz1 –
the shipping Package Don’t
Query xyz10 (highest
Delivery area until the receipt time – know.
Larry S. shipping volume).
time time the shipping area Need to
database Randomly sample
package is in time validate.
15% of the
received by the
records for each
customer
month.

DATA GATHERING, SAMPLING, AND
MEASUREMENTS
• Data gathered should be __________________ of

the process being studied.
• Poor approach to sampling

– convenience sampling
• Better approaches to sampling
– random sampling
– systematic sampling
• Before gathering the data, think about the

measurement system being used to generate the
data
– is it accurate? precise? repeatable?
COMING ATTRACTIONS
Lean Six Sigma requires good, useable data

“Oh,
“Oh, sure,
sure, he remembers
he remembers
everything. But
everything. But show me
show me
one significant insight he’s
one significant insight he’s
been
been able
able to
to draw
draw from all
from all
that data.”
that data.”

Source: The Chronicle of Higher Education
Harald Bakkan & Mischa Richter
5 - 21
MEASURING PROCESS PERFORMANCE
Type of Summary
Graph Metric, Measure
• Histogram • Mean, Standard Deviation

(numeric)
Variables
• Run chart • Defects per million (dpm)

• Control chart • Sigma level Cpk Analysis
• Scatter plot • Sigma capability

Type of Data
LSL USL
• Cp and / or Cpk
Nor mal Distr ibution
Mean = 20.8 Histogram
Std Dev = 3.3966
KS Test p-value = .5030
8
5
# Obs er va tions
0
15.0 t o <= 17. 4 to <= 19.8 to <= 22.2 t o <= 24. 6 t o <=
17 .4 19.8 22. 2 24.6 27.0
5.52
8.73
9.8
10.9
15.2
16.2
17.3
20.5
21.6
22.7
23.7
26.9
29.1
28
6.59
7.66
11.9
13
14.1
18.4
19.4
24.8
25.9
30.2
31.2
32 3
Class
• Pareto diagram • First pass yield (FPY)

(discrete)
Attribute
• Run chart • % “good” (or “bad” )

• Control chart • Defects per unit (DPU)
• Defects per million
Outgoing Product Quality
9000
8000
opportunities (dpmo)
• Sigma Capability
7000
6000
dpm
5000
4000
3000
2000
1000
0
1996 1997 1998 1999 2000 2001 2002 2003
Year
Goal: Turn data into information

PARETO DIAGRAMS
DATA
Types of Count Proportion
Shipping Errors (frequency)
Address 11 .11
Ship Early 9 .09
Ship Late 37 .37
Method 6 .06
Contents 8 .08
Packing 29 .29
100 1.00
Types of Errors and Frequency of Occurrence
Information about Shipping Errors
40
35
30
Frequency
25
20
15
10
0
Ship Late Packing Address Ship Early Contents Method
Type Shipping Error

CREATING PARETO CHARTS WITH SPC XL
SPC XL > Analysis Diagrams > Pareto Chart
Options window
for chart details
A Pareto Report Pareto Chart
is produced along 100 1

with the Pareto 90 0.9
Chart 80
70
0.8
0.7
60 0.6
Count
Group Count % of Total 50 0.5

Ship Late 37 0.37 40 0.4
Packing 29 0.29 30 0.3
Address 11 0.11 20 0.2
Ship Early 9 0.09 10 0.1

0 0
Contents 8 0.08 Ship Late Packing Address Ship Early Contents Method
Method 6 0.06 Shipping Errors

OTHER EXAMPLES OF PARETO
CHARTS
120%
1800
1600 100%
1400
Monetary Loss ($)
80%
1200
1000 60%
800
40%
600
400
20%
200
0 0%
Packing Ship Late Ship Early Address Contents Method
Pareto of Shipping Errors Using Monetary Loss
Improper Time Management

% of Time
Fire Paper Meetings Travel Other Quality

Fighting Work Improvement
(Fire Prevention)
Pareto of Activities by Time Spent

CASE STUDY: DISK DRIVE PARETO
ANALYSIS
You are the Quality Engineer responsible for a
diskette drive that is used in the assembly of a
personal computer. You receive a report listing the
error code of each unit rejected by the manufacturing
process for the first five days of production. This
report is shown below. Based on the report you
decide that the defect level must be reduced.
To understand what is causing the problems with
the diskette drives, analyze this report using Pareto
Diagrams and recommend what must be done to
improve the diskette drive quality.
DATA
Thursday, June 8
Code Code Code Code Code
410 410 413 410 412

465 465 465 413 410
412 410 410 412 440
410 412 410 410 410
218 841 440 218 410
440 410 841 413 412
410 413 413 841 413
413 410 465 410 465
Friday, June 9
841 413 841 410 410

412 410 410 412 413
410 410 218 413 410
218 412 410 218 465
465 410 465 465 413

CASE STUDY: DISK DRIVE PARETO
ANALYSIS
Monday, June 12
410 841 465 465 440
218 413 413 413 218
Tuesday, June 13
440 410 440 465 841
413 412 412 413
181 465 218 440
Wednesday, June 14
465 412 413 841 440
440 413 412 465 410
413 465
30
27 INFORMATION
25
19
20
16
15
12
10 9
8 8
0
410 413 465 412 440 841 218 181

INFORMATION
Days of the Week June 8

40
40 15
16
35
14
30 25 12
25 10
20 8 7
13 12
15 6 5 5
10
10 4 3 3
2
5 2
0 0
Th F T W M 410 413 465 412 440 841 218
June 9 June 12
16 16
14 14
12 12
10 9 10
8 8
6 6
4 4
3 3 4 3
4
2 2 2
1 1 1
2 2
0 0
410 413 465 412 218 841 413 465 218 410 440 841
June 13 June 14
16 16
14 14
12 12
10 10
8 8
6 6
3 4 3 3
4 2
2 2 2 2
1 2 1 1
2 1 1 1
0 0
440 413 465 412 410 841 218 181 413 465 412 440 410 841

ANALYSIS
KNOWLEDGE PRODUCING QUESTIONS

• The majority of the failures occurred on the first
two days of production. What was different about
these days that caused so many more failures
than on the other days?
• The majority of the failures were 410 error codes.

In addition, most of these were concentrated in the
first two days of production. By the third day the
frequency of this error was the same as the
others. What was happening at the beginning of
the week that caused so many 410 errors?
KNOWLEDGE
• After analyzing the units that had failed with the
410 error code, it was discovered that the diskette
drive design allowed a manufacturing worker to
unknowingly damage the unit. As the workers
became more experienced at installing this unit
they made this mistake less frequently. This
reduced the number of 410 errors. Since the
product could still be damaged by an
inexperienced person, a simple design change
was made to prevent the damage from occurring.

EXERCISE
• Individually
• Given your sample of M&Ms, build a Pareto

diagram by color
• Before you destroy the evidence, be sure you

have an accurate count of each color and the total
number of M&M’s in the bag
Brown =
Red =
Yellow =
Blue =
Orange =
Green =
Total

HISTOGRAMS
n = 50 DATA
18 16 30 29 28 21 17 41 8 17
32 26 16 24 27 17 17 33 19 18
31 27 23 38 33 14 13 26 11 28
21 19 25 22 17 12 21 21 25 26
23 20 22 19 21 14 45 15 24 34
INFORMATION
16
14
12
Number of Cars
10
0
6 to <12 12 to <18 18 to <24 24 to <30 30 to <36 36 to <42 42 to <=48
Gas Mileage (mpg)
Histogram of Gas Mileage Data

EXERCISE
• Individually
• Using the following rules, build a histogram that will

describe the 50 data points shown on the previous page
Number of Data Points: n =
Range of Data: R = Xmax - Xmin =
Number of Classes: k = n =
Class Width: w = R/k =

FREQUENCY DISTRIBUTION
Class # of Observations
• Using the axis system on the next page, scale the

horizontal axis and build the histogram for this frequency
distribution. Compare this histogram with the one on the
previous page.

HISTOGRAM EXERCISE (cont.)
20
18
16
Number of Cars
14
12
10
Gas Mileage (mpg)

CREATING HISTOGRAMS WITH SPC XL
SPC XL > Analysis Diagrams > Histogram
Options window
for chart details
Normal Distribution
Mean = 22.88
Std Dev = 7.7266 Histogram
16
14
12
# Observations
10
0
8.0 to <= 13.0 to <= 18.0 to <= 23.0 to <= 28.0 to <= 33.0 to <= 38.0 to <= 43.0 to <=
13.0 18.0 23.0 28.0 33.0 38.0 43.0 48.0
Class

HISTOGRAM EXAMPLES
The histogram provides a visual representation of the
distribution of data. Here are some shapes you might see.
Classic “bell-shaped” curve
Skewed distribution
Rectangular (uniform) distribution
Parabolic distribution
Exponential distribution
Bi-modal distribution

BOX PLOTS
• Another method for graphically depicting data
• Especially useful when comparing 2 or more sets of

data (side by side) or when the data sets are small
• Breaks up data into quartiles (fourths)
Maximum value in the data set
Third Quartile
Interquartile Range
+ Median
First Quartile
Minimum value in the data set
SPC XL > Analysis Diagrams > Box Plot

Box Plot
50 Titles
Gas Mileage
45 Data
40
35
Q1
30
Min
25 Median
+ Max
20 Q3
15
10

EXERCISE
• Individually
• Enter the two statapult® data sets (from round 1

and round 2) into two separate columns (or rows)
in your Excel spreadsheet
• For each set, look at a histogram. What does the

shape look like?
• For purposes of comparison, highlight both

columns (or rows) of data and create side-by-side
box plots of the data sets. What do you see
visually?

RUN CHART
DATA
171, 156, 230, 147, 167, 183,
225, 190, 250, 245, 205, 210,
260, 270, 290
INFORMATION
350
300
250
Time (hrs)
200
150
100
50
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Chronological Order
Run Chart for Time Between

Service Outages

SOFTWARE EXAMPLES
50
40
NUMBER
30
OF CSUs
DESIGNED 20
10
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
PDR NOW PDR PDR(R1)
CONTRACT MONTHS
TOTAL CSUs = 50 PLANNED (REV 1) = 30
ACTUAL = 30 TOTAL REDESIGNED = 7
PLANNED = 33
Design Progress
20
16
ERRORS
KLOC 12
1980 1981 1982 1983 1984 1985 1986 1987 1988
PROJECTS OVER TIME

PROPORTION OF MISROUTED MAIL
0.18
0.16

0.14
0.12
0.10
0.08
Proportion
0.06
0.04
0.02
1
3
5
7
9
0
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
49
51
53
55
57
59
61
Day
5 - 40
#
CREATING RUN CHARTS USING SPC XL
SPC XL > Quality Tools > Run Chart
Run Chart
170
160
150
Data can be stored in
140
separate columns.
130
Y Axis
Highlighting 2 or more
120
columns will overlay the
110
run charts.
100
90
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
X Axis
Run Chart
170
160
150
Data can be stored in one
140
column. This may be a
130
Y Axis
nice visual way to show

120
“before project” and “after
110
project” results.
100
90
80
10
13
16
19
22
25
28
31
34
37
40
43
46
1
X Axis

EXERCISE
• Individually (statapult® data)
• Using SPC XL > Quality Tools > Run Chart, and

putting both sets of data (from round 1 and round
2) in the same column (or row), generate a run
chart for the statapult® data.
• Try regenerating the run chart by using the data in

two separate columns and overlaying the charts.

CONTROL CHART
1.05
1.04
UCL
1.03
1.02
1.01
Average (oz)
Centerline
1
0.99
0.98
LCL
0.97
0.96
0.95
0.94
1 2 3 4 5 6 7 8 9 10 11 12
Sample Number
Note: Control Limits are not the same as Spec Limits
Question: What is the difference between control limits and spec

limits?
COMING ATTRACTIONS

A SIMPLE MEASUREMENT SYSTEM
ANALYSIS EXAMPLE
255 y + 3s = 254.7 = UCL

Credit Score
254
253
252
251
250
y
249
248
247
246
245 y − 3s = 244.7 = LCL
244
If the same person had scored the same credit application

(not knowing at the time of scoring it was the same
application) on 30 different occasions, the differences in
resulting scores shows us the variation in repeatability
COMING ATTRACTIONS

SCATTER DIAGRAM
FOR VEHICLE WEIGHT VS MILEAGE
Observation Weight (lbs) Mileage (mpg)
1 3000 18
2 2800 21
3 2100 32
4 2900 17
5 2400 31
6 3300 14
7 2700 21
8 3500 12
9 2500 23
10 3200 14
Bivariate Data (Vehicle Weight vs Gas Mileage)
35
30
25
Mileage (MPG)
20
15
10
0
2100 2300 2500 2700 2900 3100 3300 3500
Vehicle Weight

LOOKING AT CORRELATION USING
SPC XL
SPC XL > Analysis Diagrams > Scatter Plot
Scatter Chart (weight vs mileage)
35
y = -0.0152x + 63.507
30 2
R = 0.9191
25
20
mileage
15
10
0
2000 2200 2400 2600 2800 3000 3200 3400 3600
weight
SPC XL > Analysis Tools > Correlation Matrix
The table shown below displays the pairwise correlation coefficient.
Correlation Matrix
Correlation Coefficient = -0.958680615

EXERCISE
A customer service organization has been tracking the time
its customers spend on hold along with the proportion of
abandoned calls. Data for the past 15 days is given below.
a. Is there a strong relationship between the hold time and

the abandon rate? Why or why not?
b. Do you have any other observations or comments about

the data?
Abandon Rate vs Answer Speed

1 – 15 Feb 01
Date Avg Hold Time Avg Abandon Rate
1-Feb 17 0.022
2-Feb 22 0.027
3-Feb 29 0.025
4-Feb 14 0.02
5-Feb 68 0.053
6-Feb 52 0.043
7-Feb 19 0.024
8-Feb 44 0.044
9-Feb 37 0.034
10-Feb 47 0.038
11-Feb 82 0.057
12-Feb 25 0.029
13-Feb 27 0.022
14-Feb 48 0.041
15-Feb 35 0.042

CORRELATION VS CAUSALITY
Case Cumulative Cumulative

Fouls Points
1 3 4
2 8 12
3 13 19
4 18 41
5 31 52
6 42 78
7 58 105
8 64 148
9 74 218
10 67 247
11 79 401
12 86 522
Bivariate Data (Fouls vs Points)
600
500
400
Points Scored
300
200
100
0
0 10 20 30 40 50 60 70 80 90
Fouls
Scatter Diagram of Fouls vs Points Scored

CORRELATION VS CAUSALITY (cont.)
80
City Population (1000's)
70
60
50
100 200 300
Number of Storks
A plot of the population of Oldenburg, Germany at

the end of each year against the number of storks
observed in that year, 1930-1936.
Source: “Statistics for Experimenters”

by Box, Hunter, and Hunter. (1978)

EXERCISE
• Individually
• Using SPC XL, reproduce the scatter diagram for

Fouls vs Points and comment on the strength of
that relationship.
• Give an example of a spurious correlation that

you are aware of.

MEASURING PROCESS PERFORMANCE
Type of Summary
Graph Metric, Measure
• Histogram • Mean, Standard Deviation

(numeric)
Variables
• Run chart • Defects per million (dpm)

• Control chart • Sigma level Cpk Analysis
• Scatter plot • Sigma capability

Type of Data
LSL USL
• Cp and / or Cpk
Nor mal Dis tr ibutio n
Mean = 20.8
Histogram
Std Dev = 3.3966
8
5
# Obser vations
0
15.0 t o <= 17. 4 to <= 19.8 to <= 22.2 t o <= 24. 6 t o <=
17.4 19.8 22. 2 24.6 27.0
5.52
8.73
9.8
10.9
15.2
16.2
17.3
20.5
21.6
22.7
23.7
26.9
29.1
28
6.59
7.66
11.9
13
14.1
18.4
19.4
24.8
25.9
30.2
31.2
32 3
Class
• Pareto diagram • First pass yield (FPY)

(discrete)
Attribute
• Run chart • % “good” (or “bad” )

• Control chart • Defects per unit (DPU)
• Defects per million
Outgoing Product Quality
9000
8000
opportunities (dpmo)
• Sigma Capability
7000
6000
dpm
5000
4000
3000
2000
1000
0
1996 1997 1998 1999 2000 2001 2002 2003
Year
Goal: Turn data into information

MEASURES OF LOCATION
∑x i
x1 + x 2 + Κ + x n
x= i =1
=
n n
Sample Mean: x
8 11 6 7 12 7
Help Desk Question Resolution Time (hrs)
x = (8 + 11 + 6 + 7 + 12 + 7) 6 = 51 6 = 8.5 hours
4 5 6 7 8 9 10 11 12 13 14
(hours)
The Mean as a Balance Point

MEASURES OF LOCATION (cont.)
236,000
36,000
32,000 40,000 98,000
28,000
Mean: 46,000
Total Payroll: $1,058,000

# on Payroll: 23
Average Annual Pay:
1,058,000
= 46,000
23
Source: Faces of Mathematics: An Introductory Course for College

Students, 2nd Edition, by Roberts and Varberg. Harper & Row, 1982.

Assume that x1, x2, … xn, is a list of the sample data sorted in
ascending order. Then
~ ⎧middle value,if n is odd

x=⎨
⎩the averageof the two middle values,if n is even
Sample Median: ~
x
Data Set Name Variable Sample Sample Sample

Name Size Mean Median
Help Desk Question Resolution Time x 6 8.50 7.5
Gas Mileage y 50 22.88 21.5
Summary Measures of Location
SEATWORK/HOMEWORK
X: Sample #1 (8, 3, 7, 4, 5, 6, 5, 4, 5, 6, 5)
Find x and ~
x
Y: Sample #2 (8, 3, 7, 4, 5, 6, 50, 4, 5, 6, 5)

Find y and ~
y

Symmetric (bell-shaped) Distribution
Mean = Median
Skewed Distribution
Median Mean

MEASURES OF VARIATION
DON'T WORRY,
THAT ROPE IS
ONE INCH THICK
ON THE
AVERAGE.
Source: Faces of Mathematics: An Introductory Course for College

Students, 2nd Edition, by Roberts and Varberg. Harper & Row, 1982.

MEASURES OF VARIATION (cont.)
x y z
30 10 10
30 30 20
30 30 30
30 30 40
30 50 50
x = 30 y = 30 z = 30
~
x = 30 ~
y = 30 ~
z = 30
Three Different Data Sets
Range = Max - Min
Rx =
x
10 20 30 40 50
Ry =
y
10 20 30 40 50
Rz =
z
10 20 30 40 50
Data Sets with Same Center but Different Spreads

MEASURES OF VARIATION (cont.)
∑ (x − x )
n 2
i
2
s = i=1
n −1
Sample Variance: s2 = σ̂ 2(an estimate of σ2)
x (x − x ) (x − x ) 2
y (y − y ) (y − y ) 2
z (z − z) (z − z) 2
30 10 10
30 30 20
30 30 30
30 30 40
30 50 50
sx2 = sy2 = sz2 =

sx = sy = sz =
∑ (x )
n 2
i −x
s = s2 = i=1
n −1
Sample Standard Deviation: s = σ̂

SUMMARY STATISTICS USING SPC XL
SPC XL / Analysis Diagrams / Summary Stats
Count 50
Mean 22.88
Dot Plot Median 21.5
Mode 21
Max 45
Min 8
Range 37
Std Dev (Pop) 7.648895345
Std Dev (Sample) 7.726551096
Variance (Pop) 58.5056
Variance (Sample) 59.69959184
Skewness 0.683704873
Kurtosis 0.525173863
95% Conf. Interval for Mean

0 10 20 30 40 50
Upper Limit 25.07586019
Lower Limit 20.68413981
99% Conf. Interval for Mean

Upper Limit 25.80838337
Lower Limit 19.95161663

EXERCISE
• Individually
• Enter the two statapult® data sets (from round 1 and

round 2) into two separate columns (or rows) in your
Excel spreadsheet
• For each data set, use SPC XL > Analysis Diagrams >
Summary Stats to find the following
Round 1 Round 2
Mean:
Median:
Range:
Sample Variance:
Sample Standard Deviation:

MEASURES AND INDEXES OF QUALITY
(For Continuous Data when y and σ can be estimated)
σlevel = number of standard deviations between the

center of the process and the nearest spec
(this is a Z value)
⎛ USL − y y − LSL ⎞
= minimum ⎜⎜ , ⎟
⎟
⎝ σ σ ⎠
Cpk = a process capability index

= proportion of natural tolerances (3σ) between
the center of a process and the nearest spec
= minimum ⎛⎜ USL − y , y − LSL ⎞⎟ = σ level

⎜ 3σ 3σ ⎟ 3
⎝ ⎠
Cp = a process potential index

= specification width
process width
USL − LSL
=
6σ
σcapability = 3 • Cp *
* This calculation is sometimes used as a measure of potential. For the

purpose of benchmarking and comparing continuous (variables) data with
discrete (attribute) data, sigma capability is calculated differently, by converting
dpmo to an equivalent sigma capability, using a conversion table found in the
references section.

CAPABILITY ANALYSIS USING SPC XL
Suppose the time to process a claim follows a normal distribution with
a mean of 7 days and a standard deviation of 2 days. The upper
spec limit is 16 days, with a lower spec limit (imposed by
management) of 4 days.
LSL USL
σ =2
4 7 10 16
SPC XL > Analysis Diagrams > Cpk
Source data can be stored in a

spreadsheet, or you can manually
enter the mean and standard
deviation.
After inputting the

information, make
sure to click on
“Recompute
Scaling for Axis”

EXERCISE 1a
LSL USL
σ=2
10 15 y = 16 20
From this graph, calculate the following:
σlevel =
t3
f irs st!
he d fir
yt
Tr han
= by
Cpk =
Cp =
dpm =
σcapability =

EXERCISE 1b
LSL
σ=2
10 y = 16
σlevel =
Cpk =
Cp =
dpm =
σcapability =

EXERCISE 1c
LSL USL
σ=2
10 15 y = 16
σlevel =
Cpk =
Cp =
dpm =
σcapability =

EXERCISE 2
• Individually
• Using Excel and your two statapult® data sets (from

round 1 and round 2) as two different columns (or
rows)
• Assume y is the targeted distance for each data set
• Assume USL = y + 5 and LSL = y − 5
• For each data set, use SPC XL > Analysis Diagrams

> Cpk to find:
Round 1 Round 2
Cpk =
Cp =
dpm =

(For Discrete or Attribute Data)
Unit: any service or transaction for which a quality/performance

measure is desired. This must be specifically defined.
A = # of units completely
unit 1 defect (error) free without any
kind of rework or touch-up of
any kind during the process
unit 2
n units …
. B = # of units that move on
Started . but have had at least some
. Defect Defect rework or touch-up performed
REWORK REWORK on them
unit n
PROCESS STEPS
C = # of units scrapped
A +B
Yield = n
A
First Pass Yield (FPY) = n
Total # of defects (errors) observed from

start to finish over all units started
defects per unit (dpu) = Total # of units started (n)
FPY ≈ e-dpu (a very good estimate of FPY)

• Dpu is the best overall measure to focus on for

quality improvement purposes
• Defects drive cost, cycle time, and waste
• Defects cause customer dissatisfaction
• If we wish to compute a sigma capability or we

want to do benchmarking, however, we cannot
use dpu directly
• Dpu’s must be normalized, to account for

complexity before we can do benchmarking

Opportunities = number of value-added entities or

features of a part, product, or
service that must be met or done
right (it is highly dependent on
context and must be defined
precisely).
dpmo = defects per million opportunities
⎛ total # of defects ⎞
= (1,000,000 ) × ⎜⎜ ⎟⎟
⎝ total # of opportunit ies ⎠
or,
⎛ dpu ⎞
= (1,000,000 ) × ⎜⎜ ⎟⎟
⎝ opportunit ies per unit ⎠

OPPORTUNITIES
• DO NOT:
– Measure the number of ways we can mess something up
– Example: Suppose the process we’re studying is pouring

coffee into a cup. It would be silly to define the
opportunities as 200, let’s say, because I can mess up 200
ways by spilling coffee on the floor, spilling coffee on the
counter, spilling coffee on my shoe, etc.
• DO:
– Measure the complexity of a product or process
– Help us when benchmarking products or processes of

dissimilar complexity
– Depend highly on the context and must be defined precisely
– Represent value-added entities or features of a part,

product, or service that must be met or done right
• RECOMMENDATIONS:
– Keep it simple
• For products, opportunities may be the number of piece
parts making up the product
• For processes, opportunities may be the number of value-
added process steps
– Come up with a standard that makes sense and stick with it

so that improvement can be measured!

EXERCISE / EXAMPLE
A finished document (i.e., unit) contains a total of 4

critical entries (i.e., a critical entry is an opportunity for
a defect, mistake, or error). An inspection of 3,000
documents revealed a total of 195 errors, and 94% of
the documents were good (i.e., defect free). Calculate
the following:
dpu =
dpmo =
Total Number of Defective Documents =
Total Number of Defect-Free Documents =
FPY =
σcapability =
Hint: see next page to convert dpm into σcapability .

RELATIONSHIP BETWEEN
Cp, Cpk, dpm, and σcapability
Standard Deviations from Mean
OVERALL YIELD vs SIGMA
100K
(Distribution Shifted 1.5σ) Cp = 1
Cpk = 0.5
# of Parts Centered
(Steps) ±3σ ±4σ ±5σ ±6σ 10K ±1.5 Sigma Shift
1 93.32% 99.379% 99.9767% 99.99966% Cp = 1.33
Cpk = 0.83
7 61.63 95.733 99.839 99.9976 Cp = Cpk = 1
10 50.08 93.96 99.768 99.9966 1K

20 25.08 88.29 99.536 99.9932
40 6.29 77.94 99.074 99.9864
Cp = 1.67
60 1.58 68.81 98.614 99.9796 100 Cpk = 1.17
80 0.40 60.75 98.156 99.9728
100 0.10 53.64 97.70 99.966 Cp = Cpk = 1.33
150 --- 39.38 96.61 99.949

200 --- 28.77 95.45 99.932 10
300 --- 15.43 93.26 99.898
400 --- 8.28 91.11 99.864 Cp = 2
Cpk = 1.5
500 --- 4.44 89.02 99.830 1
600 --- 2.38 86.97 99.796
Cp = Cpk = 1.67
700 --- 1.28 84.97 99.762
800 --- 0.69 83.02 99.729
0.1
900 --- 0.37 81.11 99.695
1000 --- 0.20 79.24 99.661
1200 --- 0.06 75.88 99.593
3000 50.15 98.985 0.01
---
17000 --- 1.91 94.384
---
38000 --- 0.01 87.880
Cp = Cpk = 2
70000 78.820
USE FOR 0.001
150000 BENCHMARKING
--- 60.000
--- 2 3 4 5 6 7
Source: SIX SIGMA RESEARCH INSTITUTE

Motorola University
σcapability dpm (2-sided)

6.0 3.4
5.5 32
5.0 233
4.5 1,350
4.0 6,210
3.5 22,750
3.0 66,807
2.5 158,655
2.0 308,538
σcapability vs dpm
σ capability = .8406 + 29.37 − 2.221* ln(dpm)

MEASURING QUALITY FOR ATTRIBUTE
DATA USING SPC XL
• Record pertinent information in your spreadsheet:

– Number of units tested or inspected
– Number of defects found
– Number of opportunities per unit (optional)
– Text description (optional)
SPC XL > Quality Tools > Product Capability
Reference the cells where

your data is stored
Product Summary Report
Total Sigma
Characteristic Defects Units Opportunities Opportunities DPU DPO DPMO Capability
documents 195 3000 4 12000 0.06500 0.01625 16250.000 3.6382
Total 195 0 12000 0.01625 16250.000 3.6382

QUALITY AND CAPABILITY – Attribute Data
(Exercise)
• Consider the following scenario: 16 forms, each

containing 5 critical pieces of information, were
inspected for errors and inconsistencies. An “X”
indicates an error found. Data for the 16 forms is
shown below:
Form 1 2 3 4 5 6 7 8
X
X X X X
X X X
X X
Form 9 10 11 12 13 14 15 16
X
X X X
X X X X X
X
• Analyze the data. What do you conclude about the

quality and capability of the process? What other
comments or observations do you have?


MODULE 6
MEASUREMENT SYSTEM ANALYSIS

MEASUREMENT SYSTEM
ANALYSIS
– PROPERTIES OF A GOOD
MEASUREMENT SYSTEM
– IMPACT OF MEASUREMENT SYSTEM

VARIATION
– HOW TO SET UP, CONDUCT, AND

PERFORM A MEASUREMENT SYSTEM
ANALYSIS (MSA)
– Attribute data
– Variables data
– INTERPRETATION OF MSA RESULTS AND

METRICS
– Effectiveness
– Repeatability
– Reproducibility
– P/Tol ratio
– P/Total ratio
– Discrimination (resolution)

MEASURE
Define
Measure

Analyze


Notebook
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

(MSA)
Product, Transaction Measurement

Or Service Variability Variability
• MSA identifies and quantifies the different sources

of variation that affect a measurement system.
• Variation in measurements can be attributed to

variation in the product, transaction or service
itself or to variation in the measurement system.
• The variation in the measurement system itself is

measurement error.

KEY MSA ELEMENTS
• The objective of an MSA is to learn as much as possible

about the measurement process in a short amount of time.
• The strategy is to include people, SOPs, data recording

devices, etc., that will usually be elements of the
measurement process.
• A random selection of parts, transactions or services

representing at least 80% of the total process variation
should be made.
• The parts, transactions or services should be measured as

independently as possible to avoid measurement bias.
• Each part, transaction or service item will be measured

multiple times (at least twice) by each person or operator
using the same procedure.
• ROT:
– Variables (continuous data)
(Number of people) X (Number of parts or transactions) ≥ 20
– Attribute (binary data)

(Number of people) X (Number of parts or transactions) ≥ 60

ATTRIBUTE DATA MSA CASE STUDY
• A Green Belt in the IT department at a company was

tasked with investigating and improving the availability of
one of its internet applications.
• A cause and effect analysis surfaced several issues

related to the “calculation” of availability which brought
the numbers into question. The Green Belt decided to
set up a measurement system analysis to look at the
effectiveness of the process for determining what was
counted as a chargeable “outage” for the purposes of
calculating system availability.
• There were standards available for determining what

constituted a chargeable outage that should count in the
availability numbers for an application. Considerations
such as the impact to users, time of day, scheduled
maintenance, etc. were applied.
• To conduct the study, 20 outage tickets (reports) were

selected, half of which should have been charged
against availability (“A”) and half of which should not
have been charged against availability (“R”). 3
inspectors were selected, representative of the project
leads and managers who must make these
determinations on a daily basis. Each person was asked
to review the 20 tickets, apply the definition of a
chargeable outage, and make the determination as to
whether the ticket should have been charged to the
availability rate for the particular application in question.
• Data from the study is shown on the next page.

USING ATTRIBUTE OR BINARY DATA
R = Reject 1 = correct response
A = Accept 0 = incorrect response
Trans- True
OP 1 OP 2 OP 3
action Standard
1 R R 1 R 1 R 1 3
2 A A 1 A 1 A 1 3
3 A A 1 R 0 A 1 2
4 R R 1 R 1 R 1 3
5 A R 0 A 1 R 0 1
6 R R 1 A 0 R 1 2
7 R R 1 R 1 R 1 3
8 A A 1 A 1 A 1 3
9 A A 1 A 1 A 1 3
10 A R 0 R 0 A 1 1
11 R A 0 R 1 R 1 2
12 A A 1 A 1 A 1 3
13 R R 1 R 1 R 1 3
14 R R 1 A 0 R 1 2
15 A R 0 A 1 A 1 2
16 R A 0 R 1 R 1 2
17 A A 1 A 1 R 0 2
18 A A 1 A 1 A 1 3
19 R R 1 A 0 R 1 2
20 R R 1 R 1 R 1 3
15 15 18 48
OP 1 OP 2 OP 3
A III II II ← rejected falsely

TRUTH
R II III ← accepted falsely
Number and Type of Mistake by Operator

INSPECTION CAPABILITY
(Attribute or Binary Data)
Effectiveness (E) is the ability of an individual to distinguish
between good (accept) and bad (reject) parts or transactions.
Number of transactions identified correctly
E=
Total number of opportunit ies to be correct
Probability of False Rejects (FR) is the likelihood of rating a

good part or transaction as bad.
Number of times good transactions are rated as bad
P(FR) =
Total number of opportunit ies to rate good transactions
Probability of False Acceptance (FA) is the likelihood of

accepting a bad part or transaction as good.
Number of times bad transactions are accepted as good
P(FA ) =
Total number of opportunities to rate bad transactions
Bias (B) is a measure of an individual’s tendency to falsely

classify a part or transaction as good or bad.
⎧B = 1 ⇒ no bias (i.e., both errors are equally likely)
P(FR) ⎪
B= ⎨B > 1 ⇒ bias toward rating good transactions as bad
P(FA ) ⎪B < 1 ⇒
⎩ bias toward rating bad transactions as good
INSPECTION CAPABILITY EVALUATION GUIDELINES

Parameter Acceptable Marginal Unacceptable
E ≥ .90 .8 - .9 < .8
P(FR) ≤ .05 .05 - .10 > .10
P(FA) ≤ .02 .02 - .05 > .05
B .8 - 1.20 .5 - .8 or 1.2 - 1.5 < .5 or > 1.5

EXERCISE
• For the “inspection” measurement system data and

capability measures shown on the previous pages,
answer the following questions:
1. What is the overall effectiveness of the

inspection system?
2. What is the false rejection rate?
3. What is the false acceptance rate?
4. What is the overall bias of the measurement

system?
5. Recommendations?

USING SPC XL FOR ATTRIBUTE
MSA ANALYSIS
• To do this analysis with SPC XL, we first set up a

template
• Select SPC XL > MSA > Create MSA Template

from the menu
• Specify Number of Inspectors (Operators) (for this

example, enter 3)
• Specify Number of Replicates (for this example,

enter 1)
• Specify Number of Transactions (Parts) (for this

example, enter 20)
• Press OK

DATA ENTRY EXAMPLE
• A blank template for recording data will appear
• Enter the data into the template, as shown below
• Enter “A” for an acceptable (or good) transaction
• Enter “R” for a reject (or bad) transaction
For Attribute data enter A for Accept

MSA Data Template
and R for Reject
Date: 2/24/01
Description:
Part Type:
USL:
LSL:
Operator 1 Operator 2 Operator 3

Part # Reference Rep 1 Rep 1 Rep 1
1 R R R R
2 A A A A
3 A A R A
4 R R R R
5 A R A R
6 R R A R
7 R R R R
8 A A A A
9 A A A A
10 A R R A
11 R A R R
12 A A A A
13 R R R R
14 R R A R
15 A R A A
16 R A R R
17 A A A R
18 A A A A
19 R R A R
20 R R R R

SPC XL ANALYSIS RESULTS FOR
ATTRIBUTE DATA
• Once the data is entered on the template, select

SPC XL > MSA > Attribute Analysis
• SPC XL Output is shown below
• Summary information, such as effectiveness, is

printed by operator as well as for the
measurement process as a whole (total)
Attribute MSA Analysis

Number and Type
Mistake By
Operator
OP 1 OP 2 OP 3 Total
A 3 2 2 7 <-reject falsely
Truth
R 2 3 0 5 <-accept falsely
Inspection
Capability
OP 1 OP 2 OP 3 Total
Effectiveness 0.75 0.75 0.9 0.8
P(FR) 0.3 0.2 0.2 0.233333
P(FA) 0.2 0.3 0 0.166667
Bias 1.5 0.666667 NA 1.4

MEASUREMENT SYSTEM STUDY
EXERCISE (Attribute Data)
• A Green Belt was tasked with improving a credit approval process.
In the measure phase of her project, she wanted to look at the
effectiveness of the current “measurement” (approval) process.
• Specifically, she wondered whether analysts, given the same sets

of guidelines, would consistently make the same decision (approve
vs. not approve) for various applicants and whether there was
significant variation analyst to analyst.
• She decided to set up a measurement system analysis. Twenty six

applications (only 13 of which should have been “approved”
according to company guidelines) were used in the study. Four
analysts each saw these “dummy” applications on 3 different
occasions during the course of the study.
• The results (data) is given below:

Operator 1 Operator 2 Operator 3 Operator 4
Part # Reference Rep 1 Rep 2 Rep 3 Rep 1 Rep 2 Rep 3 Rep 1 Rep 2 Rep 3 Rep 1 Rep 2 Rep 3
1 A R R R A A A A A A A A A
2 A A A A A R A A A A A A A
3 A A A A A A A A A A A A A
4 A R A A R A A R R R A A A
6 A A A A A A A R R R A A A
8 A A R A A A A A A A A A A
10 A R A A A A A A A A A A A
12 A A A A A R A A A A A A A
13 A A A R A R A R R A A A A
14 R R R R R R R A A A R R R
15 R R R R R R R R R R R R R
19 R R A R R R R R A A R R R
22 R R R R R R R A R A R R R
23 R R R R A R R R R R R R R

MSA EXERCISE
• Using SPC XL, set up an MSA template and enter the data
from the previous page.
• Analyze the data.
• Based on your analysis, what do you conclude about the

overall effectiveness of the credit approval process in making
the “correct” decision?
• Does there appear to be significant variability from analyst to

analyst?
• Is there a bias toward either “good” applications being

rejected or “bad” applications being approved?
• Any other observations or comments or recommendations?

DESIRED PROPERTIES
OF ANY MEASURING SYSTEM
(Continuous Data)
1. Accuracy - Ability to produce an average measured
value which agrees with the true value or
standard being used
2. Precision - Ability to repeatedly measure the same

product, transaction or service and obtain
the same results
3. Stability - Ability to repeatedly measure the same

product, transaction or service over time
and obtain the same average measured
value
ACCURACY
(average)
Yes No
•• ••
• • •
• • •
•
•
•• •
Yes
•
• •
• • •
(repeatability)
PRECISION
•
•
•
• •
•
•
• •
•
• •
• •
•
•
• •
•
•
• •
•
No
• •

HOW TO PERFORM A
(Continuous Data)
PURPOSE:
To assess how much variation is associated with
the measurement system and to compare it to the
total process variation or tolerances.
Q#
4
σ 2 total = σ 2product + σ 2measurement
σ 2repeatability + σ 2reproducibility
REPEATABILITY:
Variation obtained by the same person using the
same procedure on the same product, transaction
or service for repeated measurements (variability
within operator).
REPRODUCIBILITY:
Variation obtained due to differences in people
who are taking the measurements (variability
between operators).

WHY REDUCE MEASUREMENT ERROR?
σproduct
LSL USL
σtotal
Due to large σmeasurement, the area is the proportion

concluded to be defective, when in fact only is truly
faulty.
LSL USL
σmeasurement
σtotal
±3σmeasurement ±3σmeasurement
Shaded area represents the proportion of output subject

to misclassification due to measurement error.

THINKING ABOUT REPEATABILITY
AND REPRODUCIBILITY
For each of the scenarios below, look at the data and indicate which of
the following you think is true:
a. Repeatability appears to be more of a problem than reproducibility
b. Reproducibility appears to be more of a problem than repeatability
c. Repeatability and Reproducibility appear to be about the same
• Scenario 1:
Operator 1 Operator 2
Rep 1 Rep 2 Rep 1 Rep 2
Part 1 21 23 26 28
Part 2 19 18 24 24
Part 3 20 23 27 24
Part 4 19 22 21 20
• Scenario 2:
Part 1 21 23 24 22
Part 2 19 18 21 18
Part 3 20 23 24 22
Part 4 19 22 22 20
• Scenario 3:
Part 1 21 23 25 23
Part 2 19 18 15 17
Part 3 20 23 23 25
Part 4 19 22 18 20
MSA will help quantify, more exactly, the capability of the

measurement system and answer questions about
repeatability, reproducibility, and capability with respect to
the customer specs.

CREATING AN MSA DATA TEMPLATE
• Select SPC XL > MSA > Create MSA Template from

menu
• Specify Number of Operators (for this example, enter 2)
• Specify Number of Replicates (for this example, enter 2)
• Specify Number of Parts (for this example, enter 4)
• Enter Specification Limits, if any (for this example, enter 35

and 10)
• Press OK and the following template will appear.
For Attribute data enter A for

MSA Data Template
Accept and R for Reject
Date: 8/30/01
Description:
Part Type:
USL: 35.0
LSL: 10.0
Part # Reference Rep 1 Rep 2 Rep 1 Rep 2
1
2
3
4

DATA ENTRY EXAMPLE
• Randomizing the parts, operators, and replicates, collect

the measurements and enter the data into the template as
shown below.
• This is the example data shown in scenario 1 on page 6-

16.
• NOTE: This example is for demonstration purposes only.

We are violating the R.O.T. which says (Number of
operators) • (Number of parts) > 20. We are limiting the
number of parts to 4 and the number of operators to 2 in
this case to keep the example reasonably small to make it
easier to follow along.
For Attribute data enter A for

MSA Data Template
Accept and R for Reject
Date: 8/30/01
Description:
Part Type:
USL: 35.0
LSL: 10.0
Part # Reference Rep 1 Rep 2 Rep 1 Rep 2
1 21 23 26 28
2 19 18 24 24
3 20 23 27 24
4 19 22 21 20

CONTINUOUS DATA
ANALYSIS OPTIONS
X – R Analysis –select this option from SPC XL > MSA
• Can be done with only 1 operator or 1 replicate
• Simple, straightforward and could be done without software,

though not recommended
• Cannot estimate (part • operator) interaction contribution to

reproducibility
• Highly sensitive to outliers
• Variance is estimated using converted ranges
ANOVA (Analysis of Variance) –select this option from SPC XL >

MSA
• Must have at least 2 operators, 2 replicates and 2 parts
• Mathematically complex so that software is necessary
• Can estimate (part • operator) interaction
• Is much less sensitive to outliers than the X – R method
• More precise estimates of variance than the X – R method

provides
Recommended Approach: ANOVA, if possible (need at least 2

parts, 2 operators, and 2 replicates)

COMMONLY USED MEASURES TO
EVALUATE MEASUREMENT SYSTEMS
1. Precision-to-Tolerance Ratio (P/TOL)
P/TOL = 6σmeas (Specification Limits are needed)

USL − LSL
ROT: If P/TOL ≤ .10 : Very Good Measurement System

P/TOL ≥ .30 : Unacceptable Measurement System
2. Precision-to-Total Ratio (P/TOT)
σmeas
P/TOT =
σ total
ROT: If P/TOT ≤ .10 : Very Good Measurement System

P/TOT ≥ .30 : Unacceptable Measurement System
3. Discrimination or Resolution
(# of truly distinct measurements that can be obtained by the
measurement system)
⎛ σproduct ⎞
=⎜ ⎟ × 1.41
⎝ σmeas ⎠
ROT: Resolution ≥ 5 represents an adequate measurement

system

SPC XL ANALYSIS
(Interpreting Results)
• For our example data, what is the recommended
analysis method and WHY?
• Let’s perform the analysis

– Choose SPC XL > MSA > ANOVA Analysis
MSA ANOVA Method Results

Source Variance Standard Deviation % Contribution
Total Measurement (Gage) 10.0625 3.172144385 83.71%
Repeatability 2.3125 1.520690633 19.24%
Reproducibility 7.75 2.783882181 64.47%
Operator 5.79166667 2.40658818 48.18%
Oper * Part Interaction 1.95833333 1.399404635 16.29%
Product (Part-to-Part) 1.95833333 1.399404635 16.29%
Total 12.0208333 3.467107344 100.00%
USL 35
LSL 10
Precision to Tolerance Ratio 0.76131465
Precision to Total Ratio 0.91492535
Resolution 0.6
BIAS ANALYSIS
Reference Bias
Not Available
• P/Tol is too high

• P/Tot is too high
• Resolution is unacceptable
• Reproducibility is significantly larger than Repeatability
and appears to be the biggest problem with this
measurement process

INTERPRETING ANOVA
TABULATED RESULTS
Measurement System Variance Components
Z Axis 4
Interaction
Operator
3
Reproducibility
Repeatability
Part-to-Part
X Axis Category

INTERPRETING ANOVA
TABULATED RESULTS
MSA- Xbar Chart
30
25
20
Part Average
Operator 1
Operator 2
15 UCL = 25.963
Center = 22.438
LCL = 18.913
10
0
1 2 3 4 1 2 3 4
Part Number

INTERPRETING ANOVA
TABULATED RESULTS (cont.)
MSA- Range Chart
Operator 1
Part Range
4
Operator 2
UCL = 6.126
3
Center = 1.875
LCL = .
0
1 2 3 4 1 2 3 4
Part Number

INTERPRETING ANOVA
Operator By Part
30
25
20
Measurement
Operator 1
15
Operator 2
10
0
1 2 3 4
Part #

INTERPRETING ANOVA
Sigma Product vs Sigma Total
Sigma Total
Sigma Product
LSL
USL
8.27 13.27 18.27 23.27 28.27 33.27

Measurement

INTERPRETING ANOVA
Misclassification Due To Measurement Error
dpm Potentially
Misclassified = 389,571.196
Sigma Total
LSL Sigma Meas
USL Sigma Meas
LSL
USL
0.48 5.48 10.48 15.48 20.48 25.48 30.48 35.48 40.48

Measurement

INTERPRETING X - R
METHOD RESULTS
Results from X - R analysis, although not as statistically accurate, do
not differ a great deal from the ANOVA analysis in this case and are
shown below. Only one analysis (ANOVA or X - R) need be done
when evaluating a measurement system. ANOVA is always the
preferred method if it can be done.
MSA XbarR Method Results
Source Variance Standard Deviation % Contribution
Repeatability 2.5682117 1.602564103 21.32%
Total 12.0456006 3.470677255 100.00%
USL 35
LSL 10
Resolution 0.8
BIAS ANALYSIS
Reference Bias
Not Available
4
Z Axis
Reproducibility
Part-to-Part
XAxis Category Repeatability

POOR MEASUREMENT SYSTEM
CAPABILITY ACTIONS
1. If repeatability is the dominant source of variability, it is an indication
every person is having trouble measuring, indicating inadequate
SOPs or equipment that needs to be repaired, replaced, or adjusted.
2. If reproducibility is the dominant source of variability, we should

examine the differences between operators to determine whether it
is a training, skill, or following an SOP type of problem. Inadequate
SOPs or failing to follow SOPs could be an issue.
3. If P/TOL > .30, evaluate the specification limits. Are they

reasonable?
4. If P/TOT >.30, check the items that were part of the measurement
system study. Do they represent at least 80% of the total process
variability? If they do and if the measurement system is already
state-of-the-art and performing up to its specifications, we may have
to live with it. One possible work-around is to take an average of 4
measurements, say, and use that value as the recorded entry for
that item and that measurement. (note: This involves the property of
the Sampling Distribution of the Mean, an advanced topic, which
says that the standard deviation of sample averages is reduced by a
factor of n , where n is the sample size. If n = 4, 4 = 2 , and so we
have cut the variability in half.) This technique should be used only
as a short-term approach to a gage study and should not be used as
a mask for an inherently poor measurement system.
5. If the measurement system is suspect (P/TOL or P/TOT close to .30)

but the process is operating at high capability (Cpk > 2), then the
measurement system is likely not a problem and we may be able to
continue to use it.
Things to
Check

MEASUREMENT SYSTEM STUDY
(Continuous Data Example/Exercise)
• A Human Resources employee was concerned about the
accuracy of the current process for merit increase evaluation.
She had just recently finished Lean Six Sigma training and
was interested in using the MSA tool to measure the
repeatability and reproducibility of this process.
• She selected 3 managers from different departments to

participate in the study. She created 6 “employee profiles”,
each slightly different, to be evaluated during an on-going
evaluation training study. Two of the 6 profiles represented
“low end” to “average” employees, two represented
“average” employees, and the remaining two represented
“high end” employees.
• The managers were asked to evaluate the employee profiles

and to determine the amount of merit increase each should
receive (maximum allowance was 5%). Each manager
reviewed each of the six profiles a total of 3 times over the
course of several weeks.
• Data is given below:
USL: 5.0
LSL: 0.0
Operator 1 Operator 2 Operator 3

Part # Reference Rep 1 Rep 2 Rep 3 Rep 1 Rep 2 Rep 3 Rep 1 Rep 2 Rep 3
1 3.8 3.6 3.8 3.4 3.8 3.8 5 4.8 5
2 3.8 3.8 3.7 3.3 4.3 3.7 4.7 4.8 4.6
3 2.8 2.8 2.7 2.8 2.9 2.3 2.9 3 3.4
4 2.5 2.6 2.7 2.4 2.7 2.4 2.6 2.7 2.5
5 1.3 1.2 1.2 1 1.5 1.3 1.6 1.6 1.4
6 1.2 1.3 1.3 1.3 0.8 1.2 1.5 1.3 1.6

MSA EXERCISE
• Using SPC XL, set up an MSA template and enter the data from
the previous page. For the purpose of this exercise, use an upper
spec limit = 5.0 and a lower spec limit = 0.0 in order to get metrics
such as P/tol ratio.
• Duplicate the analysis output on the following pages. Analyze the

data.
• Based on your analysis, is there any concern about the

consistency of the merit evaluation process?
• Is there more variation within a manager or between the

managers?
• Which manager, if any, seems to have the biggest problem with

repeatability?
• Any other observations or comments?
• What would you recommend?

RESULTS OF ANOVA ANALYSIS
MSA ANOVA Method Results

Source Variance Standard Deviation % Contribution p Value
Repeatability 0.04166667 0.204124145 2.43%
Operator 0.08228395 0.286851792 4.80%
Oper * Part Interaction 0.05617284 0.2370081 3.28% 0.0001
Total 1.7144321 1.309363242 100.00%
USL 5
LSL 0
Resolution 4.1
BIAS ANALYSIS
Reference Bias
Not Available
1.6
1.4
1.2
Variance 0.8 Interaction

Operator
0.6
0.4
0.2
0
Part-to-Part
Reproducibility
Repeatability
Category

RESULTS OF ANOVA ANALYSIS (cont.)
MSA- Xbar Chart

6
4 Operator 1
Profile Average
Operator 2
Operator 3
3
UCL = 3.028
Center = 2.704
2
LCL = 2.38
0
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
Profile Number
MSA- Range Chart
1.2
0.8
Operator 1
Profile Range
Operator 2
Operator 3
0.6 UCL = .815
Center = .317
LCL = .
0.4
0.2
0
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
Profile Number

Operator By Part
4
Measurement
Operator 1
3 Operator 2
Operator 3
0
1 2 3 4 5 6
Profile #

Sigma Product vs Sigma Total
Sigma Total
Sigma Product
LSL
USL
-1.224 -0.224 0.776 1.776 2.776 3.776 4.776 5.776

Measurement
Misclassification Due To Measurement Error

dpm Potentially
Misclassified = 350,208.941
Sigma Total
LSL Sigma Meas
USL Sigma Meas
LSL
USL
-1.273 -0.273 0.727 1.727 2.727 3.727 4.727 5.727

Measurement

STATAPULT® MSA EXERCISE
• In teams with a statapult®.
• Select one person from your team to be the shooter. The

remaining team members will be “operators”, whose job it is
to measure how far the ball travels in the air to the point of
impact on the floor.
• The shooter will shoot a variety of pull back angles to gain

as much variability in the distance as possible. The range
should vary at least between 60 inches and 170 inches. A
total of at least 10 shots should be taken, each of which will
be measured by all the “operators”.
• No foil or powder or any other material for marking the ball

may be used. The purpose of this exercise is to show the
measurement variability without the aid of marking
techniques.
• The “operators” may not point at the landing spot or speak

to each other during the exercise. They will only view each
shot as they see it and secretly record the distance of each
shot on a separate piece of paper.
• Enter the data from all “operators” into SPC XL (remember

to create a MSA template first) and use SPC XL / MSA /
XbarR analysis to determine the reproducibility amongst the
“operators” and the Precision to Total ratio. Evaluate the
measurement system.

STATAPULT® MSA EXERCISE
Product# Op. 1 Op. 2 Op. 3 Op. 4 Op. 5

1
2
3
4
5
6
7
8
9
10
Total

STATAPULT® MSA EXERCISE 2 (Optional)
• Class as a whole for data collection; Analysis individually

using SPC XL
• Each team will be considered to be an “operator.” Their task

is to measure the circumference (to the nearest millimeter)
of a variety of catapult balls. Each “operator” will measure
each of several products twice. That is, n = 2. Each
“operator” will devise its own measuring technique and
document it with a measurement system process flow
diagram.
• One team will serve as the control group that supplies the
products to each of the “operators” for measurement. The
control group will receive the measured value from the
“operator” and place it in the appropriate cell on the
accompanying (next page) spread sheet.
• Once the data is available, each person will use SPC XL

> MSA to conduct a measurement system study. Evaluate
the measurement system and comment on the repeatability
and reproducibility of the “operators.” Precision-to-tolerance
ratios can only be obtained if specs are given.

STATAPULT® MSA EXERCISE 2 (cont.)
Product# Op. 1 Op. 2 Op. 3 Op. 4

1
2
3
4
5
6
7
8
9
10
Total


MODULE 7
ANALYZING THE CAUSES OF

POOR PERFORMANCE
ANALYZING THE CAUSES OF
POOR PERFORMANCE
– TECHNIQUES FOR IDENTIFYING

POTENTIAL CAUSES OF VARIATION
– cause and effect
– brainstorming
– process observation
– IDENTIFYING WASTE
– 7 classic types of waste
– Cost of Waste
– VALUE AND NON-VALUE ADDED

ACTIVITIES
– ANALYZYING WORK
– Takt time
– Cycle Time
– Operator Loading

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
ANALYZE
Define

Measure
Gage for
Ma

Au
nu
Tire Press.
Tire Press.
to
al
Estimate of
Ra
Re

Front or
All
dia
gu
WX

lar
l
Engine Warm-
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type
Passengers Snow
Terrain
Pr
A
Rain
Hu
ec
Te
ipi
mi
s
Wi
mp
ain
tat
dit
e
t
Fla
nd
tud
era
ion
y
unt
Driver
tur
Alti
Mo
Fuel Additive
Acctng
HR
Legal
Functional
Area
Benefits
problems
Frequency or
Financial Measure

potential causes
Improve

Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
GAP BETWEEN CURRENT
PERFORMANCE AND GOAL
• Display / Summarize in a simple fashion

– graphs
– tables
• Make it visible to the team!
• Examples
Cycle Time for Equipment Changeover
70
60
Cycle Time (minutes)
50
40
30
20
Goal: 20 minutes
10
0
6-Aug
13-Aug
20-Aug
27-Aug
3-Sep
10-Sep
17-Sep
24-Sep
1-Oct
8-Oct
15-Oct
22-Oct
2-Jul
9-Jul
16-Jul
23-Jul
30-Jul
Week of
Measure Baseline Goal
cycle time 53 min. 20 min.
dpm 1,000 500

IDENTIFYING POTENTIAL CAUSES
• Where’s the gap? What’s the problem?

– cycle times too long
– changeover times too long
– excessive variation
– high defect rates
– excessive waiting
– etc.
• As a team, identify the potential causes of the

problem
• Sources of information
– brainstorming
– observation
– process map(s)
– discussion with operators, team, etc.
– data

THE CAUSE AND EFFECT DIAGRAM
• Brainstorm all the possible variables (or causes) that the

team believes may affect the output (or contribute to the
problem)
• Helps identify data that may be needed
Measurement Manpower Materials
Causes or
Sources of
Variation
Effect or
Response
(CTC)
or problem
statement
Miscellaneous Methods Machines involving the
CTC

THE CAUSE AND EFFECT DIAGRAM
(Example)
What are the possible causes of variation in gas mileage?
MEASUREMENT
MEASUREMENT METHOD
METHOD MACHINE
MACHINE
M
Gage for
an
Au
Tire Press.
ua
to
Tire Press.
l
Estimate of
R
R
Al
eg
ad
Front or
ul
i
W
al
Rear Drive
ar
Speedometer
X
Engine Warm-
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Oil Type Ice
Passengers
Pr
Terrain Snow
ec
H
A
Te
B
Rain
um
ip
m
it a
W
id
pe
ns
in
t io
ity
at
ra
d
tai
ud
n
Fl
tu
Driver
un
re
tit
Mo
Al
Fuel Additive
PEOPLE
PEOPLE MATERIALS
ENVIRONMENT Gary Greenbelt
Update: 2/4/2004
This is a living document

which should be reviewed
and updated.

PARTITIONING THE VARIABLES
C = Constant
• To hold a variable Constant requires Standard Operating
Procedures (SOP’s) and Mistake Proofing to eliminate
errors and reduce variation in the CTCs.
• Holding a variable constant doesn't just happen. It must
be engineered into the process.
N = Noise
• Noise variables are those that are not being controlled or
held constant.
• Mistake Proofing is needed to change an "N" variable to
a "C" variable.
X = Experimental
• These are key variables that can be controlled and held
constant at different levels or settings for the purpose of
determining the effect of this variable on the CTC.

STEPS FOR CREATING A CAUSE AND
EFFECT DIAGRAM
Method 1
1. Describe the problem or CTC (output) being discussed and place
it at the “head” of the cause and effect (or fishbone) diagram
2. Determine appropriate* cause categories (the major “bones”)
3. Brainstorm potential causes (or sources of variation) within each
category, which are believed to affect the problem or CTC, and
place on the diagram
Method 2
1. Describe the problem or CTC (output) being discussed and place
it at the “head” of the cause and effect (or fishbone) diagram
2. Brainstorm potential causes (or sources of variation) and place
on sticky notes
3. Cluster the causes into related groups and name the groups.
Use the group names as the major “bones” on the cause and
effect diagram and place the causes within each group
*Typical cause categories (“bones”) for production processes

• the 6M’s: manpower, machine, materials, methods, measurement, Mother
Nature (environment)
*Typical cause categories (“bones”) for non-production
processes
• procedures/policies, people, materials, equipment, environments, other
*Be flexible! Use what makes sense for your process. For
example, you might find it easier to use process steps for
the “bones” . There is no one best way.
Goal is to get as many variables as possible identified.

More variables = More likely to discover the real root causes

BRAINSTORMING
• Structured:
– Go around the room and have each team member,
in turn, give an idea
– Accept each idea without analysis or judgment
– If a person is out of ideas, they may pass their turn
onto the next person
– Continue, around the room, until everyone is out of
ideas
– As a group, review all the ideas. This provides an
opportunity for clarification and to discard duplicate
ideas. Often, the review may stimulate the
generation of new ideas.
• Unstructured:
– Identical to the previous process except that ideas
can be given at any time by anyone without the
need to follow a structured rotation of the team
members
Structured brainstorming tends to maximize participation and

generates more ideas, while unstructured brainstorming is
quicker to use. Sometimes a combination of the two is useful.

OBSERVING THE PROCESS
AND IDENTIFYING WASTE
• As you and your team analyze the process and
brainstorm causes of problems, watch what really
happens and document what is observed
• Observations can be made on a special data

collection form designed by the team, or simply
annotated on the process map
• Typical data and/or observations to record include

things such as:
– What kind of waste is present?
– How long does a particular step take?
– What problems or issues were observed?
– What defects were observed?
– Any opportunities for hard and/or soft benefits?
Sample Process Observation Data Collection Form

Step Description Start and Type of waste, if Number Quantity Improvement
stop time any, observed of people processed ideas
1
2
3
4
5

PROCESS OBSERVATION (EXAMPLES)
• Simple data collection forms, such as the following,

make process observation much easier and insures that
the right information is collected

SEVEN CLASSIC TYPES OF WASTE
• Overproduction
• Waiting
• Transportation
• Inventory
• Processing
• Motion
• Defects

SEVEN CLASSIC WASTES VS
ACTUAL VALUE ADDED
Actual value added time that a customer is

concerned about is usually only 2 to 5% of the
total time manufacturing a product
Actual
Value
Transportation Added Wasted
Motion
Processing
Product
Defects
Excess
Inventory
Waiting Time Overproduction

THINKING DIFFERENTLY ABOUT
COST OF WASTE
The shop floor supervisor thinks

in terms of equipment, people
and production.
Accounting and purchasing

departments think in terms
of costs, expenses, and
ROI (dollars).
$$$
Senior management thinks

in terms of market share
and company growth.
$$$hares

OVERPRODUCTION WASTE
Producing more than demand requires. Or,
using more area, space, equipment, etc. than is
required to perform a process. (sometimes
referred to as “excess resources” waste)
Examples
– Preparing reports that are not reviewed or used or
acted upon
– Producing unneeded copies
– Unnecessary data storage
– Purchasing 16 web servers when only 8 are needed
– Employing 32 underwriter when only 18 are required
for forecast estimates
Work Examples? / What can you do?/ Improvement

Tactics?

OVERPRODUCTION WASTES CAUSE…
Extra Handling
Extra Inventory Extra Space
Extra Machinery Negative Extra Interest

Results Charges
Extra People Extra Overhead
Extra Defects

WAITING WASTE
Idle people or machines waiting on the

preceding or following operation,
materials, or information.
Examples
– An aircraft waits for de-icing before it can take off
– A waiter must wait before he can process a credit card
– An underwriter must wait for a credit check before
completing a loan application
– Items are processed monthly, not as work comes in

Tactics?

TRANSPORTATION WASTE
Excess movement of materials or information that does not

add value for the customer. Extra steps in the process /
distance traveled.
Examples
Hiring Process – Applicants folder is
hand carried from one department
to another for approval.
Before Receiving Area

HR Legal
DOCK
Functional Benefits
Area
After
Warehouse
HR Legal
Functional Benefits
Area

Tactics?

INVENTORY WASTE
An excess of raw materials. Inability to maintain

throughput and manage flow. Excessive
inventory hides production and service problems
and does not add value for the customer.
Examples
– Procurement orders 250 lbs of steak when only 100 lbs is
required
– Hospital orders enough gauze pads to last for 17 months
– Multiple copies of a document are stored in every desk
Work Examples?/ What can you do?/ Improvement

Tactics?
“Inventory velocity has become a passion for us. To achieve maximum

velocity, you have to design your product in a way that covers the largest
part of the market with the fewest number of parts. For example, you don’t
need nine different disk drives when you can serve 98% of the market with
only four.”
- Michael Dell

PROCESSING WASTE
Requiring additional but unnecessary

process steps (sign-offs, etc.) Making
parts better than our customer demands,
or adding unnecessary features or steps
that mask defects. Additional steps are
sometimes added over time as a knee- -Two cooling devices
jerk reaction to an isolated problem. - Six legs
- Two cup-holders
Example
Document Change Approval Process
Before After
Change Request Change Request
Technical Review Draft Approved
Initial Draft Created Final Document

Created
Draft Approved Final Approval
Final Document
Created
Final Approval

Tactics?

MOTION WASTE
Excess movement, walking, or handling required

to perform a task
Examples
– Bank teller who must walk from her primary location to an
alternate location in order to deposit payroll taxes
– A waiter who must go to two different locations to place food
and drink orders

Tactics?
COST $28,000
ROI
3.2 months
Kaizen Team Changes And $95,000
More Sellable
Added Rotating Table Goods/Yr.
This is GREAT!!
$3,000/once
Added 2nd Person

$25,000/yr
Reduced time by
62%
Increased Production
$95,000/yr

DEFECTS (REWORK) WASTE
Processes which must be

accomplished/re-accomplished
because of errors. Items that are
defective and thrown away, or re-
worked at no cost to the customer.
Rework ties up capacity.
“Rework is free work!”
Quote from a line worker in a
Examples steel mill
– Waiter makes an error with a food order

resulting in wasted food and another meal being
prepared
– Customer omits information on a loan request
resulting in sending the application back to the
customer and reprocessing

Tactics?

THE 7 WASTES CAN BE FOUND
EVERYWHERE!
Examples by Area:
Engineering Software Marketing Administration

Waste:
Overproduction • Producing more • Producing more sets • Printing more • Producing unused or
prototypes than needed of documentation marketing material unneeded reports
than needed than needed
Waiting • Waiting for information • Waiting for test • Waiting for people to • Waiting for purchase
from a supplier scripts to be complete reviews requisition approval
• Waiting for available test developed • Waiting for information
equipment
Transportation • Delivery of test parts from • documentation loops • Paperwork and • Delivery of paper and other
a supplier documentation loops materials to a copy center
Inventory • Stockpiling more test • Pending requests for • Marketing materials • Multiple copies of a
parts than required software tests not mailed document stored in every
• Excess test equipment desk
Processing • Adding features to a • Developing • Excess approval • Project plans needing

design that the customer application features steps and/or handoffs reviews and signatures
does not want that the customer from 18 people
doesn’t want
Motion • Extra walking and • Having to perform • Searching for people • Walking back and forth
reaching into bins to extra keystrokes to to get feedback from stockroom to get
retrieve test parts accomplish a task paper for a copy machine
Defects • Incorrect tolerances • A logic error in coding • Not supplying the • An inaccurate report which
• Parts which do not meet correct information to must be corrected
specification the customer
What types of waste do you see in your everyday work?

THE 7 WASTES ARE CONTRIBUTORS
TO THE COST OF POOR QUALITY
Where Does the Cost of Poor Quality

(Waste or Muda) Come From? …
Internal Failure Costs
External Failure Costs
Appraisal Costs
Prevention Costs
Lost Opportunity Costs

THE COST OF POOR QUALITY (cont.)
INTERNAL FAILURE COSTS
Failure costs occurring prior to product/service

delivery or shipment.
Some examples:
• Over-Production
• Waiting
• Transportation The 7 Classic Wastes
• Inventory
• Processing
• Movement
• Defects
• Design Corrective Action
• Rework-Design Change
• Scrap-Design Change
• Production Liaison Costs
• Purchasing Failure Costs
• Purchased Material Replacement Costs
• Supplier Corrective Action
• Rework of Supplier Rejects
• Uncontrolled Material Losses
• Material Review/Corrective Action
• Disposition Costs
• Troubleshooting/Failure Analysis
• Investigation Support Costs
• Operations Corrective Action
• Failure to Transfer from Design to Manufacturing
• Rework
• Repair
• Reinspection/Retest Costs

EXTERNAL FAILURE COSTS
Failure costs occurring after product/service delivery

or shipment to the customer
Some examples:
• Customer Complaint
Investigation
• Returned Goods
• Retrofit Costs
• Recall Costs
• Warranty Claims
• Liability Costs
• Penalties
• Customer/User Goodwill
• Premium Freight (overnight)
• Pricing Errors

APPRAISAL COSTS
Costs associated with measuring, evaluating or auditing
products or services to ensure conformance to quality
standards and performance requirements
Some examples:
• Document Checking
• Purchasing Appraisal Costs
• Receiving/Incoming Inspection and
Test
• Qualification of Supplier Product
• Source Inspection and Control
Programs
• Manufacturing Appraisal Costs
• Planned Inspections, Tests, Audits
• Financial Audits
• Checking Labor
• Inspection and Test Materials
• Set-Up Inspections and Tests
• Depreciation Allowances
• Measurement Equipment Expense
• Maintenance and Calibration Labor
• Outside Endorsements and
Certifications
• External Appraisal Costs
• Field Performance Evaluations
• Special Product Evaluations
• Review of Test and Inspection Data
• Other Quality Evaluations
• Special Tests
• Process Control Measurement
• Laboratory Support
PREVENTION COSTS
The costs of all activities specifically designed to prevent
defects (poor quality) in products or services. These occur
prior to production and are aimed at preventing defects
before they occur.
Some examples:
• Marketing Research
• Customer Perception Surveys
• Contract/Document Review
• Design Quality Progress Review
• Design Support Activities
• Product/Service Design Qualification
Test
• Field Trials
• Supplier Quality Reviews/Ratings
• Purchase Order Technical Data Review
• Supplier Quality Planning
• Operations Prevention Costs
• Operations Process Validation
• Operations Quality Planning
• Operational Support Quality Planning
• Operator Quality Education
• Operator SPC/Process Control
• Quality Administrative Salaries and
Expenses
• Quality Program Planning
• Quality Performance Planning and
Reporting
• Quality Education
• Quality Improvement
• Quality System Audits
• Training without ROI
LOST OPPORTUNITY COSTS
Costs of missed opportunities.
Some examples:
• Delayed Market Entry
• Lost Sales
• Lost Customers
• Additional Shifts, Lines, Plants Due to Poor Productivity
(i.e., Low Capacity)
• Failure to Enter a New Market Because of …
• Competitor Obtains Patent Rights Before We Do
• Reliability Issues
• Safety Issues
• Ergonomic Issues
• Legal Issues

ACTIVITY TYPES
When reviewing your process maps, look for

and identify these 3 types of activities
Value Added Non-Value Added Non-Value Added

Value Added Non-Value Added Non-Value Added
Activities But Required Activities
Activities But Required Activities
• Activities where • Activities causing • Activities which

• Activities where • Activities causing • Activities which
the product or no value to be consume time
the product or no value to be consume time
service is created but and resources but
service is created but and resources but
transformed into a which cannot be create no value in
transformed into a which cannot be create no value in
state required by eliminated based the eyes of the
state required by eliminated based the eyes of the
the customer. on current state customer.
the customer. on current state customer.
of technology or
of technology or • • Pure
PureWaste
• Activities which,
• Activities which,
when asked, the
thinking.
thinking. Waste. .
when asked, the • Example –
customer is willing • Example –
customer is willing Regulatory
to pay for. Regulatory
to pay for. requirements
requirements

VALUE ADDED VS. NON-VALUE ADDED
• Which processes are value added in the eyes of the

customer?
• Any activity that transforms or converts information or

raw materials to meet customer requirements is
value added!
Time Value Map
Traditional efforts focus on increasing
95% Non-Value Added 5% Value Added
Lean tools focus on identifying non-value added wastes

EXERCISE
• In small groups
• Consider the process flow, shown on the next page, which

depicts the process of going to the grocery store to pick up a bag
of ground coffee.
• Review each step and identify the value added (VA) steps. (all
others are non-value added (NVA) or non-value added but
required (NVA-R).
• What examples of waste do you observe?
• On a process observation sheet, what additional data might be

useful to collect?
VA time
• Calculate the process cycle efficiency: =
Total Lead Time

PROCESS FLOW
10am 10 minutes 0.5 minutes 1 minute
drive to the drive into the find a parking
start grocery store parking lot space
1 minute
0.5 minutes 0.1 minutes 0.05 minutes
move any
park the car get out of car lock the car shopping carts
in my way
1 minute 0.5 minutes

1 minute talk with kids 1 minute
proceed into
walk to the selling items in find the coffee
the grocery
grocery store front of the aisle
store
store
1 minute 0.1 minutes

0.5 minutes 2 minutes
locate the secure the bag
walk to the wait in the
coffee of my of ground
checkout checkout lane
choice coffee
1 minute
0.5 minutes
1 minute 1 minute
talk with kids
exit the
checkout and selling items at walk back to
grocery
pay for coffee the door of the car
store
store
1 minute
move 0.1 minutes
someones 0.05 minutes 0.1 minutes
put coffee bag
shopping cart unlock car get into car
in the car
away from my
car
0.5 minutes
0.5 minutes 10 minutes 10:36 am
wait for other
cars to move drive out of the drive back
from behind parking lot home
finish
me

ANALYZING WORK
• In our processes, we must challenge ourselves to:

– analyze the work being done
– understand the types of activities being performed

(VA, NVA, NVA-R)
– strive to eliminate or reduce the non-value added

work
• Some of the tools for analyzing work include:

– process observation
– process flow / maps
– physical flow diagrams
– time value map
– brainstorming
– Takt time

ANALYZING WORK (TIME ANALYSIS)
• TAKT time is the time allocated to produce one part,

based on time allocated to the production of that part and
the demand rate. It is the minimum pace that must be
kept to meet customer demands.
• TAKT time is
• determined by the customer
• important for understanding work balance and flow
• is NOT cycle time!
Total Available Production Time

TAKT Time =
Demand for the product (in units)
• Available production time should consider the length of

shifts, number of shifts, meals, breaks, personal time,
setup, cleaning time and any other time which interferes
with the actual production of the part.

TAKT TIME EXAMPLE #1
Assume:
One 8 hour shift
One 1 hour lunch break
Average customer demand 400 units per day
3,600 sec
Total Work Time = 8 hours • = 28,800 sec
hour
3,600 sec
Non-Working Time = 1 hour • = 3,600 sec
hour
Total Available
28,800 sec – 3,600 sec = 25,200 sec
Production Time =
25,200 sec
TAKT Time = = 63 sec/unit
400 units
• In order to meet customer demand, we must produce a

functional part every 63 seconds!

TAKT TIME EXAMPLE #2
Assume:
Two 8 hour shifts
Each shift is given 1 hour for meals
Each shift is given 3 breaks @ 10 minutes each
Average customer demand 200 units per day
3,600 sec
Total Work Time = 2 • 8 hours • = 57,600 sec
hour
Non-Working Time = 1 hour for meals
+ 30 minutes in breaks
= 90 minutes per shift
• 2 shifts
= 180 minutes per day • 60 sec = 10,800 sec
minute
Total Available
Production Time = 57,600 sec – 10,800 sec = 46,800 sec
46,800 sec
TAKT Time = = 234 sec/unit
200 units

TAKT TIME EXERCISE
• Two 8 hour shifts

• Shift #1 is given one hour for lunch and 2 15-minute
breaks
• Shift #2 is not given a meal break, but is given 4 15-
minute breaks
• It takes Shift #1 15 minutes to prepare the equipment in
the morning for operation. It takes shift #2 30 minutes to
shut down and clean up at the end of the shift.
• Customer demand is 3,000 units per month
• There are 20 workdays in the current month
Calculate daily TAKT time (in seconds)

WHY IS TAKT TIME IMPORTANT?
• Takt time represents the rate at which completed units
must come off the production line to keep pace with
customer demand.
• Takt time is the foundation for planning and

understanding work balance and flow. It tells us the
targeted work content time for completing an operation
within the process.
• Takt time is a key measure used to determine the

number of operators, machines, loading, etc.
Takt time = 50 seconds

cycle time
Op 1 Op 2 Op 3 Op 4

UNDERSTANDING CYCLE TIME
• Cycle time is NOT TAKT time
• Three types
– Operator Cycle Time
– Machine Cycle Time
– Processing Cycle Time

OPERATOR CYCLE TIME
• The amount of time required for a single operator to

complete one full iteration of his/her work cycle
Machine #3 Machine #2 Machine #1 Inbound
Step A Operator Step A

#2
Machine #4 Step B
Operator #3 Operator #1
Step C Step B Step C
Machine #5 Machine #6 Machine #7 Outbound
• If it takes Operator #1 82 seconds to perform operations at

Machines #1, #6, #7, and return to Machine #1, then
Operator #1 cycle time is 82 seconds.
• If it takes 78 seconds for Operator #2 to perform operations
at Machine #2, then Operator #2 cycle time is 78 seconds.
• If it takes Operator #3 90 seconds to perform operations at
Machine #3, #4, #5, and return to Machine #3, then
Operator #3 cycle time is 90 seconds.

DETERMINING THE NUMBER OF
OPERATORS REQUIRED
Cumulative Operator Time

Total Operators Required =
TAKT time
From previous example:

Cumulative operator time:
Operator 1 = 82 seconds
= 250 seconds total
Assume TAKT time = 81 seconds
250 sec
Total Operators Required = = 3.09
81 sec
ROT for Partial Operator

Portion of
Operator ROT
(fraction)
Do not add another operator. Look for waste and/or assign
<.25
project to reduce time
Consider adding another operator. First determine if a quick

.25-.4
project can resolve timing.
Add another operator, but look to eliminate with continuous

>.4
improvement
OPERATOR BALANCE (LOAD) CHART
• Comparison of operator cycle time to TAKT time
• Identifies where opportunities exist for improvement
• Identifies problematic areas
Seconds
100 90 sec
82 sec 78 sec
80 TAKT Time = 81 sec
60
40
20
Operator A Operator B Operator C
Goal: Want cycle time for every operator to be less

than or equal to Takt time

OTHER EXAMPLES
100 TAKT Time
80
60
40
20
Operator A Operator B Operator C Operator D
100 TAKT Time
80
60
40
20
Operator A Operator B Operator C Operator D

“Least Operator”
100 TAKT Time
80
60
40
20
Operator A Operator B Operator C

OPERATOR LOAD CHART EXERCISE
• Given the process map shown below and a TAKT time
of 20 min./transaction, draw the operator load chart and
record your observations and recommendations.
Op. 1 Op. 2 Op. 3 Op. 4

assemble analyze prepare update
info info response records
cycle time: cycle time: cycle time: cycle time:

5 min. 22 min. 8 min. 4 min.
• Load Chart
• Observations and Recommendations

TIME OBSERVATION FORM
Process: Date: Plant:
Avg.
Step No. Description 1 2 3 4 5 6 7 8 Max. Min. Comments
Time
1 Pick up part

2 Attach item #1
3 Attach item #2
4 Pass to next station
7 - 47
TOTAL
HOW TO COMPUTE
OPERATOR CYCLE TIME
• Select an “average” operator
• Perform 7 cycle analysis – observe the operator perform the

task at least 7 times
• Do not include measurements with special cause variation.
• Be aware of the Hawthorne effect
• If you decide to sample multiple operators, use one form per

operator
• Pick logical steps to break up the process work

Process: Cumulative time Date: Plant:
Avg.
Time
1 Pick up a part 1 27 1:00 1:29 2:06 2:34 3:03

2 Attach item #1 11 40 1:10 1:47 2:19 2:45 3:13
3 Attach item #2 24 57 1:23 1:59 2:31 3:00 3:28
4 Pass to next station 25 59 1:26 2:04 2:32 3:01 3:31
7 - 49
TOTAL
Cumulative time
Process: Date: Plant:
Avg.
Time
1 27 1:00 1:29 2:06 2:34 3:03
1 Pick up a part 3 1 1.86
1 2 1 3 2 2 2
11 40 1:10 1:47 2:19 2:45 3:13

2 Attach item #1 18 10 12.15
10 13 10 18 13 11 10
24 57 1:23 1:59 2:31 3:00 3:28
3 Attach item #2 17 12 13.86
13 17 13 12 12 15 15
25 59 1:26 2:04 2:32 3:01 3:31
4 Pass to next station 5 1 2.29
1 2 3 5 1 1 3
7 - 50
TOTAL 25 34 27 38 28 29 30 43 24 30.14
Aggressive Very Moderate

Aggressive
WHICH TIME TO USE
• Select the operator cycle time based upon where you are
in process improvement.
Type Calculation When to Use

Very Find the minimum time Use only when
Aggressive required to perform trying to improve an
each step (see min existing cell that has
column). The sum of already been
these minimums is the Kaizened several
very aggressive times.
operator cycle time.
Aggressive Find the time required Use when improving
for the minimum an existing flow
observed cycle. system.
Moderate Calculate the average Recommended only
time required to perform when moving from
each step (see average batch and queue
column). The moderate operations to flow.
cycle time is the sum of
the averages.

MACHINE CYCLE TIME
• The time required for a machine to complete one full cycle
• Measured from start of cycle to when the part is available

and the machine can accept another part
• The intent is to measure how long the machine needs to

cycle a part.
• The time the operator sets up, cleans, or calibrates the

machine is not part of machine cycle time.
• Usually, the time required for one cycle has low variability.
However, if there is a great deal of variability in the
machine cycle time, take the average of seven readings.

PROCESSING CYCLE TIME
• The machine cycle time plus the time required for the
operator to load, unload, start, and perform any tool
change time
• A single operator can participate in multiple processing

cycle times.
Machine #3 Machine #2 Machine #1 Inbound
Step A Operator Step A

#2
Machine #4 Step B
Operator #3 Operator #1
Step C Step B Step C
Machine #5 Machine #6 Machine #7 Outbound
• Example: If it takes Operator #1 10 seconds to unload,

load, and start Machine #1 and Machine #1 cycle time is
47 seconds, then the processing cycle time for Machine
#1 is 57 seconds.

PROCESS CAPACITY CHART
• Comparison of processing cycle time to TAKT time
• Identifies problem areas
100
83 sec
TAKT Time = 81 sec
80
60
Machine Time
40 Operator Time
20
Machine Machine Machine Machine

#1 #2 #3 #4
• How would you interpret this chart?
TAKT Time
Actual Capacity = • Demand
Processing Cycle Time
81
= • 200 = 195
83


MODULE 8
TECHNIQUES FOR
NARROWING THE FOCUS
TECHNIQUES FOR
NARROWING THE FOCUS
– DATA COLLECTION CONSIDERATIONS

– Sampling
– Confidence Intervals and Sample Size
calculations using SPC XL
– TECHNIQUES FOR NARROWING THE

FOCUS WHEN DATA IS LIMITED
– Voting
– Nominal Group Technique
– Effort / Impact Analysis
– Pairwise Comparsion
– IPO (prioritization) matrix
– 5 WHY’s

ANALYZE
Define

Measure
Gage for
Ma

Au
nu
Tire Press.
Tire Press.
to
al
Estimate of
Ra
Re

Front or
All
dia
gu
WX

lar
l
Engine Warm-
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type
Passengers Snow
Terrain
Pr
A
Rain
Hu
ec
Te
ipi
mi
s
Wi
mp
ain
tat
dit
e
t
Fla
nd
tud
era
ion
y
unt
Driver
tur
Alti
Mo
Fuel Additive
Acctng
HR
Legal
Functional
Area
Benefits
problems
Frequency or
Financial Measure

potential causes
Improve

Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
PRIORITIZING AND
NARROWING THE FOCUS
• As the team measures and analyzes the
process, the goal is to identify the key input
variables and process steps to focus on
Radio

Gage for Ride Brakes Tire Rotation
Tire Press.
M
ut
an
o
ua
Estimate of
l
Full Tank Maint Sched
. Tune-Ups Weight Tire Type
Front or
R
A
ad
eg
ll
Type Fast Start Up Time
ia
ul
l
ar
X
Air Conditioner
GAS
Driver
B
Gas Type MILEAGE
A
Highway or City
Training
Weather Conditions
ns
# of Oil Type Ice
tai
e
Passengers Snow
ud
t
un
Rain
Fla
tit
Mo
Pr ity re
Al
H
um
W
ec
Te
Driver
in
ip
id
m
d
ita
pe
tio
Fuel Additive
ra
n
tu
. . . .50 - 100 potential X’s
. . . . . . .30 - 50 potential X’s
. . . . . . . . . . .10 - 30 potential X’s
-- Data
- Graphical analysis
- Statistical tools
. . . . . . . . . . . . . .the critical few

(2 - 10)
Key inputs
and Process steps
(causing problems, affecting cost
and/or performance,
and generating waste)

TOOLS AND TECHNIQUES FOR
NARROWING THE FOCUS
“Not knowing the difference between opinion and

fact makes it difficult to make good decisions.”
B
Marilyn Vos Savant
Columnist in Parade Magazine and
Guinness World Book record holder for
highest IQ
• To help verify root causes / narrow the focus

– Data! (Pareto chart of causes, frequency of
occurrence, scatter plots, etc.)
– In the early stages, if data is unavailable or very

limited:
• Team Voting
• Nominal Group Technique
• Effort/Impact Assessment
• Pairwise comparison
• IPO (prioritization) matrix
• 5 Why’s
Use what makes the most sense for your project!

USING DATA TO NARROW THE FOCUS
(Example: Pareto Charts)
Lean Six Sigma: A Tools Guide Why? What are the
causes?
Address
problems Ship too
Packing early
problems Ship late
shipping
errors
Ship method
Contents
wrong
“Show me the Data!”
DATA
Types of Count Proportion
Shipping Errors (frequency)
Address 11 .11
Ship Early 9 .09
Ship Late 37 .37
Method 6 .06
Contents 8 .08
Packing 29
100
.29
1.00 A Pareto chart
Types of Errors and Frequency of Occurrence
turns data into
information
40
35 focus on the
30 causes that occur
most frequently!
Frequency
25
20
15
10
0
Type Shipping Error
SPC XL > Analysis Diagrams > Pareto Chart

DATA GATHERING, SAMPLING, AND
MEASUREMENTS
• Data gathered should be __________________

of the process being studied.
• Poor approach to sampling

– convenience sampling
• Better approaches to sampling

– random sampling
– systematic sampling
• Before gathering the data, think about the measurement

system being used to generate the data
– is it accurate? precise? repeatable?
Lean Six Sigma requires good, useable data

SAMPLING AND
CONFIDENCE INTERVALS
Population Sample
Parameters Statistics
(μ, σ) ( x , s)
• Suppose we measure the time it takes for a

customer service representative to answer a call.
• We use the data from our sample to estimate the

average call time. How good is our estimate?
• Confidence intervals provide error bounds or an

estimate of uncertainty for a population parameter
based on our sample data.
Confidence Interval =
Sample Estimate ± Margin of Error

PARENT AND CHILD DISTRIBUTION
Sample 1
POPULATION
Sample 2
Sample 3
• We refer to the distribution of individual values from the

population as the “parent” distribution. This can have
any distribution.
• We refer to the distribution of sample averages as the

“child” distribution.
σx
n
X
μx μx
X
• There is an important theorem in statistics, called the

Central Limit Theorem (CLT), which says that no matter
what the parent distribution looks like, as long as the
sample size is big enough, the child distribution (of
sample averages) will be approximately normal

CONFIDENCE INTERVAL FOR POPULATION
MEAN, μ (Continuous Data)
⎛U⎞ ⎛ s ⎞
⎜⎜ ⎟⎟ = x ± Z ⎜ ⎟
⎝ ⎠
L ⎝ n ⎠
Where
U = upper confidence limit
L = lower confidence limit
x = sample average
Z = 2 (for 95% confidence) or
3 (for 99% confidence)
s = sample standard deviation
n = sample size
Computational Template for Confidence Limits
EXAMPLE:
Over the course of a week, we randomly select and time 16
customer service calls. We find that the average time is 15.6
minutes with a standard deviation of 2.1 minutes. What is a
99% confidence interval for the true average service time?

USING SPC XL FOR CONFIDENCE INTERVALS
FOR POPULATION MEAN, μ (Continuous Data)
• SPC XL will produce a more exact interval. Rather than

using Z = 2 or 3 from our 68/95/99 rule of thumb, it uses
the exact values from a t-distribution (similar to Z, but
adjusted slightly to account for the fact that we’re using
sample data to estimate σ)
• SPC XL > Analysis Tools > Confidence Interval > Normal
Normal Confidence Interval (Mean)
User defined parameters

Sample Size (n) 16
Sample Avg 15.6
Sample Standard Dev 2.1
Confidence Level 99.00%
Confidence Interval
Lower Limit Upper Limit
14.05296861 17.14703139
• Exercise: Suppose we sampled 64 calls and got the

same average time and sample standard deviation. How
does this affect our confidence interval?
• Exercise: Suppose we’re studying the average cycle

time to pay invoices. We randomly sample 30 invoices,
and find the average is 7.8 days with a standard
deviation of 1.4 days. Using SPC XL, construct a 95%
confidence interval for the true average cycle time for
paying these invoices.

USING SPC XL TO DETERMINE SAMPLE
SIZE FOR POPULATION MEAN, μ
(Continuous Data)
• Confidence interval calculation can be worked backwards
to determine an appropriate sample size
• Suppose in the previous example, we wanted to estimate

the true average service call time to within ±1 minutes
with 99% confidence
• Steps to determine sample size:

1. Decide on the level of confidence you want, typically
95% or 99% (in our example, 99%)
2. Specify the desired width of the confidence interval
(Upper bound – Lower bound = 2h, where h = half
interval width) (in our example, h = 1)
3. Find an approximation for the population standard
deviation from historical data, small pre-sample, etc.
(in our example, use the previous sample standard
deviation of 2.1 minutes)
4. Select SPC XL > Analysis Tools > Sample Size >
Normal
Sample Size to Estimate the Mean of a

Normal Distribution

Estimated Standard Dev 2.1
Half Interval Width 1
Results
Estimated Sample Size (n) 29.3

CONFIDENCE INTERVAL FOR POPULATION
PROPORTION, π (Binary Data)
⎛U⎞ pq
⎜⎜ ⎟⎟ = p ± Z
⎝L⎠ n
Where
U = upper confidence limit
L = lower confidence limit
p = proportion of “defectives”
(or category of interest) in the sample
q = 1 – p (q is the proportion of “non-
defectives”)
Z = 2 (for 95% confidence) or
3 (for 99% confidence)
n = sample size
Computational Template for Confidence Limits
EXAMPLE:
You work in a finance office and are in charge of processing
travel vouchers submitted by several different organizations.
You sample 100 vouchers and find 8 to have discrepancies
or errors. Find a 95% confidence interval for the true but
unknown proportion of vouchers containing errors.

USING SPC XL FOR CONFIDENCE INTERVALS
FOR POPULATION PROPORTION, π (Binary Data)
• Again, we can use SPC XL to produce a more exact

answer.
• If the number of “defectives” and the number of “non

defectives” are least 5, select SPC XL > Analysis Tools >
Confidence Interval > Binomial
• If either is less than 5, select SPC XL > Analysis Tools >

Confidence Interval > Exact Small p
Binomial Confidence Interval (Proportion)

Sample Size (n) 100
Number Defective(x) 8
Confidence Interval
Lower Limit <p< Upper Limit
0.026827589 0.08 0.133172411
• Exercise: Suppose in the previous example we had

sampled 1,000 travel vouchers and found 80 to have
discrepancies or errors. Find a 95% confidence interval
for the true but unknown proportion of vouchers
containing error.

USING SPC XL TO DETERMINE SAMPLE
SIZE FOR PROPORTIONS (Binary Data)
• Confidence interval calculation can be worked backwards

to determine an appropriate sample size
• Suppose in the previous example, we wanted to estimate

the true proportion of travel vouchers with discrepancies
to within ± 1% with 95% confidence. Also, assume that
the historical proportion of vouchers with discrepancies is
no more than 2%.
• Steps to determine sample size:

1. Decide on the level of confidence you want, typically
95% or 99% (in our example, 95%)
2. Specify the desired width of the confidence interval
(Upper bound – Lower bound = 2h, where h = half
interval width) (in our example, h = .01)
3. Find an approximation for the proportion of interest,
p. It can come from historical data. If no estimates
are available, then use p=.50 to provide a worst case
(conservative) sample size estimate. (in our example,
use 2% (.02))
4. Select SPC XL > Analysis Tools > Sample Size >
Binomial
Binomial Sample Size

Proportion defectives (p) 0.02
Half Interval Width 0.01
Results
Estimated Sample Size (n) 752.9
EXERCISE
Using the data from your sample of M&M’s, use SPC XL to

find the 95% confidence limits for the true (but unknown)
proportion of M&M’s that Mars Corporation produces for
each color.
Brown Yellow Red Orange Blue Green
X
p=
n
Lower
Bound
Upper
Bound
What do you notice about the width of your confidence

intervals?
If you wanted to estimate the true percentage of blue M&M’s

to within ± 8% with 95% confidence, what sample size would
be required? (note: you can use the proportion (p) of blue
above for your estimated proportion of blue M&M’s)

OTHER TOOLS FOR
NARROWING THE FOCUS
• In the early stages, data may be MEASUREMENT
Gage for
METHOD

MACHINE
Carburetor
Tire Press.
Transmission
M
ut
an
o
ua
Estimate of
Weight Tire Type
l
Full Tank Maint Sched
. Tune-Ups
Front or
Engine Warm-
R
ad
Type
ll
Up Time
eg
Fast Start
ia
ul
l
– unavailable
ar
Air Conditioner
GAS
Gas Type MILEAGE
B
Driver
A
Highway or City
Training
Weather Conditions
Al tains
# of Oil Type Ice
e
ud
Passengers Snow
un
tit
t
Rain
Fla
Mo
Pr
Driver
um
ec
in
id
Te
d
ip
ity
m
ita
Fuel Additive
pe
tio
ra
n
B
tu
re
– very limited
– difficult to collect
• Sometimes, data may be available but the team

is not sure where to begin
• In the early stages, if data is unavailable or very

limited, these techniques can help the team
narrow the focus for data collection and further
investigation
• Team Voting
• Nominal Group Technique
• Effort/Impact Assessment
• Pairwise comparison
• IPO (prioritization) matrix
• 5 Why’s
Use what makes the most sense for your project!

VOTING
• Start with Cause and Effect

• Give all team members a certain number of
votes
• Focus on variables/causes with the most
number of votes
Why are we having late
shipments? What
things might be
impacting or causing
Damage from that?
Quality suppliers shipping
issues company
(defects)
Late
Shipments
Address
Database
incorrect
errors
Results after voting

xx
Damage from x
xxxx Quality suppliers
shipping
issues company
(defects)
Late
Shipments
Address
incorrect Database
errors
xxx
x

NOMINAL GROUP TECHNIQUE
Person
Cause Total Ranking
1 2 3 4 5
A: Address Incorrect 4 2 1 2 3 12 #2
B: Quality Issues 1 1 3 1 1 7 #1
C: Damage, Supplier 2 5 4 5 2 18 #4
D: Shipping Company 3 3 2 3 4 15 #3
E: Database errors 5 4 5 4 5 23 #5
• Each person in the group rank orders the set of

causes / variables.
• Rank orderings are summed for each cause
and ranked by their total.

EFFORT / IMPACT ANALYSIS
• Start with the cause and effect diagram
• For each controllable cause, assess the level of

effort required to address it along with the
anticipated impact on project results
1 = low impact with high effort

2 = low impact with low effort
3 = high impact with high effort
4 = high impact with low effort Impact
Low High
Low 2 4
Effort
Required
• Example: High 1 3
1
2
Damage from
Quality suppliers shipping
3 issues company
(defects)
Late
Shipments
Address
4 incorrect
Database
errors 3

PAIRWISE COMPARISON
ct y
re er an
or s li p rs
I n c
sue u pp Com err o
ss Is
e,
S g se
ddre
ality
ag ippin ba
t a
A:
A Qu Da
m
:S
h
Da
B: C: D E:
A: Address Incorrect
B: Quality Issues B
C: Damage, Supplier A B
D: Shipping Company A B D
E: Database errors A B E D
• Each cause is compared against all other causes

• Pick the winner of each comparison (e.g., comparing Cause #E
with Cause #B above, Cause #B was the winner)
• Identifies which cause takes priority in each pairwise comparison
• Count the # of occurrences of each cause in the matrix and bar
chart (Pareto) the results
4
3
2
1
0
B: Quality A: Address D: Shipping E: Database C: Damage,
Issues Incorrect Company errors Supplier

IPO (PRIORITIZATION) MATRIX
• Used previously to help select / prioritize project
opportunities
• Same approach can be applied to identify inputs to focus

on for data collection and further analysis
Primary project
measure (CTC)
INPUTS OUTPUTS
(Requirements) (Critical to Customer)
Supplier quality Percentage of late shipments (10)
Database Accuracy
Shipment Cost (9)
Address information
Shipping
Process Number of Damaged Shipments (8)
Internal Quality issues
Shipping company Percentage of shipping errors (6)
(Customer Priority)
*Rate the relationship for each:

Strong: 9
Moderate: 3
Weak: 1
Nonexistent: Blank or 0
Critical to Customer "Outputs"

(Priority)
Shipment # of Damaged % of Shipping
Late Shipments Cost Shipments Errors
Weighted
(10) (9) (8) (6) Sum
Process "Inputs"
Supplier Quality 1* 1 16
Database Accuracy 9 1 3 117
Address Information 9 1 3 117
Internal Quality Issues 9 1 1 1 113
Shipping Company used 3 9 9 1 189
Total: 552

PRIORITIZING USING THE IPO MATRIX
APPROACH (Class Exercise)
• Suppose our family is trying to decide where to go on

vacation.
• The critical to customer “outputs” (with priority rating)

are
– Low cost (8)
– Lots of different things to do (10)
– Educational experience (6)
– Offer Escape from Reality (9)
• Our family brainstorms the following vacation options:

– Disneyland
– New York City
– Colonial Williamsburg
– Uncle Joe’s in Bowling Green Ohio
• As a class, discuss and complete the IPO matrix below:

(Priority)
Educational Escape
Low Cost Lots to Do Experience Reality
(8) (10) (6) (9)
Weighted
Process "Inputs" Sum
Disneyland
New York City
Colonial Williamsburg
Uncle Joe’s
Total:
OTHER USES FOR THE IPO
(PRIORITIZATION) MATRIX
• Selecting between project alternatives (Define

phase)
(Priority)
Doable in High Apply
2-4 Likelihood across
High ROI months of Success business
( ) ( ) ( ) ( )
Process "Inputs" Weighted Sum
Project #1
Project #2
Project #3
Project #4
Total:
• Deciding between different solutions/options

(Improve Phase)

Put your criteria for
List solution (Priority) decision making here,
alternatives Use Improve along with an
here Low existing Easy to Cycle importance rating
Cost resources Implement Time
( ) ( ) ( ) ( )
Solution #1
Solution #2
Solution #3
Solution #4
Total:

“5 WHY” ANALYSIS
• Simple yet effective tool to develop increased

understanding of an issue
• Involves asking “Why?” (up to) 5 times to uncover the

root cause of the issue (along with other contributing
factors) so that we can eliminate or minimize the impact
of the issue
• Example:
Problem: The cylinder is moving too slowly.
Why is movement too slow?

Because there’s a clog in the strainer.
Why is the strainer clogged?

Because oil is contaminated.
Why is the oil contaminated?

Chips and debris enters through holes and
gaps on the upper plate of the tank.
Solution: Prevent chips from scattering, close

holes and gaps.

“5 WHY” AND CAUSE AND EFFECT
(Partial) Cause and Effect Diagram
Ignition Weather Fuel

No fuel flow
Loose
Fu i l u
connection
cl
og
el r e
fa
Below -200F
ge
pu
d
m
p
Water in fuel
My car
Clutch not won’t start
Bad starter Dead battery depressed Car not in
Loose Park/neutral
connection
Wrong key
Mechanical Electrical Operator
(Partial) 5 WHY Analysis
Left lights on Forgot
Dead
No electrical Battery Never Check; Too busy
power No routine
maintenance
Lazy
Car
won’t
start
Never learned
Clutch not Not familiar or had access
Operator depressed with manual to car with
Error (manual) transmission manual
transmission
Car not in
park or Forgot
Other neutral
branches

EXERCISE (Optional)
• In small groups
• Suppose your statapult® team is working on a project to

reduce the time it takes to complete a statapult® launch.
You know that the average launch time has been
running around 20 seconds (from the time the ball
leaves the cup until the time the next shot leaves the
cup) and you have a goal to reduce that time to 10
seconds.
• You don’t have any historical data on the cycle times for
individual steps in the launch process. Your team has
brainstormed reasons (causes) for why it takes so long
to launch the statapult® ball (shown on the next page)
and you’re not sure where to focus or where to start your
data collection efforts.
• Using one of the techniques just discussed for helping

narrow the focus (voting, nominal group technique,
pairwise comparison, etc.), determine the top 3 causes
your team would focus on.
• What are your top 3?
• Which of the techniques did you use?

EXERCISE (cont.)
• Brainstormed list of causes for excessive launch

times on the statapult®
– too hard to sight in the angle

– distracted by team mates
– ball retriever took too long to return the ball
– measurement process took too long
– didn’t pay attention to cycle time (didn’t know it was
important)
– took time to stabilize the base
– couldn’t find team members when it was their time
to launch
– ball kept falling out of the cup
– had trouble catching the ball to load it


MODULE 9
DRAWING CONCLUSIONS FROM

SAMPLED DATA
DRAWING CONCLUSIONS FROM
SAMPLED DATA
– ELEMENTS OF THE ANALYZE AND

IMPROVE AND PHASE
– USING HYPOTHESIS TESTS FOR

COMPARING DATA SETS
– HOW TO CONDUCT AND INTERPRET

HYPOTHESIS TESTS
• Comparing means
• Comparing standard deviations
• Comparing proportions
• Comparing classification data
– HOW TO APPLY RULES OF THUMB (ROT)

WHEN COMPARING DATA SETS
– USING SPC XL FOR HYPOTHESIS

TESTING

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
ANALYZE
Define

Measure
Gage for
Ma

Au
nu
Tire Press.
Tire Press.
to
al
Estimate of
Ra
Re

Front or
All
dia
gu
WX

lar
l
Engine Warm-
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type
Passengers Snow
Terrain
Pr
A
Rain
Hu
ec
Te
ipi
mi
s
Wi
mp
ain
tat
dit
e
t
Fla
nd
tud
era
ion
y
unt
Driver
tur
Alti
Mo
Fuel Additive
Acctng
HR
Legal
Functional
Area
Benefits
problems
Frequency or
Financial Measure

potential causes
Improve

Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

HYPOTHESIS TESTING
• When searching for causes/variables that may have

a significant effect on the process CTC (in terms of
average, standard deviation, and/or proportion
defective), a statistical hypothesis test can be very
valuable
• Hypothesis Tests can be used to:

– Compare two (or more) sets of data
– Identify the critical causes/variables (Analyze
phase)
– Validate a significant improvement (Improve
phase)

HYPOTHESIS TESTING
• A method for looking at data and comparing results

– Method A vs. Method B
– Option 1 vs. Option 2
– Before vs. After Lean Six Sigma Project results
• Helps us make good decisions and not get fooled

by random variation:
– “Is a difference we see REAL, or is it just random
variation and no real difference exists at all?”
• We set up 2 hypotheses
– H0 is called the null hypothesis (no change, no
difference)
– H1 is called the alternate hypothesis
– Example: H0 : μ1 = μ2 vs. H1 : μ1 ≠ μ2
after project before project
20 22 24 26 28 30 32 34 36 38 40 42 44
– Based on the data we collect, we must decide in favor

of either H0 or H1. Which does the evidence support?

NATURE OF HYPOTHESIS TESTING
H0: Defendant is Innocent

H1: Defendant is Guilty
Since verdicts are arrived at with less than 100%

certainty, either conclusion has some probability of error.
Consider the following table.
True State of Nature
H0 H1
Conclusion is Conclusion results

H0
Correct in a Type II error
Conclusion
Drawn
Conclusion results Conclusion is
H1
in a Type I error Correct
Type I or II Error Occurs if Conclusion Not Correct
The probability of committing a Type I error is defined as

α (0 ≤ α ≤ 1) and the probability of committing a Type II
error is β (0 ≤ β ≤ 1). The most critical decision error is
usually a Type I error.

2-SAMPLE HYPOTHESIS TEST
Processing
Shift
time
Total Processing Times (in hours)

Shift y1 y2 y3 y4 y5 y6 y7 y8 y9 y s
1 2.8 3.6 6.1 4.2 5.2 4.0 6.3 5.5 4.5 4.6889 1.17
2 7.0 4.1 5.7 6.4 7.3 4.7 6.6 5.9 5.1 5.8667 1.08
Note: All times include wait time and approval times in addition to actual
value added time spent completing the form
• In a shipping process, a team member suspects that one

of the important contributors to delivering on time is the
amount of time it takes to complete the shipping
requisition. He further suspects that there might be a
significant difference in the average amount of time
spent completing the requisitions between first and
second shift.
• The team decides to conduct a test, to see if “shift” really

is an important variable. In other words, is there a
significant difference in the average completion times
between first and second shift.
• A random sample of data, shown above, was collected

over a 2 week period, for shipments with the same
complexity level.

2-SAMPLE HYPOTHESIS TEST (cont.)
The graphical interpretation of the hypotheses to be

tested are:
H0 : μL = μH
H1 : μL ≠ μH
H H H H H HH H H
L L L L L L L L L
2 3 4 5 6 7 8
H0: μL = μH
versus
H H H H H HH H H
L L L L L L L LL
2 3 4 5 6 7 8
H1: μL ≠ μH

RULE OF THUMB Q#9-11,
13
Tukey Quick Test for Detecting a

Significant Shift in Average
To determine if a significant shift in average has occurred, a test

developed by John Tukey in 1959, (Technometrics, 1, 31-48) and
popularized by Dorian Shainin, (World Class Quality, Bhote, K.R.,
1988) is called the Tukey Quick Test or End Count Technique. To
perform this test:
1. Arrange all of the data on a scale such that each of the

two groups is represented by a different symbol. Refer to
the previous example, where L = shift 1 and H = shift 2.
2. Starting from the left, count the number of similar symbols

until an opposite symbol is encountered.
3. Likewise, starting from the right, count the number of

similar symbols until an opposite symbol is encountered.
4. Summing the two counts yields the End Count.
Note: If the leftmost and rightmost symbols are the same, then the End
Count is zero. The significance associated with a given End Count can be
found using the following table:
Confidence there exists a

End Count Significance significant shift in average
6 .10 .90
7 .05 .95
10 .01 .99
13 .001 .999
The previous example illustrates an End Count of 7 (3 on left and 4 on

right), giving approximately 95% confidence in concluding H1: μL ≠ μH.

USING SPC XL TO TEST FOR
DIFFERENCES IN AVERAGES (t-test)
• The statistical test for detecting a shift in average is

called the t-test. SPC XL will run this test and report a
“p-value”
• p-values come from the data and indicate the probability

of making a type I error
• SPC XL > Analysis Tools > t Test matrix (mean)
The results below represent the p-values from a 2 sample t-test. This
means that the probability of falsely concluding the alternative
hypothesis is the value shown (where the alternate hypothesis is that the
means are not equal). Another way of interpreting this result is that you
can have (1-pvalue)*100% confidence that the means are not equal.
t Test Analysis (Mean)

P-value = 0.041
• Rule of Thumb:
– If p-value < .05, highly significant difference
– If .05 < p-value < .10, moderately significant
difference
– (1 – p-value) • 100% is our percent confidence that
there is a significant difference in the averages
RULES OF THUMB
Quick Test for Detecting a Significant
Shift in Standard Deviation
smax
1. If > 2.72 then there is a significant shift in standard
smin
deviation.
smax
A “gray zone” occurs when 1.5 ≤ ≤ 2.72 . If this is the
smin
case, then ROT #2 should be applied.
s2max n1 + n2
2. If > 10, then there is a significant shift in
s2min 2
standard deviation. This is a more discerning test than

ROT #1 above since it involves the sample sizes, n1 and
n2. However, neither n1 nor n2 should be greater than 60,
and the smaller n should not be less than 70% of the
larger n.
Q#9-11,
13
Note: If the sample sizes are the same, s (standard deviation) can be
replaced by R (range).

USING SPC XL TO TEST FOR DIFFERENCES
IN STANDARD DEVIATION (F-test)
• The statistical test to detect differences in standard

deviation is the F-test. SPC XL will run this test and give
us p-values
• SPC XL > Analysis Tools > F-Test matrix (StdDev)
The results below represent the p-values from a 2 sample F-test. This
means the probability of falsely concluding the alternative hypothesis is
the value shown (where the alternate hypothesis is that the variances are
NOT equal). Another way of interpreting this result is that you can have
(1-pvalue)*100% confidence that the variances are not equal.
F-Test Analysis (Std Dev)

P-value = 0.813
• Rule of Thumb:
difference
there is a significant difference in the standard
deviations
EXERCISE
Using the Quick Tests and associated rules of thumb

on the tabulated data below, conduct tests for
different averages and different standard deviations.
Method Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10 Y11 Y12 Y13 Y14 Y S
L 201 209 215 221 211 213 217 205 218 208 203 214 212 215 211.6 5.8
H 218 225 217 222 223 220 222 216 221 224 224 221 220 219 220.9 2.7
200 210 220 230

EXERCISE
Using SPC XL > Analysis Tools > t-Test (and F-Test),

conduct the formal t-test and F-test for the two
samples on the previous page to find the
corresponding p-values. Compare these results with
the results you obtained using the Rules of Thumb.

EXERCISE
• In Teams
• Conduct a two-sample hypothesis test involving

some aspect of the Statapult®. Each sample size
should be at least 17 (e.g., compare two different
rubber bands, two different balls, two different hold
times, two different release methods, etc.).
• Use the Rules of Thumb and/or t and F-tests to

determine if there are any significant shifts in
average or standard deviation.
• Brief the class on what you tested and the results

of your test.
What we tested: ____________________________
What we learned:
Significant difference in average distance?

Yes or No (p value = ______)
Significant difference in variation?

Yes or No (p value = ______)

TESTING FOR DIFFERENCES IN
TWO PROPORTIONS
(Auditing Example)
An audit of a sampling of reports from two different

credit teams showed that for Team#1, 11 out of 65
reports had errors. Team #2 delivered 60 reports out
of which 5 had errors. Are the proportion of defective
reports (i.e., reports containing errors) being
produced by the two teams significantly different?
Team #1: proportion of reports with errors = _______
Team #2: proportion of reports with errors = _______

USING SPC XL TO TEST FOR
DIFFERENCES IN PROPORTIONS
• SPC XL will give us p-values
• SPC XL > Analysis Tools > Test for Proportions
Test of Proportions
Number Defective Group #1 (x1) 11

Size of Sample #1 (n1) 65 user
Number Defective Group #2 (x2) 5 inputs
Size of Sample #2 (n2) 60
Results
Proportion Sample #1 (p1) 0.16923

Proportion Sample #2 (p2) 0.08333
p-value 0.15096
• Rule of Thumb:
difference
there is a significant difference in the two
proportions

BANKING CASE STUDY (EXERCISE)
Two branches of a bank process automobile loans. Auditors

who reviewed the loans from both banks for a period of one
month found that Branch #1 processed 122 loans, 12 of
which had errors, and Branch #2 processed 107 loans, 6 of
which had errors. A bank Vice President is putting heat on
the Branch #1 Manager claiming that Branch #1 is
generating significantly more errors in automobile loan
processing (9.8%) than is Branch #2 (5.6%). Do we have at
least 95% confidence that there truly is a difference in error
rates at the two branches?
Suppose we had collected data for a longer period of time

and exactly the same error rates occurred. For branch #1,
there were 244 loans processed, 24 of which had errors.
For branch #2, there were 214 loans processed, 12 of which
had errors. How does collecting twice as much data
change your confidence that there is truly a difference in
error rates at the two branches?

EXERCISE
• Individual or Team
• Using the M&M data, test the following

hypotheses: (note: π = true proportion)
1. H0: πred = πyellow

H1: πred ≠ πyellow
2. H0: πred = πblue

H1: πred ≠ πblue

SUMMARY PROCEDURES FOR 2-SAMPLE
HYPOTHESIS TESTING USING SPC XL
Test for a Shift in the Mean
1. Define test: H0: μ1 = μ2 (Process Means are the same)

H1: μ1 ≠ μ2 (Process Means are different)
μ1 μ2
2. Decide on α (0.05, 0.01, or 0.001)
3. Run your experiment and put data on a clean spreadsheet
4. Highlight the data
5. Select: SPC XL / Analysis Tools / t Test Matrix (Mean)
6. If P-value < α, then conclude H1 [with 100*(1- P)% confidence]
If P-value > α, then you cannot reject H0
Special Notes:
a. For a one-tailed test (e.g., H0: μ1 < μ2 vs. H1: μ1 > μ2 ) then divide the P-value by 2 prior to Step 6.
b. You cannot use this test if the data values are “paired” (e.g., measuring the same part using two
different instruments).
Test for a Shift in the Standard Deviation σ1

1. Define test: H0: σ1 = σ2 (Process Standard Deviations are the same) σ2
H1: σ1 ≠ σ2 (Process Standard Deviations are different)
2. Decide on α (0.05, 0.01, or 0.001)
3. Run your experiment and put data on a clean spreadsheet
4. Highlight the data
5. Select: SPC XL / Analysis Tools / F Test Matrix (StdDev)
Special Notes:
a. For a one-tailed test (e.g., H0: σ1 < σ2 vs. H1: σ1 > σ2 ) then divide the P-value by 2 prior to Step 6.
b. You cannot use this test if the data values are “paired” (e.g., measuring the same part using two
different instruments).
Test for a Shift in the Proportion
1. Define test: H0: π1 = π2 (Proportions are the same)

H1: π1 ≠ π2 (Proportions are different)
2. Decide on α (0.05, 0.01, or 0.001)
3. Run your experiment and collect data for Group 1 and Group 2 (sample size, number “defective”)
4. Increase sample size for a group if the number “defective” is less than 5.
5. Select: SPC XL / Analysis Tools / Test of Proportions
Special Notes:
a. For a one-tailed test (e.g., H0: π1 < π2 vs. H1: π1 > π2 ) then divide the P-value by 2 prior to Step 6.
b. Do not use this test if either number of “defectives” observed (i.e., X1 or X2) is too small.
Rule of Thumb: The number of defectives in each group must both be 5 or greater.

HYPOTHESIS TESTING WITH
CLASSIFICATION DATA
• Suppose we can classify data according to two

different criteria
Q#10,
13
Product A Product B
East Coast
Midwest
West Coast
Defect Defect Defect

Type A Type B Type C
Supplier 1
Supplier 2
Supplier 3
Supplier 4
• A count is recorded in each cell of the table
• This is called a (r x c) contingency table with r

rows and c columns.
• A Chi-square test for independence will help us

answer the following question:
– “Is there a significant relationship between the
row variable and the column variable?”

CHI-SQUARE TEST FOR
INDEPENDENCE
A company has four product lines. Customers return the

products for three basic reasons: (1) cosmetic defects, (2)
hardware failure, and (3) software incompatibility. The
customer return data is summarized below:
Reason
for
Return Cosmetic Hardware Software
Total
Defect Failure Incompatible
Product
Line
1 125 43 12 180
2 136 55 8 199
3 186 75 20 281
4 78 33 10 121
Total 525 206 50 781
Are "product line" and "reason for return" independent, or is

there some relationship between product line and reason for
return?

CHI-SQUARE TEST FOR
INDEPENDENCE EXAMPLE
Step 1: H0: Product Line and Reason for Return are

Independent
H1: Product Line and Reason for Return are NOT
Independent
Step 2: α = 0.05
Step 3: Under the null hypothesis, the expected frequencies,

Fi,j, for each cell are given by:
fi• • f•j ⎡ f2• • f•3 199 • 50 ⎤
Fi, j = ⎢F2,3 = = =12.74 ⎥
N ⎣ N 781 ⎦
where fi• and f•j are the observed marginal frequencies
for the ith row and jth column, respectively.
Comparing expected frequencies (Fi j) with observed

frequencies (fi j) we see the resulting table:
Total
125 43 12
1 180
(121.00) (47.48) (11.52)
136 55 8
2 199
(133.77) (52.49) (12.74)
186 75 20
3 281
(188.89) (74.12) (17.99)
78 33 10
4 121
(81.34) (31.92) (7.75)
781
Total 525 206 50
(N)

CHI-SQUARE TEST FOR
INDEPENDENCE EXAMPLE (cont.)
The test statistic is given by

4 3 (f − Fij )
2
χ = ∑∑ = 3.604
2 ij
i =1 j =1 Fij
Step 4: Degrees of freedom: df = (r - 1)(c - 1)

= (4 - 1)(3 - 1) = 6
χ 26
.95 .7301
.05 .2699
χ (.295 , 6 ) = 12.59 χ 2 = 3.604

P = .7301
Step 5: The p-value is the area determined in Step 4.

P = .7301 is greater than α = 0.05.
Step 6: We fail to reject H0 and conclude that Product Line and

Reason for Return are independent.
Notes: 1. Another consideration might be to test if the proportion of

returns is the same across product lines.
2. The Chi-Square test for independence assumes that the sample

size is sufficiently large, such that
- All expected cell frequencies are ≥ 2
- At least half of the expected cell frequencies are ≥ 5

SPC XL OUTPUT
SPC XL > Analysis Tools > Independence Test Matrix
Exercise: Give an example from your process where

you might use this test. That is, specify two
qualitative variables for which you might
want to investigate their independence or
lack thereof?
EXERCISE
A company was studying the effectiveness of 3
competing cold remedies. 180 people were randomly
selected to participate in the study and divided into 3
groups of 60. Each group was given a different brand
of cold medicine and after 2 days, asked to report on
the effectiveness of the medicine.
The following data was collected:

Brand A Brand B Brand C
No relief at all 15 16 31
Some relief 35 35 23
Total relief 10 9 6
Based on this data, what do you conclude about the

effectiveness of the three cold remedies? Are they
equally effective? Why or why not?

TEST FOR INDEPENDENCE OF
TEAM VS M&M COLOR
From the M&M exercise, count the total number of

M&Ms of each color for each of your class teams and
enter those numbers in the appropriate cells below.
Color
Team Brown Red Yellow Blue Green Orange
Using the Test of Independence from SPC XL (SPC

XL > ANAYSIS TOOLS > INDEPENDENCE TEST
MATRIX), determine if the distributions of colors are
independent of team.
What would you expect the answer to be?


MODULE 10
IMPROVING THE PROCESS AND

WORK FLOW
IMPROVING THE PROCESS
AND WORK FLOW
• IN THIS SECTION, WE WILL DISCUSS:
– ELEMENTS OF THE IMPROVE PHASE
– CONCEPTS FOR REDUCING SETUP AND

CHANGEOVER TIMES
– Single Minute Exchange of Dies (SMED)
– COMPARISON OF BATCH VS. SINGLE PIECE

FLOW
– CELLULAR MANUFACTURING
– Principles of Cell Design / Layout
– APPLYING THE PRINCIPLES OF THE

MEASURE, ANALYZE, AND IMPROVE
PHASES FOR A SIMULATED PROCESS
– MISTAKE PROOFING THE PROCESS USING

FMEA AND ERROR-PROOFING (Poka Yoke)

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
IMPROVE
Define
170

160
150
140
130
120
110
100
90
80
1
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
Month Number
Analyze

SOPs
Improve
necessary
Control
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

IDENTIFY IMPROVEMENT
OPPORTUNITIES AND GENERATE
SOLUTION ALTERNATIVES
• Based on the teams analysis, brainstorm
improvement opportunities and test out solutions
to attack the root cause(s) of the problems
• Start with the critical few causes / variables

– Example: Critical cause = incomplete forms
– Possible solutions:
• automate
• create a checklist for filling out the form
and improve training
• use popup menus, dropdown lists
• simplify the form
• etc.
• If there are too many ideas, use the IPO

(prioritization) matrix to narrow the candidate
solutions

CONSIDERATIONS WHEN
IMPROVING THE PROCESS
• Review any “parking lots” you created during earlier

team meetings (process mapping, looking for waste,
building the cause and effect, etc.) for ideas and
possible solutions
• Include any “low hanging fruit” items and make sure to

address these, even if they are not associated with most
critical causes
• Make sure to include the Champion (Sponsor) and key

stakeholders
– buy-in
– criteria for deciding between solution alternatives
• Some additional tools for reducing cycle time and

improving the work flow include:
– SMED (for quick changeovers)
– Batch vs. Single Piece Flow
– Cell Design / Layout

QUICK CHANGEOVERS
Single Minute Exchange of Dies (SMED)
Traditional thinking says:
• Long changeovers lower machine availability
• Startup and alignment after changeovers

always create scrap at first
• Because changeovers take so long, we need

to run longer lot sizes

QUICK CHANGEOVER VIDEO
Take notes as you watch the video.

Think about the following :
• Quick changeover is not about working faster, it’s

about __________________________________.
• What are 5 common steps for implementation?
• What examples stood out to you? What concepts for

reducing changeover and setup times did you find
important?
• How might this apply in your area?
VIDEO

WHAT HAS BEEN DONE?
Hitachi Automotive Products Associated Spring

Reduced changeover from Reduced changeover from
1.5 hours to 20 min. 3 hours to 20 min.
Steelmill: Three 500 Ton presses completely

changed over:
3 1/2 hours to 30 minutes
Automotive OEM injector

molders: all molds and
inserts changed, reduced
from 1 1/2 hours to 9 minutes
Machine shop screw machine:

tooling changed and new
set-up, reduced from
1 hour to 8 minutes
The Wiremold Company

Reduced changeover from Changeover times are usually
2 hours to 10 min. measured from the last good
piece of run A to the first good
piece produced in run B.
Powers and Sons
Reduced changeover from
8 hours to 2 hours

BENEFITS OF REDUCING
CHANGEOVER TIME
Higher Productivity – Reduced changeover time means

increased machine efficiency
Flexibility – Faster response to changing customer demands
Employee benefits – Simpler and safer setups with less

chance of errors
Standardization – Because the right tools and equipment

are at the right place at the right time, everyone wins
Productivity Start-up Scrap

Flexibility
Time Time

CHANGEOVER WORK ALLOCATION
Internal steps – These steps can only be performed

with the equipment shut off
External steps – These steps can be performed

while the machine is still in operation
Changeover time – The time from when the last

good part comes off the machine until the first good
part comes off after the changeover
1st
EXT.
2nd
Changeover
Time
EXT. 3rd
INT.
EXT.
INT. INT.
How will you know if there is a problem with

changeover times?
Where do you start?

GETTING STARTED …
BREAKING THE PROCESS DOWN
• Use a team of operators, tool makers, and engineers
• Observe and/or videotape the process as it is currently

done
• Map the process and challenge each step. Is it

required? Is there a better way? Eliminate or reduce
whenever possible. Record problems that occur
frequently.
• Categorize each step as internal or external.
• Wherever possible, convert internal steps to external

steps.
• Explore methods to reduce time for fastening and

connecting (rapid connections), standardizing sizes and
shapes, minimizing installation adjustments, etc.
• Develop checklists, templates, and SOP’s for setups and

actions to be done

SET-UP/CHANGEOVER OBSERVATION FORM
Step Is Time
Step # Description Int/Ext Start Stop 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320
1
2
3
4
5
6

7
8
9
10
11
12
13
14
15
16
17
18
19
20
Total Steps Int ____ Total
Time _________
Total Steps Ext____
10 - 12
BUILDING A SMED CHECKLIST
1. Determine the need for a checklist first.

2. Determine who will use the checklist, where and
when.
3. Identify all tools, tooling, dies, jigs, molds, etc. to be
used.
4. Identify if they are clean, in operating condition, and
pre-positioned for installation.
5. Identify all paperwork needed, i.e., drawings,
information about settings, and tags or forms for
removed equipment.
Clean Ready for Pre-positioned Not Available

Installation
Paperwork
Drawings _____ _____ _____ _____
Setting Specifications _____ _____ _____ _____
Control Logs _____ _____ _____ _____
Tags or Movement forms _____ _____ _____ _____
Equipment
Molds/Dies _____ _____ _____ _____
Fixtures/Jigs _____ _____ _____ _____
Transport trays/racks _____ _____ _____ _____
Carts/dollies _____ _____ _____ _____
Hand tools/accessories _____ _____ _____ _____
Ancillary machines _____ _____ _____ _____
Supplies/Apparel
Pens/pencils/paper _____ _____ _____ _____
Gloves or Masks _____ _____ _____ _____
Helmets/shoes _____ _____ _____ _____
Coats/aprons _____ _____ _____ _____
Cleaning supplies _____ _____ _____ _____

EXERCISE
• During the Measure phase of a Lean Six Sigma project,
it was noted that a particular machine experienced long
setup/ changeover times.
load machine unload

start parts parts and finish
deburr
time = 2 min. time = 12 min. time = 4 min.
dpm = 21,000 c/o = 1.8 hrs # ops = 2
dpm = 85,000 dpm = 11,000
• Using SMED techniques, the team implemented

improvements to help streamline the changeover
process.
• Use SPC XL to compare the baseline data with the post

project data. Did the teams work have a significant
impact on the changeover times? Why or why not?
Baseline After Improvements

1.1 hours 1.1 hours
1.8 hours .6 hours
2.4 hours .8 hours
2.1 hours .8 hours
1.9 hours .7 hours
1.7 hours 1.1 hours

BATCH AND QUEUE
(Physical Layout)
Storage End
Assembly
Machining
Packing
Storage
Storage
Start Storage
Storage Deburring
Inspection
Storage

SINGLE PIECE FLOW
Assembly
Inspection Deburring
Workers
Packing Machining
Customer Raw
Material

ILLUSTRATING BATCH VERSUS
SINGLE PIECE FLOW (EXERCISE)
• Break the class into two sub-teams, each team having 10
members
• The balance of the class will serve as customers or observers
• Goal: Deliver 12 sheets of paper to the customer, with each sheet
containing the signature of all 10 members
• Team One: (Single Piece Flow) Each member writes full name on
8 1/2” x 11” sheet of paper and passes sheet to next person when
finished.
• Team Two: (Batch size of 6) Each member writes full name on 6
1/2” x 11” sheets of paper and in batches of six sheets paper-
clipped together, passes each batch to next person.
• Results:
Team 1 Team 2
Total Elapsed time
to obtain 12 completed
sheets
Time until customer

receives first sheet
Average time to produce

1 sheet
Work in Progress (WIP)

(# of sheets of paper in
process at the moment
of completion)

BATCH PRODUCTION VS.
SINGLE PIECE FLOW
Lead Time = 30 Minutes
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 26 27 28 29 30
Machining x xx xx x
Deburring xx xxx x
Assembly x xx x xx
Testing xxxxxx
Packing x x xx xx
Lead Time = 10 Minutes

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 26 27 28 29 30
Machining x xx xxx
Deburring xx xxxx
Assembly x xxxxx
Testing xxxxxx
Packing xxxxxx
66% Reduction in Lead Time

WHAT IS CELLULAR
MANUFACTURING?
Starts with the following:
Frontal Wide Logical Shared

Loading Aisles Flow Duties
The “best” layout is the layout that fits your application!

U-SHAPED CELL DESIGN
• If possible, the U-shaped design is a preferred

organization for a cell.
• Parts return to the point of entry
• Reduces walking time which reduces operator cycle
time
• Facilitates communication between operators
• Visual controls can be seen by all operators
• Water spiders facilitate material requirements
Material
Out
In
Material Material

CREATING COMFORTABLE
WORK AREAS
• All tools and and equipment to be used by the operators should be
within the Secondary work area (18-24 inches) away.
• The Primary work area (14-18 inches) should be kept clear for
handling the product.
• The operator should not have to twist or turn to pick up materials
or tools.
• Assemblers or builders should have raw materials provided in front
of them (frontal loading) and dispose of bins or totes under the
work area in front of them (frontal discharge).
• Treat operators as if they are a surgeon in the Operating Room –
make their job as easy as possible to perform without interruption!
14-18 inches
18-24 inches
Work Surface Heights

Precision
Light Work
Work
Heavy 37 – 43 in.
Work 33 –37 in.
28 – 33 in.

KITES Inc. SIMULATION
(Round 1 - Baseline)
• Your company makes custom nameplates. One of your
customers, KITES Inc., has been complaining about lead
time issues
• In Round 1, you will run the simulation using the current,

“as-is” process, with each department pushing product
through until the customer has received the required
amount. Improvement rounds will follow.
• Baseline Process (Details)

– 4 operators (1 for each of 4 departments) (Each
has detailed SOPs for their work)
– Each operator may only work in their assigned
department
– 1 inspector in a quality control department
– 1 material handler responsible for all movement of
materials from supply and between departments
– 1 supply room personnel
– Traditionally, a batch size = 6 has been used
– Takt time is 20 seconds
– Simulation Time: 6 minutes, Customer Demand =
18
• Other team roles

– 1 or 2 analysts, responsible for keeping time (start
and stop time) and making process observations

(Round 1 - Baseline)
• Current Layout
Mtl Mover
Supply
Customer
Op #2 Op #3
QC
Op #4 Op #1

(Results)
after 6 minutes
Metric Round 1 Round 2 Round 3
Total # of
Nameplates
Shipped
Time till
first unit
received by
customer
Total WIP
Total
Quality
Rejects
Profit
(Loss)

FINANCIALS (Round 1)
Profit (Loss) Calculation

Amount Qty Total (+) Total (-)
Revenue (units) + $20/unit
Expenses
Support Personnel - $2/min
Operators - $2/min
WIP - $3/unit
Defects - $10/unit
Net Profit/Loss

(Round 1 Results) (cont.)
• What worked well?
• What didn’t work well?
• What types of waste exist?
• Where are the biggest opportunities for improvement?

(Round 2)
• Goals
– improve responsiveness to customer
– make more money
– use some of the tools we have learned
• As a class (or in teams)

– map the current process (logical flow and physical
flow)
– analyze the waste
– ensure you understand the customer requirements
– brainstorm ideas for improvement
– you may select and implement only 3 improvement
ideas for this round
• Constraints
– must still use batch size of 6
– each operator must remain in their department
• Re-run the simulation and record results in the

“round 2” column of the scorecard


Expenses
Operators - $2/min
WIP - $3/unit
Defects - $10/unit
Net Profit/Loss

(Round 3) (Optional)
• Goals
– improve responsiveness to customer
– make more money
– use some of the tools we have learned
• In teams,
– complete further analysis and make additional
improvements to the process
– you may conduct any time studies, as needed
– ensure you understand the customer requirements
– brainstorm ideas for improvement
– select and implement 3 improvement ideas
• Additional rules
– no requirements for batch size
– operators may be cross trained
– work may be redistributed


Expenses
Operators - $2/min
WIP - $3/unit
Defects - $10/unit
Net Profit/Loss

(Summary)
• What did you learn, or what did the simulation

reinforce?

PILOTING SOLUTIONS AND
VALIDATING IMPROVEMENT
• Develop a plan to test out and validate your

ideas
– Who will be responsible?
– What do they need to do, create, etc.?
– When will it happen?
– How long will it take? How much data is
needed?
• Conduct the pilot test and evaluate the results
• Validate (verify) the improvement

– Graphs (Visual)
• Run charts; Control charts
• Histograms; Box plots
– Measures (Statistical)
• Compare process average, standard deviation, Cp,
Cpk, etc.
• Hypothesis testing

CHOOSING A HYPOTHESIS TEST
• As a reminder, there are several different

hypothesis tests
• The appropriate test depends on what you’re

trying to compare. Do you remember the tests?
What do you Exact

want to “Quick Test”
compare? Statistical Test
2 averages
μ1 μ2
2 std deviations
σ1
σ2
2 proportions
%1 vs %2
Count/classified
data

(Example)
Processing
Automate?
time
• To reduce the time spent processing shipping

requisition forms, a team brainstormed several
solutions and decided to:
– Eliminate several non-value added steps in the
process
– Automate portions of the form
• A pilot test provided the following results on shift 1

(the cycle time, in hours, is shown)
Before any After

improvements improvements
(new method
Shift 1 (Before) Shift 1 (After) and PF)
2.8 1.1
3.6 0.8
6.1 1.4
4.2 1.5
5.2 1.2
4 0.7
6.3 0.8
5.5 0.5
4.5 0.9

(Example) (cont.)
• Graphical Evidence
Average cycle time

Run Chart appears to have
decreased, and the
7 process variation appears
reduced
6
5
Cycle Time
before
3
2
after
1
0
1 3 5 7 9 11 13 15 17
X Axis
Shift in process,
with lower variation
(i.e., “smaller” box)
Box Plot of Cycle Times
5
Q1
4 Min
Median
3 Max
Q3
2
0
Shift 1 (Before) Shift 1 (After)

(Example) (cont.)
• Numerical Evidence
Summary Statistics and Dot Plot of Data

Shift 1 (Before) Shift 1 (After) before after
Count 9 9
Average cycle time Average cycle time
Mean 4.688889 0.98889
= 4.69 hours = 0.99 hours
Median 4.5 0.9
Mode 0.8
Max 6.3 1.5
Min 2.8 0.5
Range 3.5 1
Std Dev (Pop) 1.10598819 0.314269681
Standard Deviation Standard Deviation
Std Dev (Sample) 1.173078 0.33333 = 1.17 hours = 0.33 hours
Variance (Pop) 1.223209877 0.098765432
Variance (Sample) 1.376111111 0.111111111
Dot Plot of Cycle Times
after Shift 1 (Before)

Shift 1 (After)
before
0 1 2 3 4 5 6 7
t-test shows a
statistically
significant
difference in the
t Test Analysis (Mean) average cycle
P-value = 0.0000001 times
F-test shows a
statistically
significant difference
in the standard
deviations
F Test Analysis (Std Dev)
P-value = 0.001814

UPDATE PF AND DEVELOP/REVISE SOPs
• When improvements have been validated,

make sure you update your process flow (PF)
and develop and/or revise any necessary
standard operating procedures (SOPs)
• SOPs
– Should document consistent and reliable methods
for performing tasks and/or holding variables
“constant”
• When developing SOPs

– Involve the people who perform the work. They will
be the ones implementing the improvements
– Use visual methods when possible, including
photographs, diagrams, checklists, templates, etc.
– Keep the task descriptions clear and simple
– Present the work steps in the sequential order in
which they must be performed. Avoid too many
references to other documents or loop-back
references
Standard Work Routine
Step
Work Steps Operator Time Units
Work Sequence Layout
checklists
#
1 Retrieve semi-finished board Op #2 1 Soldering

10s
& Inspection
2 Mount into jig on Table #1 Op #2 15s 1 Machine
3 Select components from bins 1,2&3 Op #2 10s 1

Assembly Assembly
4 Insert components onto board Op #2 120s 1 Table #2 Table #1
5 Place board on next station Op #2 10s Op #1
3 2
6 Solder board Op #1 240s
7 Perform detailed inspection Op #1 480s
8 Place board on next station Op #1
9 4 1
Op #2
10
Completed Blank
Components Boards
Cart Cart

MISTAKE PROOFING AND FMEA
• To help when developing SOPs, and as a final

walk through of the new process, you can use
FMEA (failure mode and effects analysis) for
ensuring the process is mistake-proofed
• FMEA is a quality planning tool used to

– Identify and eliminate potential product and process
failures or defects
– Identify, analyze, prioritize, and document potential
failures modes, their causes and effects, and to
prevent future failures
• There are two ways to accomplish an FMEA

– A simplified FMEA (using structured brainstorming)
– A more formal FMEA (using a spreadsheet, scoring,
and such

MISTAKE PROOFING THE PROCESS
(FAILURE MODE AND EFFECT ANALYSIS)
What is it?
FMEA is a systematic method for identifying, analyzing,
prioritizing, and documenting potential failure modes,
their effects on a system, product or process
performance, and the possible causes of failure.
Types of FMEAs
– Process: used to analyze processes and

identify potential process failure
modes (machines, tools, work
stations, product lines, testing,
measurement systems, service
and administration)
– Design/Product: used to analyze products and

identify potential failure modes
early in the development cycle
(components, subassemblies)
– Defect: used to analyze and prioritize

defects (errors, mistakes) to
prevent reoccurrence (in
products and processes)

WHY FMEA?
Fumbling for his recline button,

Ted unwittingly instigates a disaster.
Source: The Far Side

WHY FMEA? (cont.)
GAZETTE TELEGRAPH Sunday, Dec. 3, 1995

NATION
Physician fined, barred for 6 months

Florida doctor amputated wrong foot
Associated Press
MIAMI – A doctor who Dr. Edward A. Dauer said. "But lawyer called the punishment
amputated the wrong foot of a doctors are not God, and fair and declined to say if he will
diabetic patient and cut off hospitals are not heaven." appeal.
another patient's toe without her The hearing officer had State and federal
consent was fined $10,000 on proposed a two-year punishment investigators found a series of
Saturday and barred from and a $15,000 fine and procedural problems at
practicing medicine for six recommended that Sanchez be University Community Hospital
months. directly supervised by another contributed to the wrong-foot
The 15-doctor state Board surgeon during a five-year amputation on Willie King in
of Medicine voted unanimously probation. February. King, 52, later had his
to reduce the sentence Sanchez can apply the 140 gangrenous other foot amputated
recommended last month by a days he has been suspended at another hospital. He and
state hearing officer. The board under an emergency order, so he Sanchez came to a private
cited Dr. Rolando Sanchez's could be back in the operating settlement.
previously unblemished record room by early 1996. He will be In the second
and noted that the diabetic's on probation for two years with amputation several months later,
other leg would probably have indirect supervision. the board found that Sanchez
been amputated anyway. He has 30 days to appeal to failed to warn the patient that
"I'm not saying what he did a state appeals court. His amputation was a possibility.
wasn’t wrong," board member
A8 GAZETTE TELEGRAPH Friday, Aug. 16, 1991
Army: Patriot computer needed break

Documents show problem The reason for the Patriot's failure will never be
known for sure. But they said the most likely explanation
was known before attack was a previously unknown glitch in Patriot computer
software.
Associated Press Army technicians had determined two weeks before
WASHINGTON – The Patriot missile, a high-tech star of the attack that the Patriot computer was vulnerable to
the Persian Gulf War, might have averted its most costly losing track of incoming Scuds when the computer was
failure if its operators had used a low-tech tactic: a periodic kept running for long periods, said Army reports
rest for the missile system's computer, Army documents obtained by The Associated Press.
show. No alert bulletins were sent to Patriot operators in the
The Army credits the Patriot with shooting down nearly field because the technicians viewed this as a minor
every Scud missile fired by Iraq that threatened U.S. or problem and did not think Patriot computers would be
allied positions in Saudi Arabia and Israel. But the Patriot's kept running for more than several hours at a time.
biggest failure came Feb. 25 when it allowed a scud to slam As it happened, the Patriot battery into whose
into a U.S. barracks in Saudi Arabia, killing 28 soldiers. coverage area the Scud fell had been running its
computer for 100 consecutive hours at the time of the
attack, according to Army reports.

WHY FMEA? (cont.)
A8 GAZETTE TELEGRAPH Friday, Aug.16, 1991
Transplant patient gets wrong heart in operation

Hospital hopes to find another donor Lab technicians discovered the mixup when
Associated Press they checked a tissue sample hours into the
lengthy operation, Trunkey said. By then it
PORTLAND, Ore. — A man received the
was too late to stop the procedure.
wrong heart during a transplant, a hospital
The heart was beating on its own
admitted Thursday. Doctors hoped his body
Thursday, and the patient was in critical but
wouldn't reject the organ before another donor
stable condition.
heart with his blood type can be found.
His body could have rejected the heart
The unidentified Oregon man with type O
immediately. "The heart could have turned
blood received a type A heart Wednesday at
blue and stopped," said Dr. Jeffrey Hosenpud.
University Hospital. The cause of the mistake
Doctors were cautiously optimistic the man
was under investigation.
would survive long enough to receive the
It was the first heart mixup at University,
right type heart after he passed his first day
but officials said eight other people worldwide
without serious complications. It has been
are known to have received mismatched heart
placed at the top of the hospital's transplant
transplants. At least one survived.
list.
Patients with type 0 blood often receive
It will likely take three or four days to
type A kidneys intentionally, and "these
track the mistake, but hospital officials were
kidneys do relatively well," said Dr. Donald
sure the error came there, rather than at the
Trunkey, chairman of University's surgery
hospital where the heart was removed from
department. "But a kidney transplant is
the donor, Trunkey said.
obviously not as serious as a heart transplant.
"Everyone involved feels very badly this
There is no dialysis for a heart that fails."
has happened,” the hospital said in a
statement.
GAZETTE TELEGRAPH Friday, Oct.6, 1997
Boy's death costs Chrysler $262 million

Faulty Latch on van leads to jury verdict
Associated Press
CHARLESTON, S.C. — A jury Wednesday The $12.5 million in actual damages and
ordered Chrysler Corp. to pay $262.5 million $250 million in punitive damages awarded
to the parents of a 6-year-old boy killed when to the parents of Sergio Jimenez II was the
he was thrown from the family minivan in an largest award against the automaker, a
accident because a rear latch was defective. company spokesman said.

MAJOR COMPONENTS OF FMEA
• Product Components or Process Steps
• Failure Modes
• Potential Effects and Severity of Failure Modes
• Potential Causes and Likelihood of Occurrence
• Current Design or Process Controls in Place and

their Effectiveness
• Risk Priority Number
• Recommended Actions
2 APPROACHES FOR DOING FMEA
9 KISS (“Texas-Style”)
9 Formal (spreadsheet)

“TEXAS STYLE” (KISS) FMEA
• A formal FMEA can get complicated and is not always

needed
• A simplified FMEA thought process is shown below.
• Brainstorm:
– How the product or process can fail?
– What actions are needed to prevent these failures
from happening (i.e., mistake proofing)?
– Who is accountable for these actions and when they
will be completed?
Possible Action(s) Needed to Who takes Action and

Errors Prevent When

PROCESS/PRODUCT FMEA FORM – FORMAL APPROACH
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Product Failure Failure Effects S Causes O Controls D R Actions Plans pSEV pOCC pDET p
Component Mode (Consequence of E (of Failure C E P R
or Process Failure Mode) V Mode) C T N P
Step N

104 -- 45
45
RISK PRIORITY NUMBER (RPN) =
SEVERITY x OCCURRENCE x ESCAPED DETECTION
SCORING: (Product and Process FMEAs)

Score 5 1
Category 4 3 2
(Very Bad) (Good)
Severity (SEV) Severe High Moderate Minor Negligible
consequence consequence consequence

of failure of failure of failure
Occurrence (OCC) Very high High Moderate Low Very low

probability probability probability
cause of cause of cause of
failure mode failure mode failure mode
will occur will occur will occur
Escaped Very high High Moderate Low Very low

Detection (DET) probability probability probability
failure will failure will failure will
escape escape escape
detection detection detection
before before before
reaching the reaching the reaching the
10 - 46
“customer” “customer” “customer”
SAMPLE PROCESS FMEA: SELF-SERVE GAS
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Process Failure Failure Effects S Causes O Controls D R Actions Plans pSEV pOCC pDET p
Step Mode (Consequence of E (of Failure C E P R
Failure Mode) V Mode) C T N P
N
Find gas All pumps Must wait for pump 1 Friday / pm 3 Fill on Thursday 1 3
pump busy rush hour
Find another 3 Impatient / 3 More exercise 2 18
station hurry

Gas cap on Must move car 2 Poor 2 None 5 20
other side planning
of vehicle
Get correct Unavailable Must move car 1 Poor 1 None 5 5

hose planning
Buy different 2 Impatient / 3 Exercise 5 30
octane hurry
Initialize Pump won’t Can’t pump gas 2 Previous 1 Choose better 1 2
pump reset person has station
not paid
2 Attendant 1 Choose better 1 2

failed to station
reset
Pump won’t Must pay cash 4 Credit card 1 Replace card 1 4
read credit dirty
card
Dispense Auto shut- Spill gas on 5 Pump 1 None 5 25

fuel off fails ground failure
5 Auto tank 5 None 5 125
design
10 - 47
Pay for gas on 2 Auto tank 5 None 5 50
ground design
SAMPLE PRODUCT FMEA:
AUTOMOBILE GLASS
p
S O Current D R p p p
Product Potential Failure Potential S O D R
Failure Mode E C Design E P Actions Plans
Components Effects Causes E C E P
V C Controls T N
V C T N
Item on Engineering
Does not squirt Obscured vision Add fluid
Windshield No fluid in periodic to review
washer fluid on for driver and 4 5 3 60 low level 4 2 1 8
Washer reservoir maintenance
windows passengers light by 7/1/01
checklists
Screen
4 Blocked line 2 1 8 None
installed
Review Design
Pump doesn't eng'r'g
4 1 None 5 20 historical 4 1 3 12
work review by
failure data
6/1/01

Disconnected
4 2 Pressure fit 2 16 None
line
Unsightly Impact by object Consider Materials to

Plastic
Windshield Cracked / broken Distorted view 3 like a rock or 4 2 24 tempered review by 3 4 1 12
lamination
Broken glass in car gravel glass 6/1/01
Consider Materials to
Accident/ Plastic
3 4 2 24 tempered review by 3 4 1 12
collision lamination
glass 6/1/01
Torquing due to Road test for
3 1 2 6 None
poor fit prototypes
Water in car Sealant 1000
Leaks during rain Streaking on inside 2 Poor seal 2 Rubber gasket 2 8 None
windshield 75
Seal degrades Accelerated
2 3 2 12 None
over time ageing tests
Impact by object
Unsightly
Side Windows Cracked / broken 2 like a rock or 2 None 5 20 None
Broken glass in car
gravel
High customer Lab failure
Review of Engineering
Won't open or dissatisfaction Broken tests for
3 3 3 27 historical to review 3 2 2 12
close No access to mechanism accelerated
failure data by 7/1/01
outside air ageing
Ice build up in
3 1 Door seals 2 6 None
mechanism
3 Failed motor 1 None 5 15 None
3 Blown fuse 1 None 5 15 None
10 - 48
Loose
Crimp
3 connection to 2 2 12 None
connector
electric motor
SOME MISTAKE PROOFING
(POKA YOKE) IDEAS
(and other ways to simplify and prevent defects)
• Checklists
• Templates/Pictures
• Detailed action and work sequence
• SOP’s
• Use colors, color coding
• Others?
• Build quality into processes

• Inadvertent errors and defects
can be eliminated
• Eliminate the “but’s” in
statements like “We know that it
is not right, but…..”
• Don’t think up excuses, think
about how to do it right!
• Teamwork is the key to effective
improvement ideas

EXERCISE (OPTIONAL)
• In small groups
• Select a work-related or project-related process or

product you’re familiar with
• Practice creating a formal process or product

FMEA. For the purpose of this exercise, focus
only on two process steps or product components.
• What worked well? What did you struggle with?

PROCESS / PRODUCT FMEA FORM
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Product Failure Failure Effects S Causes O Controls D R Actions Plans pSEV pOCC pDET p
Component Mode (Consequence of E (of Failure C E P R
or Process Failure Mode) V Mode) C T N P
Step N

10 - 51

MODULE 11
CONTROLLING PROCESS
PERFORMANCE
(REALIZING AND HOLDING THE GAINS)
CONTROLLING PROCESS
PERFORMANCE
– BASIC CONCEPTS OF STATISTICAL

PROCESS CONTROL (SPC)
– DIFFERENCE BETWEEN PROCESS

CONTROL AND PROCESS CAPABILITY
– CONSTRUCTION, INTERPRETATION, AND

APPLICATION OF CONTROL CHARTS
• Types of charts
• Control Limits
• Out of Control Symptoms
• Subgrouping Strategy
– ELEMENTS OF AN EFFECTIVE CONTROL

STRATEGY
– Control plans
– 5S for Workplace Organization
– Visual controls
– PROJECT DOCUMENTATION

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
CONTROL
Define
UCL
CENTERLINE
success
LCL
Measure
Performance No
project goals Analyze
achieved?
Yes
Determine control plan for Improve

holding the gains

actions or follow-up; validate business Control
impact

team, etc.

Gate Review No Go
?
Done
Celebrate

STATISTICAL PROCESS CONTROL (SPC)
“LISTENING TO THE
VOICE OF THE PROCESS”
CONTROL CHART
Q
U Control limits extended
A
L
I Control Region
T captures variation
Y natural to the
original process
M
E
A Control limits Process trend
S established
U
Process has shifted
R
Special cause identified
E
S
TIME

TWO TYPES OF VARIATION
• COMMON CAUSE VARIATION

– Natural, inherent (random) variation in a process
– If only common causes of variation are present, the

output of a process is predictable and stable over
time
• SPECIAL CAUSE VARIATION

– Abnormal or assignable cause variation
– Examples: differing incoming materials, breakdown

in equipment, new untrained associates
– If special causes of variation are present, the output

of a process is unpredictable and not stable over
time

CONTROL CHART APPLICATIONS
• Many Uses
– In the Measure phase
Checking to see if a process is “in control” and
stable before relying on metrics like Cp and Cpk
– In the Analyze phase

When studying your baseline data to look for
trends, shifts in the process, or outliers
– In the Improve phase

As graphical evidence that you’ve made a change
(improvement) to the process. You should see a
signal on the control chart.
– In the Control phase

When you’re determining a control plan and
method for monitoring a process on an ongoing
basis
Must be some
special cause
of variation
Copyright Air Academy Associates 6

PROCESS EXAMPLE
INPUTS OUTPUTS
MACHINES (computers, software, …)

PROCESS Number paid late
MATERIALS (invoice type, marker types, …)
MANPOWER (training, experience, shift, …)

Number paid early
METHODS (speed, filing method, …)

Invoice
Processing
MEASUREMENT (days, dollars,, …)
Number rejected
ENVIRONMENT (PO available, space, …)

(Accounts Payable)
11 - 7
EXAMPLES OF WHERE SPC HAS
PROVEN SUCCESSFUL
Work Area Process Output(s)
hospital health care operating room delays, surgery time,
recovery time, number of patients,
surgery costs, lab costs, time to
process lab results
secretarial office typing cycle time, number of errors, number
of pages per day, number of phone
calls, time to answer a call
customer response to time to return phone calls, number of
service complaints complaints, time to repair, number of
repeat complaints
barber shop hair styling number of clients, time to wash, time
to dry, time to cut, cycle time, waiting
time, number of re-cuts
assembly shop assemble parts number of completed parts per day,
number of errors, cycle time
pharmaceutical manufacture weight of tablet, number of defects,
firm drugs cycle time, dimensions of tablets
maintenance scheduled or time to repair, time between failures,

shop unscheduled number of failures, number of repeat
maintenance repairs
engineering engineering time to complete, number of
designs errors, number of engineering change
proposals
software design, coding, time to complete, number of bugs,
development testing number of design changes, time
required to test
training employee number of trainee complaints in each
section training of several types, amount of required
retraining
finance processing time to process vouchers, number of
center travel vouchers voucher errors
IT department network speed packet delay time, server up time,

server response time, server load
human resources hiring number days advertised position,

number applicants, time to hire, cost
per hire

CONTROL VS CAPABILITY
PROCESS CONTROL
IN CONTROL
(SPECIAL CAUSES ELIMINATED)
TIME
SIZE
OUT OF CONTROL
(SPECIAL CAUSES PRESENT)
PROCESS CAPABILITY
LOWER SPECIFICATION LIMIT

UPPER SPECIFICATION LIMIT
IN CONTROL AND CAPABLE

(VARIATION FROM COMMON
TIME CAUSES REDUCED)
SIZE
IN CONTROL BUT NOT CAPABLE

(VARIATION FROM COMMON
CAUSES EXCESSIVE)

CONTROL VS CAPABILITY (cont.)
IN CONTROL?
OUT OF CONTROL IN CONTROL
YES
CAPABLE?
LSL USL LSL USL
NO
LSL USL LSL USL
Which processes need to be fixed?

CONTROL CHARTS
Provide evidence of whether a process has been operating

in a state of statistical control and to signal the presence of
special causes of variation so that corrective action can be
taken.
IMPLEMENTATION OF CONTROL CHARTS
22
Gas Mileage*
20
18
1 5 10 15 20 Week
•To keep this example as simple as possible, gas mileage values
are rounded to integer values.
20 Weeks of Gas Mileage Data for a Single Vehicle
20
∑ y [21 + 19 + 20 + Κ + 22] 400

i
y= =
i=1
= = 20
20 20 20

CONTROL CHARTS (cont.)
Gas Mileage
y
22
20
18
1 5 10 15 20 Week
∑ (y − y )
2
32
s 2
= = = 1.684
n −1 19
Variability in the Gas Mileage Data
Freq. of Weeks
18 19 20 21 22 Gas
Mileage
y
Histogram of Gas Mileage Data

CONTROL CHARTS (cont.)
μ ± 1σ ≅ 68.27%
μ ± 2σ ≅ 95.45%
μ ± 3σ ≅ 99.73%
Percent of Areas Under a Normal Curve
Gas Mileage
(mpg)
UCL = 23.9
24
22
20 ȳ = 20
18
LCL = 16.1
16
1 5 10 15 20 Week
Control Chart Based on Gas Mileage Data

MEASUREMENT
MEASUREMENT METHOD
METHOD MACHINE
MACHINE
M
Gage for Au an
Ride Brakes Tire Rotation t o
ua
Tire Press. l
Tire Press.
Estimate of

R R
ad Al eg
Front or ia l W ul
l X ar
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type Pr
Passengers H ec Snow
Terrain
A
Te
B
m
um ip Rain
W id ita
s
in pe it y tio
d
e
ra
at
n
a in
tu
Fl
ud
nt
Driver re
tit
u
Al
Mo
Fuel Additive
PEOPLE
PEOPLE MATERIALS
ENVIRONMENT
11 - 14
CONTROL CHARTS
• VARIABLES DATA
– Most powerful
– Two separate charts for looking at variability

• within a subgroup (e.g., range chart)
• between subgroups (e.g., X chart)
– Most common charts: X -R charts, IMR charts
– Note: when using variables control charts, the

range chart should be analyzed first to determine if it
is “in control”, since the average range is used to
determine the control limits for the X chart.
• ATTRIBUTE DATA
– Provides a record of quality history with data that is
often already being collected
– Most common charts: p charts (proportion

defective) and c charts (counts of defects)
Variables data: Attribute data:

• Length of cracks • Number of cracks
• Diameter of cracks • Proportion of units with cracks

WHAT KIND OF CONTROL CHART
SHOULD BE USED?
Need to
construct a
control chart
n = sample size or
subgroup size
Is n fixed Variables
x − S chart
No Variables (measurements) Attributes
and or attributes?
n < 10? (go-nogo)
Yes
n=1 2≤n≤9
IMR
x − R chart
chart
Defectives Defects Defects

or
defectives?*
Constant Constant
Yes No Yes No
sample area of
size? opportunity?
np-chart p-chart c-chart u-chart
* A unit is either defective or it is not (Binomial).

A unit may contain several defects (Poisson).

ILLUSTRATING THE X-bar R CONTROL CHART
14
13
12
11
10
9
8

7
6
5
4
3
2
1
9
8
7
6
5
4
3
2
1
Time 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 26 27 28 29 30
X
R
11 - 17
SUMMARY OF OUT-OF-CONTROL
SYMPTOMS
UCL
ZONE A
ZONE B
ZONE C
X ZONE C
ZONE B
LCL ZONE A
CONTROL CHART BROKEN INTO ZONES
• One or more points are outside the control limits

• 7 consecutive points are on the same side of the centerline
• 7 consecutive intervals are either entirely increasing or entirely
decreasing
• 2 out of 3 consecutive points are in the same Zone A or beyond
• 4 out of 5 consecutive points are in the same Zone B or beyond
• 14 consecutive points alternate up and down repeatedly
• 14 consecutive points are in either Zone C (i.e., in center third)

EXAMPLES OF
OUT-OF-CONTROL SYMPTOMS
UCL UCL
A A
B B
C C
X X
C C
B B
A A
LCL LCL
one or more pts are outside the control limits 7 cons. pts are on one side of the centerline
UCL UCL
A A
B B
C C
X X
C C
B B
A A
LCL LCL
7 cons. increasing or decreasing intervals 2 out of 3 cons. pts are in same Zone A or beyond
UCL UCL
A A
B B
C C
X X
C C
B B
A A
LCL LCL
4 out of 5 cons. pts are in same Zone B or beyond 14 cons. pts that alternate up and down
UCL
A
B
C
X
C
B
A
LCL
14 cons. pts in either Zone C (i.e., in center third)

CONTROL CHARTS USING SPC XL
(XbarR Chart Sample Data)
• TO ASSIST IN SELECTING THE PROPER CHART

– SPC XL > Control Charts > Control Chart Wizard
• TO ASSIST WITH DATA ENTRY (Optional)

– SPC XL > Control Charts > Create Template
• EXAMPLE:
– Create a template for an XbarR Chart (subgroup size = 3)
– Open the data file XbarR data.xls in the SPC XL directory
(or go to SPC XL > About SPC XL (Help) > Tutorial > Control
Chart Tutorials > XbarR and XbarS Charts > Open data file)
– Copy and Paste the Sample Data into the template you
created
Enter the data in the column provided below. XAxis labels are optional. When you
are ready to create the control chart select "SPC XL" - "Control Charts" - "XBarR
Chart..." from the menu.
XAxis Labels Data (Subgroup size = 3)

6/15/00 15 11 11
6/16/00 18 15 16
6/17/00 14 13 12
6/18/00 14 13 18
6/19/00 14 16 18
6/20/00 14 16 19
6/21/00 19 19 16
6/22/00 10 19 17
6/23/00 11 12 10
6/24/00 16 20 17
• CREATE THE APPROPRIATE CONTROL CHART

– SPC XL > Control Charts > XbarR Chart

CONTROL CHARTS USING SPC XL
(XbarR Chart Sample Data)
• EXAMPLE OF STACKED CONTROL CHARTS

SUBGROUPING STRATEGY
• Rational Subgroups
– Subgroups selected from a small region of

space, time, or product (homogeneous
conditions within the subgroup)
– Variation within a subgroup is representative of

common cause variation
– Changes in the process performance

(assignable causes such as procedures not
followed, differences in shifts, etc.) will appear
as differences between subgroups
– Typically consist of consecutive pieces, although

random samples can be used
– Subgroup sizes of 4 to 6 are common in the

parts industry, while individual measurements
(hence IMR charts) are common in chemical
processes and administrative areas
• Subgrouping Strategy Can Have a Profound Effect

on the Control Chart and What Can Be Learned

RECALCULATING CONTROL LIMITS
• CONTROL LIMITS SHOULD NOT BE

RECALCULATED EVERY TIME A NEW DATA
POINT IS PLOTTED
• REASONS FOR RECALCULATING LIMITS
– Initial (trial) limits based on a small amount of

data and better process data is now available
– A deliberate process change is made
– Process improvement activities have resulted in

demonstrated and sustained changes in the
average, standard deviation, or both

EXERCISE (Optional)
• Class as a whole
• Using one catapult, with the settings held

constant, each person will launch 3 times.
Consider each person’s 3 shots as a subgroup
of size n = 3 in an x − R chart.
• Use SPC XL > Control Charts > x − R chart to

evaluate process control.
• If we consider the specification limits to be x ± 5

inches, calculate Cpk.

WORKING WITH INDIVIDUAL VALUES
• Sometimes it doesn’t make sense to subgroup

your data
– Its expensive
– Data is generated only once per month or quarter,
so its infrequent
– There is no logical basis for grouping
• In these cases an IMR chart is used

– I = individual values
– MR = moving ranges (simply the difference between
successive points)
• Subgrouping helps to normalize the data, since

you’re plotting averages. When an IMR chart is
used, the individual values may or may not follow
a “normal” distribution. So, for an IMR chart, you
should always
– Plot a histogram of the data and verify normality
– If the data is not normal, then only apply rules 1-3
when checking for out of control symptoms (ignore
any rule violations for rules 4-7)
– If the data is normal, then it is okay to apply all 7
rules, as usual

INDIVIDUALS CHART FOR pH READINGS
11
10.5
10

9.5
8.5
7.5
7 Series1
Center = 8.044
Individuals Value (pH)

6.5
UCL = 10.5267
6 LCL = 5.5613
5.5
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
15-Minute Time Intervals
11 - 26
MOVING RANGE CHART FOR pH READINGS
3.5
3
Series1

2.5 Center = 0.9333
UCL = 3.052
LCL = 0
2
1.5
Range Value (pH)

1
0.5
0
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
15-Minute Time Intervals
11 - 27
EXERCISE (IMR CHART)
The average number of days to process an expense

statement is an important metric at the Travelers R Us
corporation. For the past three months, the average number
of processing days has decreased and the manager of the
accounting department has told everyone to “continue
whatever they’ve been doing because they’ve obviously made
some kind of improvement to the process and they’re
performing much better!”. Data for the past 15 months is
shown below. Is there any evidence to suggest that things
have improved? Why or why not?
month days to process

1 29
2 23
3 31
4 28
5 22
6 22
7 23
8 32
9 24
10 24
11 20
12 23
13 30
14 28
15 26

EXAMPLE OF ATTRIBUTE DATA
(c-chart)
Each unit in a production process is tested and inspected

and the total number of defects (scratches, dents, etc.) is
recorded. Data for the last 20 units is shown below:
Unit Count of Defects
1 4
2 0
3 3
4 1
5 5
6 3
7 4
8 0
9 3
10 2
11 5
12 1
13 2
14 6
15 8
16 7
17 6
18 8
19 7
20 9 The control chart indicates a
shift (7 points in a row on the
same side of the center line).
The process does not appear
stable, and should be
investigated to find out why.

INSTRUCTOR SUBSTITUTION RATES
(FALL SEMESTER) (p-chart)
0.12 Series1
Center = 0.0531
UCL = 0.0846
LCL = 0.0216
0.1

0.08
0.06
proportion
0.04
0.02
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
11 - 30
Week Number
PARETO CHART
80 REASONS FOR SUBSTITUTIONS
70
60

50
40
Percentage
30
20
10
0
TDY Meetings Leave Obs/Trng Medical Misc Retreats
11 - 31
p-CHART FOR SUBSTITUTIONS
0.12
Fall Semester
0.1 Spring Semester

0.08
0.06
UCL = .0497
0.04
Proportion of Substitutions
Center = .0248
0.02 New Policy
0 LCL = 0
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 26 27 28
Week Number
11 - 32
EXERCISE
p (proportions)-chart
c (count)-chart
• Break into teams. Each team will be given an M&M color

for which it will construct a p-chart.
• Collect data from all members in the class regarding the

number (x) of a given color each had in his/her bag and
the total number (n) in that bag.
• Use SPC XL > Control Charts > p-chart to investigate if

Mars Corporation has its process of putting that
particular color into the bags under control.
• p-chart data is put into the Excel spreadsheet in two

columns (or two rows): one column for x and a second
column for the corresponding n.
• Using only the x column (or row) generated above in the

Excel spreadsheet, use SPC XL > Control Charts > c-
chart to generate a c-chart (count chart) for that
particular color of M&M. Evaluate “color” control using
the c-chart.
• Which would be the better chart to use in this case?

CONTROL CHARTS - EXERCISE
Individually, or in small groups, discuss each of the scenarios below
and identify what type of control chart (X-bar R chart, IMR chart, p
chart, or c chart) would be appropriate:
1. Each week, between 1,000 and 2,000 travel reimbursement

requests are processed. If any information is incomplete or
incorrect, the request is returned to the submitter for correction.
The accounting department is interested in tracking the
percentage of incomplete/incorrect requests, because they are
concerned that the number of returned requests has been
growing.
2. The cycle time for processing change requests is measured

and recorded for each change request. The team suspects
there may be trends or patterns based on the day of the week
being processed, so they decide to look at a sample of 8
requests each day.
3. For each board, the cycle time to complete a particular test is

recorded. The team wants to monitor the test times to ensure
they are not increasing.
4. A product is inspected for scratches and other types of

damage during final inspection. Every occurrence of damage
is written up and the total number of problems is recorded for
each unit of product.

CONTROL CHARTS – EXERCISE (cont.)
5. A finance center takes daily samples of process vouchers with

a subgroup size n = 1,000. The following data depicts the total
number of errors in the subgroup of 1,000 for each of the past
25 days.
Day 1 2 3 4 5 6 7 8 9 10 11 12 13
Errors 25 24 23 23 27 25 21 27 39 23 22 23 27
Day 14 15 16 17 18 19 20 21 22 23 24 25
Errors 21 24 20 24 26 20 18 30 21 22 23 24
Comment on the state of process control.

PROCESS STABILITY, CAPABILITY,
AND FLEXIBILITY
Stable - Process that has only

common cause variation
Capable - Process that is stable and meets

current customer requirements
Flexible - A process which is capable of

meeting future customer
requirements

PROCESS IMPROVEMENT
PF / CE / CNX
Collect Data
Is Identify and
Process N remove special
Stable cause variation
Systematically
Is N investigate
Process common cause
Capable and improve the
process (DMAIC)
Is N Innovate or
Process totally
Flexible redesign
Standardize (SOP)
and Replicate

STRENGTH OF CONTROL CHARTS
• Provides a systematic and efficient method of gathering

data and transforming it into INFORMATION
• Allows us to make decisions based on FACTS rather
than intuition or memory
• Identifies non-random impacts to a process
• Warns of system degradations prior to making defects
• Forms a basis for improving a process within the
specifications
• Gets everyone involved
• Build expert data base
• Shows evidence of process improvements
LIMITATIONS OF CONTROL CHARTS
• Processes in control are not necessarily capable

• Problem identification takes time
• Control charts signal problems but not their reasons
• Analysis of past data that “happens to be available” may
not provide a clear understanding of the process

HOLDING THE GAINS
(CONTROL PLANS)
• A control plan is a means for holding the gains and

sustaining process improvements
• It documents information about the critical variables in the

process which we’ve learned about during our Lean Six
Sigma project
• It provides an action plan for what to watch, how often, in

what manner, by who, and corrective actions
• It empowers local process owners
• Its intent is to institutionalize the improvements

CONTROL PLAN (EXAMPLE FORMAT)
Process name: Prepared by: Orig. Date:
Process Owner: Approved by: Revision Date:
Customer: Core Team:
Performance Measurement Reaction Plan

Departments Process Step Indicators Requirements Msmt. Sample How Where (Corrective SOP
Description (Target, specs) Method Size Often? Who? Record? Action)
KPIV KPOV Ref.
Need accountability … owners of the metrics!
• Major Components
• Critical Process Operations and Steps
• Key Process Output Variables (primary and
secondary CTCs) with Targets and Specs
• Key Process Input Variables (KPIVs)
• Measurement System Methodology and Analysis
• Monitoring Methods and Responsibilities

PROCESS FLOW AND CONTROL
SYSTEM (PFCS) EXAMPLE
(Underwriting Process)
Process Description: Process Customer:
Underwriting Correspondent
Process Flowchart
Position
Mail
Step Coordi
Underwriting Corresp. Cap Mkts -nator
/Time
Obtain Loan
File from
B
U/W Queue
X2
Receives Docs
NT Obtain NT No for Suspended
Yes
Required? from Cap Approved? C /Conditioned
Markets Loan
X7
Underwrite Yes
Loan Resubmit
to U/W
X1 X3 X4 X5
No Suspend Fax Rec. Needs
Decisionable? Needs to Letter &
Loan seller Submit Docs
C
Yes
B
No Deny Lead Yes Counter No
Approved?
Loan Agrees? Offer?
X9
Update
System
Yes
No
Yes
Notify Retain
Submit Offer
Seller of File Copy
to Seller
Declination
Yes Receive
Fax
Conditions? Condition Send Original
Conditions Request & X6
to seller
Provide Docs
to Seller
X8
No Acceptable?
No B
Input in Yes
System
Fax
Receive Y1
Approval to
Approval
seller

PROCESS FLOW AND CONTROL
SYSTEM (PFCS) EXAMPLE (cont’d.)
(Underwriting Process)
CTC Outcome Quality Indicator(s):
Timeliness, Profitability, Accuracy Y1 = Time to complete loan U / W
Process and / or Misc.
Outcome Indicators Checking Information
Control Contingency Include: Contingency
Process Control Charts Plans, Abbreviations
Limits Checking Item Frequency Responsibility Plan
Procedures Remarks,
Outcome Indicators Charts Specs / Actions Req’d etc.
Targets for Exception
Y1 = Time to complete loan U / W. Every Loan Daily U / W Lead
IMR Chart
X1 = Avg cycle time of initial Every Loan Daily U / W Lead

Decisions = Initial decision cycle Reviewed
time divided by total loans reviewed.
IMR Chart
X2 = % Loans requiring NT = # Every Loan Daily U / W Lead

Loans requiring NT divided by # Requiring NT
credit packages received.
IMR Chart
X3 = Suspense Rate = # Loans Every Daily U / W Lead

suspended divided by # loans Suspended
received. Loan
IMR Chart
X4 = Average days in Suspense Every Loan in Daily U / W Lead

Queue. Suspense
IMR Chart
X5 = % Loans in Outstanding Every Loan in Daily U/W

Suspense > 10 days. Suspense
IMR Chart
X6 = # Days in Condition Queue = Every Loan with U/W

date final approval – date Conditions Daily
conditioned divided by # loans
removed from the Condition Queue.
IMR Chart
X7 = Average # of Conditions Every Loan with Weekly U / W Lead

iterations = # of re-evaluations of Conditions
Conditioned Loans divided by every
loan submitted.
IMR Chart
X8 = Counter offer cycle time = date Every Loan with Weekly U / W Lead
counter offer accepted / rejected – Counter Offer
date offer made.
IMR Chart
X9 = % Overturned Decisions = # Every Rejected Weekly U / W Lead

overturned divided by # rejected. Loan
IMR Chart Overturned

5S … ORGANIZING THE WORKPLACE
• 5S is a methodology to implement and maintain

a clean and well-organized workplace to maintain
high performance
• Based on the old adage “a place for everything

and everything in its place”

5S VIDEO
Take notes as you watch the video.

Think about the following :
• Where was 5S used? Give some examples.
• Why is 5S important? What are the benefits

(outcomes)?
• How might this apply in your area? Comments or

observations?
VIDEO

THE 5S SYSTEM
Organize the workplace …

1. Sort - Identify what is needed
(keep) and what is not (remove)
2. Set in Order - Determine a place for needed

S items … identify the storage
A place with clear markings
F
E 3. Shine - “Clean” the office and
T maintain the equipment
Y
4. Standardize - Office procedures, systems,
and policies
5. Sustain - Review regularly the SOPs

and insure compliance
Provide documentation so that the work environment

is clearly defined, self-regulating, and self-improving.
Makes the “hidden factory” more apparent.

GETTING 5S STARTED
SORT
• Identify those items that are needed and those

that are not
• Keep what is needed and remove what is not

(out of date manuals and documentation, etc.)
• Red-tag & put unneeded items in a holding area

for further review
• Determine methods to keep the unneeded items

from returning to the work area.
Red-tag
Do you see….
• Unnecessary things cluttering up
your workspace
• Tools or materials left around
• Paperwork lying around that isn’t
required
• Unnecessary duplicates of
anything
• Tools or office equipment that isn’t
properly classified and stored

RED TAG CHECKLIST
5S SORT Conditions
1. Record the date found
2. Record the condition from
Obsolete (Unneeded)
choices at right
3. Record location found
Defective (Fixable)
4. List name or description of item
5. Record your name or initials
(Unfixable)
6. Tag items to be moved
7. Place items, other than trash, in
Scrap (Reclamation)
Red Tag Holding Area
8. Set specific date for
final disposition Trash
Not needed in this area (Still useable elsewhere)
Occasionally needed here (Used weekly)
(Used monthly)
Date_______________
Seldom used here
Condition___________
Location____________
Item_______________
Name______________

5S(cont.)
SET IN ORDER
• Organize what’s left! For the items needed,

determine a place for them (“a place for
everything and everything in it’s place”)
• Identify the storage place for those items with

clear markings
• Allow space only for what is needed
• Identify how everyone will know the standards
In your workplace….
• Are aisles and storage
places clearly marked?
• Are workstations and
equipment clearly identified?
• Do current reports have a
place that is identified?
• Is anything stored around
fire extinguishers?

SET-IN-ORDER CHECKLIST
5S Set-in-order
1. Determine need first
2. Determine good home location. Ask workers
for input to determine safe, easy to use,
location.
3. Mark location, with visual factory principles in
mind.
4. Brief all operators, all shifts, to aid understanding
and compliance.
5. Include check of these locations into the work
area checklist.
Work-in-process/Materials Needed Unneeded No Home Has Home

Finished goods _____ _____ _____ _____
Shipping materials _____ _____ _____ _____
Raw materials _____ _____ _____ _____
Partial assemblies _____ _____ _____ _____
Spare parts _____ _____ _____ _____
Equipment
Molds/Dies _____ _____ _____ _____
Fixtures/Jigs _____ _____ _____ _____
Carts/dollies _____ _____ _____ _____
Supplies/Apparel
Documents/charts _____ _____ _____ _____
Helmets/shoes _____ _____ _____ _____
Coats/aprons _____ _____ _____ _____
Cleaning supplies _____ _____ _____ _____

5S (cont.)
SHINE
• Make sure work area is clean and free of debris
• Keep equipment (copiers, printers, computers,

etc.) clean and maintained
• Establish a schedule for maintaining the shine
In your workplace….
• Look carefully at the floor,
aisles and around
machines. How much dirt,
oil, and dust can you find?
• Is there paperwork lying
around?
• Do the computers, desks, or
chairs require cleaning?
• Is there food or garbage just
lying around?

SHINE CHECKLIST
5S Shine
1. Use Shine Checklist
2. Determine items or areas to be cleaned
3. Mark location, with visual factory principles
in mind.
4. Brief all operators, all shifts, to aid
understanding and compliance.
5. Include check of these locations into the
work area checklist.
Open Areas Clean Dusty Oily/Greasy Worn Paint

Floor area _____ _____ _____ _____
Stairs _____ _____ _____ _____
Underside, if open _____ _____ _____ _____
Behind equipment _____ _____ _____ _____
Top of equipment _____ _____ _____ _____
Ceilings & Aisles _____ _____ _____ _____
Equipment
Machine surfaces _____ _____ _____ _____
Fixtures/Jigs _____ _____ _____ _____
Carts/dollies _____ _____ _____ _____
Supplies/Apparel
Tool boxes/Tools _____ _____ _____ _____
Drawers/Cabinets _____ _____ _____ _____
Trash/Wastes bins _____ _____ _____ _____
Cleaning Equipment _____ _____ _____ _____

5S (cont.)
STANDARDIZE
• Identify actions to be taken to ensure

cleaning and inspection continues. (make
sort, set in order, and shine a daily habit)
• Document the process and procedures

with checklists, guidelines, flowcharts,
diagrams, photographs, etc.
• Ensure that everyone is properly trained

on the correct procedures and that these
procedures are the current best practices
• Assign and document everyone’s

responsibility towards achieving this step

5S (cont.)
SUSTAIN
• Ensure the first four steps are sustained

as part of the daily routine
• Employ processes to monitor

compliance such as regular inspections
using checksheets
• Ensure that the targets and expectations

are clear to everyone
• Identify the areas that need

improvement and place importance on it
• Ensure that the procedures are simple to

understand and to follow
• Reward and recognize success

5S CLASS EXERCISE (OPTIONAL)
• Use the five steps to 5S this room.
• On a flip chart capture the following:
– The steps and output from each step in

the process
– A layout map of the before and after
– Minimum and maximum quantities for

what is to be kept in the room.

VISUAL FACTORY MANAGEMENT
Definition:
Innovative method of providing
valuable information to everyone!
Employee involvement and management’s

support is CRUCIAL!!!

BENEFITS OF VISUAL CONTROLS
Visual controls are all around us,

and we use them every day to stay
within recommended guidelines.
Speed limit signs, directional signs,
yield signs, parking signs, signs are
everywhere.
We stay out of harm’s way by using

crosswalks, turn signals, and
staying within lanes for turnoffs.
How many visual controls do you

have at home and work?

BENEFITS OF VISUAL CONTROLS
(cont.)
• Cleaner and Safer Workplace

Replace A
fter Use
• Decreased Inventories
• Less Wasted Time
• Improved Morale
• Improved Product Quality
• Make Working Areas User

Friendly by:
• Answering questions
• Identifying equipment,
materials and locations
• Describing actions and
procedures
• Providing safety
warnings and
precaution information

ARE THERE VISUAL CONTROLS
AVAILABLE?
• Are types of required lubrication identified?
• Where and controlled by whom?
20 30
10 FULL
40
0
• Are pressure specifications posted

and understood by the operators
and maintainers?
• Where and controlled by whom?

USING VISUAL CONTROLS
HARD TODAY’S
HAT GOAL
AREA 6 0 0
PRODUCTION RESULTS
TODAY’S
SCHEDULE SOPs
RUN
1 3 7
QUALITY RESULTS
LOW
STATION 3
OPERATOR
FIRE SLEDGE
EXSTINGUISHER HAMMER CLEAN
GLOVES
DIRTY
GLOVES
WARNING
HOT STEEL

USING VISUAL CONTROLS (cont.)
Pareto of Suggestions Implemented
For May 2000
9
Make sure the metrics that
Quantities by Type are posted are large
enough to be seen from a
reasonable distance.
Managers and workers do
4 not want Visual Controls
that they have to squint to
3
see.
2
Provide trend information

so everyone will know if
Mech. Elect. Pneumatic Safety
they are doing better or
Defects Found Before Shipping
worse at a glance.
For May 2000
Package Not
Sealed Post those items that the
Loose Wires
6
1
Cover
workers can control:
Scratched • Absenteeism
2
• Improvements Identified
• Safety Incidents
• Production Per Worker
Missing Instructions Packet
19 • Defect/Rework Rates

VISUAL BOARD EXERCISE (OPTIONAL)
• Divide the class into 2 groups
• One group to develop a visual board for the

implementation of Lean Six Sigma at their site. Where is
it to be placed, who owns and needs to keep it updated?
What should be in it?
• The other group is to develop a standardized visual

board for a Lean Six Sigma project. This is to be
documented so that a copy can be given to the rest of
the class.

DOCUMENTING THE PROJECT
• PROJECT NOTEBOOK…BE SURE TO INCLUDE

– Updated Project Summary
– Demonstration of Lean Six Sigma tools used
– Demonstration of (sustained) gains
– Financial and business benefits
• FORMAL PROJECT REPORTS…MAY BE ONE OR

MORE OF THE FOLLOWING
– Executive Summary (one or two pages)
– Written report including
• Business Case (Why This Project?)
• Summary of activities during each phase (Define,
Measure, Analyze, Improve, and Control)
• Key lessons learned
• Results
– PowerPoint or other electronic presentation
• COMMUNICATION OF RESULTS
– Company intranet and/or project database
– Newsletters
– Team meetings
– Department meetings
– Emails
– SOP’s
– Word of mouth

CREATING A PROJECT STORYBOARD
• A project storyboard is a nice, high level

summary of the project
• It helps to summarize project actions and

provides a pictorial record of the project
• To create a storyboard
– Use the DMAIC structure
– Describe each of the main steps (Define, Measure,
Analyze, Improve, and Control) and include 2-4 bullet
points for what you did/learned during that phase
– Include several graphics which will communicate the
key messages

PROJECT STORYBOARD (cont.)
Example Format
Stage Brief Description Supporting Graphics (examples)

Define • Discuss project, Charter
• purpose VOC
charter, business • goals data
De A
De t. B
C
De .
case, etc.
pt.
pt.
tg
p
Mk
TEAM
IPO
Person 1 *
2 *
3 *
4 *
Current State
Measure • Discuss current Pareto of
Problems
state (baseline VALUE ADDED NON-VALUE ADDED
analysis)
16
Data over time 14
12
histogram
Number of Cars
10
0
6 to <12 12 to <18 18 to <24 24 to <30 30 to <36 36 to <42 42 to <=48
Gas Mileage (mpg)
1
Analyze • Describe problems, Damage from
suppliers
2 scatterplot
Quality shipping
3
causes, and how issues
(defects)
company
Late
Shipments
they were verified 4

Address
incorrect
Database
errors 3
• Analysis and 40
focus on the
interpretation of data
35
30 causes that occur

most frequently!
Frequency
25
20
15
10
0
Type Shipping Error
Implementation plan
Improve • Analysis of results ID TASK 1 2 3
DAY #
4 5 6 7
Before After
from tests (pilots) 1.0
1.1
1.2
Assess the measurement
system capability
Brief the team
Select items to be
included in the study
Determine the sample
1.3
size
Take data
• Improvements made
1.4
(measurem ents)
1.5 Analyze the data
Determine actions to be
1.6
taken
. .
. .
. .
and criteria used

5.2
Collect Data on Recent
Customer Com plaints
(Priority)
IPO Matrix Use Improve
Low existing Easy to Cycle
of Solutions Cost resources Implement Time
• Implementation plan ( ) ( ) ( ) ( )
Solution #1
Solution #2
Solution #3
Solution #4
Total:
Control • Summary of key

learnings and next 170
Number of Hours between Unscheduled Maintenance Activities
New process flow 160
steps Future
150
140
130
improvement
• Updated SOPs, 120
110
100
PFs, etc. 90
80
17
41
13
21
23
25
27
31
33
37
45
1
11
15
19
29
35
39
43
47
Month Number


MODULE 12
DMAIC SUMMARY
LEAN SIX SIGMA PROJECT MASTER
STRATEGY (DMAIC)
… the Roadmap
customer and the
business? Clearly
define the problem,
project, and team.
Define
Pr
oj thi
ec s
t s wa
uc y
ce
ss
…
Measure
Define the current

to assess
performance.
Analyze
What needs
improvement?
Identify the root
I mprove
Make the changes

and analysis.
Control
repeat problems.

DMAIC ROADMAP
Define Project
Define
Tollgate Review
Understand Current
Process (baseline)
Measure
Tollgate Review
Analyze Causes of
Poor Performance
Tollgate Review

Work Flow
Improve
Test and Confirm
Solutions
Tollgate Review
No
OK
Yes
Control
Measure Success
No
OK
Yes
Celebrate
DEFINE
Define




Improve
Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go
Yes
MEASURE
Close Down Project

MEASURE
Define
Measure

Analyze


Notebook
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

ANALYZE
Define

Measure
Gage for
Ma

Au
nu
Tire Press.
Tire Press.
to
al
Estimate of
Ra
Re

Front or
All
dia
gu
WX

lar
l
Engine Warm-
Air Conditioner
GAS
MILEAGE
Training
A
Weather Conditions
# of Ice
Oil Type
Passengers Snow
Terrain
Pr
A
Rain
Hu
ec
Te
ipi
mi
s
Wi
mp
ain
tat
dit
e
t
Fla
nd
tud
era
ion
y
unt
Driver
tur
Alti
Mo
Fuel Additive
Acctng
HR
Legal
Functional
Area
Benefits
problems
Frequency or
Financial Measure

potential causes
Improve

Control

necessary
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes
IMPROVE
Define
170

160
150
140
130
120
110
100
90
80
1
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
45
47
Month Number
Analyze

SOPs
Improve
necessary
Control
No
Toll
Gate Review No Go
Kill Project?
Go/No Go
?
Go Yes

CONTROL
Define
UCL
CENTERLINE
success
LCL
Measure
Performance No
project goals
achieved? Analyze
Yes
Determine control plan for

holding the gains Improve

actions or follow-up; validate business
impact Control

team, etc.

Gate Review No Go
?
Done
Celebrate

THE DEFINE PHASE - Tollgate
Define
Measure
Analyze
Improve
Control Define Phase Completion Checklist

Do you have a problem statement which describes the
problem without offering solutions?
Do you, your sponsor, and the team agree on the
project, problem statement, goals and measures and do
you have clear descriptions of each?
Do your project goals pass the “S.M.A.R.T.” test?
Do you have one or at most two clearly identified
primary project measures? What, if any, secondary
measures have you identified?
Do you have clearly identified project boundaries,
including what is in and out of scope?
Have you identified the key stakeholders? What are
their main concerns and what is your plan for getting
buy-in?
Do you have a good, cross-functional team?
Have you created a high level process map (SIPOC
diagram)?
Do you understand who your customers are, and what’s
most important to them?
Have you met with Finance to understand ways to
quantify the project benefits?
Have you updated your project/knowledge notebook and
any enterprise project tracking database?
Copyright Air Academy Associates 12 -8

THE MEASURE PHASE - Tollgate
Define
Measure
Analyze
Improve
Control
Measure Phase Completion Checklist
Do you and your team have a good understanding of the

current “as-is” process?
Have you mapped the value stream and / or mapped the
sub-processes within the value stream, at a high level
with 5-10 steps and at lower levels as needed?
Do you have a clear plan for what needs to be measured
and have you defined the measures precisely?
Do you have a data collection strategy and have
accurate, representative baseline data?
Do you have concerns about the accuracy or reliability of
the measurement system being used for collecting your
data? If so, do you have a plan for addressing these
concerns?
Do you understand the customer requirements for all
CTCs? Have you assessed the capability of the current
process in terms of meeting these requirements, using
measures such as Cp, Cpk, FPY, etc.?

THE ANALYZE PHASE - Tollgate
Define
Measure
Analyze
Improve
Control
Analyze Phase Completion Checklist
Have you reviewed the process flow (PF) to identify non-

value added steps, bottlenecks, missing steps,
sequencing problems, and other sources of waste and / or
variation?
Did you create a physical process map and / or time value
map if applicable?
Do you know the ratio of value added to non-value added
time, wait time, etc.?
Did you calculate TAKT time and perform any sizing
calculations?
Did you brainstorm all the potential causes of the
problems, focusing on quantity of variables rather than on
a particular area or jumping to solutions too early?
Did you partition the variables using C, N, X? Did you
look for opportunities to turn “Ns” into “Cs”?
Did you screen the list of potential variables to focus on
the critical few?
Did you use data rather than emotion, and can you
explain how you chose the critical few causes / variables
to focus on? What are the critical few? What are the root
causes of the problems?
Do you understand the key areas to be improved from
your analysis?

THE IMPROVE PHASE - Tollgate
Define
Measure
Analyze
Improve
Control
Improve Phase Completion Checklist
Did you explore alternative solutions? What were they?

How did you decide on the best solutions? Did you get
buy-in from your project Sponsor?
Did you pilot or test the solutions and evaluate the
results to confirm improvement in your primary project
measures (CTCs) without degradation to your secondary
measures?
If the results of the pilot were not acceptable, what
actions did you take?
How did you measure the improvement? (graphical?,
statistical?)
Did you focus on making value flow at the pull of the
customer?
Did you investigate and achieve any setup reductions (if
applicable)?
Did you determine any changes needed to the physical
layout or flow?
Did you revise your process flow, update the SOPs, and
mistake-proof the new process?

THE CONTROL PHASE - Tollgate
Define
Measure
Analyze
Improve
Control Phase Completion Checklist
Control
Did you involve and train the team, process owner,

and other key stakeholders in their new SOPs /
standard work?
Did you complete and achieve sustainability in 5S
activities?
Do you have a plan in place to hold the gains and
ensure long-term accountability?
Did you implement any visual controls?
Did you implement any methods to manage the work
flow?
Did you prepare a project storyboard, documenting
your team’s work?
Did you communicate lessons learned, best practices,
future project ideas, and so on?
Did you work with your Finance organization to
quantify the project benefits?
Did you review the project work and
recommendations with your Sponsor?
Did you ensure that the customer was pleased with
the improvements?
Have you updated your project/knowledge notebook
and any enterprise project tracking database?

DMAIC TOOLS – MEMORY JOGGER
Tools I might use:
DEFINE
9 What’s important to the business? customer?
9 What are the right projects to be working on?
Select the right project and define the business
case
9 Get a team formed and begin to understand the
voice of the customer
9 Identify key stakeholders
9 Define the process
MEASURE
9 How well are we doing? What does the
current process look like and how is it
performing? (baseline) (current state)
9 What about our measurement process? Is it
accurate, repeatable, and reproducible?
9 What data do we need to collect?
9 What is our current process capability?
ANALYZE
9 What are the problems? Bottlenecks?
9 What are the sources of waste? non-value
added tasks?
9 What are the key input variables and output
variables in the process?
9 Where should we focus? What are the root
causes of the problems?
IMPROVE
9 Identify ways to optimize performance and
minimize variability / waste
9 Investigate and identify candidate solutions
9 Test out solutions and mistake proof the
process
9 Re-assess the capability of the new process
9 Define new process flows, SOP’s, etc.
CONTROL
9 Measure Success. Quantify the gains. ($,
cycle time, performance)
9 Put a plan in place to hold the gains; determine
accountability; monitor the process
9 Train team on updated process flow and SOP’s
9 Communicate results, share lessons learned,
document future plans and recommendations
9 Create a storyboard/summary of the project

MODULE 13
REFERENCES, GLOSSARY OF TERMS,

AND COURSE EVALUATION FORMS
AN ACTIVITY AND LEAN SIX SIGMA
TOOLS AND METHODOLOGY
CROSS REFERENCE GUIDE Most
Commonly
Used During:
D M A I C
E E N M O
F A A P N
Where I S L R T
in N U Y O R
Guide? E R Z V O
If you want to… Then consider using… (Module) E E E L
Understand the Lean Six Sigma Strategy and DMAIC, Knowledge Notebook 1
• • • • •
Roadmap
Plan for documentation Knowledge Notebook 1, 3 • •
Lead effective meetings and enhance teamwork “PACER”; team building 3 • • • • •
Capture the Voice of the Customer QFD, House of Quality 4 • •
Define and analyze key project stakeholders Stakeholder analysis 3 •
Define project goals and work on problem SMART goals, problem 3
•
statements and scoping statements, project charter
Define and quantify the Cost of Poor Quality COPQ, 7 wastes 2, 7 • •
Define and understand your process IPO diagram, SIPOC, value 2, 4, 5
stream map, time value and
• • •
physical flow maps,
PF/CE/CNX/SOP
Describe variables data graphically Histogram, box plot, run chart 5 • • •
Describe categorical data graphically Pareto diagram 5 • • •
Summarize variables data numerically Mean, std dev., sigma level, 2, 5, 8
• • •
Cp, Cpk, confidence interval
Summarize attribute data numerically FPY, dpu, dpmo, sigma 2, 5, 8
• • •
capability, confidence interval
Prioritize data IPO Matrix, FMEA, Pareto, 3, 8
Effort Impact analysis, voting, • • • •
pairwise comparison
Look for relationships with variables data Correlation coefficient, scatter 5
• •
diagram, regression
Look for relationships with count data Chi Square Test for 9
• •
independence
Test for significant shifts in the mean t-test 9 • •
Test for significant shifts in the standard deviation F-test 9 • •
Test for significant shifts with proportion data Test for proportions 9 • •
Assess the capability of a process Cp, Cpk 5 • • •
Evaluate your measurement system to see if you MSA 6
• •
have integrity in the data
Look for trends in the data over time Run chart 5, 11
• • •
Control chart
Reduce changeover time SMED 10 • •
Reduce extraneous variation and waste PF/CE/CNX/SOP 2, 7 • • • •
Organize the workplace and set up visual controls 5S, work cell design, visual 10, 11
• •
controls
Determine sample size for estimating means and Sample size (normal and 8
• • •
proportions binomial distributions)
Develop a plan for sustaining the gains made on Control plans 11
•
your project
Mistake proof your process FMEA; poka yoke 10 •
Determine if your process is stable and repeatable Control charts 11 • • •
Copyright Air Academy Associates 13- 1

REFERENCES
BASIC STATISTICS
¾ Kiemele, Mark J., Schmidt, Stephen R., and Berdine, Ronald J., Basic
Statistics: Tools for Continuous Improvement, 4th ed., Air Academy Press,
1997.
¾ Moore, David S., The Basic Practice of Statistics, 2nd ed., W. H. Freeman
and Company, 2000.
DESIGN OF EXPERIMENTS
¾ Barker, Thomas B., Quality By Experimental Design, 2nd ed., Marcel
Dekker Inc., 1994.
¾ Schmidt, Stephen R. and Launsby, Robert G., Understanding Industrial
Designed Experiments, 4th ed., Air Academy Press, 1998.
¾ Taguchi, G. and Konishi, S., Taguchi Methods: Orthogonal Arrays and
Linear Graphs, American Supplier Institute, Inc., 1987.
¾ Advanced DOE:
¾ Box, George E. P., Hunter, William G., and Hunter, J. Stuart, Statistics for
Experimenters, John Wiley & Sons, 1978.
¾ Box, George E. P. and Draper, Norman R., Empirical Model-Building and
Response Surfaces, John Wiley & Sons, 1987.
¾ Hicks, Charles R., Fundamental Concepts in the Design of Experiments, 5th
ed., CBS College Publishing, Holt, Rinehart, and Winston, 1999.
¾ Montgomery, Douglas C., Design and Analysis of Experiments, 5th ed.,
John Wiley & Sons, 2000.
SIX SIGMA / KNOWLEDGE BASED MANAGEMENT

¾ Adams, Murray, Kiemele, Mark J., Pollock, Lee, and Quan, Tom, Lean Six
Sigma: A Tools Guide, Air Academy Press, 2003.
¾ Schmidt, Stephen R., Kiemele, Mark J., and Berdine, Ronald J., Knowledge
Based Management, Air Academy Press, 1996.
¾ Chowdhury, Subir, The Power of Six Sigma, Dearborn Trade, 2001.
¾ Joiner, Brian L., Fourth Generation Management, McGraw Hill, 1994.
¾ Tayntor, Christine B., Six Sigma Software Development, Auerbach
Publications, 2003.
BENCHMARKING
¾ Camp, Robert C., Benchmarking, ASQ Quality Press, 1989.

REFERENCES (cont.)
CHANGE MANAGEMENT, TEAMS, and the HUMAN SIDE
¾ Blaha, Robert B., Beyond Survival: Creating Prosperity Through People, Air
Academy Press, 1995.
¾ Hutton, David W., The Change Agents’ Handbook, ASQC Quality Press, 1994.
¾ Johnson, Spencer M.D., Who Moved My Cheese?, G.P. Putnam’s Sons, 1998.
¾ Palmer, Brien, Making Change Work, ASQ Quality Press, 2004.
LEAN MANUFACURING AND LEAN THINKING

¾ Adams, Murray, Kiemele, Mark J., Pollock, Lee, and Quan, Tom, Lean Six
Sigma: A Tools Guide, Air Academy Press, 2003.
¾ Goldratt, Eliyahu M., The Goal, Second Revised Edition, North River Press,
1992.
¾ Lareau, William, Office Kaizen (Transforming Office Operations Into a
Strategic Competitive Advantage), ASQ Quality Press, 2003.
¾ Sharma, Anand and Patricia E. Moody, The Perfect Engine, The Free Press,
2001.
¾ Womack, James P. and Daniel T. Jones and Daniel Roos, The Machine that
Changed the World: The Story of Lean Production, Rawson Associates, 1990.
¾ Womack, James P. and Daniel T. Jones, Lean Thinking, Simon and Schuster,
1996.
QUALITY (GENERAL)
¾ Brassard, Michael and Ritter, Diane, The Memory Jogger II (Pocket Guide of
Tools for Continuous Improvement and Effective Planning), Goal/QPC, 1994.
¾ Deming, W. Edwards, Out of the Crisis, MIT Center for Advanced Engineering
Study, 1986.
¾ Ishikawa, Kaoru, Guide to Quality Control, Asian Productivity Organization,
1982.
¾ Juran, J. M., Juran on Planning for Quality, The Free Press, 1988.
¾ Lam, K. D., Watson, Frank D., and Schmidt, Stephen R., Total Quality: A
Textbook of Strategic Quality, Leadership & Planning, Air Academy Press,
1991.
¾ Walton, Mary, The Deming Management Method, The Putnam Publishing
Group, 1986.
SPC
¾ Grant, Eugene L. and Leavenworth, Richard S., Statistical Quality Control,
McGraw Hill, 1988.
¾ Wheeler, Donald J. and Chambers, David S., Understanding Statistical
Process Control, 2nd ed., SPC Press, 1992.

REFERENCES (cont.)
FMEA
¾ Automotive Industry Action Group (AIAG), Potential Failure Mode and
Effects Analysis – Reference Manual, 3rd ed., 2001.
¾ IEC Standard Publication 812, Analysis Techniques for System Reliability –
Procedure for Failure Mode and Effects Analysis, 1985.
¾ Stamatis, D. H., Failure Mode and Effect Analysis, ASQC Quality Press,
1995.

¾ Automotive Industry Action Group (AIAG), Measurement Systems Analysis
(Third Edition), 2002.
¾ Barrentine, Larry B., Concepts for R&R Studies, ASQ Press, 1991.
QUALITY FUNCTION DEPLOYMENT (QFD)

¾ Cohen, Lou, Quality Function Deployment: How to Make QFD Work for
You, Addison-Wesley, 1995.
ADVANCED STATISTICS/TOOLS
¾ Anderson, T. W., An Introduction to Multivariate Statistical Analysis, 3rd ed.,
¾ Draper, Norman R. and Smith, H., Applied Regression Analysis, 3rd ed.,
¾ Johnson, Richard A. and Wichern, Dean W., Applied Multivariate Statistical
Analysis, 4th ed., Prentice-Hall, Inc., 1998.
¾ Hocking, Ronald R., The Analysis of Linear Models, Brooks/Cole Publishing
Co., 1985.
¾ Myers, Raymond H., Classical and Modern Regression with Applications,
2nd ed., Duxbury Press, 1990.
¾ Neter, John, Wasserman, William, and Kutner, Michael, Applied Linear
Statistical Models, 4th ed., Richard D. Irwin, Inc., 1996.
¾ Ostle, Bernard and Malone, Linda C., Statistics in Research, 4th ed., Iowa
State University Press, 1987.
¾ Seber, G. A. F., Linear Regression Analysis, 2nd ed., John Wiley & Sons,
2003.

GLOSSARY OF TERMS
Reference
Term/Tool What It Stands For Description
(module)
Data which is not measured on a continuous
attribute data 2 good/bad, pass/fail data
scale. (Qualitative).
A graphical tool which provides a visual

depiction of a data set. It summarizes both the
box plot 5
location and dispersion of the values in a data
set breaking the data into quartiles.
A type of control chart for count data, based on

c-chart 11 “count” chart
the Poisson distribution.
Shows the possible causes (or process inputs)
CE 2 Cause and Effect Diagram for a given effect (or process output).
Generated via brainstorming.
All of the activities associated with switching the
materials, operating settings or tooling on a
changeover 10
piece of equipment so that it can produce a
different part or perform a different task.
Says that for most cases, regardless of the
shape of a distribution of a population, sample
averages drawn from the population will behave
CLT 8 Central Limit Theorem
like they come from a normal distribution. This
allows us to use the normal distribution in a
variety of situations.
This is the way we partition the input variables
on a cause and effect diagram. Constants (or,
Constant, Noise,
CNX 2 controlled variables) require standard operating
Experimental
procedures to identify the way these variables
will be held constant.
Variability in a process that is truly random;
common cause 11 common cause variation
inherent variability.
A chart showing a process parameter over time.
Used to determine if the process is stable. Also
control chart 11
useful in identifying shifts, trends, and special
causes of variation.
Costs associated with scrap, rework, etc. Five

main categories are: internal failure costs,
COPQ 2, 7 Cost of Poor Quality
external failure costs, appraisal costs,
prevention costs, lost opportunity costs.
A measure of the linear relationship between

correlation 5
two variables.
Capability Index (potential)

Measures the potential capability of a process
Cp 5 USL − LSL without considering centering. A Six Sigma
Cp = process will have a Cp value or 2 or larger.
6σ

GLOSSARY OF TERMS (cont.)
Reference
(module)
Capability Index (actual) Measures the actual capability of a process,
taking into account the centering of the process.
Cpk 5 ⎡ USL − y y − LSL ⎤
Cpk = min ⎢ , A Six Sigma process will have a Cpk value of
⎣ 3σ 3σ ⎥⎦ 1.5 or larger.
A strategy for process improvement and gaining

Define, Measure, Analyze,
DMAIC 1 knowledge. It involves using the right tools at
Improve, Control
the right time to answer the right questions.
A systematic method for varying the inputs of a

process in order to observe the outputs and
DOE Design of Experiments
learn about the relationships, which factors are
most critical, etc.
A measure of quality for discrete or attribute

defects per million data which is used for benchmarking purposes.
dpmo 5
opportunities A Six Sigma process will have no more than 3.4
defects per million opportunities.
A measure of quality for discrete or attribute

dpu 5 defects per unit
data.
A 5-step process to achieve and maintain a

sort, set in order, shine, clean and well organized workplace and to
5S 11
standardize, and sustain improve workplace efficiency. “5S plus 1” refers
to “safety”, throughout the entire 5S process.
A simple yet effective tool that can be used to

identify the root cause of an issue by asking the
5 Why’s 8 question “Why?” up to five times. The use of
this tool will help develop increased
understanding of an issue.
A systematic way of looking at how a process or

product can fail and working to identify ways to
Failure Mode and Effect bullet proof it so we don't have escapes to the
FMEA 10
Analysis customer. FMEA's help us prioritize where to
focus attention based on the severity and
likelihood of failures
The percentage of units completing a process

FPY 2, 5 First Pass Yield
the first time with no rework or repair.
A bar chart that depicts the frequencies of

histogram 5 numerical or measurement data. It shows the
distribution of data.
Statistical procedure for testing a claim about a
hypothesis test 9 Examples: t-test, F-test population parameter or for comparing 2 means,
standard deviations, or proportions

Reference
(module)
Individual Moving Range A control chart for variables data with individual
IMR 11
Chart measurements (no subgroups)
A diagram for describing all activities as a

Input-Process-Output process, showing the inputs (sources of
IPO 2
Diagram variation) and outputs (measures of
performance) of a process.
A matrix for helping prioritize a list of input

Input-Process-Output
IPO Matrix 3, 8 variables (based on their relationship to
Matrix (simple QFD)
customer requirements).
Continuous improvements in an activity over

Kaizen time which reduce waste. Such improvements
are usually small and incremental in scope.
A Japanese term which means “card signal”.
More generally, it is used to describe any visual
Kanban
signal used to indicate the need for material
replenishment by an upstream activity.
A strategy for obtaining the knowledge about
our products and processes which involves
Knowledge-Based asking and answering the right questions. KBM
KBM 2
Management is an extension of the quality philosophies of
Deming, Juran, and Ishikawa with a focus on
the "hows".
Air Academy Associates’ approach to training.
A method of focusing on problem solution rather
KISS 2 Keep It Simple Statistically
than statistical theory. (kiss”ING” involves
numerous graphics!)
Limits which are used to determine whether a
process is stable (in statistical control). They
Lower Control Limit and
LCL and UCL 11 are calculated from the data and typically set at
Upper Control Limit
the +/- 3 sigma limits (note: control limits are
NOT spec limits).
A business strategy combining the power of
both traditional lean and traditional Six Sigma
Lean Six Sigma 1 methods. Focuses on reducing waste and
variation, enhancing customer value, and
improving business profitability.
Lower Specification Limit The bounds of acceptable values for a given
LSL and USL 2 and Upper Specification product or process, as specified by the
Limit customer.
A measure of the center of a distribution. Also

mean (μ) 2, 5 Mean (pronounced "mu")
known as the average or balance point.
The center of a distribution. If calculated from

median 5 data, it is the value that has 50% of the data
above and 50% below.

Reference
(module)
Method for analyzing the variation present in a
Measurement System
MSA 6 measurement system and assessing its
Analysis
capability to make good measurements.
A probability distribution for continuous data,

normal distribution 5
characterized by the smooth, bell-shaped curve.
Activities which consume time or resources but
NVA 2, 7 non-value added do not directly contribute towards meeting the
customer’s requirements.
A data point which appears to be different or
unusual (not consistent with the rest of the data
or distribution). Outliers should be carefully
outlier 2
examined. Before throwing out an outlier, we
should understand the cause and be able to
control or prevent it from happening via an SOP.
A control chart for attribute data obtained by
p-chart 11 “proportions” chart
plotting the proportion of failures/successes.
A bar chart for attribute (or categorical) data that
Pareto diagram 5 is presented in descending order of frequency.
It is especially useful for prioritization.
A flowchart of a process listing the process

steps. Can be done on a "as is" basis as well
as "should be" and "could be". Process
PF 2 Process Flow
flowcharts help us identify non-value-added
steps, potential input and output variables,
ensure consistency, etc.
A powerful methodology that can be used to
remove up to 70-80% of extraneous process
PF/CE/CNX/SOP 2
variation. It is the first line of defense against
variation.
Sometimes referred to as a “spaghetti diagram”.
It is a layout diagram of the work area that
Physical process illustrates the path followed by materials or parts
2, 5
map through the facility. It can be used to highlight
excessive amounts of material transport or
employee movement.
A method for making processes immune to
Poka Yoke 2, 10 Mistake-proofing
operator or user error.
A series of activities, administrative or

production-oriented, which are usually
Process 2 performed in a particular sequence to
accomplish a specific objective or produce a
desired product.

Reference
(module)
Precision to Tolerance A measure of the capability of the measurement

Ratio process. It is the proportion of specs that is
P/Tol 6 consumed by the measurement process.
6*σ
P/Tol = Values less than .10 indicate a good
USL − LSL measurement process.
Precision to Total Ratio

A measure of the capability of the measurement
σ process. It is the proportion of the total
P/Total 6 P/Total = msmt standard deviation that is consumed by the
σ total
measurement process standard deviation.
The probability of making a type I error

(concluding the alternative hypothesis is true
p-value 9 significance level
when the null is actually true). It provides the
exact level of significance of a hypothesis test.
A systematic process for integrating customer

Quality Function
QFD 4 requirements into every aspect of the design
Deployment
and delivery of products and services.
A measure of spread in a set of data. It is the

R 5, 11 Range
largest value minus the smallest value.
A statistical technique for determining the

regression mathematical relationship between the output of
a process and one or more input variables.
In an MSA, repeatability is the variation

obtained by the same person using the same
repeatability 6 repeatability instrument on the same product or service for
repeated measurements. It measures the
variability due to the gage or equipment.
In an MSA, reproducibility is the variation

reproducibility 6 reproducibility obtained due to differences in people (or test
equipment) that are taking the measurements.
ROT Rule of Thumb A guideline or quick test we can apply.
In an FMEA, this is the column which indicates

high risk process steps or elements of a design.
RPN 10 Risk Priority Number
It is equal to Severity X Occurrence X Detection.
Higher numbers are worse.

Reference
(module)
The square of the correlation coefficient. It
Also called coefficient of represents the strength of a regression model
R-squared metric
determination by measuring the proportion of variation in the
data which is accounted for by the model.
The percentage of units completing an entire

RTY 2 Rolled Throughput Yield sequence of process steps, all the way through,
with no rework or repair.
A two-dimensional chart plotting a variable over

run chart 5
time.
A two-dimensional plot for displaying bivariate

scatter diagram 5
data. It helps to see relationships.
A measure of process capability which is

correlated to the defect rate of a process.
Sigma capability values are tabularized or may
sigma capability 2, 5 be calculated using the formula in module 5.
Alternately for continuous data, sigma capability
is also used as a measure of potential for the
process (=3*Cp).
⎡ USL − y y − LSL ⎤ For continuous data, sigma level is the number

sigma level 2, 5 min ⎢ , ⎥ of standard deviations between the mean and
⎣ σ σ ⎦
the nearest specification limit.
Single Minute Exchange of A series of techniques used to facilitate the

SMED 10
Dies rapid changeover of any process or equipment.
Details how (and to what level) a variable

Standard Operating
SOP 2 should be held constant in a process. SOP's
Procedure
help drive out unnecessary variation.
A layout diagram of the work area that illustrates

the path followed by materials or parts through
Also known as a “physical the facility. It can be used to highlight excessive
Spaghetti diagram 2, 5
process flow” amounts of material transport or employee
movement while following the steps described in
a process flow chart.
A methodology for studying a process (using
control charts) to determine if the process is
SPC 11 Statistical Process Control stable and repeatable. Control charts help
identify special causes of variation, shifts and
trends.
Nonrandom variation that can be detected by
special or assignable
special cause 11 the use of control charts and good process
causes of variation
documentation.

Reference
(module)
A measure of variability or spread in a
standard deviation standard deviation
2, 5 distribution. We try to reduce variation using
(σ) (pronounced "sigma")
PF/CE/CNX/SOP, statistical analysis, etc.
The average time to produce a unit in order to
Takt time 7
meet the current rate of customer demand.
A comprehensive and coordinated maintenance
program designed to maximize equipment
effectiveness by minimizing downtime and
optimizing output in terms of speed and quality.
Total Productive
TPM This approach is founded on the ability of well
Maintenance
trained equipment operators to proactively
identify and correct small maintenance issues
before they become significant and lead to
breakdowns.
A sequence of processes or activities to include
all of the actions (both value added and non-
value stream 2, 5 value added) required to bring a product or
service from concept to raw materials all the
way to the customer.
Features, steps, etc. that the customer
considers desirable. Value is measured in
VA 2, 7 value-added
terms of what an external customer is willing to
pay for.
A dimensional measurement can be obtained,
variables data 2 measurement data and is only limited in value by the resolution of
the measurement system.
Any activity or process that does not add value
to the product or service from the perspective of
the customer. The seven “classic wastes”,
waste 2, 7 originally identified by Taiichi Ohno, include:
defects, over processing, overproduction,
unnecessary movement, unnecessary
transportation, inventory, and waiting.
Material which has been partially processed, but

WIP 2, 7 Work in Process has not yet reached the state of approved
finished product.

DPMO TO SIGMA CAPABILITY
CONVERSION TABLE
(Selected Values)
sigma sigma
dpmo dpmo
capability* capability*
500,000 1.5 6,210 4.0
460,172 1.6 4,661 4.1
420,740 1.7 3,467 4.2
382,089 1.8 2,555 4.3
344,578 1.9 1,866 4.4
308,538 2.0 1,350 4.5
274,253 2.1 968 4.6
241,964 2.2 687 4.7
211,855 2.3 483 4.8
184,060 2.4 337 4.9
158,655 2.5 233 5.0
135,666 2.6 159 5.1
115,070 2.7 108 5.2
96,801 2.8 72 5.3
80,757 2.9 48 5.4
66,807 3.0 32 5.5
54,799 3.1 21 5.6
44,565 3.2 13 5.7
35,930 3.3 9 5.8
28,716 3.4 5 5.9
22,750 3.5 3.4 6.0
17,864 3.6
13,903 3.7
10,724 3.8
8,198 3.9
* includes 1.5σ shift in average
Conversion Approximation: sigma capability = 0.8406+ 29.37- 2.221 • ln(dpmo)

COURSE EVALUATION SHEET
YOUR IDEAS AND COMMENTS WILL HELP US IMPROVE OUR WORKSHOPS. THANKS FOR YOUR TIME
AND EFFORT IN COMPLETING THIS COURSE EVALUATION.
WORKSHOP NAME: LEAN SIX SIGMA BLACK BELT (WEEK 2) DATE: _______________________
NAME: _________________________________ JOB TITLE: ___________________
1. WHAT DID YOU LIKE MOST ABOUT THIS COURSE?
2. WHAT DID YOU LIKE LEAST ABOUT THIS COURSE?
3. PLEASE RATE THE FOLLOWING AREAS:
POOR EXCELLENT
A. COURSE OBJECTIVES 1 2 3 4 5
B. INSTRUCTOR 1 2 3 4 5
C. INTERACTION AND DISCUSSION 1 2 3 4 5
D. TOPICS COVERED 1 2 3 4 5
E. PARTICIPANT'S GUIDE 1 2 3 4 5
F. GROUP EXERCISES 1 2 3 4 5
G. MEETING ROOM ACCOMMODATIONS 1 2 3 4 5
1 2 3 4 5
H. WORKSHOP DURATION
TOO LONG OK TOO SHORT
I. OVERALL WEEK RATING 1 2 3 4 5
4. DO YOU HAVE SUGGESTIONS THAT WOULD MAKE THE COURSE MORE EFFECTIVE?
5. TO WHOM WOULD YOU RECOMMEND THIS WORKSHOP? (PLEASE INCLUDE NAME, ADDRESS,
AND PHONE NUMBER. WE WILL NOTIFY THESE PEOPLE OF FUTURE OFFERINGS.)
6. ANY OTHER COMMENTS (use back if needed):

COURSE EVALUATION SHEET
YOUR IDEAS AND COMMENTS WILL HELP US IMPROVE OUR WORKSHOPS. THANKS FOR YOUR TIME
AND EFFORT IN COMPLETING THIS COURSE EVALUATION.
WORKSHOP NAME: LEAN SIX SIGMA BLACK BELT (WEEK 1) DATE: _______________________
NAME: _________________________________ JOB TITLE: ___________________
1. WHAT DID YOU LIKE MOST ABOUT THIS COURSE?
2. WHAT DID YOU LIKE LEAST ABOUT THIS COURSE?
3. PLEASE RATE THE FOLLOWING AREAS:
POOR EXCELLENT
A. COURSE OBJECTIVES 1 2 3 4 5
B. INSTRUCTOR 1 2 3 4 5
C. INTERACTION AND DISCUSSION 1 2 3 4 5
D. TOPICS COVERED 1 2 3 4 5
E. PARTICIPANT'S GUIDE 1 2 3 4 5
F. GROUP EXERCISES 1 2 3 4 5
G. MEETING ROOM ACCOMMODATIONS 1 2 3 4 5
1 2 3 4 5
H. WORKSHOP DURATION
TOO LONG OK TOO SHORT
I. OVERALL WEEK RATING 1 2 3 4 5
4. DO YOU HAVE SUGGESTIONS THAT WOULD MAKE THE COURSE MORE EFFECTIVE?
5. TO WHOM WOULD YOU RECOMMEND THIS WORKSHOP? (PLEASE INCLUDE NAME, ADDRESS,
AND PHONE NUMBER. WE WILL NOTIFY THESE PEOPLE OF FUTURE OFFERINGS.)
6. ANY OTHER COMMENTS (use back if needed):

END OF FIRST DAY CRITIQUE
1. What did you like best about the session today?
2. What did you like least about today’s session?
3. Is the pace of the course too fast, too slow, or about right?

06-Lssbb-Pg1a Vol 1 (Mod. 1-13) (Ingles)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

06-Lssbb-Pg1a Vol 1 (Mod. 1-13) (Ingles)

Uploaded by

Copyright:

Available Formats

LEAN SIX SIGMA

Texts: Knowledge Based Management

1650 Telstar Drive, Ste 110

All rights reserved. Do not reproduce.

I hereby acknowledge that this and all other Air Academy

1. Rate your level of Lean Six Sigma knowledge on a scale of

2. What does Lean Six Sigma mean to you? (25 words or

3. What is your commitment level on a scale of 0 to 10 to the

Copyright Air Academy Associates

Expected Yearly ROI from Project:

What are 3 important pieces of information you’d like to share about

Copyright Air Academy Associates

WEEK 1 Module 1 Lean Six Sigma...The Journey Begins

Module 2 Lean Six Sigma Fundamentals

Module 3 Defining the Project and Managing Change

Module 4 Understanding the Voice of the Customer

Module 5 Measure…Making Sense out of Data Using

Module 6 Measurement System Analysis

WEEK 2 Module 7 Analyzing the Causes of Poor Performance

Module 8 Techniques for Narrowing the Focus

Module 9 Drawing Conclusions from Sampled Data

Module 10 Improving the Process and Work Flow

Module 11 Controlling Process Performance / Realizing

Module 12 DMAIC Summary – Putting it All Together

Module 13 References, Glossary of Terms, Course

Copyright Air Academy Associates i

WEEK 4 Module 22 Quick Review of DMAIC and DOE

Copyright Air Academy Associates ii

- To understand the nature of a Lean Six Sigma culture and

Copyright Air Academy Associates iii

- Be able to conduct a Measurement System Analysis (MSA)

Copyright Air Academy Associates iv

- Gain confidence applying the Lean Six Sigma tools and

Copyright Air Academy Associates v

- Know how to recover from experiments that do not confirm

Copyright Air Academy Associates vi

1. Identify and Define opportunities within their business unit

2. Measure and Analyze a process using the Lean Six Sigma

3. Improve a process using Lean Six Sigma tools and techniques.

4. Control a process by using Lean Six Sigma tools and Realize

5. Repeatedly apply the Define – Measure – Analyze – Improve –

6. Communicate at all levels of an organization.

7. Change the culture of an organization.

Copyright Air Academy Associates vii

• Start and stop times

• You are responsible for learning

• Questions are good

• Importance of applying right away

• Organization / Your Role

• Background in Lean and/or Six Sigma

Copyright Air Academy Associates viii

LEAN SIX SIGMA...

• IN THIS MODULE, WE WILL DISCUSS:

– THE WHAT AND WHY OF LEAN SIX SIGMA

– KEY ELEMENTS AND INFRASTRUCTURE

– ROLES AND RESPONSIBILITIES

– LEAN SIX SIGMA PROJECT MASTER STRATEGY

– COURSE PHILOSOPHY AND INSTRUCTIONAL

Copyright Air Academy Associates 1-1

WORLD CLASS QUALITY

VARIATION and WASTE … the enemies of

Copyright Air Academy Associates 1-2

• WORLD CLASS CULTURE