LSSGB Course Handout

Certified Green Belt
Lean Six Sigma

Accredited by: Council of Six Sigma Certification, USA
Patronized by: Pivotal Resources Inc., USA
This programme is conducted by

00 Intro 2
Define Phase 11
01 Define phase 12
02 Intro to Minitab 27
Measure Phase 49
03 Process Mapping 50
04 Process Capability 63
Analyze Phase 74
05 C & E Tools 75
06 Graphical tools 83
07 Test of Hypothesis 90
08 Scatter, Correlation & Regression 103
Improve Phase 116
09 Design of Experiments 117
10 2K factorial DOE 130
Control Phase 146
11 Intro to Control phase & Control Plan 147
12 Control Charts 153
13 Project Report 170
Certified Green Belt
Dr. Sumit Shandilya

Executive Director & Chief Strategy Officer
INTRODUCTION
00 Intro 1-2 © July 23, Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved. 2
2
Six!Sigma!Practically- A!Management!Philosophy
Focus on your biggest/ chronic problems . . ?

Yes or No?
D
Assign it to most competent person. . . ?

M
Yes or No?
A
Support with all the tools/ resources. . . ?

Yes or No?
I
C
And guarantee uninterrupted time. . . ?

Yes or No?
12
© Quality Growth Services Pvt Ltd. Unpublished proprietary work available only under license. All rights reserved.
So!Leaders!Must!Understand!WHY!It!Works
qAlignment!with!the!strategic!priorities!of!the!
business,!generally!with!a!sharp!focus!on!financial!
results.
D
² Top!Executive,!Champions
M
qDedicated,!well!trained!problem!solvers:!
A
² Black/!Green!Belts,!Green!Belts
I
C
qwell!architected!support!system
² Six!Sigma!Infrastructure!
qThe!powerful!Tool!Box
²RDMAICR!process!of!variable!reduction.
00 Intro 13 3
Six!Sigma!Statistically?
Accuracy = Centering Precision = Spread
D
M
A
I
C
Spread Spread
Customer target
Customer target
Reduce Variation Focus on Target

Off target and variation = process variability = DEFECTS
14
Six Sigma Philosophy
00 Intro 1 - 15 © Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 15 4
Six Sigma Focus
REDUCE VARIATION
1 - 16 © Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 16
Six Sigma Level Performance

When!variation!is!reduced!to!the!extent
that!it!becomes!half!of!specification!tolerance,!
It!is!called!a!six!sigma!(6s)!level!performance
m
LSL USL
ss sss s
Process!Variation=!0.50!x!SpecificationTolerance
00 Intro 1 - 17 © Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved.17 17 5
Sigma Level- Statistically : A performance Measure
Sigma Level Defect % Yield

(Short Term) Opportunity (Long Term)
(Per Million)
(Long Term)
2 308,537 69.15%
3 66,807 93.32%
4 6,210 99.38%
5 233 99.98%
6 3.4 99.99966%
1 - 18 © Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved.18 18
Six!Sigma!for!Black/!Green!Belt
Expectation
D
M
A
Gap? Six Sigma Project

I
C
Performance
00 Intro 19 6
Products
Customer
CTQ – Critical to Quality
CTD – Critical to Delivery Expectation
CTP – Critical to Price
D
M
Price/
Quality Delivery
Value
A
Cycle
Defects Cost
I
Time
C
Processes
CTP1 – Critical to Process 1
Organization
CTP2 – Critical to Process 2 Performance
CTP3 – Critical to Process 3
20
q
Practical Problem Real Problem

D
M
(Define) (Measure-Analyze)
A
I
C
Practical Solution Real Solution

(Control) (Improve)
00 Intro 21 7
DMAIC- Funnel!:!Project!Methodology
q
Practical Problem Project Charter, Tree Diagram,
Define Pareto,Base Line,
Entitlement, Macro-map
Project Goal
D
PFD, VSM, MSA

All Xs CTC-CTP
Measure
M
Capability
A
C-E Matrix, Box plot, Main effect plot,Interval

I
Plot Interaction Plot, Multivari Analysis,

Analyse Suspect Xs Scatter plot, Hypothesis Test
C
DOE, Piloting , Validation

Improve Key Xs
Control Control Xs PFMEA, Control Plan, Error proofing, SPC
23
Process!Performance:!Six!Sigma!Lingo
Best performing similar process within or

outside the organization:
D
Benchmark
M
A
I
Historical best performance of the process:

C
Entitlement
Current Performance Level of Process:

Base Line
00 Intro 24 8
Six!Sigma!Ultimately
$
D
(ROI)
M
Customer
Satisfaction
A
I
Yield Improvement
C
Defect Reduction
25
Six!Sigma!:!From!where!it!came?
Period/ Year Event
1983 Started by Motorola Bob Galvin, CEO

D
1988 Formation of Six sigma consortium

M
1989-90 Tried by IBM/ DEC Initiative failed

A
I
1993 Structured by Allied signal- Black/ Larry Bossidy

Green belts
C
1995 Adopted & Structured by GE Jack Welch- The

real promoter
1997 onwards Rapid growth
1999 ASQ started training
00 Intro 26 9
Key!roles!and!responsibilities
LEVEL ROLES TRAINING
Executive §Determine the scope ,identify goals 1 day
§Identify priorities, drive the use of 6s
D
Champion §Maintain strategy and projects link 1-3 days

M
§Recognize and define projects

§Select Black/ Green Belts
A
§Removes roadblocks ,Review project activity

§Report project activity to executive
I
Master BB §Instructor of BB/GB, Training material developer BB + Addl trg

C
§Support Exec./Champions
§Leader of Large scoped projects
Black Belt §Lead Project team/,Practitioner of DMAIC 4-5 weeks
Green Belt §Lead/ support Project team/Practitioner of DMAIC 1 week
Process Owner §Owner of solution/ Pivotal team member 1 day
27
END!OF!MODULE
31
00 Intro 10
Certified Lean Six Sigma Green Belt
DEFINE PHASE
Define Phase 11
Lean!Six!Sigma!– GB
Define!Phase
Define
Objectives 1. Define
• Ensure the right priorities
• Clarify the issue/opportunity
• Define the process and customer(s)
• Set up the project for success
01 Define phase 12
Define Road Map
Issues, Opportunities,
Strategic Goals
Initial Charter:
Problem/Opportunity,
Business Case, Objective
` Initiate Change
1. Define Management Efforts
S I P O
C
Document Targeted
Process
Clarify Customer Ensure Alignment with
& Requirements other Projects
Selecting & Prioritizing Efforts
01 Define phase 13
Identifying Potential Improvement Needs
Sources of Project Ideas & Examples
• Strategic plan & goals

- Free Cash Flow
• Measures and assessments
- Performance against KPIs
- Visual Management
• Complaints, requests, inefficiencies, successes
- Common/recurring SOED issues
- Learnings from other projects
• Innovative ideas
Project Filtering
Example
Low
A. Poor space utilization in buildings
B. Purchasing system errors
C. Accounting errors
D. Inventory systems do not
Level of Effort
match physical counts

E. Increasing damages and line
faults
F. Obsolete materials being used
G. Website is not user friendly
H. Employee turnover is too high
I. Increasing customer complaints
J. Office supplies costs are
increasing
High
Low High
Impact of Benefit
01 Define phase 14
Project Selection Criteria
1. Define
Meaningful
• Impact on key customers
• Impact on strategic plan
• Financial benefit — costs AND revenue
- Immediate
- Long-term
• Urgency or sequence
Manageable
• Resources needed
• Expertise available
• Organization “buy-in”
• Likelihood of success
• Learning experience opportunity
• Project duration
Key Handoff
From Leaders to Team 1. Define
What leaders do (or should):

• Create broad description of the issue (concern, challenge,
opportunity, problem)
• Ensure it’s a priority & aligns with other initiatives
• Provide support—commitment, guidance, resources
What Green & Black Belts & teams do (or should):

• Review & refine issue, define specific objectives
• Ensure the effort is really warranted (object if needed)
• Identify overlaps and/or synergies with other efforts
• Follow “DMAIC” to get results as quickly and effectively as possible
(“shortcuts” are okay)
01 Define phase 15
Solution “Parking Lot”
• Capture ideas at the

beginning and throughout
D, M & A
• Keeps team from getting Initial Solution Ideas (Parking Lot)

sidetracked
• Helps ensure good ideas • Eliminate non-essential data from

don’t get lost requisition form
• Set-up customers with a pre-paid

• Acknowledges early “purchase card” for rush orders
“creative thinking”
• Automate process end-to-end
• Train all staff in current

procedure
The Project Charter
01 Define phase 16
Getting Started
1. Define
Project Charter
1. Business Case
2. Problem/Opportunity
3. Scope
4. Metrics
5. SMART Objective
6. Estimated Benefits
7. Constraints & Risks
8. Team Members & Time
9. Milestones
Project Charter
Project Title: <<short & clear name for the project, no technical terms>>
Location of Process: <<region, site, department, etc.>>
Business Case Estimated Benefits

<<What is the business rationale for this project? What will be <<Hard and/or soft>>
the impact on the company and/or customers – in what area, if e.g. reduced transportation costs (x) or increased customer
it is successful? What are consequences of not doing it?>> satisfaction (x)
e.g. more competitive; enhance capabilities; increase reliability
Constraints & Risks
Problem/Opportunity Statement <<Anything that may impact the project scope, timing, results>>
<<What is the specific issue, pain or gap to address? What is the e.g. resource availability; technical challenges; overlap with other
observable evidence? How is the problem impacting the business projects; specific expertise required
currently or what gaps/opportunities exist today to address?>>
Team Members and Time Allocation per Week
Project Scope
Name Time Allocated
IN: <<What will be addressed in the project – region, customer
group, product group…>> Project sponsor
OUT: <<What will not be addressed in the project?>>
Project leader
Process owner
Metrics
Controller
Unit Baseline Target
Other team members
e.g. Availability days 90% 95%
Milestones
e.g. complaints # x x
Kick-off: << DD/MM/YY>>
Define: << DD/MM/YY>>
SMART Objective Measure: << DD/MM/YY>>
<<The objective should be Specific-Measureable-Attainable- Analyze: << DD/MM/YY>>
Relevant-Time related. Describe a goal to attain not the solution
Improve: << DD/MM/YY>>
to develop. >>
e.g. Reduce complaints by x% till end of 2015 Control: << DD/MM/YY>>
Targeted end date: << DD/MM/YY>>
Rev: N°___________
Define
01 Define phase Sponsor, Process Owner and Leader Sign-off: ____________ _______ Date :____________ 17
Business Case
Rationale and context for this project/problem
• Concise description of project’s like or

relevance to the business road map,
goals, site/functional targets, etc.
• Impact (benefit/relevance) of potential improvement
• Impact (cost or risk) of not improving
• Should be verified/accepted by leadership

• May be applied to multiple
projects
ü E.g. Short-term “Rapid
Improvement” and longer-term
“Redesign”
Problem/Opportunity Statement
Essential Description of “What’s Wrong”
Should Include: Should Avoid:

• Concise description of pain, gap, • Assumptions
challenge stated as facts
• Observable evidence • Root causes
• Impact (potential or actual) • Blame
Examples …
Good: Not Good:

Custom-made orders for Fashion- We need to relocate production to
Athletic footwear are not meeting be nearer the major Fashion-
commitment dates. This is creating Athletic customers.
significant dissatisfaction in a new and
critical market and jeopardizes short- Sales people are not doing an
term sales and longer-term market effective job managing
leadership expectations.
01 Define phase 18
Problem/Opportunity Statement Elements
• (What ) (Where) (When) (How Much) (How do you know)

• What is the problem (Pain)?
• Where is this problem being faced?
• Since when this problem is?
• How much is the problem in quantitative term?
• How do you know that this is the problem?
• Has the performance degraded from the past?
• Is performance not improving?
• Is some target not being met ?
• Is someone else doing better?
Problem/Opportunity Statement Elements

The thing The suspected pain or gap
Sales leads are not being handled promptly.
Evidence
Initial/ Prospective customers have called to complain
Anecdotal that no one has contacted them.
Adding ___% of leads in past two months were not

Data followed up for two weeks or more.
Impact
Initial/
Anecdotal We risk losing business and hurting profits.
Adding Slower follow-up appears to coincide with a recent

Data 12% decline in new customers
01 Define phase 19
APPLICATION: Problem/Opportunity Statement
In your project teams
1. Draft (or review and refine) the

Problem/Opportunity Statement for your project*
2. Check to make sure it’s well written and focused. For
example:
• Are the core elements there: Description, Evidence, Impact?
• Is the description objective, observable, verifiable?
• Have you avoided including a solution or blame?
3. Note any key questions about your project direction that

need to be addressed
20
*This is still early and the Problem Statement will likely
evolve over the course of the project.
Scope: Why is it Important?
Leaders & sponsors think:

“They can do this.”
Team considers and thinks:

“We can do this.”
Reality and time reveal:

“You can do this.”
01 Define phase 20
Scope
What’s in? / What’s Out?
1. State what’s to be accomplished,

analyzed, discussed—or what’s in scope
2. State what’s NOT to be accomplished, analyzed,

discussed—or what’s out of scope
Remember & balance the “2 Ms”:

Meaningful & Manageable
Scope Tool
In/Out of Frame
Example: The Mercy Hospital Patient Records Team
In-Scope Major IT
Reasons System
for Changes
Admits Format Cost/Effici
of Records ency of
Design of Record Design of Treatment
Forms Handling On-line Codes
Primary Tools
Care
Processes
Who
Needs
Ongoing Who
Records
Process Receives
Measures Records
Clinical
Practices
Out of Scope
01 Define phase 21
Metrics
How will we observe/quantify improvement?
• Linked directly to the pain and results

(Problem and Objective)
• Used to validate and baseline the Problem,
and assess results
- More on this in Measure section: “Project Y”
What would be likely Metrics for this situation?

Problem: Custom-made orders for Fashion-Athletic footwear are not
meeting commitment dates. This is creating significant dissatisfaction
in a new and critical market and jeopardizes short-term sales and
longer-term market leadership.
SMART Objective: Reduce incidence of late (after commit date)
custom footwear deliveries to <__ %. Implement improvement
by end Q3. Confirm project results end Q4.
SMART Objective (What do we want to achieve?)

Specific, Measurable, Attainable, Relevant, Time-bound
Should Include: Should Avoid:

• Action verb: Reduce, Increase • Solution or
action to achieve
• What’s to be changed; description
results
of intended results
• Vague outcome
• Target benefit
• Open-ended timeframe
• Date for implementation
and results • Project plan or action steps
Example …
Good: Not Good:
Reduce late (after commit date) Orders have to be submitted
custom footwear deliveries to <__ both electronically and via fax.
%. Implement improvement by Eliminate causes of late orders
end Q3. Confirm project results as soon as we can.
end Q4.
01 Define phase 22
Estimated Benefits: Hard or Soft
Initial Projections based on Problem & Objective
How will project results convert into financial impact?
“Hard” Financial Benefits: “Softer” Benefits:

Revenue / volume / price increase Avoidance of lost revenue
Improved yield / reduced waste Avoidance of future cost
Materials / supplies usage Increased speed to market
Overtime labor reduction Improved customer satisfaction
Staffing reduction (headcount) Increased capacity or efficiency
Inventory reduction Increased employee satisfaction
Soft impacts generally drive hard ones, but not in a proven, 1:1 ratio...
Constraints & Risks
• Constraints – Limitations on
resources and/or capabilities that
may influence the Design project and/or solutions
E.g. Space, money, people, talent, time ...
• Risks – Potential problems or challenges that may

arise during the course of the project
E.g. Anticipated team members are pulled out in favor of
another project. Technical problems with equipment delays
testing of new solutions. Etc. ...
01 Define phase 23
Team Members and Time
Who are Key Contributors?
Core roles Secondary roles

• Project Sponsor • Subject Matter Experts
• Project/Team leader - Info Technology
- HR
• Process owner
• External partners
• Controller
• Vendors/suppliers
• Team members
• Regulators
• Customer “advocate”
The Improvement Team

Important Considerations... 1. Define
How large a group? The right stuff (a balance)

- Three to six: the “Core Team” - Willing to challenge status quo
- Appropriate representation - Confronts & manages
(across, up, down) obstacles, staying positive
- Comfortable
seeking/evaluating facts
Practical factors - Can communicate the Vision
and value of effort
- Availability, Location,
Commitment, Knowledge
Name Time Allocated
- Include & communicate
time commitment up front!! Project sponsor Melanie Q 5 hrs/month
Project leader Esteban P 20hrs/month
Process owner H Ying Hui 5 hrs/month
Controller Graeme L 8 hrs/month
Other team members TBD 8-12 hrs/month
01 Define phase 24
Milestones & Preliminary Plan
Some Key Questions:
Milestones
Target date for each
DMAIC phase Kick-off: 15 July 2016
Define: 2 Aug 2016
Measure: 18 Aug 2016
Analyze: 3 Sep 2016
Improve: 15 Oct 2016
Control: 15 Dec 2016
Targeted end date: 31 Dec 2016
Preliminary Plan
• Key tasks/actions, with assigned
responsibilities and dates
Example: Preliminary Plan

WHO’S TARGET ACTUAL
RESPONSIBLE COMPLETION COMPLETION
DATE DATE
ACTION/MILESTONE
DEFINE Kumi T. March 1 March 2

• Complete draft of Charter
José R. March 5 March 5
• Complete initial customer & process identification
MEASURE
Chris B. March 9 March 12
• Prepare data collection plan: Output & Process
measures Mary W. & Pat H. April 1 April 2
• Complete initial data collection
Kumi T. April 9 April13
• Present refined Charter to SLT
ANALYZE
Chris B. April 12
• Confirm Process Improvement strategy (vs. April 13
Redesign)
Mary W. April 20
April 21
• Develop initial cause hypotheses
• Complete cause verification José R. May 3
IMPROVE Mary W. May 5

• Develop draft solutions list
José R. May 9
• Select solution for piloting
José R. May 20
• Pilot implementation complete
Chris B. & Pat B. June 10
• Full implementation complete
CONTROL
Pat H. June 24
• Measure & report results to SLT
01 Define phase 25
Customer
Project Charter Focus
Project Title: Perfume Trial Oil Cycle Time Reduction CI Culture
Location of Process: Oil Development & Storage

Operational
Estimated Benefits Excellence
Business Case
Speed and efficiency of development is key to our ability to • TBC -10 Minutes average daily for approximately 40-60
respond to new business opportunities and stay ahead of employees (approx. 1900 man hr per annum)
competitors in new fragrance options. The ability to track, locate • Increased focus on creative activities resulting in higher win
and use oils is key to this efficiency and best use of perfuming rates
talent. Team Members and Time Allocation per Week
Problem/Opportunity Statement Name Time
Allocated
Movement and storage or trial perfume oils is inconsistent and
hard to trace main warehouse for last 3 years. Average tracing Project sponsor Martin Van Wyck 1 Hr
time is 45 mts. This creates extra work for perfumers seeking
trial oils and delays completion of development process. Project leader Ely Costa 6
Project Scope Process owner Ely Costa – Vicky Hui 6-2 hrs
IN: Primary: New creations from Perfumery team and
Controller Morgan Fairchild 2 Hr
Secondary (to be considered): library oil replenishment
OUT:registered oil recompounding, oils from another site Other team Maurice Cayer, Janet 1 Hr
OUT: <<What will not be addressed in the project?>> members Estabridis, Gustav Schnell
Metrics
Constraints & Risks
Unit Baseline Target Project running during summer months with heavy vacation
conflicts. Funds to purchase central storage equipment. System
Perfumer time Minutes Avg 15 Less than
changes may take longer than expected.
looking for oils Minutes 5 minutes
Milestones
daily a daily a
(involving 40-60 Kick-off: 07 /28/XX
person person
employees daily)
Define: 08/04/XX
SMART Objective Measure: 08/18/XX
Analyze: 09/01/XX
Reduce number and duration of oil searches and improve
traceability and storage of trail oils from…………to ……………. Improve: 09/15/XX
Implement solutions by end of 15 Sept 20XX and validate results Control: 09/28/XX
by 31 Oct 2015. Targeted end date: 09/30/15
Rev: N°___________
Define Sponsor, Process Owner and Leader Sign-off: ____________ _______ Date :____________
End of Module
Contact
01 Define phase Confidential business and proprietary information of Givaudan, may not be copied or distributed to anyone without the express written permission of Givaudan
26
Minitab Introduction
and Basic Statistics
Data Entry and Data Organization
QGS-Business Excellence Division
8 July 2023 Edition-1 1
Learning Objective
qMinitab Introduction, data entry, data
Management
qPerformance Analysis Tools
02 Intro to Minitab
27
MINITAB
INTRODUCTION,
DATA ENTRY &
DATA MANAGEMENT
surendra@qgspl.com
Module Objectives
Throughout this presentation, we will review:

m The structure of MINITAB, including main windows, menu
structure, tool bars and shortcut keyboard commands
m Types of data, data entry and data organization techniques
m Correct data structure for analysis in MINITAB
m How to create and interpret basic graphs in MINITAB
m Moving output from MINITAB into other programs
02 Intro to Minitab
© Quality Growth Services Pvt Ltd. surendra@qgspl.com
Unpublished proprietary work available only under license. All rights reserved.
28
Minitab Windows: Types of windows
MINITAB has four types of

windows. They are:
l Session window
l Data window
l Project Manager window
l Graph window
MINITAB’s Default
windows
l Session Window (top)
l Data Window (bottom)
Minimized window
l Project Manager window
surendra@qgspl.com
Minitab Windows: Toolbars
q MINITAB Standard Toolbar:

m Provides shortcuts to performing frequently used commands
m Buttons of the Standard Toolbar change as the active window is changed to reflect the
unique capabilities of each window
m Use your mouse to determine the use of each button, by placing the pointer on the
button
m A brief description of functionality for each button will appear.
q Project Manager Toolbar:

m Provides shortcuts to the various folders in the Project Manager
m Project Manager Toolbar buttons do not change as MINITAB’s active window changes
02 Intro to Minitab
29
Data Windows Details
Column Name Row

The column name row is located
just above row 1 of the worksheet.
Column names may be up to 31
characters and may contain spaces.
Data Entry Arrow

There is a data entry arrow above row 1, which dictates the direction the cursor will
move after the Enter key is pressed.
Data Formats
When you enter data into MINTAB, the program will read it as one of three formats:
hNumeric
hText (alphanumeric)
hDate/Time
surendra@qgspl.com
Minitab Projects and Worksheets

} Differences between projects and worksheets
qProjects can contain some or all of the following elements:
} Data windows (1 or more)
} Graphs
} Session window
} Project Manager
qProjects also retain:

} The last settings used in a dialog box
} The position of all elements in the project
02 Intro to Minitab
30
Minitab Projects and Worksheets
} Differences between projects and worksheets
qProjects:
} Can be opened using the toolbar
} Cannot be opened into other projects
} Cannot be opened using Open Worksheet
q Worksheet files (.MTW) contain:
²Data window (1)
q Worksheets cannot:
²Retain the last settings used in the dialog box
²Retain any windows except the data window
²Be opened using the toolbar
surendra@qgspl.com
Using the Standard Toolbar
Edit Last
Command
Click this button
to return to the
last dialog box in Project Manager Brings the
which you made Toolbar, opens Brings the Project
any changes. the Project Session Brings the Manager to
(Ctrl+E) Manager Window window to current the front,
and the identified the front, and worksheet to and makes
folder makes it the the front, and it the active
active makes it the window.
window. active window. (Ctrl+I)
(Ctrl+M). (Ctrl+D)
02 Intro to Minitab
31
Minitab Windows
Maneuvering Through the Windows

q Currently, four different window types are open in MINITAB.
They are:
m Graph
m Session
m Data
Close all Graphs
m Project Manager
Session window
Project Manager
Data window
surendra@qgspl.com
Data Entry
Common Method
Not preferred by Minitab
02 Intro to Minitab
32
Data Entry
Minitab way: Single column
How to do?
surendra@qgspl.com
Data Menu
Menu that Reorganizes or Redefines Data Structure
q Data Menu:
m Data ØStack ØStack Blocks of Columns
m Data ØCode ØNumeric to Text
m Data ØChange Data Type ØText to Date/Time
m Data ØStack ØStack Rows
m Data ØExtract from Date/Time ØTo Text
02 Intro to Minitab
33
Reorganizing Data
Tables vs. Variable Columns

Preferred Data Structures:
m Data set contains response for each location and each month,
and is a continuous variable
m Both X’s (location and month) in this case are discrete
variables
surendra@qgspl.com
Stack Blocks of Columns
Reorganizing Data from Excel Tables into a Column Format

q Stacking Data:
q Use either Stack Blocks of Columns
or Stack Rows to rearrange data
²Rearrange data shown on the
previous page into a single column
format
²Create two X (input) columns of data
and one Y (response) column of data
02 Intro to Minitab
34
Stack Blocks of Columns
1. Data ØStack ØBlocks of Columns.

2. Complete the dialog
box as shown:
3. Click
surendra@qgspl.com
Change Data Type

Changing from Text to Date/Time
1. Choose Data ØChange Data Type ØText to
Date/Time.
2. Fill in the dialog box as shown below:
3. Click
02 Intro to Minitab
35
Change Data Type
Minitab Output
surendra@qgspl.com
MODULE TOPICS
§ Descriptive Statistics
§ Store Descriptive Statistics
§ Graphical Summary
§ Bar Chart
§ Pie Chart
§ Time Series Plot
02 Intro to Minitab
36
Bar Chart
1. Choose Stat ØGraphsØBar Chart

2. In the drop down list, select A
Function of Variables
3. Slect One Y, Simple
4. Click OK
surendra@qgspl.com
Bar Chart
1. In the new dialogue box, select the

followings:
§ Function- Sum
§ Graph Variables- Sales
§ Categorical Variable- Month
2. Click OK
02 Intro to Minitab
37
Pie Chart
1. Choose Stat ØGraphs Ø Pie Chart

2.Select the following:
§ Radio button-Chart values from a
table
§ Categorical variable- Region
§ Summary variable- Sales
3. Click OK
surendra@qgspl.com
Time Series Plot
1. Choose Stat ØGraphsØTime Series

Plot
2.Select the following:
§ Simple and press OK
§ In the series panel (new dialogue
box), select Sales
3. Click OK
02 Intro to Minitab
38
Understanding Statistics
What is statistics?
26
What is statistics?
q Six Sigma Belt- Common Tasks
mTo predict behaviour of the process
mTo compare the behaviour with the target
mTo identify the causes and link with the process
behaviour etc…
A lot of information are needed

about the process
02 Intro to Minitab 27 39
What is statistics?
The set of information
derived from the Sample data
to estimate
the process (Population)
are called Statistics

28
Population vs Sample
q Population :
Set of all possible outcomes of the process
q Samples:
Subsets of populations
q Data is obtained using samples because
m we seldom know the entire population
m Population statistics are desired, but often not
available
m Samples approximate the population
q Descriptive statistics apply to samples
Statistics- Practically
30
Statistics- Some examples
m Minimum
m Maximum
m Count
m Range
m Mean
m Median
m Sigma etc…
MEAN: Centre of Gravity
The Mean of ‘n’ values is the total of the
values divided by ’n’
x1 + x2 + x3.....xn
x=
n
In Standard Mathematical Notation it is
i -n i -N
xi xi
x= n estimates
N m=
i -1 i-1
The symbol S” means ‘sum of’
Median
• The median is the middle number if the data

set has an odd number of observations
• Otherwise it is the average of the two
middle numbers
Examples
Odd data set: 1, 2, 3, 4, 5, 6, 7 The median is “4”
Even data set: 1, 2, 3, 4, 5, 6, 7, 8 The median is “4.5”
Organize the data from low values to high when determining the median
Median
• The 1/4th data point is called
‘1st Quartile’ (Q1) or ‘25th Percentile’
• The 3/4th data point is called
‘3rd Quartile’ (Q3) or ’75th Percentile’
• Difference of Q3 and Q1 is called ‘Inter Quartile Range’
Examples
Odd data set: 1, 2, 3, 4, 5, 6, 7
The median is “4”
Q1 is 2 and Q3 is 6
Median is also known as 2nd Quartile or 50th Percentile
34
Mode
The most frequently occurring value in a data set
Examples
Data Set: 45 47 49 51
46 47 49 52
47 47 50 53
47 48 51 54
Mode is: 47
Standard Deviation
The average distance between the individual

numbers and the mean.It is designated by “s”
2 2 2
Ù ( x1 - x ) + ( x 2 - x ) ....( x N - x )
s =
N
s =
å i
( x - x )2
estimates s = å i
( x - x ) 2
n -1 N
Variance
s 2
=
å (x - x) i
2
n -1
•Square of Standard deviation is called VARIANCE
•Numerator is called ‘ Sum of Squares’
•Denominator is called ‘Degrees of Freedom’
02 Intro to Minitab
44
Descriptive Statistics
Opening a worksheet file

Step 1: Choose File Ø Open Worksheet.
Step 2: Open the folder containing the reference files.
Step 3: Select Sales1.mtw.
surendra@qgspl.com
Display descriptive statistics dialog box

Step 1: Stat Ø Basic Statistics Ø Display Descriptive Statistics.
Step 2: In the List box,
double-click Sales.
Step 3: Click Statistics. Variables Box
Step 4: Select desired statistics
List
Box
02 Intro to Minitab
45
Step 5: Click OK
Session Window Output
surendra@qgspl.com
Store Descriptive Statistics
1. Choose Stat ØBasic Statistics ØStore

Descriptive Statistics.
2. In Variables: enter Sales.
3. In By Variables: enter Location
Quarter.
4. Click Statistics and check Sum.
5. Uncheck the defaults Mean; N non
missing.
6. Click OK in each dialog box.
02 Intro to Minitab
46
Graphical Summary
1. Choose Stat
ØBasic Statistics
ØGraphical
Summary
2. In Variables:
enter Sales.
3. . Click OK
surendra@qgspl.com
HISTOGRAM
Histograms give a graphical view of the
distribution of the values
D
It reveals the amount of variation that any

M
process has within it.

A
5
I
C
FREQUENCY 3
64 65 66 67 68 69 70 71 72
02 Intro to Minitab
47
Important Population Attribute
For a normally distributed process:
68.26%
D
M
95.44%
A
I
C
99.73%
-4 -3 -2 -1 0 1 2 3 4
Verify the above characteristics with Hamburger Data
44
Summary
q Minitab Windows
Five windows which allow you to:
m Store data (Data window)
m Generate analyses (Session window)
m Generate graphs (Graph window)
m Summarize the Data window (Columns folder)
m Store commands used to generate output (History folder)
q Data Entry
m Column names can be up to 31 characters
m Can be typed, imported, copied/pasted or brought in from a
stored Minitab file
m Preview and Options are available to customize data
imported from non-Minitab files
MEASURE PHASE
Measure Phase 49
Lean!Six!Sigma!– GB
Process!Mapping
1-1 1
Process Analysis
Process!step!analysis
Process!output!analysis!(CTC)
Process!input!analysis!(CTP)
– Process!Flow!diagram!+!CTC-CTP!map!(I-O!sheet)
– Value!stream!map!(Lead!time)
03 Process1 -Mapping
3 3 50
Process!Mapping!–
CTC-CTP!Analysis
Pivotal!Resources,!Inc
Module Objectives
Show!where!process!mapping!fits!into!the!strategy!for!
success
Show!the!connection!between!process!mapping!and!C&E!
matrices,!FMEAs,!and!control!plans
Define:
– Actual!process!vs.!Documented!process
– Value-added!!vs.!Non-value-added!(VA!vs.!NVA)
– Process!inputs!and!outputs!(X’s!&!Y’s)
Demonstrate!
– Macro-map
– Process!flow!diagram!(PFD)
– CTC-CTP!Detailed!map
Work!through!a!process!mapping!exercise
5 5 51
Types of Process Maps
SIPOC/Macro-map.
– A!macro-map!is!a!high!level!process!flow!chart.
– It!usually!shows!several!steps!on!both!sides!of!the!project!scope.
Process!flow!diagram!(PFD).
– A!process!flow!diagram!is!a!detailed!map!of!every!step!in!the!
process,!including!the!hidden!factory.
– The!process!flow!diagram!uses!standard!symbols!for!readability!
and!a!common!nomenclature.
Input/Output Worksheet.
– A!detailed!table!that!contains!all!of!the!process!details:!set!
points,!input!and!output!specifications,!etc.
– The!worksheet!also!contains!estimates!of!dpu,!RTY,!COPQ,!cycle!
time,!etc.
1-7 7
SIPOC Diagram ….Shared, High Level View of the Process
Supplier Input Process Output Customer
S I ® P O ® C
X
X “Requirements”
(Key Process
X Indicators)
X
www.qgspl.com Unpublished proprietary work

03 Process Mapping Confidential and proprietary business information of Givaudanwww.iipeindia.com available only under license.
All rights reserved.
52
S I P O C

www.iipeindia.com available only under license.
Why Use A SIPOC Diagram?

“Bird’s-eye view” of a process 1. Define
• Clearly defines boundaries/scope
• Identifies major outcomes & activities
• Avoids over-detail, confusion
• Helps plan & target measurement
• Focuses cross-functional alignment & analysis

03 Process Mapping www.iipeindia.com available only under license.
53
6σ
Macro-map SIPOC
Suppliers Inputs Process Outputs Customers Requirements
•CNC lines •Cut glass •Tempered Glass •Final customer •Size as per final
•VQ lines •Consumables gauge
•Glass supplier •BCP ink • Appearance as per
-AIS Roorkee limit sample
D
-AMG • Curvature & Cross
-AIS Taloja Curvature as per
M A
•Ink supplier standard
-Ferro
-Johnson
•Print screen
maker
I
C
Start
End
Cut Glass Logo Glass Glass Heating Quenching Unloading
Inspection Packing
Receiving Marking Loading & Bending
Inference:- There are 7 steps on I-01 Process

www.qgspl.com Unpublished proprietary work Page No.: 15
© Quality Growth Services Pvt Ltd. Unpublished proprietary work available only under license. All rights reserved. All rights reserved.
Bearing Assembly Example
• Background:
• An engineer for the Inspection Yields
Bearing Assembly
process in a Bearing 100
Manufacturing plant 98
noticed that the process 96
yield has gone down.
Yield
94
• A Six Sigma expert was 92
assigned to identify the 90
causes of these 88
differences and to 86
Week Week Week Week
reduce the defects.
1 2 3 4
Shift 1 Shift2

54
A Simple Macro-Map
Target Process
Receipt of Pick and Place Insert Inner Fix The Inner

Components Balls in Outer Shell Race Race
Fail
Readjust Fail
Inspect Inspect
Inner race and Balls
Pass
Pass
Ship

Detailed Process Mapping
A graphic representation of the flow of tasks

decisions, and pathways in a process:
• Standard map or “flowchart”
• Cross-functional or
“swim lane” map

55
Flow Chart : Which one is easy to understand
• In the start of the activity, first of all we clean the 1. Clean the outer shell using SOP no.
outer shell using SOP no. 954-21, then we check 954-21,
it using the same SOP. If it is found OK, then 2. Check for any contamination using
grease is injected inside the shell using SOP no. the SOP 954-21.
954-32 thereafter the ball is inserted inside the 3. If NG go to step 9
shell using SOP 954-32. It is again checked for
4. Inject grease inside the shell using
contamination if any using SOP no. 954-44 and
SOP no. 954-32
OK products are shipped.
5. Insert the ball inside the shell using
• In case during checking after cleaning the outer
SOP 954-32.
shell, there is any contamination, it is recleaned
using IPA and at the next step it is hold to 6. Check for contamination if any using
evaporate the IPA. In case failure happens during SOP no. 954-44.
checking after ball insertion, it is reworked and 7. Ship OK products
step no. 1 onwards are repeated. 8. Rework NG using step 1.
9. Clean using IPA and go to step 3
Start
SOP 954-21 SOP 954-44
SOP 954-32 SOP 954-32
D D
SOP 954-21 Check for
Inject Grease
Clean Outer Shell Inspect Balls Insert Balls Any Ship
in Shell
Contamination
SOP 954-21 Stop

Fail for contamination Fail
Remove Balls & 21
Clean again
Flow Chart
Flowchart: What is it?

§ A tool that graphically represents the steps of a process
Why use it?
§ To analyze the
§ number of steps
§ the time required for each step,
§ to detect bottlenecks,
§ unnecessary steps,
§ repetitions, and
§ other obstacles
When to use it?
§ Describe activities,
§ Identify/scope the problems,
§ identify the causes of problems
Unpublished proprietary work

22
03 Process Mapping available only under license.
56
Detailed Mapping Benefits ...
• Provides a “picture” of simple or
complex processes
- Helps focus on how work is really done (As-Is)
- Allows others to better understand your process
- Shows connections and dependencies: Supplier/Customer,
Upstream and Downstream
• Helps determine what and where to measure

• Essential to clarifying issues & finding root causes
o How work is done and why will explain the cause
• Key to planning for new or improved processes

available only under license.
Flow Chart
Different icons/symbols to indicate the different
types of actions in the process:
§Beginning/ End : §Activity Flow path :
§Activities : §Connector :
§Decision points : §Data step :
§Movement :
§Storage :
§Delay : Unpublished proprietary work

24
57
Flow Chart
How to create?
1. Observe for a few repetitions the process to be
charted.
2. Write down all the steps taken and decisions made in
the process.
3. Mark the path of the Flowchart from the beginning to
the end
4. Return to the beginning of the path and repeat Step
3 for any paths that branch off from the main path
(at the decision points)
5. Record the last step at the bottom of the page, draw
an ellipse around it, and connect the primary path
and any branching paths to the last step
6. Review for accuracy
available only under license. 25
Chart the Documented Process
Start
Check for
Inject Grease
in Shell
Contamination
Stop
Remove Balls &
Clean again

26
58
Flow Chart
Following optional but important

information may also be indicated:
1. Data generation steps
2. value added/ non value added activities
3. Cycle time
4. Existence/ non-existence of SOPs

Chart the Documented Process
Start
SOP 954-21 SOP 954-32 SOP 954-32 SOP 954-44
D D
SOP 954-21 Check for
Inject Grease
in Shell
Contamination
SOP 954-21 Stop

Remove Balls &
Clean again

28
59
Identify the Non-value-added Steps
• Label each process step as value-added

(VA) or non-value-added (NVA).
vA value added process changes the product in
a way that is meaningful to the customer.
Check for
Inject Grease
in Shell
Contamination

Remove Balls &
Clean again

Identify the Non documented Steps
Clean Balls Allow Solvent

w/ IPA to evaporate
Fail, first time (14%)
Check for
Inject Grease
in Shell
Contamination

Remove Balls &
Clean again

30
60
“Swim Lane” Map/Flowchart - Example
Production [Compounding & Packing]
Receive PO <999kg
Planner PO Size
Send RM
FG ready in
from
Warehouse warehouse
pallets
Small Batch
Compounding
Compounder
Large Batch
Compounding
>1000kg
Small Batch
Packing
Packer
Large Batch
Packing

Construct the CTC-CTP Map

List!the!steps!from!the!PFD!w/!SOP’s!and!VA!or!NVA
VA or I/O C or
Process Step NVA SOP # Inputs & Outputs Type Specifications N Capability Equipment
Clean Outer Shell NVA 954-21
Inspect Balls NVA 954-21
Clean Balls with IPA NVA 954-21
Insert Grease in Shell VA 954-32
Insert Balls VA 954-32
Check for Contam. NVA 954-44
Ship NVA
Remove Ball & Clean NVA

Evaporate Solvent NVA
32 32 61
Construct the Detailed Map
Identify!the!X’s!and!Y’s,!specs.,!C!or!N,!capability!and!equipment.
VA or I/O C or
Process Step NVA SOP # Inputs & Outputs Type Specifications N Capability Equipment
Clean Outer Shell NVA 954-21 Cleaned Outer shell Output No visible contam. IPY=92% Ultra-Wash
Solvent type Input IPA C
Solvent Contam. Conc. Input < .05% N
Soak Time Input 300 s C
Bath Temp. Input 110 F C
Scrub Brush Type Input Accord # 45 C
Scrub Brush Age Input < 5 weeks N
Insert Balls VA 954-32 Quantity of Balls Output 32+ 1 IPY=94% Smith/Olsen

Level of Balls Output Below Shell Height IPY=96%
Cleaned Outer shell Input No visible contam. C
Viscosity of Grease Input 2+0.2 cms (Penit). C
Room Humidity Input 40 ± 2 % N
1 - 33 33
Glossary
Process!map: Control!variable:
– A!graphical!representation!of!an!actual – A!variable!or!an!input!which!can!be!readily!
process. changed!to!measure!the!effect!on!an!output!
(Y),!knob!settings,!materials,!etc.
Input/Output!Worksheet
– A!table!with!information!for!process!inputs! Noise!variable:
and!outputs!that!is!beneficial!to!improving! – A!variable!or!input!that!is!difficult!to!control,!
the!process. ambient!temperature!or!humidity,!machine!
differences,!customer!state!differences,!etc.
Value-added!step:
– An!operation!that!transforms!the!product!in! Documented!process:
a!way!that!is!meaningful!to!the!customer. – The!process!as!defined!by!the!controlling!
documents!(process!specifications,!work!
Customer: instructions,!training!manuals,!etc.).
– Defined!by!the!scope!of!the!project.!!The!
customer!may!be!the!end!user,!another! Actual!process!and!the!hidden!
internal!organization!or!a!later!process!step. “factory”:
Output!(Y): – The!process!as!it!is!in!the!“factory”,!
– Output!à any!item!or!feature!on!a!product! including!rework,!equipment!preferences,!
that!is!defined!as!critical!by!the!customer. undocumented!inspection!criteria,!
“experience”,!work-arounds,!etc.
Input!(X):
– Input!à item!that!has!an!impact!on!an!
Ideal!process:
output. – The!process!as!you!would!like!it!to!be,!
simple,!efficient,!defect-free,!and!without!
any!non-value-added!steps.
36 36 62
Process Capability
Capability : Some common Metric

qCp/Cpk
qSigma Level/ Z-Score
qDPU
qIPY/ RTY etc…
04 Process Capability 2 63
Measure Types
Continuous/Discrete
Continuous: Any variable measured on a continuum or scale

that can be infinitely divided. Common types include weight,
time, money, or temperature.
Discrete or Attribute: A count of items, a characteristic or
category. Common types include counts of orders or units,
location/geography, frequency of an action or event.
Continuous Discrete/Attribute
• Number of orders filled
• Time to delivery
• Number of days lead
• Cost or price time
• Percent of on-time • Type of product
orders*
• Number of ingredients
* Discrete converted to a continuous “score” or percentage
3-3
q In training room, the students are required to

report between 09:00+10 mts
qAverage reporting time of 3 batches are as
follows:
mBatch-1 : 09:00
mBatch-2 : 09:10
mBatch-3 : 09:15
Batch 1: Batch 2:
08:00, 09:00, 10;00 08:50, 09:10, 09;30
Average: 09:00 Average: 09:10
Batch 3:
09:15, 09:14, 09:16
Average: 09:15
5
A Simple Definition
q Capability metrics report how well process output features meet
customer specifications
q EXAMPLE:
m PPM (Parts Per Million) reports the number of defective parts expected in a
group of 1,000,000, and thus the defective and compliance rate.
² Example: 100,000 PPM reflects 90% compliance to specifications
m Cp (sē-pē) reports the ratio of the specification range to the natural range of
process output
² Example:
v Temperature Spec range: 68 to 72 0F
v Process Temperature Range: 69 to 71 0F
v Cp = (72-68)/(71-69) = 4/2 = 2.0
Capability Calculation : Cp/ Cpk
q Complaints are occasionally received by customers about the
temperature level in a waiting room.The environmental control system
for the room is designed to maintain temperature between two user-
defined values. Since the system was installed, temperature levels
were recorded hourly for each of ten hours per day.One of the
business partners who was trained in Six Sigma collected a year’s
worth of data from the records for analysis.
q A technician is in charge of a Heat Treatment Shop. Material is

Annealed before Heat Treatment. One critical parameter for annealing
is the Soaking Temperature. Soaking Temperature, along with cycle
time determines the Physical properties of Steel. It is important to
maintain a consistent range of temperatures in the Soaking Furnace.
7
Step 2 – Short term variation - example

q Use Minitab to estimate short
term variation:
m Stat > Quality Tools >
Capability Analysis
(Normal)
Why Measure Capability?
q Enables allocation of resources based on DATA!
(This is not common!)
q Quantifies the defect rate
q Identifies opportunities for improvement
q Assessing capability allows the organization to predict its
true quality levels for all its goods and services
q Identifies the nature of the process problem – centering or
spread
9
Step 3 – Short term variation example
q Use Minitab to estimate short term variation:

m Stat -> Quality Tools -> Capability Analysis (Normal)
Step 4: Compute Cp
Process Potential Capability(Cp):
Product Tolerance USL - LSL
Cp = =
Process Tolerance 6s st
Product Tolerance
q Example:
m A process mean is 325,
standard deviation is 15,
an upper spec. limit is 380 and
lower spec. limit is 270
-5 -4 -3 -2 -1 0 1 2 3 4 5 q What is Cp?
q What is Cp if the mean is 355 but
-3s +3s the standard deviation does not
Process Tolerance change?
11
A Problem With Cp
qWhich is the better process?
qWhat is the difference in Cp between the two
processes?
qWhat can be done to make Cp more effective as
a process capability statistic?
-5 -4 -3 -2 -1 0 1 2 3 4 5 -5 -4 -3 -2 -1 0 1 2 3 4 5
Step 4: Meet Cpk – Process Performance
Cpk = min{ Cpl, Cpu} q Example:

m - LSL m A process mean is 355,
Cpl = standard deviation is 15,
3s st an upper spec. limit is 380
USL - m and lower spec. limit is 270
Cpu = q What is the Cpk?
3s st
q What is the Cp?
-5 -4 -3 -2 -1 0 1 2 3 4 5
13
Cpk for Chamber Temperature

USL - LSL 1265 - 1235
Cp = = = 1.01
6 ´ S st 6 ´ 4.86
é X - LSL USL - X ù
Cpk = min ê , ú
ëê 3 ´ s st 3 ´ s ú
st û
é1246.1 - 1235 1265 - 1246.1ù

= min ê ,
ë 3 ´ 4.86 3 ´ 4.86 úû
= 0.74
123 124 125 126 127
Defects vs. Defectives
q What is a defective?
m A unit that contains at least one defect.
q What is a defect?
m An output of a process that does not meet a defined
specification:
² concentration, time, accuracy, satisfaction index, etc.
Name: Bozo the Cloun

Address: 35 Mane St
City: Circustoun, Kanses
15
DPU: Defects Per Unit

qDPU =
Number of defects found in a group of units
Number of units in the group
qDPU = Average Number of Defects Per Unit

qExample:
mEach document has 5 key sections
mA group of 100 documents are inspected, 500
sections
m18 sections were not correct
mDPU = 18/100 = 0.18
Process Metrics the 6s Way
q IPY6s -- Six Sigma in-process yield.
IPY6s =1 - dpu IPY6s = e- dpu

m Six Sigma counts defects not defectives.
m Defects cause Rework!
q RTY -- rolled throughput yield.

m RTY is the probability of any unit having zero defects.
m RTY is the product of all subordinate process yields.
RTY = IPY 1´ IPY 2 ´ IPY 3 ´ ...´ IPYk
Remember: Focus on the defects.

17
Calculating DPU from IPY

q Defects per unit (dpu):
100%
dpu = Total defects observed 80%
Total units processed 60%

IP Y
40%
20%
q For one defect opportunity per unit: 0%

0 0.5 1 1.5 2 2.5 3 3.5 4
m IPY = 1-dpu dpu
m dpu = 1-IPY
q For multiple defect opportunities per unit:
m IPY = e-dpu
m dpu = -Ln(IPY)
DPMO: “Normalizing” for Complexity with Opportunities
q DPU is determined by two key components:

m Process propensity for noncompliance
² (Tendency to “goof up”)
m Number of opportunities for noncompliance in a “unit”
² (Number of ways to “goof up”)
m Example:
² Unit = one shift’s worth of production, dpu = 6
² Unit = one day’s worth of production, dpu = 12 (two shifts)
q Opportunities for defects focus on:

m noncompliance to design intent
m noncompliance to customer expectations
q Defects Per Million Opportunities (DPMO) “Normalizes” for Complexity
defects/Unit
DPMO = ´ 1,000,000
Opportunities/Unit
19
X-Ray processing flow and specifications
Check in Retrieve Take Send Results

And Referral The To
Register Documents X-Ray Physician
Primary Generate/Send
Physician Request for
Visit X-Ray
Cycle time specifications:

Check in: 15 minutes
Retrieval: 2 minutes
X-Ray: 30 minutes
Send Request: 2 business days
X-Ray Calculations Example
Check in Retrieval X-Ray Send Results Final Test

Units=345 Units=182 Units=160 Units=182 Units=450
defects=12 defects=37 defects=15 defects=2 defectives=50
Opps/Unit=3 Opps/Unit=2 Opps/Unit=2 Opps/Unit=1
IPY = ____ IPY = ____ IPY = ____ IPY = ____ FTY = ____
dpu = ____ dpu = ____ dpu = ____ dpu = ____ dpu = ____
RTY = ____ RTY = ____ RTY = ____ RTY = ____ RTY = ____
Process Totals:
dpu = ________
Exercise: Calculate the missing quantities and RTY = ________
select the first area of focus for the Green DPPM = ________
Belt project. (Hint: Assume all of the DPMO = ________
opportunities counts are PASSIVE in this SigmaLevel = ________
example.)
21
X-Ray Calculation Results

Check in Retrieval X-Ray Send Results Final Test
Units=345 Units=182 Units=160 Units=182 Units=450
defects=12 defects=37 defects=15 defects=2 defectives=50
Opps/Unit=3 Opps/Unit=2 Opps/Unit=2 Opps/Unit=1
IPY = 96 . 6 IPY = 81 . 6 IPY = 91 . 1 IPY = 98.9 FTY = 88.9

dpu = 0 . 035 dpu = . 203 dpu = . 094 dpu = .011 defectives / unit = .11
RTY = 96 . 6 RTY = 78 . 8 RTY = 71 . 8 RTY = 71.0 RTY = 71.0
Process Totals:
dpu = å dpu k = 0.035 + 0.203 + 0.094 + 0.011 = 0.343
RTY = Õ IPY = (. 966 ) ´ (. 816 ) ´ (. 911 ) ´ (. 989 ) = . 71
k Where should the
DPPM =
defectives
´ 1,000,000 =
50
´ 1,000,000 = 111,111 Green Belt focus
# produced 450
his energy?
dpu .34282
DPMO = ´ 1,000,000 = = 42,852
Opp. / unit (3 + 2 + 2 + 1)
ANALYZE PHASE
Analyze Phase 74
Tools for Determining Potential Cause & Effect
Module Objectives
q Introduce Fishbone Diagrams
q Introduce Cause and Effect Matrix Diagrams
05 C & E Tools
© Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved. Pg 2 75
Common & Special Causes
Type of Cause will Often Appear in Data
COMMON
• Variation/effects created by normal factors in the process
• Random, normal, chronic, more frequent
Response: Look closer at process data, procedures, normal
environment
SPECIAL
• Variation/effects created by factors outside normal operation
• Non-random, unusual, sporadic, less frequent
Response: Avoid over-reacting, follow process and data to
pinpoint unusual conditions/causes
One Method: Fishbone Diagrams

q Fishbone diagram:
m A method of systematically identifying
all of the potential causes that may
be contributing to a problem (effect)
q Method for construction of a fishbone diagram:

1. State the problem, place it in a box on the right side
2. Draw a horizontal arrow to the problem box
3. Write the traditional main categories of factors* or suspected
categories above and below the line, and connect them to the main
line
4. Within each of the main categories, brainstorm and list all of the
detailed factors that may be causing the problem
5. Further refining: list all of the inputs in each of the detailed factors
05 C & E Tools
Processing? Transactional?
Policies Place
4 P’s
Man Machine Method

People Procedure
6 M’s
Measurement Materials Mother Nature

(environment)
© Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved. Pg 5
Start With a Process Map

q Refer to a macro-map from an expense accounting example
Finance Finance
Account Charge Personnel Distributes Departme
Documentation Enter Departmental nt Heads
Sent to Finance Charge on Expense Review
System Reports Report
Departmental
Heads Request
Change
05 C & E Tools
Fishbone Diagrams in Minitab Fishboneexample.mtw
q Open Minitab.
q Select STAT > Quality Tools > Cause-and-Effect…
Fishbone Diagrams in Minitab

q Enter the column headings in six columns in the Minitab
worksheet.
q Enter Causes under the appropriate heading.
05 C & E Tools
Correcting Expense Account Posting
q Create a simple fishbone from the Expense macro-process map.
Cause-and-Effect Diagram
Measurements Material Personnel
Error filling out form

Bad forms
Error reading form
O utdated guides
Error key punching
D ifferent dept
definitions
C omputer glitch
O utdated procedures
Env ironment Methods Machines
Since Jan 2019 late arrival

of employees in HA Lab
has gone up
05 C & E Tools
05 C & E Tools
C-E MATRIX DIAGRAMS
Prioritization -- Cause & Effects Matrix

q C&E Matrix
m A matrix that numerically relates inputs (X’s) to Effect and
Frequency of occurrence
m Segregates the causes to common and assignable.
q Benefits
m The team agrees on the cause priorities.
m The C&E matrix priorities are customer driven.
?
05 C & E Tools
C_E Matrix
High
A. Cause-1
B. Cause-2
C. Cause-3
D.Cause-4
Frequency
E. Cause-5
F. Cause-6
G. Cause-7
H.Cause-8
I. Cause-9
J. Cause-10
K. Cause-11
L. Cause-12
Low M.Cause-13
Low High
Impact
© Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved.
C&E Matrix -- Summary

q Use the C&E Method sequentially as a brainstorming
tool.
q Macro-C&E
m Use all of the outputs within your project scope.
m Use only the process steps for the inputs.
m Rank order the process step as it affects the output.
q Micro-C&E
m Start a C&E matrix with the same outputs and the top 3-4
process steps from the macro-C&E.
m Include all of the inputs from each of the process steps for
a micro-C&E.
05 C & E Tools
Graphical!Methods
Quality!Growth!Services
Module Objectives
u Introduce!a!few!methods!of!graphical!analysis
– Box!plots
– Main!effects!Plot
– Interval!Plot!and!
– Interaction!plots
– Scatter!Plot
06 Graphical tools
1 - 2© Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 2 83
Process Variation
Intentional Unwanted
y = f Process ( x1 , x2 ,..., xk ) + f Noise (n1 , n2 ,..., nk )
u Noise!variation!from!discrete!inputs
– Different!employees,!systems
– Different!days,!shifts
– Different!batches,!inputs
u Noise!variation!from!continuous!inputs
– Time!of!day
1 - 4© Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 4
Common Classification of Noise Variables

2 2 2 2
s Noise = s Positional + s Cyclical + s Temporal
u Positional!(within!part!variation)
– Variation!within!a!single!office
u Cyclical!(part-to-part!variation)
– Variation!between!offices
u Temporal!(time-to-time!variation)
– Shift-to-shift,!Day-to-Day,!Setup-to-setup
– Variation!not!accounted!for!by!Positional!or!Cyclical!
06 Graphical tools
Graphical Analysis – Example
u A!company!is!facing!long!cycle!time!of!testing!of!
product.
u The!company!expects!its!Experts!to!test!120!items!
every!2!hrs.!
u The!following!data!is!collected:
– Number!of!items!tested!per!2!hr period
– Month
– Day!of!Week
– Name!of!the!Expert
– Time!of!Day
Exercise: Determine Capability

u Using!Minitab,!analyze!the!Items!checked!data!in!
LoanApps.mtw for!process!capability.
– Remember,!the!expectations!are:!!
120!applications!reviewed!per!2!hr period.
u What!is!the!capability?
u Are!these!good!or!bad!values?
06 Graphical tools
Capability Analysis
u Run!Stat!>!Quality!Tools!>!Capability!Analysis>Normal…
Process Capability of Apps_Rev
LSL
P rocess D ata Within
LS L 120.00000 Overall
Target *
USL * P otential (Within) C apability
S ample M ean 152.61944 Cp *
S ample N 1440 C PL 0.33
S tDev (Within) 32.89894 C PU *
S tDev (O v erall) 33.51791 C pk 0.33
C C pk 0.33
O v erall C apability
Pp *
PPL 0.32
PPU *
P pk 0.32
C pm *
75 100 125 150 175 200 225 250

O bserv ed P erformance Exp. Within P erformance Exp. O v erall P erformance
P P M < LS L 144444.44 PPM < LS L 160719.67 P P M < LS L 165228.34
PPM > USL * PPM > USL * PPM > USL *
P P M Total 144444.44 PPM Total 160719.67 P P M Total 165228.34
Box Plot Comparisons
u Construct!a!Box!plot!for!Lines!by!Loan!Officer
– In!Minitab!select!Graph!>!Box!plot…
06 Graphical tools
Box Plot Output
u Notice!that!box!plots!do!NOT!require!“numeric”!x!
variables.!!This!plot!will!accept!“categorical”!text!
variables.
Boxplot of Apps_Rev vs LO
250
200
Apps_Rev
150
100
Fred Joe Mary Sam

LO
1 - 10
© Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 10
Recall: The Box Plot Interpretation

u Whisker!definition: Whisker
– Upper!and!lower!whiskers!
extend!from!the!quartile! Third Quartile
lines!to!the!highest!and!
lowest!values!inside!the!
upper!and!lower!limits,!
respectively.
u Outliers
– Points!beyond!the!upper!and!
lower!limits Median
First Quartile
Upper Limit: Q3 + 1.5*(Q3-Q1) Whisker

*
Lower Limit: Q1 - 1.5*(Q3-Q1) Outliers
*
06 Graphical tools
1 - 11
© Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 11 87
Box Plot Comparisons
u Construct!a!Box!plot!for!Lines!by!Loan!Officer!and!Time!
of!Day
– In!Minitab!select!Graph!>!Box!plot…
1 - 12
Main Effects Plots
u Plots!averages!of!the!data!at!the!factor!levels
– Stat!>!ANOVA!>!Main!Effects!Plot…
Main Effects Plot (data means) for Apps_Rev

LO Time of Day
200
190
Mean of Apps_Rev
180
170
160
150
140
130
Fred Joe Mary Sam EarlyAM LateAM EarlyPM LatePM
06 Graphical tools
1 - 13
Interaction Plots
u An!interactions!plot!is!a!plot!of!
means!for!each!level!of!a!factor!
with!the!level!of!a!second!factor!
held!constant.!!
u Parallel!lines!indicate!no!
interaction!is!present.
Interaction Plot (data means) for Apps_Rev
EarlyAM LateAM EarlyPM LatePM
LO
200
qStat > ANOVA > Interactions Plot… Fred
Joe
180
Mary
Sam
LO 160
140
120
Time of Day
200
EarlyAM
LateAM
180
EarlyPM
LatePM
160
Time of Day
140
120
Fred Joe Mary Sam
1 - 14
The (w)Rap!
u List!your!key!takeaways! u How!can!I!use!this!info!in!
from!this!module. my!project?
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
06 Graphical tools
1 - 18
Test of Hypothesis
Introduction
Module Objectives
q To introduce statistical hypothesis testing
q To define null and alternative hypotheses
q To demonstrate a method for constructing a hypothesis

test
q To explain test of hypothesis for mean
07 Test of Hypothesis 4 90
The Purpose of Inferential Statistics
q The purpose:
m To draw conclusions (inferences) about populations from
measurements of samples
q The truth about populations:

m Populations have parameters which are fixed, and known
only to God.
m Statistics is the science of guessing, from sample data, what
God knows about population parameters.
q Our initial guesses, before we collect data, are called

hypotheses.
5
Hypotheses
q Hypotheses:
m Hypotheses are statements about population parameters, not
statements about samples.
q The Null Hypothesis:
m Abbreviated as H0
m Usually H0 is a statement of no effect or no difference.
m We reject or fail to reject H0 based on statistical evidence.
q The Alternative Hypothesis:
m Abbreviated as Ha
m A statement about a population parameter that is suspected of
being true, if H0 rejected
Order in the Court
q Hypothesis testing is like
the Indian legal system
where a defendant is True State of Nature
assumed innocent until Did Not Commited
proven guilty. Commit Crime A Crime
mH0: Defendant is
Commit Crime
Did Not
Guilty
innocent (Assumed) Correct
Goes
Verdict
mHa: Defendant is guilty Free
Verdict
(Must be proved)
Commited
A Crime
Innocent
Correct
Person
Verdict
Convicted
?
7
Hypothesis Testing Errors

q Type I error
m Rejecting the null hypothesis
when it is, in fact, true
m Also known as producer risk
m The probability of a Type I error
is denoted by a. (0< a <1)
q Type II error True State of Nature
m Accepting the null hypothesis, Ho Ha
when it is, in fact, false
Conclusion Drawn
Correct Type II
Ho
m Also known as consumer risk Decision Error

m The probability of a Type II
error is denoted by b (0< b <1)
Type I Correct
Ha
Error Decision
Means Tests
Module Objectives
q To demonstrate four hypothesis tests for means
m One sample z/t test
m Two sample t test
m Multisample Test- One way ANOVA
q To practice hypothesis testing for means
Testing Means
mtarget
q Testing for one mean Ho: m = mtarget
m Z-test for large samples or Ha: m < mtarget
when s is known m > mtarget
m t-test for smaller samples or m ¹ mtarget
when s is unknown m
q Testing for two means Ho: m1 = m2

m 2-sample t-test Ha: m1 < m2
m1 > m2
m1 ¹ m2
m1 m2
q Testing for three or more means
m 1-way ANOVA Ho: m1 = m2 =…= mn
m Covered in a later module
Ha: @ least one is
different from
another m1 m2 m3 m4
12
Comparing Two Means

q Fundamental analysis q The 2-sample t-test is
question: “Is there a real the hypothesis test for a
difference between these two
groups?”
difference in means.
Population 1 Population 2
-6 -4 -2 0 2 4 6 8
Comparing two means is one of the most common HT’s.

MeanExamples.mtw
Exercise – 2-sample t-test
q A hose manufacturing company has used an inventory
planning system for many years to manage the inventory
of production consumables and maintenance spares. A
new system has recently been installed for evaluation
with the hope of reducing average inventory.
• Should the standard system be replaced by the new

system? The changeover would be expensive.
• Conduct a 2-sample t-test in Minitab. Use a = 0.05.
(Hint: Stat > Basic Statistics > 2-sample t…)
14
Exercise – 1 sample test

Mean Vs. Target (t-test)
q An Engineer has designed a new formulation to improve the
breaking load.
q He checked the breaking load of 15 samples made with new
formulations. Want to know if the new formulation has impact
on breaking load. The average breaking load from old
formulation was 85.5 kg/ sq mm
1. Conduct a 1-sample test to test this hypothesis. Use a = .05.
2. The data are in worksheet MeanExamples.mtw. (Column:

Trained) (Hint: Stat > Basic Statistics > 1-sample t…)
1-sample t-test – Minitab Results
Test of mu = 85.5 vs mu < 85.5
Variable N Mean StDev SE Mean
Trained 15 80.79 8.93 2.31
Variable 95.0% Upper Bound T P
Trained 84.86 -2.04 0.030
What is the conclusion?

16
2-sample t-test – Minitab Results

Two Sample T-Test and Confidence Interval
Two-sample T for New System vs Std System
N Mean StDev
New System 10 44.30 4.64
Std System 10 46.50 5.40
Estimate for difference: -2.20

95% upper bound for difference: 1.71
T-Test of difference = 0 (vs <): T-Value = -0.98 P-Value = 0.171 DF = 18
Both use Pooled StDev = 5.04
Analysis of Variance
ANOVA One-way –
Hypothesis testing for multiple means
m1 m2 m3 m4
Introduction to ANOVA
ANOVA is a general technique that can be

used to test the hypothesis that the
means among two or more groups are
equal, under the assumption that the
population are normally distributed.
Introduction to ANOVA
One way ANOVA:
If in the experiments if the effect of various
levels of a single factor is to be compared,
it is called one way ANOVA.
e.g If the effect of various (levels) of

Temperature (Factor) i.e. 800C, 1000C,
1100C, 1200C to be compared on resistors
on a heating oven.
21
One Way ANOVA

q In ANOVA, the variation in Response is
partitioned in to 2 components,
m Factor effect (Treatment effect) and

m Error effect.
How is it done?
One Way ANOVA
q One way ANOVA form two mean sum of squares,
m one for treatments (MST) and
m the second for error (MSE).
qThese are typically displayed in the tabular form, known
as an ANOVA Table.
SOURCE ANOVA TABLE
df SS MS F
Treatment l-1 SS(T) SS(T)/(l-1) MS(T)/ MSE)
Error n-l SSE SSE(n-l)
TOTAL n-1 TSS
23
ANOVA Introduction
q Example:
m A rubber manufacturing company wants to compare
effects on rapid modulus with four new formulations
against the existing formulation.
m Ten different batches were sampled for each formulations
and the average RM was computed.
q Problem:
m Were any of the formulations different from each other and
the existing formulation?
How can we analyze this data?

Case Studies IntroANOVA.mtw
q Case 1
Trial TmtA TmtB TmtC TmtD TmtCntrl1
1 6.6 7.0 6.1 7.5 5.5 q Case 2
2 6.4 7.0 5.9 7.5 5.7
Trial Tmt1 Tmt2 Tmt3 Tmt4 TmtCntrl2
3 6.5 7.0 6.0 7.6 5.5
1 6.0 5.5 6.4 6.4 7.6
4 6.6 6.8 6.1 7.4 5.4
2 4.3 8.3 5.3 7.1 4.9
5 6.5 7.1 5.9 7.4 5.5
3 4.7 4.8 6.7 8.8 6.3
6 6.6 7.0 5.8 7.4 5.5
4 6.7 6.8 6.9 8.8 5.3
7 6.5 6.9 6.0 7.6 5.4
5 5.0 8.2 6.8 6.0 2.2
8 6.6 6.9 5.9 7.7 5.5
6 4.9 8.2 5.0 7.4 6.5
9 6.7 6.9 6.1 7.5 5.5
7 7.6 6.3 7.2 9.1 5.7
10 6.5 6.9 6.1 7.6 5.7
8 7.9 7.8 6.1 5.5 6.0
Mean 6.6 7.0 6.0 7.5 5.5
9 7.2 6.6 6.4 9.4 7.4
10 8.2 10.1 5.6 8.6 3.0
Mean 6.3 7.3 6.2 7.7 5.5
Are the treatments different? Are the they different now?

25
ANOVA in Minitab – A Quick Demo
q In Minitab select Stat>ANOVA>One-way (Unstacked)…
H 0 : m1 = m 2 = ... = m j
H a : @ least one m j is different

Minitab ANOVA Results – Case 1
What do you
One-way ANOVA: TmtA, TmtB, TmtC, TmtD, TmtCntrl1
think this
Analysis of Variance means?
Source DF SS MS F P
Factor 4 24.65720 6.16430 643.60 0.000
Error 45 0.43100 0.00958
Total 49 25.08820
Individual 95% CIs For Mean
Based on Pooled StDev
Level N Mean StDev ----------+---------+---------+------
TmtA 10 6.5500 0.0850 (*)
TmtB 10 6.9500 0.0850 (*)
TmtC 10 5.9900 0.1101 (*)
TmtD 10 7.5200 0.1033 (*)
TmtCntrl 10 5.5200 0.1033 (*)
----------+---------+---------+------
Pooled StDev = 0.0979 6.00 6.60 7.20
Accept or Reject Ho? What is Ho?

27
Minitab ANOVA Results – Case 2

What do you
One-way ANOVA: Tmt1, Tmt2, Tmt3, Tmt4, TmtCntrl2
think this
Analysis of Variance means?
Source DF SS MS F P
Factor 4 31.51 7.88 3.89 0.009
Error 45 91.25 2.03
Total 49 122.76
Individual 95% CIs For Mean
Based on Pooled StDev
Level N Mean StDev --+---------+---------+---------+----
Tmt1 10 6.250 1.457 (------*-------)
Tmt2 10 7.260 1.561 (-------*------)
Tmt3 10 6.240 0.729 (-------*-------)
Tmt4 10 7.710 1.412 (------*-------)
TmtCntrl 10 5.490 1.748 (-------*------)
--+---------+---------+---------+----
Pooled StDev = 1.424 4.8 6.0 7.2 8.4
Accept or Reject Ho?

In Summary
q ANOVA is a powerful tool for testing equality of means
q ANOVA compares the variability between group means
to the average variability within a group
m ANOVA p-values are significant when the group-to-group
differences are too large to be explained by the within group
variability
q The initial assumptions for ANOVA:
m Each subgroup is normally distributed
m The variances of all the subgroups are equal
q The ANOVA table is the standard way of reporting
ANOVA analysis
29
The (w)Rap!
q Key Takeaways q How can I use this in my
m ANOVA is a powerful tool for project?
determining the differences m ______________________
between multiple means m ______________________
m ANOVA compares the
between-group variability to the
m ______________________
within-group variability m ______________________
m The assumptions of ANOVA m ______________________
are subgroup normality and m ______________________
equal variances
m ______________________
m Others:
m ______________________ m ______________________
m ______________________ m ______________________
m ______________________ m ______________________

Scatter!Plot,!Regression!and!
Correlation
Quality!Growth!Services
Module Objectives
u To!introduce!simple!linear!modeling
u To!demonstrates!methods!in!Minitab!for!
constructing!models
u To!practice!regression!techniques
u To!explain!the!difference!between!correlation!&!
regression
08 Scatter,1 Correlation & Regression

-©2Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 2 103
Introduction to Regression
u Used!for!quantitative!variables!(X’s!and!Y’s)!
u For!review:!!What!is!the!focus!of!Six!Sigma?
Y = f (x )
Q. What!does!this!equation!represent?
A. A!mathematical!model!of!a!process
u Purpose!of!Regression:!!to!predict!Y!from!a!setting!of!x
u Examples:
– Revenue!=!f(sales!rep!experience,!#!of!client!visits,!length!of!client!visit)
– Loan!Default!Rate!=!f(credit!rating,!monthly!income,!age)
– Time!=!f(#!tasks,!complexity,!queue!length!)
1 -©3Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 3
A Simple Linear Model

Independent
Variable
y = b 0 + b1 x
24 Dependent
22 Variable
20
18
Intercept Slope
16
is the value of = Rise/Run
14 y(x=0) 6 =6/3=2
12
10
8 3
y = 4 + 2x
6
4
2
0
0 1 2 3 4 5 6 7 8 9 10

Constructing a Linear Model
Investigate Potential
Higher-Order Model
No
Collect Data Plot the Seems or Yes (technology

(y & x) Points Linear? available)
Yes (technology not available)

Calculate Draw a “best
Slope and fit” line
Intercept (eyeball)
Check Fit Use Least

Statistics Squares Fit
Collecting Data (y & x) – A Few Thoughts

u Make!sure!the!process!settings!cover!the!likely!
operation!range!(but!not!too!far).
– Too!great!a!range!à Points!outside!the!normal!range!may!have!too!
great!an!effect!on!the!model.
– Too!small!a!range!à Error!term!may!dominate!the!fit.
u Take!several!replicates!at!each!input!setting!(x).
– Replicate!runs!help!increase!the!model!accuracy.
u Randomize!runs!whenever!practical.
– Run!order!is!often!significant!factor.
– The!output!(y)!at!different!inputs!(x’s)!is!not!always!independent!of!
previous!settings.

Analyzing Coffee Shop Revenues
u An!engineer!is!trying!to!find!the!
causes!of!Low!breaking!strength.!
One!of!the!suspect!causes!are!the!
number!of!twists!in!the!linking!rod! Twists x Strength y
Data!are!collected!related!to! 1 500
twists!and!stength 3 1750
u Construct!a!simple!linear!model! 6 1800
for!Strength!as!a!function!of!the! 7 2000
2 600
Number!of!twists.
4 1400
12 4200
15 5400
5 1700
9 3000
Plot the Data
6000
6000
5500
5500
5000
5000
4500
4500
Strength
4000
4000
($)
3500
3500
Expense
3000
3000
2500
2500
2000
2000
1500
1500
1000
1000
500
500
00
0 1 2 3 4 5 6 7 8 99 10 11
11 12
12 13
13 14
14 15
15 16
16
DaysTwists
in Confinement

Draw a “Best Fit” Prediction Line–Eyeball
6000
5500
5000
4500
Y
4000
3500
3000
2500
2000
1500
1000
500
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
X
The Error Term Graphically
6000
5000
e10
4000 e9
3000
e8
2000 e3 e5 e7
e4 e6
1000 e1
e2
0
0 2 4 6 8 10 12 14 16
08 Scatter,1 Correlation
-©10
& Regression
Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 10 107
Residuals versus Fits
u Regression!Model!and!Residual
YActual = b 0 + b1 X + e
uYactual =!Ypredicted +!e

Error (residual): Difference
between the actual value and
the predicted value at some X
Observed Value: The
actual value at some X
The Line (fit): is the
predicted value at some X
1 -©11
Quality Growth Services Pvt. Ltd. Unpublished proprietary work available only under license. All rights reserved. 11
The Fit Term Graphically
6000
5000 X
4000 X
3000 X
X
2000 X
X
X
1000 X
X
X
0
0 2 4 6 8 10 12 14 16
-©12
& Regression
The Method of Least Squares
u The!true!“best!fit”!model!minimizes!the!total!
deviation!of!the!observations!from!the!model.
åe i
2
= Minimum
u The!equations,!derived!with!calculus,!for!calculating!
the!coefficients!using!the!method!of!least!squares!are:
b1 » slope =
n å x y - (å x )(å y )
i i i i
nå x - (å x ) 2 2
i i
b 0 » y - intercept = y - b1 x
1 -©13
Regression in Minitab
u Open!a!blank!Minitab!worksheet!and!enter!the!food!
service!data.
u From!the!main!menu!select:
Stat!>!Regression!>!Regression…
-©14
& Regression
Minitab Session Output
The!regression!equation!is Does This

Look
Y!=!63!+!339!X Familiar?
Predictor!!!!!!!!Coef StDev T!!!!!!!!P
Constant!!!!!!!!!63.5!!!!!!!185.5!!!!!!!0.34!!!!0.741
Twists!!!!!!339.30!!!!!!!24.15!!!!!!14.05!!!!0.000
What does Ho: b1=0 Ho: b0=0

this tell
us? Ha: b1¹0 Ha: b0¹0
S!=!324.3!!!!!!!R-Sq =!96.1%!!!!!R-Sq(adj)!=!95.6%
1 -©15
Regression in Minitab – Another Way

u From!the!main!menu!select:
Stat!>!Regression!>!Fitted!Line!Plot…
-©16
& Regression
Minitab Output – Fitted Line Plot
Fitted Line Plot

Revenue = 63.5 + 339.3 Takeoffs
6000 S 324.346
R-Sq 96.1%
R-Sq(adj) 95.6%
5000
4000
Revenue
3000
2000
1000
0
0 2 4 6 8 10 12 14 16
Takeoffs
1 -©17
Minitab Storage
u Store!“Residuals”!and!“Fits”
-©18
& Regression
Exercise – Linear Models in Excel
u In!a!random!sample!of!nine!invoice!processing!
employees,!their!years!of!experience!(x)!and!
the!number!of!transactions!processed!per!hour!
x y
(y)!were!recorded.!!
4.2 2.8
u Use!the!data!to!obtain!a!least!squares!estimate
3.8 2.5
for!the!model!!y!=!b0 +!b1x.!!Check!the!adequacy
4.8 3.1
of!the!model!through!residuals!analysis. 3.4 2.1
4.5 2.9
u Answer: 4.6 2.8
4.3 2.6
y predicted = 0.086 + 0.616 x 3.7 2.4
3.9 2.5
1 -©19
Correlation
uCorrelation:
– Measures!the!strength!of!the!linear association!between!
two!variables
– Will!not!measure!strong!non-linear!relationships
– DOES!NOT measure!cause!&!effect
– Can!be!used!as!a!first!step!to!regression
– Should!be!used!in!conjunction!with!graphical!techniques
uThe!measure!of!the!strength!of!the!linear!
association!in!a!correlation!analysis!is!“R” –
the!correlation!coefficient.
-©20
& Regression
Properties of R
u -1<R<1
– R!>!0!indicates!a!positive!linear!relationship
– R!<!0!indicates!a!negative!linear!relationship
u R2 gives!the!proportion!of!total!variability!of!the!y!
values!that!can!be!accounted!for!by!the!independent!
variable!x.
u Correlation!coefficients!are!not!additive.
– If!y!=!f(x1)!with!R1 and!y!=!f(x2)!with!R2,
Then!for!y!=!f(x1,!x2),!the!total!variability!¹ R12 +!R22
– x1!&!x2! may!be!highly!correlated!and!contribute!the!same!information!!
for!correlation!with!y.
1 -©21
Decision Guidelines on Correlation

u Decision!points!for!
correlation!are!based!on!
sample!size.
Decision Decision
u The!following!table!can!be! n Point n Point
used!as!a!guideline!for! 5 0.878 18 0.468
different! 6 0.811 19 0.456
7 0.754 20 0.444
sample!sizes!of!n: 8 0.707 22 0.423
9 0.666 24 0.404
2 10 0.632 26 0.388
D.P. = 11 0.602 28 0.374
n 12 0.576 30 0.361
13 0.553 40 0.312
14 0.532 50 0.279
u For!other!sample!sizes!the! 15 0.514 60 0.254
decision!point!is! 16 0.497 80 0.220
17 0.482 100 0.196
approximately:
-©22
& Regression
Correlation Calculation
u Calculating!R: u In!the!same!notation:
SS xy SS xy
R= b=
SS xx SS yy SS xx
u where: u Then!b and!R!are!related!

this!way:
( )2
SS xx = å xi - x = å xi -
2 (å xi )2
n
SS xx
SS yy = å (y - y ) = å y -
2 (å y ) 2 i
2
R= b
i
n
i
SS yy
( )( )
SS xy = å xi - x yi - y = å xi yi -
(å x )(å y )
i i
1 -©23
Calculating R & R2 in Minitab & Excel
u Minitab!Calculations:
– Stat!>!Regression!>!Regression… – Session!window!output!
reports!R2! (See!Slide!23)
– Stat!>!Regression!>!Fitted!Line!Plot… – R2 is!reported!on!the!
graph!under!the!model!(See!Slide!25)
– Stat!>!Basic!Statistics!>!Correlation… – Session!Window!
Output!(R)
Correlations!(Pearson)
Correlation of C1 and C2 = 0.980, P-Value = 0.000
u Excel!Calculations:
– On!the!Add!Trendline >!Options menu!select!“Display!R-
squared!value!on!chart”!!(See!Slide!21)
-©24
& Regression
Other Models
u Both!Minitab!and!Excel!have!the!capability!to!fit!higher-order!
models,!that!is,!models!with!squared!and/or!cubed!terms.
u The!Minitab!Fitted!Line!plot!has!choices!for!Linear,!Quadratic!
or!Cubic!fits
u Excel!can!also!fit!these!other!models:
– Polynomial!(to!sixth!order):!!
y!=!a +!b1x6 +!b2x5 +!b3x4 +!b4x3 +!b5x2 +!b6x!
– Logarithmic:!! y!=!a +!b*ln(x)
– Exponential:!! y!=!a ebx
– Power: y!=!a xb
u Both!Minitab!&!Excel!calculate!the!coefficients!and!provide!
estimates!of!R2 for!each!model
1 -©25
The (w)Rap!
u List!your!key!takeaways! u How!can!I!use!this!info!in!
from!this!module. my!project?
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
– _____________________ – ______________________
-©27
& Regression
IMPROVE PHASE
Improve Phase 116

Introduction to
Designed Experiments
IIPE - QGS-Business Excellence Division
Module Objectives
q To introduce the concept of statistically designed
experiments
q To discuss terminology for statistically designed
experiments
q To provide a tool for planning designed experiments
09 Design of Experiments 5 117

Experiments – Defined
An experiment is a test or series of
tests in which purposeful changes are made to
input variables of a process or system so that
changes in the output responses can observed
and identified
Douglas C. Montgomery
Y=f(x)
8
General Model of a Process or System

Noise Inputs
(Discrete)
Controllable Key Process

Process
Inputs Outputs
Noise Inputs
(Continuous)
What kind of inputs do you have?

Some Purposes of Experimentation
q Determine
m Which variables are most influential on the output
m Where to set the inputs to produce the output at the most
desirable level
m Where to set the influential inputs to reduce the variability in
the output
m Where to set the controllable inputs so that the effects of the
uncontrollable inputs are minimized
q Find the defining equation (y=f(x)) in order to optimize

the process
10
Consider Golf As an Experiment

q What is the output of a round of golf?
m A low score (a low number of strokes or
shots)
q What are some controllable inputs?

m Type of club & ball
m Walking and carrying the clubs or riding
in a cart
m Number of beers consumed while playing
q What are some uncontrollable inputs?
m Swing consistency
m Weather – wind, rain, sun, temperature
How can we test the inputs?

“Best Guess” Approach
q Used often in industry
q Procedure
m Select the “best guess” combination of factors
² Ping clubs, Titleist ball, riding in a cart, four beers
m Run a test (play a round)
m Compare the output with expectations (Score: 94 – not
good)
m If deficient, change one factor to a different level – Retest
m Repeat as necessary
q Drawbacks
m If the first guess is wrong, more trials are necessary –
inefficient and time consuming
m If first guess is acceptable, experimentation stops and “best”
solution may never be found
12
OFAT Approach – One-Factor-At-a-Time
q Often used when there is little q For the factors beers & walking
prior knowledge or riding:
q Procedure
m Select a combination of
Main Effects Plot (data means) for Score_2
factor levels for a Trans Beers
89
baseline
m Change the levels of one 88
factor at a time through-

Mean of Score_2
87
out the range of interest
of the factor 86
m For each factor, select

85
the optimum level
84
Ride Walk 0 4
Where should we set each level?

Drawbacks of OFAT
q Major drawback à OFAT fails to consider interactions
m Interaction – the failure of one factor to produce the same
effect at different levels of another factor
q Another drawback Interaction Plot (data means) for Score_2
m OFAT is always 92 Trans

Ride
W alk
91
less efficient than 90
statistically- 89
designed
Mean
88
experiments 87
86
85
84
83
0 4
Beers
14
The Solution – Factorial Experimentation
q The correct approach to dealing with several factors is

to conduct a factorial experiment.
m a.k.a. DOE (Design Of Experiments)
q Factorial experimentation
m Factors are varied together instead of one at a time
m Experiments are a complete set of experimental runs and
are analyzed when all of the runs are complete
DOE is the Six Sigma Experiment Method!

Factorial Experimentation -- Example
q Consider two factors of the golf q If a third factor is added, balls

example: beers & carts (Titleist or Topflite), the
q A factorial experiment would design becomes:
be set up as follows:
m Each level of each factor would
be run at each level of the other
Beers
4
Beers
0
w r
0 Carts
w r
Carts
Can you have more than three factors?

16
Step 1: State the Practical Problem
q Step 1 is not as easy as it sounds

q A clear problem statement is as necessary in
experimentation as it is in every other part of Six Sigma
q What would be the problem statement for our Golf
Experiment?
The average number of strokes

of my golf rounds is too high. I
average 90. I want to average
less than 85.

Step 2: State the Experiment Objective
q The Experiment Objective is much more specific than the problem
statement.
m Example:
² Problem Statement: Activity time for order processing has too much
variability
² Objective: Reduce the order processing time variation from ± 10%
to ± 5%
m Example:
² Problem Statement: The number of incorrect invoices is too large
² Objective: Reduce the number of incorrect invoices from 2% to
0.1%.
What would be an objective for Golf?

19
Step 2: State the Experiment Objective

q For our Golf example:
Objective: Reduce the average

number of strokes of my golf game
by five strokes.
q Other objectives:
m Increase the length of my drive by 30 yards
m Increase the number of times my drive stays in the fairway
m Increase the average number of greens reached in
regulation

Step 3: Select the Output Response(s)
q Output Responses: q Definition Questions

m Looking for the “Y” of Y=f(x) m What is important –
m Response is selected first centering or variation?
m Response should be related m How much change do we
to the Six Sigma project goals want to detect?
m Sometimes (often) there are m Is the measurement system
more than one output adequate?
response m How stable is the
response?
m Are there secondary or
consequential responses?
How important is the measurement system?

21
Step 3: Select the Output Response(s)
q Examples of experimental responses:

m Marketing: response rate to ads, brand recognition
m Sales: close rates
m Invoice process: correct invoices, days to pay
q For our golf example:

m Major response: Total number of strokes
m Other possible responses: Distance from tee and distance
from the fairway center (for club & ball experiments)

Step 4: Select the Input Factors
q Input factor – one of the inputs of a process whose
influence upon a response is studied in the experiment
m Quantitative (variable) inputs: Temperature, pressure, time,
etc.
m Qualitative (discrete) inputs: operators, machines, plants,
batches, catalysts, etc.
q Which factors to select?

m Use the Six Sigma tools!
²Process Map, C & E Matrix, FMEA
²Multi-vari analysis, Hypothesis tests
23
Step 4: Select the Input Factors

q For the golf example:
Factors: Type (brand) of clubs

Type (brand) of balls
Walking or Riding
Number of beers
What other factors influence golf scores?

Step 5: Select the Factor Levels
q Levels: The values (settings) of the input variable
m Example: Temperature is an input
²Levels: 125, 150, 175
m Example: Operator is an input
²Mary, Beth, Tom, Saunders
q For the Golf example:
Factor Level
Clubs Ping, Callaway
Balls Top Flite, Titleist
Transportation Walking, Riding
Beers 0, 4
25
Step 5: Select the Factor Levels

q Considerations for factor levels:
m How big of a change do I want to see?
m What is a normal range of variation?
m How far can I go and still be safe?
m What are the machine/process limitations?
m What type of experiment is this?
²Screening – Use widely-spaced levels
²Optimization – Use narrow levels, determined by previous
experiments.
q How many levels?

m Depends on resources, experimental purpose
m Two-levels is very convenient as shown in a later module

Step 6: Select the Experiment Design
q Simple Comparative Experiments

m Test of two means
²1- & 2-sample t-tests
²Paired t-tests
m 1- & 2-variance tests
m 1- & 2-proportion tests
q Experiments with a Single Factor:

m The Analysis of Variance
q Statistically Designed Experiments – DOE’s
DOE is a Six Sigma Deliverable!

27
A Form to Help With Planning
Effective experiments require a plan.

A Few Tips for Experimentation
q Use your non-statistical knowledge of the problem
m Especially valuable for choosing factors and levels
m Using statistics is no substitute for thinking about the problem
q Keep the design and analysis as simple as possible
m KISS – Keep it Simple, Stupid!
m Complex experiments and analyses are often fraught with
errors
q Know the difference between statistical and practical importance
m Just because a change in a process produces a statistically
significant difference, does not mean it is important
q Experiments are usually iterative
m Our knowledge increases with time. Expect to use several
experiments to arrive at the optimum process.
m Rule of thumb: Use not more than 25% of resources in 1st
Expt.
29
The Final Report

q Always write a separate report on a DOE
m DOE’s usually involve many people and consume lots of
resources
m Most people would like to know the results prior to the close
of the project
m Reporting/Presenting a DOE helps teach the principles of
DOE to a wider group. Remember critical mass & culture
change
q To help, there is an outline for a DOE Final Report:

DOEOutline.doc
Always report the DOE results!

Barriers to Effective Experimentation
Problem not clear
Objectives not clear
Inadequate brainstorming
Results of the experiment unclear
DOE is too costly
DOE is too time consuming
Lack of understanding DOE strategies
Lack of understanding DOE tools
Not confident during the early stages
Lack of management support
Need for instant results
Lack of adequate coaching / support
31
09 Design of Experiments 129

2k Factorial Designs
Module Objectives
q To demonstrate two-level factorial designs
q To review DOE terminology and notation
q To practice DOE design & analysis
10 2K factorial© Quality
DOE Growth Services Pvt Ltd. 2 130
The Top Five Reasons for Using 2k designs
1. And the number one reason for using Factorials is:
Y = f(x)
2. They are easy to understand and analyze (Minitab has many
routines for 2k designs).
3. They form the basis of fractional factorial designs, the subject of
the advanced techniques.
4. They can be augmented to form composite designs when more
detailed information is necessary.
5. They require relatively few runs per factor studied.
5
2k Factorial Designs -- Notation

q 2k designs are experiments where all factors have only two levels.
q Notation:
m In General: How many factors and levels in a 2 x 2 x 2?
How many treatment combinations in a full factorial?
m How many factors and levels in a 2 x 2 x 2 x 2 x 2?
How many treatment combinations in a full factorial?
What does 25 equal?
m How many factors and levels in a 27 experiment?
How many treatment combinations?
q 2k
m How many factors and levels in a 2k factorial?
m How many treatment combinations?
Adding to the DOE Method DOE Recipe
1. State the practical problem and experimental objective

2. State the factors and levels of interest
3. Select the appropriate sample size
m Stat>Power and Sample Size>2-Level Factorial Design
4. Design the experiment
m Stat>DOE>Factorial>Create Factorial Design…
5. Run the experiment/Collect the data
6. Construct the ANOVA table for the full model
m Stat>DOE>Factorial>Analyze Factorial Design…
7. Re-run a reduced model by eliminating non-significant (high-p-
value) or low Sum-of-Squares effects (Pareto or Normal Plot)
8. Investigate the residual plots to ensure model adequacy
m Stat>Regression>Residual Plots…
7
DOE Method (continued) DOE Recipe
9. Investigate the significant interactions (p-value < 0.05) – highest

order first
• Stat>DOE>Factorial>Factorial Plots or
• Stat>ANOVA>Interaction Plots
10. Investigate significant Main Effects (p-value < 0.05)
• Stat>DOE>Factorial>Factorial Plots or
• Stat>ANOVA>Interaction Plots
11. State the mathematical model obtained (Y=f(x)) – calculate e2 ‘s and
evaluate for practical significance
12. Translate the model into real process terms. Draw conclusions and
make recommendations.
13. Replicate the optimum conditions. Plan next experiment &
institutionalize the change
DOE Method Demonstration
q A Black Belt wants to determine the impact of different
levels of advertising on product recognition.
q He designs an experiment using:

m Two different Television profiles (two vs. three in a week),
m Two Radio profiles (daily vs. three times per week), and
m Two different Newspaper profiles (weekend only vs. daily).
q He has identified enough “similar” market cities to run

one replicate.
9
A Few Basics
q In a 2k experiment:
m One factor level is designated “low” and coded as a –1
m The other level is designated as “high” and coded as a +1
q Standard order: Table of contrasts
TV Radio Press
-1 -1 -1 PRACTICAL
INTERPRETATION
1 -1 -1
-1 1 -1
1 1 -1
-1 -1 1
1 -1 1
-1 1 1
1 1 1
• Exercise
• Create a 24 Factorial Design Matrix
• How many runs are needed?
Main Effects
q In a 2k DOE:
m The Main Effect for a factor is the
average of all the data when the factor
is at the high level minus the average
of all the data when the factor is at the
TV Radio Press Recognition
low level
-1 -1 -1 43
m Or: = Response
MainEffect - Response 1 -1 -1 45
High Low
-1 1 -1 45
1 1 -1 49
-1 -1 1 43
q For our experiment the Main Effect 1 -1 1 46
for TV is: -1 1 1 45
1 1 1 49
45 ++49 ++ 46 + 49 43 ++45 ++43 ++45
TV == --
4 4
== 47.25 - 44 = 3.25
11
Graphing Main Effects

Main Effect for TV
Response TV = High = 47 .25
48
47
Response
46
+3.25
45
44
43 Response TV = Low = 44
42
Low (-1) High (+1)
Level (TV)
MainEffect TV = Res TV = High - Res TV = Low

= 47 .25 - 44 = 3.25
Calculating Main Effects from Contrasts
q Multiply the response by the contrast for that factor,
sum the values, divide by n (the number of values for
any level)
TV Radio Press Recognition Rec. x TV Rec. x RadioRec. x Press
2x/wk daily weekend 43 -43 -43 -43
3x/wk daily weekend 45 45 -45 -45
2x/wk 3x/wk weekend 45 -45 45 -45
3x/wk 3x/wk weekend 49 49 49 -49
2x/wk daily daily 43 -43 -43 43
3x/wk daily daily 46 46 -46 46
2x/wk 3x/wk daily 45 -45 45 45
3x/wk 3x/wk daily 49 49 49 49
Sum 13 11 1
n 4 4 4
Sum/N 3.25 2.75 0.25
13
Contrasts and Calculations for Interactions
q How do you calculate the contrasts for the interactions?

m Multiply them together!
TV Radio Press T*R T*P R*P T*R*P Recognition

-1 -1 -1 1 1 1 -1 4343
1 -1 -1 -1 -1 1 1 4545
• Calculate the -1 1 -1 -1 1 -1 1 4545
1 1 -1 1 -1 -1 -1 4949
magnitude of
-1 -1 1 1 -1 -1 1 4343
the inter- 1 -1 1 -1 1 -1 -1 4646
actions the -1 1 1 -1 -1 1 -1 4545
same way, 1 1 1 1 1 1 1 4949
13 11 1 3 1 -1 -1 Sum
too.
4 4 4 4 4 4 4 n
3.25 2.75 0.25 0.75 0.25 -0.25 -0.25 Sum/n
Which ones are significant?

New Product Promotion Example
q Step 1: State the practical problem

m The Black Belt would like to improve recognition for a new
product.
q Step 2: State the factors and levels of interest

m TV Profile: 2x/week, 3x/week
m Radio Profile: Daily, 3x/week
m Newspaper Profile: Weekend only, Daily
q Step 3: Select the sample size

m Given: 1 replicate (How many runs?)
15
Step 4: Design the Experiment
q In Mintab: Stat>DOE>Factorial>Create Factorial Design
Step 4: Design the Experiment (cont.)
q Select the Factors button:
Why doesn’t Minitab need the level values?

17
Step 4: Design the Experiment (cont.)

q Select the Options… button:
Step 4: Minitab Output
q Click “OK” on the main DOE Menu
• Step 5: Run the experiment/Collect the data
19
Step 6: Analyze the Full Model

q Open Minitab worksheet
m Select Stat>DOE>Factorial>Analyze Factorial Designs…
Minitab Output – Full Model
Estimated Effects and Coefficients
for Recognition (coded units)
Term Effect Coef

Constant 45.6250
TV 3.2500 1.6250
Radio 2.7500 1.3750
Press 0.2500 0.1250
TV*Radio 0.7500 0.3750
TV*Press 0.2500 0.1250
Radio*Press -0.2500 -0.1250
TV*Radio*Press -0.2500 -0.1250
Do these look familiar?

21
Minitab Output – Full Model

Analysis of Variance for Recognition (coded units)
Source DF Seq SS Adj SS Adj MS F P

Main Effects 3 36.3750 36.3750 12.1250 * *
2-Way Interactions 3 1.3750 1.3750 0.4583 * *
3-Way Interactions 1 0.1250 0.1250 0.1250 * *
Residual Error 0 * * *
Total 7 37.8750
Why did this happen?

The Pareto Diagram
Pareto Chart of the Effects

(response is Recognition, Alpha = .05)
1.412
Factor Name
A TV
A
B Radio
C Press
B
AB
Term
AC
ABC
BC
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Effect
Lenth's PSE = 0.375
What terms should be kept?

23
Step 7: Re-run With the Reduced Model

q Select Stat>DOE>Factorial>Analyze Factorial
Designs…
m Remove terms ABC, BC, and AC & C
Minitab Results – Reduced Model
Estimated Effects and Coefficients for Recognition

(coded units)
Term Effect Coef SE Coef T P

Constant 45.6250 0.1250 365.00 0.000
TV 3.2500 1.6250 0.1250 13.00 0.000
Radio 2.7500 1.3750 0.1250 11.00 0.000
TV*Radio 0.7500 0.3750 0.1250 3.00 0.040
S = 0.353553 R-Sq = 98.68% R-Sq(adj) = 97.69%
Keep reducing until all p-values < alpha

25
Step 8: Investigate the Residuals

Check the Four-in-One Residual Plot
Residual Plots for Recognition
Normal Probability Plot of the Residuals Residuals Versus the Fitted Values
99
0.50
90
0.25
Residual
Percent
50 0.00
-0.25
10
-0.50
1
-0.50 -0.25 0.00 0.25 0.50 44 46 48
Residual Fitted Value
Histogram of the Residuals Residuals Versus the Order of the Data

6.0 0.50
4.5 0.25
Frequency
Residual
3.0 0.00
1.5 -0.25
-0.50
0.0
-0.4 -0.2 0.0 0.2 0.4 1 2 3 4 5 6 7 8
Residual Observation Order
Any problems here?

Step 9: Investigate the Interactions
q Select Stat>DOE>Factorial>Factorial Plots…
Setup the Main Effects Plots, too.

27
Minitab Output – Interactions Plots
Interaction Plot (data means) for Recognition

daily 3x/week
TV
48
2x/Week
3x/Week
46
TV
44
Radio
48 daily
3x/week
46
Radio
44
Press
48
weekend
daily
46
Press
44
2x/Week 3x/Week weekend daily
What is the interpretation?

Step 10: Investigate the Main Effects
Main Effects Plot (data means) for Recognition

TV Radio
47
Mean of Recognition 46
45
44
2x/Week 3x/Week daily 3x/week
Press
47
46
45
44
weekend daily
What is the interpretation?

29
Step 11: Mathematical Model

q From the Minitab analysis:
Term Effect Coef
Constant 45.6250
TV 3.2500 1.6250
Radio 2.7500 1.3750
TV*Radio 0.7500 0.3750
(Recognition) = 45.6 + 1.6 ´ TV + 1.4 ´ Radio + 0.4 ´ (TV * Radio)
The focus of Six Sigma: Y=f(x)!

The Mathematical (Empirical) Model
q What is the value for Recognition if all of the levels were set high?
Rec = 45.6 + 1.6 ´ 1 + 1.4 ´ 1 + 0.4 ´ (1 * 1) = 49

m Low?
Rec = 45.6 + 1.6 ´ -1 + 1.4 ´ -1 + 0.4 ´ ( -1 * -1) = 43
q Where should they be set to maximize Recognition?

m They should all be set high
q What is the value when all effects are set to zero?
m 45.6
q What does this value represent?
m The grand average of the experiment, or, if all effects are
linear, the expected value when the levels are at the
midpoints
31
Using the Optimzer

q Submit the final model for optimization
q Stat>DOE>Factorial>Response Optimizer
Optimizer Output
q The Optimizer reports optimal levels and response:
Optimal TV Radio
Hi 3x/Week 3x/week
D Cur 3x/Week 3x/week
Optimal Response 0.63333 Lo 2x/Week daily
Recognit
Maximum
y = 49.0
d = 0.63333
% to Target
Settings can be manipulated for fine tuning

34
Step 11: Translate into Practical Terms
Recognition = 45.6 + 1.6 ´ TV + 1.4 ´ Radio + 0.4 ´ (TV * Radio )

q Conclusions:
m On average, if the process parameters are set neither high
nor low, the mean Recognition be 45.6
m TV & Radio advertising are both important to recognition
will, the effects are not simply additive.
m The interaction suggests a synergistic effect between TV
and Radio advertising
q Recommendations:
m To maximize Recognition set TV & Radio advertising to the
high level. The estimate for Recognition at this operating
point is 49.
• Step 13: Replicate the optimum conditions. Plan the
next experiment and institutionalize the change.
Practical results are the goal of experimentation

CONTROL PHASE
Control Phase 146

Introduction to Control and Control Plans
(Green Belt)
© July 23, Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved. 1
Module Objectives
q Develop a procedure for the control phase

m Plan
m Execute
q Discuss Project Transition Action Plans

q Explore Control Plans
q Introduce 4th Generation Management Concepts
2
© July 23, Quality Growth Services. Unpublished proprietary work available only under license. All rights reserved.
Six Sigma Summary
q What did we do?

q We measured
m Clearly defined the
problem q We analyzed
m Found the critical m Examined the data
outputs & inputs
m Identified the root cause
m Determined capability of the problem
of the q We improved
measurement systems
and m Discovered y=f(x)
processes m Optimized the process
?
3
Control Phase Questions

q What can be done to make sure the defects stay reduced?
q What actions can be taken by the team to ensure the
improvement is permanent?
q What systems can be put in place to help the functional
team maintain the gains routinely?
q Do I have buy-in from the important players?
q With whom do I leave the project?
q When does my role in the project end?
4
Keys to Control
Metrology
6
Integrating the Keys

q A control plan integrates all of the keys to control.
Preventative Sample Size

Data Process
Location Characteristics
Maintenance & Freq.
Responsible Metrology
Persons Control Plan
Process
Decision
Process Process
Rules
Control
Documentation
Flow
7
Remember Control Plans?
Process What's Input or Spec. Limits / Meas. Control Sample Who/What Where Decision Rule/
Freq. SOP's
Step Controlled? Output? Requirements Method Method Size Measures? Recorded Corrective Action
Bake Plate Omega T/C PM Maintenance
Input 112 ± .3 °C 100% Weekly ASI Adjust to spec. 9785-123
Temp. Meter Calib. Technician
PMGI 3 sites/ 1 wafer/
Coat/ Output 195 ± 45 nm SP SPC Operator Oracle None 9785-369
Thickness wafer lot
Bake
PMGI 36 sites-
PMGI Output 14 ± xx mm KLA SPC 100% Operator Oracle Rework wafer 9785-411
Width all wafers
36 sites-
Undercut Output 2.0 ± 0.5 mm KLA SPC 100% Operator Oracle 9785-615
all wafers
q A control plan is:

m A summary of all of the control activities for the process.
m A method of identifying holes in the control system.
m A list of control activities yet to be implemented.
m An input to the FMEA (mature processes).
m An output of the FMEA (new processes).
8
Process Control Plans. What’s the Point?

q The intent of process control is:
m Run the process on target
m Minimize variation about the target
m Minimize required adjustments and over-control
m ALWAYS meet customer requirements
q The purpose of a control plan is:

m Institutionalize process improvements
m Highlight areas requiring extra education
m Provide one-stop-shopping for control information
9
Developing a Process Control Plan
Process What's Input or Spec. Limits / Meas. Control Sample Who/What Where Decision Rule/
Freq. SOP's
Step Controlled? Output? Requirements Method Method Size Measures? Recorded Corrective Action
Inputs to the Control Plan:

m Final process map m Maintenance schedules and
maintenance procedures
m FMEA
m Training materials
m Process documentation
m Troubleshooting guides
m Customer requirements
m Calibration schedules
m IT capabilities
m Operating windows
m MSA’s
m SOP’s
m Capability studies
m Out of control action plans
10
A Word About Decision Rules

q Critical elements to successful decision rules:
m Who makes the decision?
m What action does he take?
m When to take action? When to
elevate the decision?
m Where to act? To what extent?
m Where is the action recorded?
By whom is it recorded?
m Why is action necessary?
Use data!
m How effective was the action?
Follow up!
11
Control Plan Buy-in
q The most important part of the control plan:
The approval block.
Approved by:
Approved by:
Approved by:
q Buy-in is critical to the project and Green Belt’s success

m Enables the improvements to be institutionalized
m Provides a valuable demonstration to the process management
m Allows for extension to other processes
m Educates process engineers
m Allows the Green Belt to move on to another project.
12

Control Charts
(Green Belt)
Module Objectives
q To present a brief history of SPC

q To provide guidelines for using control charts
q To demonstrate X-bar & R chart
q To demonstrate p, np, c and u control charts

SPC – A Short History
• When was SPC developed?
By whom?
q Dr. Walter Shewhart of Bell Labs
developed a theory of Statistical
Process Control in 1924.
q During the ’20’s Dr. Shewhart
presented his theories in a series of
lectures that were published in a
book , Economic Control of Quality
Manufactured Product (1931).
q SPC came into wide use during the
1940’s as a result of war production
efforts.
Dr. Shewhart’s Ideas
q All processes display variation.

q There are two components of variation:
¦ A steady component that is inherent to the process that
Dr. Shewhart attributed to chance and undiscoverable causes,
currently called “common causes”
¦ An intermittent component attributed to assignable

causes that could be economically discovered and removed.

The Normal Curve
-3s -2s -1s x +s +2s +3s

68.26
95.46
99.73
Control Charts
• Transformation of a normal distribution curve in the form of 3 parallel lines, where
• The middle line indicates mean and called central line (C.L.)
• The upper line indicates Mean +3 Sigma and called upper control limit (UCL)
• The lower line indicates Mean -3 Sigma and called lower control limit (LCL)
U.C.L.
+3s
C.L.
Mean
-3s
L.C.L.

Steps for Constructing an X-bar/R Chart
q Gather the Data
¦ Select subgroup size
¦ Select the frequency of collecting the subgroup data
¦ Record the raw data
¦ Calculate the average and range of each subgroup
¦ Plot the subgroup averages and ranges on the control charts
q Calculate the Control Limits

¦ Calculate the average range (R-bar) and the Process average (X-bar-bar)
¦ Calculate the upper and lower control limits for the average and range
charts
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
1 5.10 5.27 5.34 5.34 4.10 4.92 5.05 5.37 4.94 4.66 4.16 4.43 4.95 4.73 5.09 4.99 5.90 4.45
2 4.77 5.40 5.17 5.32 4.78 4.69 4.57 4.22 4.88 5.55 4.63 5.69 4.70 4.74 5.20 4.47 5.49 5.95
3 4.32 5.34 5.47 5.81 4.73 5.81 4.58 6.14 5.29 5.58 4.51 4.64 5.11 5.08 4.46 5.09 5.57 5.90
4 4.43 5.71 4.78 5.11 5.12 4.26 4.60 5.20 5.06 5.23 5.34 4.19 4.87 5.01 4.45 5.10 5.55 4.21
5 4.80 5.60 5.36 5.60 5.62 4.95 5.03 5.50 5.23 5.22 4.83 5.16 5.28 4.82 4.70 4.75 5.60 5.73
X-Bar 4.69 5.46 5.22 5.44 4.87 4.92 4.76 5.28 5.08 5.25 4.69 4.82 4.98 4.88 4.78 4.88 5.62 5.25
R 0.78 0.44 0.69 0.70 1.52 1.55 0.47 1.92 0.41 0.92 1.19 1.50 0.58 0.36 0.74 0.63 0.41 1.74

X-bar & R Chart – Minitab Output
Xbar/R Chart for Sample 1-Sample 5
1
3.0SL=5.583
5.5 5
Sample Mean
5.0 X=5.029
4.5 -3.0SL=4.475
Subgroup 0 5 10 15 20 25
2 3.0SL=2.030
Sample Range
1 R=0.9602
0 -3.0SL=0.00E+00
Step 1: Gather the Data

q Select subgroup size
q Select the frequency of collecting the subgroup data
q Record the raw data
q Calculate the average and range of each subgroup
q The following data is an example (or subset) of pin length data from
the file CCVAR.mtw which contains 25 subgroups of 5 samples
each. Subgroup Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Mean Range
1 5.009 4.829 5.644 4.689 5.660 5.166 0.971
2 4.274 4.739 5.278 4.267 5.398 4.791 1.131
3 4.878 4.978 4.867 4.764 5.878 5.073 1.114
4 4.887 5.205 5.052 5.899 5.208 5.250 1.012
5 4.280 4.241 4.619 5.414 5.639 4.839 1.398
6 4.660 5.030 4.596 4.451 4.500 4.647 0.578
7 4.544 5.105 4.666 5.795 5.266 5.075 1.251
8 5.104 4.774 4.321 4.435 4.802 4.687 0.784
9 5.266 5.400 5.340 5.710 5.600 5.463 0.444
10 5.340 5.168 5.469 4.782 5.365 5.225 0.687

Example – Constructing Charts in Minitab
CCVar.mtw
q In Minitab, open sample worksheet CCVar.mtw
q From the menu select Stat > Control Charts > Xbar-R…
q Select data
columns
C1-C5
Example: Control Chart Tests

q Press the Tests… Button
q Select the tests to run. Tests 1 & 5 are best when
searching for gross special causes.
q Press the Help
button for
explanations of
each of the tests

X-bar & R Chart – Minitab Output
Xbar/R Chart for Sample 1-Sample 5
1
3.0SL=5.583
5.5 5
Sample Mean
5.0 X=5.029
4.5 -3.0SL=4.475
Subgroup 0 5 10 15 20 25
2 3.0SL=2.030
Sample Range
1 R=0.9602
0 -3.0SL=0.00E+00
A Few Words on Control Limits

q Basis for control limits.
¦ Control limits are also referred to as natural process limits.
¦ Control limits are based on the mean and standard deviation of
the process as it is, not how we wish it to be.
There is no connection to specification limits!
¦ The limits in the previous example are calculated by using the
average range data to estimate sigma.
q Purposes of control limits.
¦ To determine when local action is necessary and to act on it
(i.e., when special cause variation exists).
¦ To display the natural variation of the process, that is, the
common cause variation.
¦ To show any obvious, unnatural patterns in the data.

X-bar & R Chart Limit Calculations
R = ( R1 + R 2 + ........Rk ) / k
X 1 + X 2 + ... + X k
X= = 50.24
k
UCL R = D 4R n 2 3 4 5 6 7 8 9 10
LCL R = D 3R D4 3.27 2.57 2.28 2.11 2.00 1.92 1.86 1.82 1.78
UCL X = X + A2R D3 * * * * * 0.08 0.14 0.18 0.22
A2 1.88 1.02 0.73 0.58 0.48 0.42 0.37 0.34 0.31
LCL X = X - A2R
d2 1.13 1.69 2.06 2.33 2.53 2.7 2.85 2.97 3.08
s X = R/d2
Interpreting Control Charts

q Control charts are powerful tools for detecting
special cause variation. If a special cause is found:
1. If the out of control condition is part of a trend, mark
the beginning and end of the trend
2. Analyze the process to determine the cause of the
OOC condition
3. Correct the condition and prevent it from recurring
4. Recalculate the control limits omitting the OOC
values from the calculations
5. Re-evaluate the chart according to the new limits
q Control limits are only recalculated when the causes
for the OOC conditions have been identified and
corrected

Interpreting Range Charts
q Range charts are interpreted first
q In Range charts, a single OOC point is often caused by
miscalculation, mis-plotting, or mis-measurement. Verify the
accuracy of the data first.
q Trends in the Range data can be caused by:
¦ Real changes in the variability of the process
¦ Changes in the measurement system
q Identify the cause of the OOC and recalculate the control limits,
excluding the OOC point
q If a point is excluded for the range calculations, exclude it from the
X-bar calculations, also.
?
Taking Action – Special Causes
q Taking action is the point of control charts.

Otherwise they are just a waste of time and money
¦ If actions are required by the data and never taken, the charts are just feel-
good wallpaper. “Look, we have SPC!”
q Actions are based on increasing process knowledge
¦ Identify the root cause of the variation and fix it
¦ Document on the chart what the problem was and the resulting fix
¦ Document the problems and actions into a troubleshooting guide so
actions can be taken locally
q Recalculate the control limits only if the special causes have been
identified and eliminated

Taking Action – Common Causes
q Compare the natural process limits to the
specification limits.
¦ Determine the process capability from the control limits
and compare to the specification limits. Is the process
capable?
q Reducing the natural process variation is the point of

most Six Sigma projects. In the Control phase, the
common cause variation should be well within the
specification limits.
q Don’t recalculate the control limits unless a known
process change explains a short term shift
?
Summary – Average and Range Charts

q Used mostly in machine-dominated processes
q Used when data can be sub grouped rationally
q Powerful tool for differentiating between special cause
and common cause variation
¦ Its power comes from the power of variable data vs.
attribute data. Remember sample size calculations?
q Easily indicates trends, process shifts, and other non-

random variation
q Enables a quick estimate of process capability

What If Subgrouping Is Irrational?
q Individuals (I & MR) Charts
¦ For use when there is no rational basis for subgrouping or
when the measurements are expensive and/or destructive.
¦ MR stands for Moving Range. The moving range is
calculated by recording the differences between successive
pairs of measurements.
¦ Control limits are calculated as follows:
UCL MR = D 4R n 2 3 4 5 6 7 8 9 10
LCL MR = D 3R D4 3.27 2.57 2.28 2.11 2.00 1.92 1.86 1.82 1.78
UCL X = X + E 2R D3 * * * * * 0.08 0.14 0.18 0.22
E2 2.66 1.77 1.46 1.29 1.18 1.11 1.05 1.01 0.98
LCL X = X - E 2R
d2 1.13 1.69 2.06 2.33 2.53 2.7 2.85 2.97 3.08
R
s X =
d2
Practice & Homework – I & MR charts

q Open Minitab data file Exh_qc.mtw.
q Construct an I & MR chart from the Weight column.
q What are your conclusions?
q Turn on all the tests.
q What are your conclusions now?
q For Homework, repeat the analysis on the Length
column.

Control Charts for Attribute Data
q Defective Data q Defect Data

¦ Package does/does not leak ¦ Missing parts in a kit
¦ Lamp does/does not light ¦ Missed contract obligations
¦ Go/no-go gauge data ¦ Errors on an invoice
¦ Correct/Incorrect Shipment ¦ Missed customer calls
q p- and np-charts q c- & u-charts

¦ p-charts for proportion non- ¦ u-chart for the number of defects per
conforming units, sample size not inspection unit, sample size not
necessarily constant necessarily constant
¦ np-charts for proportion non- ¦ c-chart for the number of defects per
conforming units, sample size is inspection unit, sample size constant
constant
p-Chart for Proportion Nonconforming

q When to use a p-chart
¦ When it is difficult or uneconomical to make a numerical measurement
¦ When it is desired to combine different types of defects into an overall
proportion
¦ When the available data are for attributes
¦ When the data come from a binomial process
q Management summaries
¦ Many management summaries are attribute forms and could benefit from
control chart analysis.
¦ Examples: scrap rates, quality audits, first-run yields, etc.
Would you like to teach management to interpret trends on their charts?


Calculating Control Limits For a p-Chart
q For each subgroup

¦ Record: n – the number of units inspected per set
k – the number of sample sets
r – the number of nonconforming units per sample
¦ Calculate p – ^
the proportion non conforming:
^ r
p=
n
¦ Calculate the control limits: 3 p (1 - p )

UCLp = p +
n
r 1 + r 2... + r k
p=
n 1 + n 2 + ... + n k LCLp = p -
3 p (1 - p )
n
Notes on p-Chart Control Limits

q LCL
¦ When the mean proportion defective is small, the LCL can be a negative number.
In this case there is no LCL and p=0 is still within the control limits.
¦ It is advantageous to pick a subgroup size that establishes a lower control limit.
This way improvements can also be detected.
q Subgroup size
¦ The previous formulae assume n is constant. As an approximation, if the
subgroups are ± 25% of the mean then substitute n-bar, the mean subgroup size
for n in the control limit formulae.
¦ To minimize confusion it is best to have constant sample sizes whenever practical.
¦ Minitab calculates p-chart control limits for each subgroup based on that
subgroup’s size.

Constructing p Charts in Minitab Pat.mtw
q Open Minitab file pat.mtw

q From the menu select Stat>Control Charts>Attributes Charts>P…
q Enter Passed
for Variables:
q Enter Cases
for Subgroups Sizes:
Interpreting p-Charts
q Attribute charts have fewer tests for special causes due to the
lower information content of attribute data.
q A point outside the 3s limit is assumed to be from special cause
variation.
q When np<5, the
likelihood of runs
below p-bar
increases.
Take care when
interpreting runs
in these cases.
Unless the subgroups are equal sizes, only the first test is used.

Minitab p-Chart Output
P Chart of Passed
1
0.40 1
UCL=0.3711
0.35
Proportion
0.30 _
P=0.2928
0.25
LCL=0.2145
0.20 1
4 8 12 16 20 24 28 32 36
Sample
Tests performed with unequal sample sizes
Charts For Defects – c-Charts and u-Charts
q When to use c- or u-charts

¦ When the data are attribute data of defects, not defectives
¦ When the data come from a Poisson process
¦ c-charts are charts constructed for number of occurrences for a constant exposure
¦ u-charts are constructed for rate of occurrence for either constant or varying
exposure
q Examples
¦ Each month, 100 invoices are audited and the total number of mistakes is
recorded.
¦ In a molding process, there is a problem with pinholes in plastic bottles. Each day,
a number of bottles are examined and the number of pinholes are recorded.
¦ Each month, the number of accidents that occur in an organization are recorded.

Calculate c-Chart Control Limits
q For each subgroup
¦ Record: c – the number of defects found
q Calculate c – the mean number of defects:

_
c 1 + c 2 + ... + ck
c=
k
q Calculate the control limits:
UCLc = c + 3 c
LCLc = c - 3 c
Constructing c-Charts in Minitab cfu.mtw
q From the menu select Stat>Control Charts>Attributes

Charts>C…
q In Variable: select cfu
q Click the OK tab

u-Charts
q When
¦ When the data are from Poisson process
¦ When the data are rates of occurrence with different sample
sizes
q Calculate Control Limits

¦ Calculate the process
average nonconformities: c 1 + c 2 + ... + ck
u=
n1 + n 2 + ... + n k
u u
UCLu = u + 3 =u+3
n n
¦ Calculate the control limits:
u u
LCLu = u - 3 =u-3
n n
Charts for Defects – Summary
q Charting defects is one step closer to the Six Sigma ideal, but these
charts still focus on outputs, not inputs.
q c- and u-charts are also charts of attribute data and suffer the same
shortcomings as p- and np-charts.
q It is easier to calculate c-charts than u-charts due to the constant
sample size.
q When the sample sizes are large enough and the number of defects is
large, attribute data behaves like variable data from a normal
distribution.

Closing Your Project
Green Belt Process Owner
Responsibility
Time
Module Objectives
• To discuss project transitions
– To introduce Project Transition Action Plans
• To discuss Realization
• To provide a template for Six Sigma Project Final Reports
• To provide guidance for managing multiple projects
2
Let’s Talk About Project Closure
• What is a closed project?
– What does it look like?
– Who is in charge?
– Who owns the results?
• Who tracks the results of a closed project?

– Process Metrics?
– Financial Savings?
• What is the Green Belt role in a closed project?
3
Project Transition – Responsibility Flow

• Green Belts must work their way out of the project.
• Responsibility shifts from the Green Belt to the Process
Owner.
• Start early.
• Have the project
team do most of the
work. Resist the Green Belt Process Owner
Responsibility
temptation to do it
all yourself.
• Complete a Project
Transition Action Plan.
Time
4
Project Transition – Keys to Success
• Get buy-in early – no surprises!
– Project team must include all the important players.
– Regular reports to process management
• Establish systems for frequent measurements.
– Process control and product monitoring
• Update the procedures and documentation.
– Document any remaining parts of the hidden factory.
• Take the time to truly educate (not just train or retrain) everyone.
– New procedures and new measurements require education.
• Monitor the results.
– Frequent measures of key project metrics
• Complete a project transition action plan.
5
Project Transition Action Plan

• Make a plan to complete your project.
– What tasks are required to complete
the project?
– Who is responsible for each task?
– When will each task be complete?
– Who owns the plan? -- Not the Green Belt!
– Who drives the plan to completion? – Not the Green Belt!
• The plan is not complete until everyone agrees to it
– Everyone who has an action item assigned must agree
– Process owners agree that the plan completes the project.
6
Project Transition Action Plan Template
Project Title:
Transition Ow ner:
Plan Date:
Action Scheduled Actual

Actions Required for Completion Ow ner Initials Com pletion Com pletion
1
2
3
4
5
Signatures:
Transition Owner
Process Owner
Master Black Belt
Black Belt
7
D à M à A à I à C à Realization
• When is Realization?
– After the project is closed!
• What is Realization?
– The twelve months after project closure
– When the profits are counted (realized)
• Who owns the project during realization?

– The Process Owner for Process Metrics
– The Finance Rep for accumulating savings
8
The Green Belt’s Swan Song
9
A Few Suggestions About Final Reports

• Start early.
– Write the Final Report as much as possible with each phase
completion.
– Projects don’t have to have every item completed before you write
the Final Report.
• Use a Project Transition Action Plan.
• Include the PTAP in the Appendix of the Final Report.
• Use PowerPoint or the Minitab ReportPad as a container

for important analyses.
– Easy pasting into the Final Report
• Get all the signatures and initials!
– Process Owner, Champion, MBB
– Anyone else who as an outstanding action
10
Final Report – Major Elements
• Executive Summary
• Project Definition
• Project Management Details
• Summary of changes made to affect improvement
• Outstanding actions required to obtain improvement
• Phase Conclusions
• Phase Deliverables
• Control Methods Applied
• Control Plan
• Key Six Sigma Elements
• Appendices
11

Let us create
the journey of skill
and success
for the future
of this nation
Feel free to visit or call
Head Office: H-13, Kirti Nagar, New Delhi -110015, India

Tele-Fax: 0091-11-25431737, 25918332, 41425273
www.iipeindia.com | www.qgspl.com | www.satisfyindia.org

info@iipeindia.com | qgs@qgspl.com | satisfyindia@gmail.com
This proposal is solely meant for M/S Zetwerk Manufacturing Businesses Pvt Ltd. It must not be modified, edited,
distributed, published or claimed for other purposes without the written consent from Quality Growth Services.
Quality Growth Services | H-13, Kirti Nagar, New Delhi -110015 | www.qgspl.com

LSSGB Course Handout

Uploaded by

Document Information

Original Description:

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

LSSGB Course Handout

Uploaded by

Copyright:

Available Formats

Certified Green Belt

Lean Six Sigma

This programme is conducted by

Dr. Sumit Shandilya

Focus on your biggest/ chronic problems . . ?

Assign it to most competent person. . . ?

Support with all the tools/ resources. . . ?

And guarantee uninterrupted time. . . ?

Reduce Variation Focus on Target

Six Sigma Philosophy

Six Sigma Level Performance

Sigma Level Defect % Yield

Gap? Six Sigma Project

Practical Problem Real Problem

Practical Solution Real Solution

PFD, VSM, MSA

C-E Matrix, Box plot, Main effect plot,Interval

Plot Interaction Plot, Multivari Analysis,

DOE, Piloting , Validation

Control Control Xs PFMEA, Control Plan, Error proofing, SPC

Best performing similar process within or

Historical best performance of the process:

Current Performance Level of Process:

1983 Started by Motorola Bob Galvin, CEO

1988 Formation of Six sigma consortium

1989-90 Tried by IBM/ DEC Initiative failed

1993 Structured by Allied signal- Black/ Larry Bossidy

1995 Adopted & Structured by GE Jack Welch- The

1999 ASQ started training

Champion §Maintain strategy and projects link 1-3 days

§Recognize and define projects

§Removes roadblocks ,Review project activity

Master BB §Instructor of BB/GB, Training material developer BB + Addl trg

Green Belt §Lead/ support Project team/Practitioner of DMAIC 1 week

Process Owner §Owner of solution/ Pivotal team member 1 day

• Ensure the right priorities

• Clarify the issue/opportunity

• Define the process and customer(s)

• Set up the project for success

Selecting & Prioritizing Efforts

• Strategic plan & goals

match physical counts

What leaders do (or should):

What Green & Black Belts & teams do (or should):

• Capture ideas at the

• Keeps team from getting Initial Solution Ideas (Parking Lot)

• Helps ensure good ideas • Eliminate non-essential data from

• Set-up customers with a pre-paid

• Train all staff in current

The Project Charter

Location of Process: <<region, site, department, etc.>>

Business Case Estimated Benefits

• Concise description of project’s like or

• Impact (benefit/relevance) of potential improvement

• Impact (cost or risk) of not improving

• Should be verified/accepted by leadership

Should Include: Should Avoid:

Good: Not Good:

• (What ) (Where) (When) (How Much) (How do you know)

• Is performance not improving?

• Is some target not being met ?

• Is someone else doing better?