You are on page 1of 105

Six Sigma

One credit course

TCE, Madurai

VV 1
Six Sigma
Single most effective problem solving methodology for business
and organizational performance.
Worlds top corporations have used this to increase their profits
Six Sigma proficiency in your resume is a pre-requisite to move into
a position in management.
As the methods and tools of six sigma have spread, it has become
easier to understand and more straight forward to implement.
The mysteries of Six Sigma have been revealed.
Simply stated, it is about applying structured, scientific method to
improve any aspect of business, organization, process or person.
Six sigma is for those who want to improve.
Do you want to improve?
VV 2
Good to Great
Goodness is not a function of larger than life leaders
What makes it great?
Disciplined people , Disciplined thoughts, Disciplined actions over extended
period of time
From the Industrial age to Information age, a paradigm shift has
come from leadership by position to leadership is a choice
In the new paradigm the greatest asset is the people in the organization with
their bodies, minds, hearts and consciousness all engaged and contributing
A trim tab in an aircraft that with the small motion allows bigger rudder to
achieve great effect
The leader acts as Trim tabs and their small actions can effect greater changes
In a knowledge economy, where 70 to 80% of value added to goods and
services comes from the knowledge work, can we imagine the results flowing
from having the entire people in the organization Six Sigma literate?

VV 3
Six Sigma
It is about engaging:
Disciplined data collection and analysis
Best possible ways of meeting customer need
Applying comprehensive, actionable description of methods and tools
Minimizing wasted resources
Maximizing profit

VV 4
Skill Level and Belts
The skill level needed to solve complex problems need varying skills in
The certification in Six Sigma is accordingly classified as below.

Skill Level Problem Type of Belt

High level of Highly Black Belt
statistical expertise complex
Medium level of Moderate Green Belt
skill complexity
Everyday routine Less Yellow Belt
work Complexity
Master Black Belts are the Trainers

VV 5
Six Sigma


VV 6
Purpose and Objective
To enable students to apply Six Sigma concepts in any of
the activities that they want to improve

At the end of the course, students will:
Be able to explain and appreciate the Six Sigma approach
Understand the application of key Six Sigma tools
Practice this for any continuous improvement
Use data and take decisions based on data

VV 7

Why Six Sigma

What is Six Sigma and Basic Statistics
DMAIC methodology in detail
Application of Six Sigma tools
Mini project discussion (Ref guide for Project)

VV 8
Code of Conduct

Active participation
Positive attitude
Be on time
Cell phones off

VV 9

List of expectations from participants

VV 10
Roles and Responsibilities
- From Faculty
Teach and guide
Help in project and additional support, if any needed
Active listening
Team work in exercises
Ask relevant questions when in doubt
Do a project using six sigma techniques
Time keeper
Any volunteers?

VV 11
Why Six Sigma?
More challenging customer expectations on
Increasing global competition
Ever growing market needs
Reduced Cycle time coupled with lesser costs
Complexity of problems
Identifying and eliminating problems
Better prediction of outputs (Big Y)
VV 12
What is Six Sigma?
Understanding and reducing variation
Identifying Critical to Quality
Shortly as CTQ
Data driven approach
All with data
Predictive design (Not the Test-in-Quality)
Iterative approach using
A structured and well disciplined methodology
A set of tools and techniques
VV 13
The select, few, measurable characteristics that
are key to customer satisfaction
Failure to meet the CTQ will be a defect
Six Sigma measures and improves our ability to
deliver CTQs with minimum variation
On-time delivery
Complete in all respects
Accuracy and reliability
Time to repair
Timeliness of doing things
VV 14
Design in Quality

Reactive Design Predictive Design

Transition to ...
Quality Quality


Evolving Requirements CTQ Flow down

Design rework/tweaking Control Critical Design Parameters

Modeling/Simulation with
Build & Test Design/Process Capability Flow up

Measurement Statistical Quality Prediction

Test in Quality Design in Quality

VV 15
Six Sigma some definitions
Sigma it is the distribution about the mean - Average squared
deviation of each data point from the mean.
The Standard Deviation ( ) is the square root of variance a
common measure for variation
Sigma value - indicates how often defects are likely to occur and a
high value indicates high rate of success
Sigma level - the number of standard deviations between the center
of the process and the closest specification limit
Six Sigma a proven method and a set of tools, control techniques,
reporting methods and management techniques that combine to form
break through improvements

VV 16
Six Sigma - some definitions
Six Sigma performance
Statistical term for a process that produces less than 3.4 defect per million
Six Sigma improvement
The key outcome of a business or work processes are improved dramatically
often by 70 % or more
Six Sigma deployment
The prescriptive rollout of six sigma methodology across the organization
with assigned practices, roles and procedures as per accepted standards
Six Sigma organization
Uses six sigma methods and tools to improve performance, customer delight
Lower costs, grow revenue, lower cycle time, reduce complexity
Increase capacity and capability
Minimize defects and errors
VV 17
What is Six Sigma?


VV 18
Six Sigma Significance

DPMO % Defects
1 691462 69
2 308537 31
3 66807 6.7
4 6210 0.62
5 233 0.023
6 3.4 0.00034
7 0.019 0.0000019
Six Sigma is 3.4 Defects in Million Opportunities
VV 19
Basic Statistics

VV 20
Basic Statistics
Statistics is the science of collecting, analyzing, interpreting and presenting data.
Central tendency refers to the middle point of a distribution
Central tendency helps to state one value that would best capture and communicate the
The mean is a simple arithmetic average
How to compute Mean:
Add all the data points and divide by the total number N
M= X1+X2+.Xn / N
M=X/N Where indicates the summation of X1.Xn

The median is the point that divides the distribution into two equal parts such that an exactly equal
number of scores fall above and below the point (50 th percentile)
How to compute Median:
Median = the( (n+1)/2) th item in the a data array (Midpoint value in a data set arranged from the lowest to highest)
the data set contains an odd number of values, the median is the middle value
If the data set contains an even number of values, the median is the average value of the two
middle values - where n= Number of items in the data array

The mode is the most frequently occurring value VV 21
Basic Statistics
Population is the total set of units about which one would like to get information.

Sample is a subset of the population, usually much smaller and drawn randomly
from the population.

Inference is the process of looking at characteristics of the sample data values like
mean or variation, to make conclusions about the likely values of the
corresponding characteristics in the population.

X i
Population mean i 1 ( N is population size)

X i
Sample mean X i 1 (n is sample size)

VV 22
Basic Statistics
The measures of variability:
The Range
Variance (already explained earlier)
Standard Deviation

Range is the difference between the highest and lowest observed values
Range (R) = Value of highest observation - Value of lowest observation
Range is easy to understand and to find
Its usefulness as measure of dispersion is limited because it considers only the
highest and lowest values and ignores the variation among of other
The range is more sensitive to outliers than the variance.

VV 23
Basic Statistics
The Variance ( ) as already explained is the Average Squared
Deviation of each data point from the Mean.
The Standard Deviation ( ) is the Square Root of the Variance.
The most common and useful measure of variation is the
standard deviation.
The Variance for a sum or difference of two variables is found by
adding both variances

Var ( X ) i 1
n 1
VV 24
Basic Statistics
Normal Curve

A probability distribution where the most frequently occurring value is in the middle and
other probabilities tail off symmetrically in both directions.
Bell shaped and characterized by mean (measure of central tendency) and standard
deviation (measure of dispersion or variability)
Area under the curve is 1 which is the probability of all outcomes
Curve theoretically does not reach zero.
Curve can be divided in half with equal pieces falling either side of the most frequently occurring value.
A normal curve indicates random or chance variation.
The peak of the curve represents the center of the process.
The area under the curve represents virtually 100% of the process is capable of producing
VV 25
Arriving at Six Sigma
Identify the CTQs Define Defect Look for Defects
Critical to Quality
Opportunities in Products or
Characteristics or Any step in the Services
the Customer process where a Count Defects
Requirements for a Defect could occur failures to meet CTQ
Product or Service in a CTQ in all process steps

Defects per
Million of
Opportunity Convert DPMO to Arrive at DPMO
Defects Per Million
Use the SIGMA
2 308,537 TABLE
3 66,807
4 6,210 Measuring & Eliminating
5 233 Defects is the Core of Six Sigma
6 3.4
VV 26
Is it

VV 27
Six Sigma Methodologies
DMAIC Approach

Define Measure Analyze Improve Control

DMADV Approach (DFSS)

Define Measure Analyze Design Verify

VV 28
Six Sigma Methodologies - DMAIC
DMAIC - Emphasis on new product or process development for
delivering robust products

Define: Identify the areas for improvement and interpret the customer needs
into Critical to Quality (CTQs) parameters
Measure: Collect data and measure the current Product / Process
Analyze: Identify the areas that affect the quality or contribute to the
problem and analyze the root cause of the problem
Improve: Evaluate and implement a consistent solution for continuous
Control: Sustain the improvements by continuously monitoring through a
control plan
VV 29
Six Sigma Methodologies - DFSS
DMADV (DFSS) - Emphasis on new product development
and delivering robust products

Define: Identify product/process performance , Critical to Quality items

Measure: Collect data and Critical To Quality (CTQ) flow down to
Subsystems and components
Analyze: Identify the areas that affect the quality and the design for Xs
(DFx) parameters . Analyze various options and do a high level design
Design: Design a robust design
Verify: Ensure that the design meets the performance and customer needs

VV 30
Customer driven design Drives quality measurement
of processes with 6 and predictability
capability. improvement early in design
Predicting design quality
up front. Utilizes process capabilities
to make final design decisions.
Top down requirements
flow down (CTQ flow Monitors process variances
down). to verify 6 customer
requirements are met.
Cross functional
integrated design

VV 31
The Basic Model/Transfer Function
Magic Equation

Transfer Function


x2 Y=f(x1,x2...)
Big Y
Y = f(X)
Improving the output (Y) is of interest
Y Could be: Identify all xs and identify the critical ones
Customer Satisfaction Understand how Y is related to xs
Productivity improvement Y is a dependent variable and x is the independent
Defect Reduction
Variable. Thus Y is the effect and x is the cause
Any activity that needs Improvement

VV 32
Define Measure Analyze Improve Control
Project Data Data Analysis To-be- Control
Charter Collection using Statistical Process Map Plan
(Goals, Roles) tools Score tools Modified Control
Voice of Card Graphical FMEA Mechanism
Customer Current (as- Analysis using Mistake
(VOC), QFD is) Process Dot plots, Proofing
Thought Map Box plots, Voice of
Process Map Voice of the Scatter plots, Process
SIPOC Process SPC and
(VOP) Control Charts,
Measurement Histograms,
System Cause & Effect
Evaluation Analysis,
FMEA, Pareto
Charts etc
VV 33
Define Phase Project Charter
Defining what needs to be improved
Define the Problem statement through a project charter
Charter, Goals, Roles:
Problem Statement (Gap, Opportunity)
what, when, where, how big, impact, benefits
Goal Statement
what is to be accomplished
Team members and Responsibilities
who will be involved , team members, support people, coach, sponsor
Assumptions and Constraints
Limitations (time, support)
Preliminary Project Plan
DMAIC start and completion time

VV 34
Define Phase VOC
A strategy and system for continuously tracking and
updating customer requirements
Customer needs and data gathering technologies
Customer surveys and interviews
Customer behavior
Customer scorecards
Quality Function Deployment (QFD)
Market research/changes
Competitor assessment etc

VV 35
Define Phase VOC
What do people actually want?
Not what technology can we add to it?
Complexity leads to unreliability and makes use non-intuitive.
Customer attributes: what they want
Grouped into similar concerns
Preserve customers words and phrases
Includes concerns of retailers, regulators
Weight relative importance of concerns
Each gets a %, total sums to 100%
Separate input from each market segment
Will it create competitive advantage?

VV 36
Define Phase QFD
Quality Function Deployment
It is used to link the VOC directly to link the internal processes
QFD Matrices are graphical displays of the results of planning process
This is a system that methodically move customer requirements to
specifications for the process and ultimately into critical to quality (CTQ)
characteristics of the product and process
Also helps in prioritizing and decision making process
QFD is developed through a series of Matrices
In its simplest form, it involves a matrix that presents customer requirements
as rows and product or service features as columns
The cell where the row and column intersect, shows the correlation between
individual customer requirement and product requirement
When this is enhanced, by correlation of the columns with one another, the
result is called the House of Quality

VV 37
Define Phase QFD

House of Quality 5. Tradeoffs

3. Product
2. Importance characteristics

1. Customer 4. Relationship 6. Benchmarks

requirements matrix

7. Technical assessment &

target values
VV 38
QFD - House of Quality
Strong Positive Correlation
Positive Matrix
: Strong Negative Design req Target Direction

Like yield,
More is better

Importance rating

Less is better
Customer Req A specific amount

whats (rows)
Like low cost,
Small size etc Relationships
Strong 9

Medium 3

Weak 1
Target values

QFD is used for translating the needs and wishes of the customer into design requirements for products and services.
Quality : Meeting Customer Expectations
Function : The means to achieve the Quality
Deployment : The implementation of the means
VV 39
Define Phase TMAP
Thought Map (TMAP) is to capture, organize and develop the
teams thoughts that leads to the goal.

Elements in a TMAP
Questions asked - What we know and dont know
Tools used
Data collected
Knowledge obtained
Decisions made
Barriers encountered

VV 40
Define Phase TMAP
Benefits of TMAP
It provides an overview of the steps required to complete a six sigma project
It provides an efficient method of communicating the progress of the project
It helps to remember what assumptions were made, what was done, and the results that led to
where are we today in the project
It documents questions and answers for review by others later on
It documents the current level of thinking on a problem
It promotes critical thoughts while working on the project

TMAP is prepared in the Define phase and is used throughout every

phase of the DMAIC roadmap

TMAP Is a Living Document

VV 41
Define Phase TMAP
Some guidelines to consider when creating TMAP
Use Post-It notes
Involve all team members brain storm
Identify actions and people responsible for it
The following is to be avoided when doing a TMAP:
Critiquing thoughts behind project
Using a cookbook method of solving a problem
Blindly applying predetermined solutions
Acting on hunches not verified (but include them)
Knowing the answer before breaking down problem
Skipping steps or analysis
Jumping to conclusions
Questions could be:
What we know and we dont know, who are the customers and how will they benefits
what are the causes of the problems etc
VV 42
Define Phase TMAP
Tabular Format

VV 43
Define Phase TMAP
1st Level Thoughts


Vacation Planning a
from family Stay &
work vacation accommodation

VV 44
Define Phase - TPM
2nd level
2nd Level Thoughts
1st level 2nd level
1st level Thoughts Thoughts
2nd Level Problem 3rd level
Thoughts Statement Thoughts
1st level 1st level
Thoughts Thoughts 4th Level
2nd Level
Thoughts 2nd level
t 3rd Level 4th level 3rd level
Thoughts Thoughts Thoughts
VV 45
Define Phase - SIPOC
Acronym for Supplier, Input, Process, Output and Customer.
SIPOC helps to identify competing customer requirements, identify gaps in
requirements and the associated process outputs or inputs responsible.
Helps in building the first controlled and organized view of work and helps to
provide the link between outputs and customer requirements
SIPOC is built from inside out, starting from the center, with the process
Identify the process and define its scope and boundary
Identify the outputs (what will be produced by the process)
Define the customers (name, title, organizational entity)
This also defines customer requirement
Define the inputs to the process (HR, materials, information)
Identify the sources of inputs

VV 46
Define Phase - SIPOC
Term Definition Actions
Process A Single Box description of the series Identify the Process - What is the definition
of activities that transform inputs into of the process?
Process The starting and ending points of the Identify the boundaries - Where should a
Boundaries process identified in the single box start and a stop boundary be placed?
Outputs The products or services delivered by the Identify the process outputs - What outputs
process that should satisfy the customer should be included?
expectations and meet or exceed the
customer requirements
Customers The people who receive the outputs and/or Identify the process customers - How to
those who put requirements on the outputs identify who are the customers?
Customer The quantifiable expectations that the Identify customer expectations of outputs -
Requirements customer has for the process outputs What are the customer requirements?
Process The quantifiable expectations that the Identify the inputs - What is the definition
Requirements process has for the suppliers inputs of an input
Suppliers Those who provide the necessary inputs to Identify the suppliers of inputs - What is the
the process definition of a supplier?
Process output The metrics used to determine how well Identify the processs expectations of the
measurements customer requirements are being met inputs - What are input requirements and
VV who specifies that? 47
Supplier Input Process Output Customer
Start 2

6 5 3 4
End 2

Ensure how outputs relate to customer requirements

The input requirements are quantifiable expectations the process puts on the inputs

VV 48
Importance of Measure
When you can measure what you are speaking about,
and express it in numbers, you know something about
it; but when you cannot express it in numbers, your
knowledge is of meager and unsatisfactory kind. It may
be the beginning of the knowledge, but you have
scarcely in your thoughts, advanced to the state of
Lord Kelvin

What gets measured gets Managed

VV 49
Measure Phase
Data For Measurement
Identify what to measure and stratification
Layers of data
Criteria for selecting measures
Accuracy, Repeatability & Reproducibility, Stability
Availability, Lead time
Develop operational definition
Identify data sources
Prepare collection plan
Efficiency and Effectiveness of Measures
Refine measures
Develop Baseline Measures and Identify Improvement opportunities
Defect Measures (Concept of DPO and DPMO)
VV 50
Measure Phase
Types of Data
Quantitative data (Variable data)
Continuous data: temperature, Diameter, Sales
Discrete or Count data: number of defects, employees hired
Qualitative data (Attribute data)
Examples: Good/Bad, Satisfied/Not satisfied, Go/No Go

The tools used for plotting and analysis are different for different types of data

VV 51
Measure Phase
Types of Variable Data




Pulse rate 36o-38oC Social class Gender
Height Ethnicity

VV 52
Measure Phase VOP
Voice of the Process (VOP) is the use of control charts and
subgrouping techniques to judge how a process is performing.
Voice of the Process Control Charts
Control charts allow our processes to speak to us by showing
variability in data and a graphical representation of that variability
and how distributions are formed
Statistical process control (SPC) - based on numerical (variable)
measurements that become a picture of the process over time
All processes are subject to variability
Natural causes : Random variations
Assignable causes: correctable (Lack of infrastructure, Skill level)
When is a process said to be under control ?
When the variability in the quality characteristics is due to natural causes only
VV 53
Value Added
WORK Unavoidable, non-
value added activities
are sometimes
described as required,
non-value added. We
must challenge
ourselves to find
VALUE NON-VALUE creative alternatives to
ADDED ADDED these activities.

Value Added:
Important to the
Transforms information
or material

VV 54
Measure Phase Process Map
Process Map
It is like a flow chart by adding more process information
It is a graphical representation of activities and events
Visually shows the complexity of the process
Goal is to capture all steps, identify their basic function and
connect it in a way that represents the process
Identifies the key process input variables (xs) that go into the
process step and the resultant key output of that step (small ys)
Inputs can be classified as essential input, noise, controllable or
standard operating procedure (SOP)
Steps are the tasks that that transform the input (xs) into outputs
VV 55
Measure Phase
Process Map (As-is in Measure phase)
Ys Step 1 output Ys Step 2 output

Step1 Step 2 Decision Step 3

xs input A xs input C xs
input B input D
input E
Identify the process boundaries - start and stop
Document all steps in the process
List all inputs (xs in each step as x for critical, C for controllable, N for noise, SOP for Std operating
List all outputs of each step
VV 56
Process Map Dos and Donts
DO map the process as it DONT map the process as you think it
actually happens happens or as you think it ought to
DO think about the process
across the entire organisation DONT restrict your process map to
the activities in your own department
DO talk to the other people
who are involved in the DONT work in a vacuum
process DONT attempt to process map before
DO define the beginning and you identify a beginning and an end
end of the process before you
start DONT get bogged down with too
DO the process map at a high much detail
level DONT struggle on your own
DO ask questions

VV 57
Process Map Results
Re-check the process by looking at the following:
Start, end points and customers should be clear
Inputs and outputs should be identified
Indicate title of person / area responsible for each
A person not familiar with the process should be
able to easily understand the flow without any
The level of detail should be adequate to describe
inefficiencies VV 58
Measure Phase Why Measure
To make a decision:
Meet standards & specifications rejecting bad parts and accepting good parts
Criteria for acceptance of new measuring equipment
Evaluate variation
Compare measurement error to specifications

Stimulate continuous improvement:

Where to improve?
How much to improve?
Is improvement cost effective?
Prevention oriented
Compare measurement error to product variation

VV 59
Measurement System Variation
What we see What it actually is


Observed Variation
Actual Process

Total Observed Variation = Process Variation + Measurement Variation

VV 60
Measure Phase - MSE
Measurement System Evaluation
A method which shows us where variation exists in the
measurement system
Total Variation = Process Variation + Measurement variation
Variability of the measurement system has to be a lot less than
process or output variation
To understand the process, we must reduce the measurement
Sources of measurement variation could be:
Operator, Machine, Set up and Measurement noise
MSE should be RAVE (Relevant, Accurate, Valid and Easy to
VV 61
The Necessity of Training Farm Hands for First

Class Farms in the Fatherly Handling of Farm Live

Stock is Foremost in the Eyes of Farm Owners.

Since the Forefathers of the Farm Owners Trained

the Farm Hands for First Class Farms in the

Fatherly Handling of Farm Live Stock, the Farm

Owners Feel they should carry on with the Family

Tradition of Training Farm Hands of First Class

Farmers in the Fatherly Handling of Farm Live

Stock Because they Believe it is the Basis of Good

Fundamental Farm Management.

VV 62
The Necessity of Training Farm Hands for First

Class Farms in the Fatherly Handling of Farm Live

Stock is Foremost in the Eyes of Farm Owners.

Since the Forefathers of the Farm Owners Trained

the Farm Hands for First Class Farms in the

Fatherly Handling of Farm Live Stock, the Farm

Owners Feel they should carry on with the Family

Tradition of Training Farm Hands of First Class

Farmers in the Fatherly Handling of Farm Live

Stock Because they Believe it is the Basis of Good

Fundamental Farm Management.

VV = 36 63
Defining Accuracy, Bias and Precision
On target

Reduced spread

Center process XXX

XX Right mean;
Small standard

X Off target
X X Right mean; XX
XX Wrong mean;
X X large standard XXX
XX Small standard
X X deviation deviation

Accuracy is a measure of how close the average result matches the truth
Precision is a measure of how repeatable the results are for multiple measurements
Bias is the difference between the average of the measurements and the truth
VV 64
Analyze Phase
Most difficult and unpredictable part in DMAIC

Data Analysis
Statistical tools
Process Analysis
Understanding of problem areas
Areas that contribute to the problem
Root Cause Analysis
Cause and Effect analysis
Fish-bone Diagram

VV 65
Analyze Phase

Analyze the data :



VV 66
Analyze Phase

Population: The total set of units (for example, people, parts,

measurements, invoices, days, ) about which we would like

Sample: A subset of the population, usually much smaller and drawn

randomly from the population, yielding data values that we can
analyze Practically, Graphically, and Analytically.

VV 67
Graphical Analysis
Some ways to analyze data graphically

Dot plots
Scatter plots
Box plots

VV 68
Analytical Analysis

Measures of Location, or Central Tendency

Mean and Median

Measures of Variation, or Dispersion, Spread

Range, Variance and Standard Deviation

VV 69
Analyze Phase - Graphical Analysis
Dot Plots:
To display variation in a process.

Quick graphical comparison of two or more processes

Shows location and spread of data and it also identifies outliers
For plotting this, put a horizontal axis for the measurements and place a dot for
each data point and when the dots are very close, it can be stacked.

20 25 30

Stopping Distance in Feet

VV 70
Analyze Phase
Box Plot: (Tukey)
The Box plot helps in visually comparing the averages and variability
of two or more data sets. It also identifies outliers.
A Box (Whisker) Plot is an easy way of comparing the mean, quartiles,
and outliers of different data sets
Aids in understanding the distribution of the data and to get a quick,
graphical comparison of two processes.
Outlier *
Maximum observation

75th percentile

Median (50th percentile)

25th percentile

Minimum observation
VV 71
Analyze Phase
Scatter Plot: Shows the relationship between one variable and another
Visually indicate strength of correlation
Communicate positive/negative sense of correlation
To Plot:
Collect more paired samples

Draw diagram: Horizontal axis, cause variable

Vertical axis, effect variable

Plot the data

Analyze data to determine correlation (straight line or tight clusters indicate

strong relationship)

VV 72
Analyze Phase
Run chart

Purpose: To track process over time to display trends to focus

attention on changes in the process

To plot

Determine what you want to measure
Determine period of time to measure and in what time
Create a graph (vertical axis= occurrences, horizontal axis= time
Collect data and plot

VV 73
Analyze Phase
Pareto Charts
Purpose: To compare how frequently different causes occur or
how much each cause costs your organization

Pareto principle: 20% of causes account for 80% of the problem

When/Why: Sort data to determine where to focus improvements

or to set priorities

How: Collect data (check sheets, surveys), total results, and

arrange data in descending order
Draw a Pareto chart outline
horizontal axis: categories of problems
vertical axis: units of measure
Draw a bar for each category to height on vertical axis
Plot the cumulative totals and calculate the percentage
Analyze results

Or you can use Excel or Minitab to do Pareto charts

VV 74
Analyze Phase
Histogram: Depicts distribution and shape of data, gaps and outliers
Characteristics of a Bar Chart and counts the number of data points in specified
To Plot:
Count the number of data points
Determine the range (R) for entire set
Divide range value into classes (K)
Determine class width (H) (H=R/K)
Determine class boundary, end point: (1st point = lowest data value + H)
Construct frequency table based on values computed in previous step
Construct a histogram based on frequency table

VV 75
Process Capability

Cp - Capability Index It is the ratio of the specification

spread (USL - LSL) to the potential process spread
(6 times the within-subgroup standard deviation). 6
Cp does not consider the location of the process mean in relation to the specification
interval. It is a measure of the capability the process could achieve if centered between
the specification limits.

Cpk is a Capability Index that equals the minimum of Cpu and Cp

Cpk considers the location of the process mean relative to the specification interval.
It is a measure of how the process is actually performing.

Cpk = Min { USL - or - LSL }

3 3
VV 76
SPC - Control Charts
What is a Control Chart?
A control chart is a statistical tool used to distinguish between
variation in a process resulting from Common Causes and
variation resulting from Special causes.
It presents a graphic display of process stability or instability over
time. Process stability is defined as a state in which a process has
displayed a certain degree of consistency in the past and is expected
to continue to do so in the future.
A control chart is a graphic recording to monitor, control and improve
process performance by studying variation and its source
Focuses attention on detecting and monitoring variation over time
Serves as a tool for on going control of a process
Helps to improve a process to perform consistently and predictably
Provides a common language for discussing process performance
VV 77
SPC - Control Charts
Why use control chart?
It is simple to use
Enables synthesis, visualizations and interpretation of data
Enables to a large extent about what is going to happen
Helps in taking correct decisions at the right time
Control Limits represent the Voice of the Process
Control Limits are representative of the amount of expected variation in the
While every process displays Variation, some processes display controlled
variation, while other processes display uncontrolled variation
Controlled Variation is characterized by a stable and consistent pattern of variation
over time (associated with Common Causes)
Uncontrolled Variation is characterized by variation that changes over time
(associated with Special Causes)

VV 78
SPC - Control Charts
Type of

Attribute Data Variable data

(Count or Classification) (Continuous Measurement)

Defects Subgroup Subgroup size Subgroup size

size = 1 is small is large
(usually 3 to 5) (usually > 10)

Constant Variable Constant Variable

subgroup subgroup subgroup subgroup size
I-MR X and R X and S
size, c >5 size size
chart chart chart

C chart U chart NP chart P chart

number of number of number of Proportion

defects defects per defective of defective
unit units units

VV 79
SPC - Control Charts
Range Chart
X Bar Chart



UCL = Upper Control Limit

Out of Control Condition LCL = Lower Control Limit

X = Mean

R = Average Range
Monitor Changes Over Time
VV 80
Control Chart - The type varies based on the type of data (Variable and attribute data)
If a point lies outside the control limits, this deviation could be due to special causes
UCL and LCL are the control limits (+ or 3 sigma Values) which is 3 process standard
deviations above or below the average

Assignable (Special) UCL

Cause Variation Process
Natural Variation LCL 3
(Common) Causes

Common causes - variation that are inherent in a process over time. They affect every
outcome of the process and everyone working in the process
Special causes - process variation that is not inherent in the process. Defined by Shewhart as
a fleeting event, not systemically related.

VV 81
Analyze Phase
It is invaluable in applications where processes or products have safety
or security issues and also equally applicable to any process (Process
FMEA) or product (Design FMEA) where failures have an impact on
customer satisfaction or business.
Key to reduce or eliminate the risk of failures
A structural way of identifying the potential ways or modes a product or process
can fail, the effect of the failure and what causes the failures
By this approach we can identify the impact of problem (Severity), the occurrence
(Probability) and how easy to detect (Detection)
This helps in reducing the risk by taking appropriate actions (corrective and
With this we can prioritize the actions to eliminate the risk

VV 82
Analyze Phase
Severity (of Effect) - importance/impact of effect on
customer requirements (1=Not Severe, 10=Very Severe)

Occurrence (of Cause) - frequency with which a given Cause occurs

and creates Failure Mode. Can sometimes refer to the frequency of
a Failure Mode (1=Not Likely, 10=Very Likely)
Detection (capability of Current Controls) - ability of current control
scheme to detect:
- the causes before creating failure mode
- the failure modes before causing effect
(1=Likely to Detect, 10=Not Likely to Detect)

A Scale of 1 to 10 or 1, 3, 7, 10 could be used but not zeros

VV 83
Analyze Phase FMEA Steps
1. For each Process Step, list ways that it can vary or go wrong (Failure
2. Identify Effects associated with each Failure Mode
3. List all Causes for each Failure Mode
4. List the Current Controls for each Cause
5. Create Severity, Occurrence, and Detection rating scales
6. Assign Severity, Occurrence and Detection ratings by going down
the FMEA in each of these columns.
7. Calculate the risk priority number for each potential Failure mode
8. Determine recommended actions, with responsible parties assigned
and estimated date of completion, to reduce High RPNs - Estimate
effect of Actions on RPN (PSEV, POCC, PDET).
9. Take appropriate actions and re-calculate all RPNs
VV 84
FMEA Common Mistakes
Confusion Between Causes & Effects (5 Whys)
Use the process steps to identify what could go wrong with that
process step.
Poor Estimation of SEV, OCC and DET
Ensure you have cross functional representation and use (or
agree to collect) as much data as is practical
Doing an FMEA on an Undocumented Process
On new processes, once a significant failure mode is identified,
someone often says we can prevent that by doing this. This
then leads to confusion on estimating the impact are you doing
it with the suggested change or without it?
Pareto on the causes not done.
More than one potential failure modes and causes appear in one cell (In the table)

VV 85

Current Failure Failure Severity Potential Prob. of Present Detection R

Process Mode Effect Failure Occurrence Control Capability P
Causes N

RPN = Severity x Probability of occurrence x Detection

These 3 factors are measured on a scale of 1 to 10
Severity can vary from no effect to hazardous or catastrophe
Occurrence can be very unlikely failure to very high and almost inevitable
Detection can be very easy detection to cannot detect or very low probability of detection
Use Pareto chart (80:20 Rule) of the FMEA RPNs and ensure which causes has a greater
VV 86
Analyze Phase
Root Cause Analysis (RCA)
RCA is used to brainstorm and identify the root causes of the problems.
Root causes can be identified by asking Why multiple times (5 times thumb rule)
Fishbone Diagram
Measurements Process Technology

Problem Statement

Low Productivity

Raw Materials People

VV 87
Root Cause Analysis
5 Whys Level of problem. Related countermeasure.

There is a puddle of oil on

Clean up the oil.
the shop floor.

Because the machine is leaking oil. Fix the machine.

Because the gasket has deteriorated. Replace the gasket.

Because we bought gaskets made of

Change the gasket specifications.
inferior material.
Because we got a good price on those
Change the purchasing policy.
Because the buyer is evaluated on Change the evaluation process
short-term cost savings. of the buyer.
Source: Jeffrey Liker, The Toyota Way, McGraw-Hill, 2004
Initial source: Peter Scholtes, The Leaders Handbook, McGraw-Hill, 1998.
Analyze Phase
Cause & Effect Matrix is used for prioritizing
process steps & inputs based on the impact on
process outputs & the customer needs
List all output variables
Rank and weight the output variables
List all input and process variables
Evaluate the strength of the relationship between output
and inputs/process variable the correlation factor
Cross-multiply weight and correlation factor
Prioritize and identify the vital few process steps / input
VV 89
Analyze Phase - DOE
A properly designed experiment reveal, quantify and confirm the
underlying Y = f(x) relationship
Some of the terms used in DOE are:
Response (output Synonymous with Y)
Factors (the input variable we set for experiments xs)
Levels (2 or more variables we set for the factors)
Replications (repeating the experiment or parts of it to understand how much variation)
Runs (series of runs with predetermined values for each factor)
Key elements in DOE:
Planning the experiment before conducting it
Exploring the effect of more than one input variable at a time
Minimizing the number of runs in the experiment
Replicating key experimental conditions to assess variation
Accounting for factors that are not included in the experiment

VV 90
Analyze Phase - DOE
The type of experimental approach depends upon the current level of
knowledge of the process

DOE terminology:
Response Variable = Dependent Variable = Y
Factors = Independent Variables or Process Variables = xs
Levels = The settings of xs (usually low and high)
Main Effects = The change in Y caused by setting an x
Interaction Effects = The combined effect of two or more inputs
Treatment = A run of the experiment at particular level settings

VV 91
Analyze Phase - DOE
Experiments can be grouped into the following
general categories:
Trial and Error
Full Factorial
Fractional Factorial
Steps in DOE:
Identify the factors to be evaluated
Define the levels of the factors to be tested
Create an array of experimental combinations
Conduct an experiment under the prescribed conditions
Evaluate the results and conclusions
VV 92
Improve Phase
Develop new methods and ideas on how to solve the problem
Use those techniques to develop workable solutions
Criteria for selection of solutions
Ease of implementation
Cost to implement
Probability of achieving the goal
Long term benefits
Buy-in from senior management
Modify the Process
Implement Solutions
Problem Prevention (Mistake proofing techniques)

Measure results and ensure success

VV 93
Improve Phase
Process Map (To-be in Improve phase)
Ys Step 1 output Ys Step 2 output

Step1 Step 2 Decision Step 3

xs input A xs input C xs
input B input D
input E
Identify the process boundaries - start and stop
Document all steps in the process
List all inputs (xs in each step as CR for critical, C for controllable, N for noise, SOP for Std
operating Procedure)
List all outputs of each step
VV 94

Current Actions Responsibility Severity Post action Post Post

Process Taken (same as Prob. of action action
earlier) Occurrence Detection RPN

Post RPN = Severity x Prob.of Occurrence x Detection

VV 95
Mistake Proofing
As FMEA helps in prevention and prediction of problems, mistake proofing emphasizes the
detection, and correction of mistakes before they become defects
It puts special attention on one constant threat to any process: Human Error
Mistake proofing is to pay critical attention to every activity in the process and to place
checks and problem prevention at each step
This is used for:
Fine tune improvements and process designs
Gather data from processes approaching six sigma performance
Eliminate the process issues and defects needed to raise the sigma level
Steps in Mistake Proofing:
Identify possible errors that might occur despite preventive actions
Determine a way to detect that an error is taking place or about to occur
Identify and select the type of action to be taken when an error is detected
Control, shutdown, warning
Coming up with methods to detect, self-correct, block/shutdown or warn of a problem
requires imagination and creativity
VV 96
Mistake Proofing
Mistake Proofing (known as Poka-Yoke)
Prevents inadvertent causes that results in failures using simple low
cost methods for prevention
It is a mechanism or procedure that prevents the mistake being made
or make it so obvious to eliminate it
Poka Yoke devices make it nearly impossible to make a mistake
Examples include:
Electrical connectors that are specially shaped to prevent reverse plug-in
Overflow prevention systems
Gas filling inlets for cars
Others could be color and shape, symbols and icons checklists, clear forms etc.

VV 97
Control Phase
The critical input xs and the output CTQs are the movers and shakers of the
process that align to the needs of customer
In the control phase, we monitor the CTQs and control the inputs, the critical xs
and when done properly this helps in sustaining the gains
The control methods may include:
mistake-proofing methods
statistical process control
any appropriate procedure like automated data collection and process execution technologies
sampling required
sampling frequency
proper feedback
For this:
Make a control Plan
Continuously monitor
Make appropriate modifications/corrections
Sustain the gains
VV 98
Control Phase
Control Plan:
A Control Plan is a tool to help manage the activity necessary for
control of the critical inputs enabling a process to continually meet
its product or service goals.
Control Plans document what to monitor, how, when, and by whom,
providing a simple and complete source of knowledge that is robust
to organization and personnel changes

VV 99
Control Phase
Control Plan Requires:
Knowing how to assure control
Understanding what questions Control Plans can help you answer
Assigning accountability
Keeping measures and standards visible
Observing out-of-control conditions
Implementing actions to bring the process back into control
Defining an action plan to sustain the goal
Knowing how to document the Control Plan

VV 100
Sample Control Plan

Process What Why Why How When Who Actions

Activity Key Input Importance Effect Measurement Measurement Responsibility Control

Variable method frequency (Measurer) actions

VV 101
Project Discussion
Determine the Big Y the area that needs improvement
Define the relevant Processes
Estimate the current performance
Find out the impact of the problem
Define a Project Charter
Use the basic tool set both soft tools (like TPM, Process Map,
FMEA, QFD, RCA) and Statistical tools
Modify the process (to-be-process)
Implement the new process
Estimate the Gain (Measure the output)
Have a control plan to sustain the gains.

VV 102
Project Discussion
Project ID
Project Title
Project Description
Project Team
Potential Areas Involved
Define Response Variables
Measurements on Response Variables
Vital Few Identified
Define Phase Duration
Measure Phase Duration
Analysis Phase Duration
Improve Phase Duration
Control Phase Duration
Congratulate Your Team

VV 103
1. The Six Sigma Way
By: Peter S. Pande, Robert P. Neuman, Roland R. Cavanagh
Tata McGraw Hill, 2006
2. Six Sigma for Dummies
By: Craig Gygi, Neil DeCarlo, Bruce Williams
Wiley Publishing, Inc, 2006
3. The Six Sigma Handbook, Thomas Pyzdek and Paul
Keller, McGraw Hill, 3rd Edition, 2010
4. Statistical methods for Quality Improvement, Hitoshi
Kume, Productivity Press (India) Pvt Ltd, First Indian
Edition, 2006
VV 104
Six Sigma Course

Thank You

VV 105