6 SigmaLecture
6 SigmaLecture
How close
are you to
meeting your
target?
What is Six Sigma (6 σ )
The goal
Minimize variation
Center the process, such that the center of the value delivered is at
least six standard deviations away from what is considered non
acceptable or the upper or lower customer specification
The degree of acceptable variation in a process depends on
knowing the target
Do you know
where your
target is?
What is Six Sigma?
Statistically
Having no more than 3.4 defects per million
Conceptually
Program designed to reduce defects
Requires the use of certain tools and
techniques
Six Sigma Quality
σ % Non- Defective
2 95.46
3 99.73
σ PPM
4 99.9937
5 99.999943 2 308,537
6 99.9999998 3 66,807
4 6,210
5 233
6 3.4
Postal System
Medical Profession
200,000 wrong drug prescriptions/yr 68 wrong drug prescriptions/yr
What does 6 Sigma Mean in Your
Life?
PPM
Tax Advice Prescription
Writing Restaurant
1,000,000 Bills
Payroll
100,000 Processing
Baggage
Handling
10,000
1,000
Best
10 In
Class Airline Defect
0 Rate
1
0
1
1 2 3 4 5 6 7
Sigma Level
Business Impacts of
Variability
Prevention Costs Internal Failure
Education and training Scrap
Quality planning Supplier caused losses
Product design Qualification tests Rework
Supplier qualification Reset / Re-inspection
Customer interface Unplanned downtime
Controlling process Trouble shooting
Appraisal Costs Engineering change notices
Incoming inspection External Failures
Maintenance and calibration Processing customer enquiries
of equipment Maintaining customer field service
Setup inspection and tests Retro fit costs
Field testing Incurring penalties / claims
Process audits Product warranty
Lost sales
Six Sigma Programs
Cut costs
Employed in
Design
Production
Service
Inventory management
Yes
Is this a
NO recurring
issue?
NO!
Do you already
NO know how to fix
a problem?
Yes!
NOT A Six-σ project You have a
POTENTIAL
Six σ project
Identified potential project
Role
Owner of the solution delivered by Six Sigma team
Typical Responsibilities
Implementation of the team solution
Assist with culture change at local level
Assist Champion with potential project identification
Co-leads realization phase with Finance Rep
Provides resources to serve as team members on projects
Training
2 to 3 days focused on understanding Six Sigma and the role he / she will play in
the deployment
Eventually trained as Green belt
Pitfall
If the Process Owner is not engaged throughout DMAIC, he will not accept the
solution when it’s time to close a project
Master Black Belt
Role
Variable from company to company
Should make Champions life easier in mature deployment
Typical Responsibilities
Instructor and / or mentor Black Belts and / or Green Belts
Training material developer
Deployment assistant to core team and Champions
Keeper of project backlog list
Driver of project closure process
Leader of larger scoped projects
Training
Certified BB + additional course work and requirements
Black Belts
Role
Practitioner of DMAIC Methodology
Responsibilities
Ideally full-time facilitator / leader of Six Sigma project team
Team should discover and recommend project solution
Executes 4 to 6 projects / year
Training
4 to 5 weeks of DMAIC training
Personal and professional development for later leadership roles
Pitfalls
The likely success of Six Sigma projects dramatically decreases without the
implementation of full-time Black Belts
If BBs are not selected from amongst the best within an organization, it sends
the wrong message.
Role
Carry the language of Six Sigma deeper into the organization
Accelerate number of employees positively affected by Six Sigma
Responsibilities
Become local advocate
Part-time implementers of smaller scope projects with direct impact to daily
non-Six Sigma duties
Assist BB with team activities and tool application to project area
Training
Varies from company to company, typically 5 to 15 days of abridged version of
DMAIC
Pitfalls
If Green Belts are not properly supported, their effectiveness diminishes
dramatically
Team Members
Role
Extend the reach of the Six Sigma language into the trenches
Responsibilities
Assist Black Belts with data collection and tool application
Provide invaluable process expertise to Six Sigma team
Assist Process Owner with the long-term implementation of solution
Training
Trained by Black Belts during team meetings
4 to 8 hours of formal classroom training in the basics
Pitfalls
Teams that are too small or too large may prevent project success
Recognition of team member contribution is critical throughout the project life
Black belt must seek out and get buy-in from Team Members
Executives and Deployment
Leaders
Role
Own the vision, direction, integration and results
Lead the culture change
Responsibilities
Determine the scope of the Six Sigma Deployment
Identify financial, project, and training related goals
Identify strategic priorities to which Champions will align projects
Drive the use of 6σ as standard problem solving methodology
Training
1 to 2 days targeted at Six Sigma’s potential impact
Pitfalls
Without visible executive support, six Sigma Struggles
Must push to ensuring full-time Black Belts
Six Sigma Process
Define
Measure
Analyze DMAIC
Improve
Control
Six Sigma Implementation: DMAIC
Framework
1. Define (D) Customers and their priorities
deficiencies in product.
If the process mean shifts (by 1.5 sigma), no more than 3.4 units per
Evaluate project
Contd….
A cause is a proven reason for the existence of the defect. Multiple causes are
common, in which case they follow the Pareto principle; i.e., the vital
few causes will dominate the rest.
Chronic problems are usually not easy to solve and require careful planning
and collection of data to confirm and analyze the input and output variables.
Measurement capability involves both the ability of the people making the
measurements and the capability of the measuring instruments .
The specification limits must be at least six sigma above and below the
process mean.
It must recognize both short term and long term variation.
Tools and Techniques of Six
Sigma
Measure Phase
• Process mapping
• Process FMEA
• Data collection planning
• MSA
• Graphs and charts
• Stratification
• Process capability analysis
• Sigma calculation
• Pareto chart(s)
• Brainstorming
• Cause-effect (or fishbone) diagrams
Analyze phase:
Contd…..
To generate information, we need to:
Present the data in a way that clearly communicates the answer to the question
Collect and Analyze Data
The great majority of worker errors fall into one of four categories:
Unintentional
Inadvertent
Unwitting
Technique Unpredictable
Conscious
Communication
Contd…..
Remedies for inadvertent errors involve two approaches:
1. Reducing the extent of dependence on human attention.
2. Helping workers remain attentive.
Unintentional
Technique errors: Specific
Unavoidable
Tools and Techniques of Six
Sigma
Analyze Phase
3. Design a remedy.
• Self-control analysis
• Process Control Plan
• Mistake proofing
• 5S
• Statistical Process Control
• Standard Operating Procedure
• Change management
Additional Six Sigma Tools:
6 process
Off
Off
spec 3 process spec
-3 -2 -1 µ +1 +2 +3
Lower Upper
Spec Spec
Limit Limit
Focus of Six Sigma Problem
Solving
Y= f(x)
To get results, should we focus our behavior on the Y or X ?
Y X1… X2
Dependent Independent
Output Input -
Effect Process
Symptom Cause
Monitor Problem
Control
Focus on X rather than Y, as done historically
Understanding Variation (Six
σ)
Traditional Philosophy : Goalpost Taguchi Philosophy : Variation is
Mentality Evil !
LSL USL LSL USL
HEADSET
Practical Statistical
Problem Problem
Practical Statistical
Solution HEADSET Solution
How Do We Improve Capability?
y = f(x1,x2, … ,xk)
Paint Thickness = (Paint Viscosity, spray Angle, spray
Pressure, Standard Operating Procedures,
… , Spray Pattern)
Types of Data
Distributions
Measures of the Center of the Data
Mean
Median
Measures of the Spread of Data
Range
Variance
Standard Deviation
Measures of the shape of the Data
Normal Distribution
Normal Probabilities
Data Mining
Types of Data
x
n
x n
x n 1
n
n
x n 1
n
Median : Reflects the 50% rank – the center number
after a set of numbers has been sorted
Does not necessarily include all values in calculation
Is “robust” to extreme scores
Why would we mainly use the mean, instead
of the median, in process improvement
efforts?
Spread of data variability Measures
x i x i
i 1 x i 1
N n
Population Sample Standard
Standard Deviation
Deviation
N n
xi x x
2 2
i
i 1 s i 1
N n 1
Spread of data variability Measures
Property 1
Property 2
The area under sections of the curve can be used to estimate the
cumulative probability of a certain “event” occurring
Cumulative
probability of
obtaining a value
between two values
Numerical Summary
Measures
X
Mean for population data,
N
X
X
Mean for population data,
n
Median
Mode
Measures of Dispersion for Ungrouped Data
Range
Variance and Standard deviation
The standard deviation is the most used measure of dispersion.
The formulas for calculating the variance and standard deviation
are as follows:
x
2
x
2
x 2
N
2
N N
x
2
x X
2
x 2
n
s2
n 1 n 1
Example
a) Mean
x 400 $33,333
n 12
th
n 1
Median is term in a ranked data set.
2 Therefore,
n 1 12 1
6.5
2 2
21 18
Median $19,500
2
Mode $21,000, $18,000, and $16,000
Solution (contd.)
x
2
x2
19,528
4002
Variance s 2 n 12 563.1515
n 1 12 1
Standard deviation s s 2 563.1515 23.73 $23, 730
Measures of Position
1. p ( xi ) 0 for all i
2.
p( x ) 1
i 1
i
The collection of pairs [xI, p (xi)], i =1,2,…, is called the probability distribution
of X, and p (xi) is called the probability mass function (or pmf) of X.
Discrete Random Variables
(contd…)
p( x)
p (1)
p (2) p (3) px
p (0)
0 1 2 3 x
The function f(x) is known as the probability density function (or p d f) of the
random variable X. The p d f must satisfy the following conditions:
f ( x)dx 1
2.
The Probability Density Function
(pdf)
x=a x=b
Discrete Distributions
Hypergeometric Distribution
Suppose there is a population of finite size N that contains a
specified number of nonconforming components D. Then,
the probability that a sample of size n containing x non-
conforming components is given by
N D
D
x n x
p( x ) 1p(x)=
i 1
i
N
x=0,1,2,...,min n,D
n
where, a a!
b b!a-b !
Discrete Distributions
(contd…)
Binomial Probability Distribution
The probability Pr of an event succeeding r times is
n!
p 1 p
r n-r
Pr or,Pr n C r p r q n-r
r!n-r !
np
r
Pr e-np
r!
Mean of Poisson Distribution
np
1 1 2 x-
2
f x e for X
2
1.
f ( x)dx 1
X 0.25
c) P( X x) P Z 0.90
0.025
x 0.25
Hence, 1.28
0.025
and, x 0.282
Tutorial
Problem I
A lot contains 140 electronic components and 20 are
selected without replacement for quality testing.
a. if 20 components are defective, what is the probability
that at least one defective component is in the sample?
b. if 5 components are defective, what is the probability that
at least one defective component appears in the sample?
Solution
b. P(X 1) = 1 – P (X = 0)
20 135 135!
0 20 20! 115! 135! 120!
P(X 0) 0.4571
140
140! 115! 140!
20! 120!
20
Problem II
A photocopying machine in an office breaks down an
average of three times per month. Using the Poisson
probability distribution formula, find the probability that
during the next month this machine will have
a. exactly two breakdowns
b. at most one breakdown.
Solution
x e (3) 2 e 3 (9)(.04978)
P(X 2) .2240
x! 2! 2
Solution (contd…)
Problem III
The diameter of a machine shaft produced in a
manufacturing company is normally distributed with a
mean diameter of 0.001 inches and a standard deviation of
0.0002 inches.
a. what is the probability that the diameter of the shaft
exceeds 0.0013 inches?
b. what is the probability that a diameter of the shaft is
between 0.0007 and 0.0013 inches?
c. what standard deviation of diameters is needed so
that the probability in part (b) is 0.995?
Solution
0.0013 0.001
a. P(x 0.0013) P Z P(Z 1.5) 0.06681
0.0002
b. P (0.0007 < x < 0.0013) = P (-1.5 < Z < 1.5)
= 2P (Z<1.5) - 1 = 0.86638
0.0007 0.001 0.0013 0.001
c. P(0.0007 < x < 0.0013) = P Z
0.0003 0.0003
P Z
0.0003
2P Z 0.0003 1 0.995 1.405
0.000214