GB Training Dmaic

D M A I C
Six Sigma
Green Belt TRAINING
GB TRG MATERIAL Slide 1 Proprietary to Wipro Ltd

How Successful Leaders see Six Sigma
Six Sigma is the most important initiative GE has

ever taken…it is part of the genetic code of our future leadership.
-Jack Welch, Former CEO,
General Electric
You can see the “before” & “after” of an organization

when Six Sigma grabs hold & takes place.
-Richard Johnson, Director of Six Sigma,
Allied Signal
Successful leaders have made Six Sigma their way of conducting business.
- Azim Premji, Chairman,
Wipro Limited

What is Six Sigma?
 Six Sigma as a metric

 At Six Sigma level, a process produces less than 3.4 defects in a million or
99.99966% Yield
 Six Sigma as a methodology

 Improving process capability to desired levels
 Six Sigma as a phi los ophy

 Business excellence through organizational transformation
 Six Sigma : Variation Based thinking

6 Sigma as a metric
The Sigma Scale provides a common scale/ metric for comparison.

Which process is performing best?
PROCESS PERFORMANCE
Call servicing 32 seconds ASA vs. goal of 35
Billing 98% accuracy
Accounts Receivable 33 days average aging vs. goal of 40
Customer Service 82% rated 4 or 5 on responsiveness
Sigma level DPMO

2 308,537
3 66,807
4 6,210
5 233
6 3.4

Six Sigma: As Methodology
An efficient & effective method :

Six Sigma is a structured application of
process improvement tools,
applied by dedicated resources on a project basis,
to increase customer satisfaction and reach strategic business goals.
Black Belt’s Statistical Black Belt’s

Expertise Tools Expertise
Business Statistical Statistical Business

Problem Problem Solution Solution
P R O J E C T

Six Sigma: As Philosophy
A philosophy :
For many companies, Six Sigma is a management philosophy.
Successful leaders have made Six Sigma their way of conducting business.
- Azim Premji, Chairman,
Wipro Limited
‘Six sigma is the best training that an organization can give to its employees,
it’s even better than sending them to Harward Business School.’
- Jack Welch, Former CEO, GE

Six Sigma is Variance Based Thinking
Consider below performance on ‘On time delivery’
Mean 0.70 days
Target 0 days
-20 -10 0 10 20
On Target !!!! Good Performance ??

Six Sigma is Variance Based Thinking
What customer sees!!
-20 -10 0 10 20
Customer always looks at variance , mean is … meaningless

Are We Looking at the Variation?
Don’t worry,
That rope is
one inch thick
on average.
“If I had to reduce my message to management

to just a few words, I’d say it all had to do with
reducing variation”
W. Edwards Deming

DMAIC Overview
DEFINE
The Problem
MEASURE
Y ( Outcome)  DMAIC applies to an existing process
Validate Measurement System that needs improvement
ANALYZE
Baseline Y and Set Goal
Identify X’s(Variation Sources)
IMPROVE
Quantify X’s
CONTROL
Control X’s

DMAIC
Establish CTQ Characteristics

CTQ (Critical to Quality)
A CTQ is a
Business CTQ
Product or Service
characteristic Customer CTQ
that satisfies a
Internal CTQ
Customer Requirement
OR
Project CTQ
Process Requirement

What are Business CTQs?
 Business CTQs drive the business goals & vision
 Typical Business CTQs are in the area of:
 Operational Excellence
 Cost Reduction
 Productivity Improvement
 Employee Satisfaction
 Customer Satisfaction
 Sales Growth
 Profitability

Define your Customer

Who is your customer?
 A customer is someone who
 Uses your product or service
 Decides to buy your product or service
 Pays for your product or service
 Gets impacted by your product or service
 Internal & External customers
 Primary & Secondary customers
 Key concepts
 Different markets / segments

 Different customer requirements
 Product / Service chain

Explore Customer CTQ

Customer CTQs
 Six Sigma begins with the customer
 Customers find it easier to define what they do not want
 Customer CTQs are defined by customers
 Sources of customer CTQs
 Survey results
 Service reviews
 Meetings

Customer Requirement – perceived by Sales Team

Customer Requirement – captured by Design Team

Customer Requirement – created by Production Team

Customer Requirement – delivered by Implementation Team

Actual Customer Requirement

Examples of Customer Responses on CTQs
A Car Purchaser A TV Purchaser

 Good Acceleration • Flat Screen
 Spacious • Good Sound
 Affordable • NTSC / PAL Compatible
 Power Steering • Affordable
 Loan Facility • Good After Sales Service
A Caller to Help Desk A Prospective Employee
• Quick Answering • Good Salary

• Courteous Response • Location Preference
• Quick Problem Resolution • Flexible Working Hours
• ESOPs
• Separate Cabin

How to Explore Customer CTQs
 Often, customer requirements are hazy
 Customer requirements must be understood clearly
 VOC is a technique to organize, analyze & profile the customer requirements

(collection of requirements is done through surveys / meetings)
VOC Table

VOC Data Source
REACTIVE SOURCES PROACTIVE SOURCES
Complaints Interviews
Problem or service hotlines Focus groups
Technical support calls Surveys
Customer service calls Comment cards
Claims, credits Sales visits/calls
Sales reporting Direct observation
Product return information Market research/monitoring
Warranty claims Benchmarking
Web page activity Quality scorecards
NPS survey outcomes can be Reactive & Proactive source

Define Internal CTQ / CBP

How to Arrive at Internal CTQs / CBPs

 Customer CTQs( needs) are often hazy & they must be converted into meaningful
internal goals that are assignable to functions
More structured methods & tools are available to convert Customer CTQs /
Business CTQs into internal CTQs / CBPs
 CTQ Drill-down
 Quality Function Deployment (QFD / House of Quality)

Example CTQ Drill Down
Improve
Profitability
Increase Offshore
Reducing
Realization Business
cost
or Billing rate Model
More
Manpower Travel & High End
Infrastr. cost New Geos Offshore
cost overheads Practices
component
Productivity Salary cost
Productivity (NLD) Projects were selected from the Business

CTQ of Profitability

One Example of A Typical CTQ Drill-down
 Example of On-time delivery CTQ
Total Units 168
Orders Fully Serviced 132 Orders Not Fully Serviced 36
A 36 B 47 C 31 D 18
A4 B 27 C1 D4
6 Late 31 late 8 late 7 Late
52 Defects 36 Defects
With the data support, team decided to take project in

Product B, Which is contributing to more than 70% defects

DMAIC
Define a Project

Theme Selection Matrix

 Theme selection matrix can be used to prioritize project themes coming from the
brainstorming exercise
 It is a structured & data-oriented prioritization exercise to select right project

 A scale of 1-9 is applied to evaluate multiple themes on four aspects. Final score is
arrived by multiplying the scores of each category
Rating Parameters for the themes (1-3-5-7-9)
Importance Importance
S.no. Selected Themes CTQ type to the to the Financial
internal E xternal benefi ts/ Final
customer customer A nnum E ffort sco re

Theme Selection Matrix: Example

 A project team wants to select 1 project themes out of 4 by using the Theme selection
matrix
Rating Parameters for the themes (1-3-5-7-9)
Importance Importance
S.no. Selected Themes CTQ Financial
to the to the Final
benefits/ Effort
internal External s core
A nnum
Customer Customer
Transition team productivity
1 Productivity
Improvement
NLD at small a/c ( less than 20
2 Productivity
resource )
NLD at large accounts ( more

3 Productivity
than 20 resources )
Improving Service Desk C-Sat/
4
performance ( CCOP Imp. ) Productivity

Scope the Project

What is Scoping?
 It’s an attempt to define what will be covered in the project deliverables
 Scoping sharpens the focus of the project team
 Sets the expectations right
 Types of Scoping
 Longitudinal Scoping
 Lateral Scoping

Longitudinal Scoping
 Longitudinal scoping is done on the length of the process
 e.g. – From the time a bug is assigned to the time it is submitted for fix
 e.g. – From the time of customer reporting the complaint till final satisfaction confirmation
 Mostly the ‘start’ & ‘end’ points are baton change points
 A macro as-is process map must be prepared to facilitate longitudinal scoping

SIPOC
Tool
P
Process
3
S
S I 1 2 5
O C I
P
O
C
Suppliers Inputs Outputs Customers

4
Process Boundary

Example of a Photocopying Process

Example - SIPOC
Process Name
Starting point: User sending a mail to Helpdesk
End point: Call logging in the Tool/Email Response to the Customer
Suppliers Inputs Process Outputs Customers

Receipt of Mail
End User Mail
from User
Call allocation to
Wipro MIT Helpdesk Team the Helpdesk
Executive
Helpdesk • Acknowledgement
End User
Executive Logs the Mail
Ticket On E-
Client – IT Dept IT infrastructure helpline*
HelpDesk sends
acknowledgement
to the Customer

Process Mapping Nomenclature
Process Decision Data
Pre-defined Terminator
Document
Process
Manual Manual
Delay
Operation Input

Lateral Scoping
 Lateral scoping is done on the breadth of the process
 e.g. – All despatches from North & South regions
 e.g. – Calls received during general shift
 One or more of the following are covered here:
 What all kinds of units the process will cover
 In what situations the process is valid
 What are the qualifiers for the transactions
 What functional domains does the process cover
 In what geographical areas the process is valid
CTQ Drill down and pareto analysis will help in effective lateral scoping

Scoping Techniques
 Write inside the box what you think project covers

Tool
 Write outside the box what you think project excludes
Example for an ‘On time Despatch’ project

I
n
s
Longitudinal i
d
e
Starts after receipt of PO from customer Any transit delays -
O
Ends at the despatch from factory u
Product damages t
s
Lateral i
d
Billing errors e
Despatches from Mumbai plant
Non-availability of customer
All despatches during shift A
Despatch through transporter X
Project boundary

Quantify Benefits

Projection of Benefits
 Six Sigma has a strong focus on money
 Management buy-in is easier for tangible benefit projects
 Tangible benefits could be of various types
 Cost Reduction / Saving
 Increase in Sales / Revenue
 Enhanced productivity
 Enhanced measurable Customer Satisfaction
 Enhanced measurable Employee Satisfaction

DEFINE THE PROBLEM STATEMENT

Defining a Problem Statement
Our customers are angry with us and thus delay paying their bills.
In the last 3 months, 12% of our customers are late, by over 45 days in
paying their bills. This represents 20% of our outstanding receivables &
negatively affects our operating cash flow.
When
Consequence
What Magnitude

D M AIC
Establish
Performance
Parameters

Introduction - MEASURE
 A robust measurement system forms the basis of any Six Sigma project
 A measurement system has two characteristics
 Design of the measurement system
 Precision of the measurement system

Map Detailed As-is Process

What do you see here…any clarity?

See the Difference Now…
Recall..One
picture is
better than
1000 words
Process Mapping is a graphic display of steps & activities that constitute a process

Benefits of Process Mapping
 Simple & visible structure for thinking through a complex process
 Enables seeing the entire process as a team
 Enables seeing that changes are not made in a vacuum and will carry through,
affecting the entire process down the line
 Magnifies non value-added areas or steps
 Identifies cycle times of each step in the process
 Helps re-examine (if needed) the scope and charter of your project

Micro Process Map of a Manufacturing Process
despatch
as reject
no
yes g/f sampling recording marking for finishing as

yes transfer to storage at installati guard
? application & properties ? next route per reqmnt. packing despatch/
recoiling concerned customer on film
G/f testing O.k as per cust. (coil, sheet, shipping
line end removal
reqmt. profile)
baking top
additional additional hot air squeezing water hot air squeezin water baking primer
as per coating additional additional ? no
inspection drying by rubber quenching drying g by as per coating
cold air cold air thickness wf t=55-65 cold air quenching cold air end
temp-100 - roll temp=100 wf t-15-18 Paint
drying drying pmt-224 visco-58-65 drying (r.t.) rubber thickness drying
110 Deg C Deg C visc-30 sec peel off
2 burner roll -
yes
continuous squeezin hot squeezin cold (room squeezin wet squeezin chromating coating baking
hot air water hot air
degreasing g by water g by g by scrubbing on back side as per g.p.
uncoiling of temp) g by concn- 0.8 - drying drying
rinsing 30 wft thickness quenching complaint
coil concn-0.3-0.4 rubber rubber water rinsing rubber by scoth rubber 1.2 temp-100 temp -90-
temp=80 Deg C visco-30 -35 PMT -224 raising by
temp-80 Deg roll roll roll bright Ph-1.2-2.0 Deg C 100 Deg C
sec
customer
START
2 stage
POINT testing of ph , concn. , squeezing
squeezing by
temp, by rubber
rubber roll visit to
of water , degreasing roll & CPC
customer for
and chromating assesment
replacement /
rework end
Make the process map aligned to your Project

objective, Don’t paste ISO process maps.

Another way of mapping processes- CFPM
 Teams can use this structure for processes that move across functions
Step
Function 1 Step 5
4
See Cycle time also.

Step 3 0.2, 0.1-
Average time & Range
0.5Hrs given
Function 2
Step 2 1.2, 1-
1.5Hrs
Step 1 Step 8 End

Function 3
4.5, 4-6
Hrs
Step 6 Step 7
Function 4
Process map is must in cycle time reduction projects


Points to Look For
 The pain areas (identified at the time of project selection) must be within the
selected scope
 Guard against analyzing the process at this stage, just map as-it-is
 Do not map the process as you would like it to be
 Take care of Parallel & Sequential activities for Cycle time reduction projects

Define what to measure and how to measure

Why Collect Data?
 Improvement can only occur if we understand where we are & where to go, supported
by a measurement system that validates both situations
 Successful organizations have a common language to communicate-- Data
 Common language- Data, promotes objectivity in decision-making process
 Have you reached where you intended to? -- only data answers that question
 It’s rightly said…If you can’t Measure it, you can’t improve it.

Develop Operational Definition
 An Operational Definition is a precise description that tells how to get a value for the
characteristic you are trying to measure. It includes what something is and how to
measure it
 Purpose:
 Removes ambiguity – common understanding of defect definition
 Identifies what to measure
 Identifies how to measure it
Aim is to have apple to apple comparison for

Baseline performance and improved performance.

Operational Definition - Examples
Try to Remove ambiguity from the statements
 I’m coming by 8 o’ clock….verify 8 am or pm
 I’m reaching Dammam by 8 o’ clock flight….verify 8 o’ clock is Arrival or Departure time
 I’m waiting for Doctor for so long…verify 5 minute or 5 Hrs
 Give me half bottle water---bottle may by 200ml to 20 Lt
 I’ve reduce patient’s waiting time…at which stage or what is the start & end point
 I’ve improve profitability…in which region/ which product

Operational Definition & Data collection plan
Operational Definition
Unit of What to Measure How to Measure
Measurement
Performance Operational Data source Sample Who will When will How will Other data that
measure (Y) definition and size collect the data be data be should be
location data collected collected collected at the
same time

Operation Definition - Example

Exp.: SLA Calculation at A/c level
 Objective: Why we need this/ Biz case
 Scope/ coverage: Reasons if not 100%
 Clarify what we are measuring: Uptime/ Response/ Resolution time, etc

 Is it aligned with the customer reqt (SLA/ Contract)
 Source of Data: IS System (Pure), Manual, Hybrid (IS System with manual
intervention )
 Sampling plan/ Freq of data collection
 Process owner
 Any stratification reqd/ recommended: SLA w.r.t. Skill calls, spare calls, EC Vs FE
 Challenges in authenticity of data: Like RE Call closure may not reflect 100%

Types of Data
 Discrete data
 Data that can take a limited number of values (Pass / Fail, OK / Not OK, W in / Loss)
 Examples
 Days in a week
 Number of ‘yes’ responses to a satisfaction survey
 Number of countries that play cricket
 Continuous Data
 Data that be expressed in either fractions or whole numbers
 Examples
 Temperature of the room You need to be careful while selecting the

data type and you need to follow the same
 Exchange rate of a currency
approach in improve phase also.
 Yield of a process
 Height of a person

Which Type of Data is Preferable?
 Large sample sizes are required to measure higher Sigma multiples for discrete data
 Since continuous data measurements can be broken down, relatively smaller
samples are required for higher Sigma calculations
 Discrete data does not allow to understand the process variation…
8:30 AM 11:00 AM
A B C D
How G ood is G ood and How bad is B ad.

Design Sampling Plan for

Establishing Process Baseline

What is Baseline?
 ‘Baseline’ refers to a reference point from where the improvement would be

measured
 For Base lining teams need to collect fresh ‘Samples’.
 Historical data may have measurement errors like operational definition issues,
which can lead to wrong projection of the Baseline

Introduction to Sampling
 We do sampling all the time
 Populations & Samples Tool
 Practical aspects – Cost & Time
 Sampling is done to study a representative portion of population
 Any term describing the characteristics of a sample is called statistic
 Any term describing the characteristics of a population is called parameter S

a
m
p
l
i
n
g
Sample
Population

How Big a Sample?
Too Big: Too small:

• Requires too • Won’t do the job
many resources
Take BB Assistance for calculating Sample Size

D M AIC
Validate
Measurement System
for ‘Y’

Validate Measurement System for ‘Y’
In transactional environment a robust

Operational definition will ensure Good
measurement system
Use ITIL guidelines & Industry best
practices for arriving at right Operation
definition ( e.g. AHT, FCR etc )
Take BB help in designing & performing GR&R

Gage Repeatability & Reproducibility

DM AIC
Establish
Process
Baseline

Compute Process Sigma Multiple Using

Collected Data

Six Sigma Metrics
Customer Specification / Unit / Defect Definition 1
Discrete Data Continuous Data
Defects
Defects per Unit Mean

(DPU) 2a 2b
Std. deviation
Defects per Million Opportunities
(DPMO)
Process Sigma Multiple - Z 3

Process Sigma Multiple for Discrete Data

Business Problem
 A project team wants to reduce the total waiting time taken in hospital (from reception
to discharge). Team has randomly collected the following data of last 3 weeks for the
waiting time for few patients/ customers :--
(Data given is in minutes)
USL decided by Management based on other hospitals study is decided at 60 Minutes max.

1 Customer Specification / Unit / Defect
 Unit: Any customer/ patient coming for treatment
 Customer Specifications: USL is 60 minutes
USL: Upper Specification Limit for a Performance Standard. Anything above this is a defect.
 Defect Definition: Any customer whose total waiting time

exceeds 60 minutes

2a Discrete Data
 Let’s treat this data as discrete
 That means I would count the number of patients who had to wait more than 60
minutes
Customer 1 5 Customer 11 50 Customer 21 49

 There are only 2 defects in 30 units

2a Defects per Unit
Number of defects found at any check-point

DPU =
Number of units processed at that check-point
 So, DPU of the given data is = 2 / 30 = 0.067

3 Sigma Multiple Calculation – Discrete Data
 Every DPMO value relates to a particular Sigma Multiple or ZST value. In this case, this
process is working at 3.00 Sigma multiple
 The same can be calculated by Given Excel template also.
ZST ZST

Process Sigma Multiple for Continuous Data

2b Continuous Data
 Let’s treat the data of the waiting time data as Continuous and calculate Base Line
Sigma
 For calculating Baseline or Sigma Multiple ZST in continuous data, we need to

understand: Mean, s td deviation & s hape (Type of Di s tribution)

3 Sigma Multiple Calculation – Continuous Data
 In this case, Mean = 57.73 minutes
 Std dev : 63.54
 USL : 60 Minutes
 ZST can be calculated by the Excel Template.
 This process is working at 1.54 Sigma multiple ( ZST = 1.54, or DPMO = 484750).
 Can you notice the difference in Z ST & DPMO values by treating the same data in
different way. What is the message?

Introduction of Normal Distribution

Introduction to Normal Distribution
Normal Distribution with Figure 3.01

Mean =100
Standard Deviation = 10 1 unit
of
standard
deviation
- +
70 80 90 100 110 120 130
 It’s a Probability Distribution, illustrated as N ( µ, σ )
 Simply put, a probability distribution is a theoretical frequency distribution
 Higher frequency of values around the mean & lesser & lesser at values away from mean
 Continuous & symmetrical
 Tails asymptotic to X-axis
 Bell shaped
 Total area under the Normal curve = 1

Prediction based on Normal Distribution
Figure 3.02
- +
µ
- 1σ + 1σ
68.26%
- 2σ + 2σ
95.46%
- 3σ + 3σ
99.73%
- 4σ + 4σ
99.9937%
- 5σ + 5σ
99.99943%
- 6σ + 6σ
99.999998%

DM AIC
Define
Performance
Goals

Improvement Paths in Process Sigma Multiple
 Mutually signed-off change in specification limits
 It’s not a process improvement
 Shifting the mean
 Reducing the variance

Benchmarking Exercise

Benchmarking
 Benchmarking is about looking at others for superior methods Tool
 Others may include suppliers, competitors, customers or even a different industry
 Benchmarking could be done both on means as well as ends

B
e
n
Types of benchmarking c
 h
m
a
 Internal r
k
i
n
 Functional g
 Competitive

Establish Process Improvement Goal p-value &

confidence band concept

Key Concepts
 If the existing process has to be improved, then the improvement goal should be
chosen only after proving statistically that it can be achieved only due to change,
& not by chance , I.e; the probability of achieving the goal with the current process
should be very low (generally <5%)
 In Minitab, Graphical summary command is used to identify the right target, that is
outside of 95% probability band of current data set.(Target can be set on mean as well
as std deviation)
Summary for June
A nderson-Darling Norma lity Test
A -Squared 351.69
P -V alue < 0.005
M ean 160.41
S tDev 483.73
V ariance 233992.60
S kew ness
K urto sis
9.161
12 3. 073
Any value, outside
N 1477
M inimum 1.77
this band will be a
1 st Q u art ile
M edian
2 3. 49
50.75 good target, as the
3 rd Q u ar ti le 9 2. 88
0 1200 2400 3600 4800 6000 7200 8400 M a xi mu m 8 35 4. 38 chances of getting
95% C onfidence Interval for Mean
135.72 185.10 it by current
95% C onfidence Interval for Median
46.95 55.05
process is <5%
95% C onfidence Interv al for StDev
95% Confidence Intervals
466.89 501.83
Mean
Median
50 75 100 125 150 175 200

DM AIC
Identify
Variation
Sources

Means & Ends of a Process
Y = f ( X1, X2, X3……Xn )
Y X
Dependent Independent
Process Output Process Input / Step
Effect Cause
Symptom Problem
Monitor Control

Generating & Prioritizing X’s
 Identifying & prioritizing X’s could be done using both non-statistical & statistical tools
 Identification tools include (non-statistical basis)

 Experience of process doers
 Brainstorming & Multi-voting Focus on identifying problem areas

 Process mapping
 Fishbone
Focus on identifying potential root causes
 Prioritization tools include
 Pareto Diagram
 Stratification
 Correlation (for a single X)
 Regression (for multiple X’s) Focus on prioritizing potential root causes
 Chi-square test
 ANOVA
 Failure Mode & Effect Analysis (FMEA)

Identification Tools
 Brainstorming & Multi-voting is a structured methodology to collect responses from a

group & arrive at a short-list
 Even though it provides a short-list, it’s not an effective prioritization tool because short-listing
is not done on any mathematical basis
 Use of Process mapping requires studying the micro as-is process mapped in step 2
earlier & walking through it
 Fishbone relates an ‘effect’ to its ‘cause’ by drawing a tree diagram

Points to Remember in Brainstorming
 All ideas are important, don’t out rightly reject any idea
 Participation should be ensured from all team members
 To ensure this, project teams could use the round-robin method of idea generation
 It’s advised to use the Black Belt as the facilitator here

Fishbone
 Also called “Cause & Effect’ or ‘Ishikawa’ diagram

Tool
 Focus of Fishbone is to arrive at the root causes of the problem areas identified
through multi-voting / process mapping
 It works on the principal of asking ‘why’ to each cause till you reach the ‘root cause’
F
i
s
h
b
o
n
e

Characteristics of a Fishbone

Characteristics of a Fishbone
• Search for the REAL

CAUSES of the problems
that you deal with.

How to Construct a Fishbone
 Review the problem statement / defect definition
 Brainstorm & identify possible causes
 Sort causes into reasonable clusters
 Choose the clusters & name them into big bones
 Check the logical flow of cause and effect

Fishbone Structure
How’s for 1st

Cause 1 Cause 2 House
Root-cause
221
Cause 11 Cause 21
Cause 22
Cause 12
Root-cause
222
Effect
Backbone
Cause 31
Cause 3
Big Bone Medium Bone

Fishbone Structure
How’s for 2nd

Root-cause
221
Cause 11 Cause 21
Cause 22
Cause 12
Root-cause
222
Effect
Backbone
Cause 31
Cause 3

Fishbone Structure
How’s for 3rd

Root-cause
221
Cause 11 Cause 21
Cause 22
Cause 12
Root-cause
222
Effect
Backbone
Cause 31
Cause 3

Fishbone Example
CRM malfunction Error in Listening
No
Lack of training maintenance
Printer error on listening skills Problem with
CRM not saving the instrument
data properly Rough
handling
Wrong Customer
Backbone Details
No credit card verification
Root-cause 1
Foul order No cross-check
Root-cause 2
Fishbone can be used in organizing the causes also.

Key Concepts
 It’s difficult to judge how many levels one needs to explore in a fishbone
 Sometimes project teams may hit the root-cause at multi-voting / process mapping
level, but they may do a further root-cause for the sake of it
 In such cases, it is always a good idea to see if there is further scope for exploring by
asking ‘why’
 Always remember that we are trying to identify controllable X’s , what can not
be controlled – can not be a root-cause
 One must be able to see the effect on ‘Y’ when values / levels of root-causes are
changed

Separate Vital Few from Trivial

Many for Further Screening

Prioritization Tools
 Some teams may do a multi-voting here, at the output of root-cause analysis
 Pareto diagram works on the 80:20 rule of 20% causes contributing to 80% of defects
 Co-relation & Regression help in identifying the movement of continuous ‘Y’ with
respect to continuous ‘X’
 ANOVA & Chi-square help in identifying the movement of continuous / discrete ‘Y’
with respect to discrete ‘X’
 Failure Mode & Effect Analysis (FMEA) identifies the process failure modes &
assigns a number to it that helps to prioritize the actionables
Success of all prioritization tools depends upon data collection

Pareto Diagram
 A graphical tool for ranking causes form most significant (Vital Few) to least significant (Trivial many).
 Based on Pareto principle, which was first defined by J.M Juran in 1950. Pareto principle was made Tool
after 19th century Italian economist V. Pareto (1897). Most effects come from relatively few causes.
Also known as 80:20 Principle.
 Pareto diagram indicates which area should be taken up first in eliminating defects and improving
operations.
 Example:Suppose a person identifies multiple root-causes of reaching his office late.

Now he is not sure where to focus so that he reduces the occurrence of P
reaching late by minimum 50%. a
r
e
t
o
 He has identified following root causes
 Woke up late
 Clothes not ready
 Breakfast not ready
 Bus not coming on time
 Traffic jam
 Bus waiting for other employees
 He collects data on how frequent each of the root cause is & constructs a Pareto

Pareto Diagram
Frequencies of root causes for reaching office late
70 100
60
80
50
t t
n 60 n
u 40 e
o c
r
C 30 e
40 P
20
20
10
0 0
t i m e d y l a t e m d y g
a t i i n
i n g o n
o t r e a u p c j a
i
f
f
o
t r e a w

c o m t e s
o h
n k e
W o T r a a s t
f
n B u s
s
u n o t
l
C B r e

a k
B
Count 25 18 15 6 5 2
Percent 35.2 25.4 21.1 8.5 7.0 2.8
Cum % 35.2 60.6 81.7 90.1 97.2 100.0

Other Prioritization tools
 So far, we have not used any statistical tool to prioritize X’s.
 Depending upon the data characteristics of Y & X, we can choose the appropriate tool
Correlation
Continuous & ANOVA
Regression
Y
Identify
opportunities for
Discrete converting ‘Y’ Chi-square
into a continuous
one or use FMEA
Continuous Discrete
X

Regression
 While correlation tells us only about the direction of movement, it does not
throw much light on degree of movement in one variable with respect to Tool
movement in another
 Regression of ‘Y’ on ‘X’ results in a transfer function equation that can be used to
predict the value of ‘Y’ for given values of ‘X’
R
Y = f ( X ) e
g
r
e
s
 ‘Y’ can be regressed on one or more X’s simultaneously s
i
o
 Simple linear regression is for one X n
 Multiple linear regression is for more than one X’s

Multiple Linear Regression (MLR)
 Multiple regression regresses ‘Y’ on more than one X’s simultaneously
 It is similar to Two-way ANOVA we have discussed in step 5, except for the difference
that X’s used in ANOVA were discrete
 The approach is similar & a linear multiple regression equation looks as follows:
Y = A + B1X1 + B2X2 + ………+ BnXn… + C
where Y = Dependent variable / output / response
X1 = First independent variable / input / predictor

X2 = Second independent variable / input / predictor
A = Intercept of fitted line on Y axis
B1 & B2 = Regression coefficients / Slopes of the fitted plane on two axes
C = Error in the model

Multiple Regression Example
 Suppose we are trying to predict rent of an apartment based on the size of the
apartment & its distance from the main commercial area. We gather the following
information as below. If you are looking for a two-bedroom apartment 2 miles from the
main area, what rent should you expect to pay?
Y X1 X2

Multiple Regression Example
 Minitab gives the following output:
Regression Analysis
The regression equation is
Rent = 852 + 1381 * Rooms

Predictor Coef StDev T P
Constant 852 1146 0.74 0.511
Rooms 1381.0 259.5 5.32 0.013
Distance -5.4 142.1 -0.04 0.972
S = 975.3 R-Sq = 91.5% R-Sq(adj) = 85.8%
Analysis of Variance
Source DF SS MS F P
Regression 2 30665301 15332650 16.12 0.025
Residual Error 3 2853449 951150
Total 5 33518750
 For a 2 bedroom apartment 2 miles away from main area, expected rent could be INR
3614 per month

ANOVA
 ANOVA is used to short-list potential discrete X’s for a continuous ‘Y’

Tool
 We can use one-way ANOVA & see the variation in ‘Y’ for one ‘X’ at a time
 We can use two-way ANOVA for more than one X
A
N
O
V
A

ANOVA Example
 A team puts great emphasis on customer satisfaction. For some weeks, the ratings
seemed to suffer & the manager tried to identify the factors that could be causing this.
He chooses two of the potential discrete factors as ‘Team’ that handles the call & ‘Shift’
in terms of A & B. Which factor is vital?
Team Shift Rating
Desktop A 3.4
DC A 4.5
Desktop A 3.2
DC A 4.8
Desktop A 3.6
DC A 4.2
Desktop A 4.1
DC A 4.6
Desktop A 3.1
DC A 5
Desktop B 3.5
DC B 4.2
Desktop B 4.6
DC B 4.9
Desktop B 3.5
DC B 4.1
Desktop B 4.3
DC B 4.7
Desktop B 3.4
DC B 3.8

ANOVA Example
Two-way Analysis of Variance
Analysis of Variance for Rating

Source DF SS MS F P
Team 1 3.281 3.281 17.38 0.001

Shift 1 0.013 0.013 0.07 0.800
Interaction 1 0.544 0.544 2.88 0.109
Error 16 3.020 0.189
Total 19 6.857
 Team that handles the call seems to be more important than shift
 In other words, two levels of the team, Desktop & DC, are significantly different in their average
ratings (between 'team' variation)
 However, error is quite high (within ‘team’ variation) which implies that there are other factors
contributing to variation in ratings

Failure Mode & Effect Analysis (FMEA)
 FMEA is a simple tool to prioritize the failure modes & actions

• Tool
F
•
 By understanding why and how we fail, we can plan for success M

E
A
 It works on the belief that proactiveness saves time
 Typically, FMEA is applied on the output of root-cause analysis, & is a better tool for
focus / prioritization as compared to multi-voting
 We shall focus on Process FMEA (Design FMEA is used in designing products)

FMEA Concept & Output
Effect
Severity
Process / Product Failure Risk Action

Cause
Characteristics
Mode Priority Plan
Occurrence Number
Control
Detectability
 Risk Priority Number (RPN) = S*O*D
 Severity, Occurrence & Detectability are measured on a scale of 1-10

FMEA Concept & Output
 Process / Product characteristics - Purpose of the product or Process
 Failure Mode - How can the product / process fail to function?
 Effects - Which effects are most severe to customer?
 Causes - Which causes are most likely to occur?
 Controls - Ability for current controls to detect causes?
 RPN - Which high risk cause we work on first?
 Action Plan - Recommended actions & responsibilities

Scales of ‘SOD’ - Severity
 Severity is the seriousness of the effect of the failure mode on the customer
 Rating Scale:

Scales of ‘SOD’ - Occurrence
 Occurrence is the probability that a specific cause will result in the particular failure mode
 Rating Scale:

Scales of ‘SOD’ - Detectability
 Detectability is the probability that a particular cause will be detected
 Rating Scale:

FMEA Table

DMA C I
Explore
Potential
Causes

Explore Potential X’s for Causation

Recall Step 6: Creating a List of X’s & Prioritization
Brainstorming Regression, FMEA
FMEA
x1 x7 = 38%
x2 x6 = 27% Vital X’s
x3 x2 = 12%
x4
x5 Exploration of x9 = 4%
x6 the y-x x10= 4%
x7 relationship x5 = 2%
x8 x1
Trivial X’s σ error
x9 x3
x10 x4
x11 x8 = 13%
x12 x11
x12
Fishbone

DMA C I
Design Counter-Measures

Counter-Measure Matrix
 Counter Measure Matrix is used to arrive at action to be taken on a root cause that
would lead to the desired result Tool
 If there are more than one counter-meas ure for a root cause, then the counter
meas ure is prioritized usi ng three parameters C
o
u
n
 Each counter-measure is given a rating on a scale of 1-7 (higher the better) & overall t
e
r
-M
score is calculated by multiplying the 3 ratings, one with the best score is chosen
e
a
s
u
er
M
a
t
r
i
x

Counter-Measure Matrix - Example
Cost optimisation levers

• CCOP
• Service Desk
• IR
• Automation

Map the New Process

DMAI C
Develop Control Mechanism

Controlling the X’s
 So far, we have identified the best settings for each of the vital ‘X’
 The key now is to ensure that the X’s don’t vary away from the targeted setting
 Process control is a crucial tool in ensuring that this Six Sigma project delivers lasting
benefits
 Maintaining X’s at their target level can be done in two ways

 Detection (reactive) Why was change not detected?
 Prevention (proactive) Why did / would change occur?

Correction Loop
Prevention
Prevention &
Detection
Mistake-Proofing Statistical Process

Control

Mistake-Proofing (Poka-Yoke)
 It is a technique to make errors difficult to happen, if not impossible

Tool
 Ground Rules of Mistake Proofing
 Intelligence does not depend upon hierarchy
 Target tasks that require constant alert / vigilance / memory M

i
s
t
a
 Look for cutting unproductive time to foster creativity k
e
-
P
 Defects are not acceptable, in whatsoever small number r
o
o
f
i
 It is best to make it impossible to do it wrong n
g

Advantages of Mistake-Proofing
 Easy to accomplish without any formal training
 Removes repetitive tasks
 Fosters creativity & value-addition
 Ensures less defects

Mistake-Proofing Techniques - Examples
Technique Prevention Detection
Cameras that don’t click Air-conditioners trip when

SHUTDOWN if the shutter is on detect slight over-heating
ATM does not accept a Quality check at each

CONTROL card inserted wrongly point in the assembly
Car security system

WARNING alerting the driver that all Smoke detectors
doors are not closed

If Mistake-Proofing is Not Possible?
 Provide guidelines
 Check-lists
 SOP’s
 Templates
 Use visuals
 Color-codes
 Shapes

Statistical Process Control (SPC)
 SPC was developed by Walter A. Shewhart in 1924

Tool
 Historically, SPC has been used to monitor & control output ‘Y’
S
 In DMAIC, we apply SPC to control X’s (remember ‘Y’ is only monitored) t
a
t
i
s
t
i
c
a
 However, sometimes applying SPC to ‘Y’ can also be beneficial in detecting trends l
P
or
c
 About SPC e
s
s
 Aids visual monitoring & controlling C
o
 Depends heavily on data collection n
t
r
o
l

Basics of Control Charts
 Control charts are useful for tracking process statistics over time and detecting the
presence of special causes
 A process is in control when most of the points fall within the bounds of the control
limits, and the points do not display any nonrandom patterns

Purpose of Control Limits
Special Cause Variation

UCL
e
s
u
a n
o
C i
t
n i
a CL
o r
m a
V
m
o
C
LCL
Special Cause Variation
Process Control’ is inherent to process characteristics as against ‘Process Capability’
which is measured as per outside targets & specifications

Purpose of Control Limits
0.135%
UCL = µ + 3σ
99.73% CL
LCL = µ - 3σ
0.135%
Out of control point
Control Limits define a probabilistic level

of occurrence of an ‘out of control’ point

Do Not Over-react to Change
e
m
i
t
UCL
g
n
i Excellent Work Change the transporter??
d
a
o
L
LCL
Distinguish between Noise & Change

Top Eight Indications of an Out of Control Process
1. A single point outside 3σ control limits
2. Nine points in a row on the same side of the centerline
3. Six points in a row, all increasing or all decreasing
4. Fourteen points in a row, alternating up and down
5. Two out of three successive points more than 2σ on the same side of the
centerline
6. Four out of five successive points beyond 1σ on the same side of the centerline
7. Fifteen points in a row within 1σ on either side of center line
8. Eight points in a row with none between 1σ on either side of center line

Choosing An Appropriate Control Chart
Continuous Data
Individual Data Points Subgroups

Pulling one sample at fixed frequency Taking periodic grouped data
Variability of individual Variability of average

I & MR characteristics over time characteristics over time
X&R

Choosing An Appropriate Control Chart
Discrete Data
Defectives Defects
P-chart U - chart
Rest of the Control Charts are in the BB training Scope

Continuous data Control Charts

I & MR Control Chart
 Average Handle Time (AHT) by an agent is a vital ‘X’ for productivity. This is monitored
periodically to keep it within desired operating range.
 Data on AHT is collected over several hours

Hour AHT ( sec )
1 65
2 69
3 67
4 66
5 63
6 70
7 71
8 68
9 64
10 69
11 63
12 68
13 84
14 81
15 68

I & MR Control Chart : Minitab Output

I-MR
I &Chart of Temperature
MR Chart of AHT
1
5 UC L=82.93
80
e
u
l
a
V
l _
a 70
u X=69.07
d
i
v
i
d
n
I 60
LC L=55.20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Observation
UC L=17.04
16
e
g 12
n
a
R
g 8
n
i
v
__
o MR=5.21
M 4
0 LCL=0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Observation

Discrete data Control Charts

P Control Chart
 A call center receiving calls will have to closed within specified SLA. Team has
collected data every hour to monitor process control
 Defective is defined as call not closed within SLA

P Control Chart : Minitab Output

P Chart of P defectives
0.4
0.3 UCL=0.2906
n
o
i
t
r 0.2
o
p
o
r
P
_
0.1 P=0.0944
0.0 LCL=0
2 4 6 8 10 12 14 16 18 20 22 24
Sample
Tests performed with unequal sample sizes

Points to Remember in Control Charts
 Do not apply SPC tools to processes that are known to be out on control
 Do not compare control limits with specification limits
 Do not ignore ‘out-of-control’ signals if your ‘Y’ is meeting the specifications & ‘X’ is
meeting the operating limits
 Ensure that observations are independent of each other
 It’s quite possible that ‘X’ is under control, but ‘Y’ is out-of-spec’s

Re-compute Process Baseline & Verify Against

Target

Validating Process Improvements
Two-Sample T-Test and CI: Cost / Indent, Status (Control Phase)
Two-sample T for Cost / Indent

SE
Status N Mean StDev Mean
Before 7 300.7 47.6 18
Improved 7 154.1 61.6 23
Difference = mu (Before) - mu (Improved)

Estimate for difference: 146.7
95% lower bound for difference: 93.8
T-Test of difference = 0 (vs >): T-Value = 4.98 P-Value = 0.000 DF = 11
I-MR charts
( stage wise )

Develop Action Plan to Sustain

Improvement

How to Sustain Process Improvement
 Control Plan
 Training & Communication
 Project handovers with completed project dockets
 Database of vital X’s
 QFD / VOC
 FMEA / Fishbone
 Six Sigma Project Audits
 Process Owners
 SOP’s

DMAIC SUMMARY: Problem Solving Flow
MEASURE ANALYZE IMPROVE CONTROL
Business Statistical Statistical Business

Problem Problem Solution Solution
P R O J E C T

Six Sigma resource in Wipro Intranet
 Q City : http://qcity.wipro.com:8080/hive/index.jsp

Six Sigma resource in KM
 KM Link : http://knetsites.wipro.com/sites/1010/Pages/Default.aspx

Process Excellence – Wipro Way

http://wiproway.wipro.com/default.asp

WI MQ – Mentor List
Name LOB Region Email id

Sachin Agarwal [ MBB ] Services Common HO - BLR [email protected]
Jyoti Prakash [ LBB ] Services Common HO - BLR [email protected]
Priya Mahadevan [ LBB ] MIT HO - BLR [email protected]
Rashmi Salgar [ BB ] ES & SI HO - BLR [email protected]
Sumathi Rajesh [ BB ] Regional MQ S2 [email protected]
Bhakti Gokhale [ BB ] Regional MQ West [email protected]

Thank You
Improvement is a Journey…
and Improvement is a Journey…
and
Tog etherTog
Weether
Will
We Will
161
•Proprietary to Wipro Limited © 2009 Wipro Ltd - Confidential

GB Training Dmaic

Uploaded by

Copyright:

Available Formats

GB Training Dmaic

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GB Training Dmaic

Uploaded by

Copyright:

Available Formats

How Successful Leaders see Six Sigma

Six Sigma is the most important initiative GE has

You can see the “before” & “after” of an organization

GB TRG MATERIAL Slide 2 Proprietary to Wipro Ltd

What is Six Sigma?

 Six Sigma as a metric

 Six Sigma as a methodology

 Six Sigma as a phi los ophy

 Six Sigma : Variation Based thinking

GB TRG MATERIAL Slide 3 Proprietary to Wipro Ltd

The Sigma Scale provides a common scale/ metric for comparison.

Call servicing 32 seconds ASA vs. goal of 35

Billing 98% accuracy

Accounts Receivable 33 days average aging vs. goal of 40

Customer Service 82% rated 4 or 5 on responsiveness

Sigma level DPMO

GB TRG MATERIAL Slide 4 Proprietary to Wipro Ltd

Six Sigma: As Methodology

An efficient & effective method :

process improvement tools,

applied by dedicated resources on a project basis,

to increase customer satisfaction and reach strategic business goals.

Black Belt’s Statistical Black Belt’s

Business Statistical Statistical Business

GB TRG MATERIAL Slide 5 Proprietary to Wipro Ltd

Six Sigma: As Philosophy

For many companies, Six Sigma is a management philosophy.

GB TRG MATERIAL Slide 6 Proprietary to Wipro Ltd

Six Sigma is Variance Based Thinking

Consider below performance on ‘On time delivery’

Mean 0.70 days

On Target !!!! Good Performance ??

GB TRG MATERIAL Slide 7 Proprietary to Wipro Ltd

Six Sigma is Variance Based Thinking

What customer sees!!

Customer always looks at variance , mean is … meaningless

GB TRG MATERIAL Slide 8 Proprietary to Wipro Ltd

Are We Looking at the Variation?

“If I had to reduce my message to management

GB TRG MATERIAL Slide 9 Proprietary to Wipro Ltd

GB TRG MATERIAL Slide 10 Proprietary to Wipro Ltd

GB TRG MATERIAL Slide 11 Proprietary to Wipro Ltd

CTQ (Critical to Quality)

characteristic Customer CTQ

GB TRG MATERIAL Slide 12 Proprietary to Wipro Ltd

What are Business CTQs?

 Business CTQs drive the business goals & vision

 Typical Business CTQs are in the area of:

GB TRG MATERIAL Slide 13 Proprietary to Wipro Ltd

Define your Customer

GB TRG MATERIAL Slide 14 Proprietary to Wipro Ltd

Who is your customer?

 A customer is someone who

 Uses your product or service

 Decides to buy your product or service

 Pays for your product or service

 Gets impacted by your product or service

 Internal & External customers

 Primary & Secondary customers