R-DMAIC-D Six Sigma--TBI

R-DMAIC-D Six Sigma

Prepared by Julian Kalac, P.Eng

Lean Six Sigma Master Black Belt

Define

Measure

AnalyzeImprove

ControlResults

Define

Measu

re

AnalyzeImprove

ControlResults

2

σ Use data-driven, measurement-based, statistical methods to Solve problems, improve performance

σ Focus: Surgical “inch-wide, mile-deep” investigation and resolution

σ Approach:

σSolve problems at the system and root cause level

σ Implement robust control plans for sustained improvements

What is “Six Sigma”?An Analytical Methodology that Focuses on Reducing Process Variation

Six Sigma as a Metric

1

)( 2

n

xxi

Sigma = = Deviation( Square root of variance )

-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7

Axis graduated in Sigma

68.27 %

95.45 %

99.73 %

99.9937 %

99.999943 %

99.9999998 %

result: 317300 ppm outside

(deviation)

45500 ppm

2700 ppm

63 ppm

0.57 ppm

0.002 ppm

between + / - 1

between + / - 2

between + / - 3

between + / - 4

between + / - 5

between + / - 6

=

4

•Identify customer

metrics

•Select performances

standards

•Select objectives

•Map the process

•Validate measurement

System (MSA)

•RESULT: Process maps

and good quality data

collected by listening to

the process

•Identify sources of variation & failure points

•Redefine and Re-prioritize

•Establish process capability

•Identify improvement opportunities

•RESULT: The critical sources of variation

identified root causes determined

•Monitor processes to prevent

recurrence of variation, defects

and non-value work

•Maintain performance levels

•find more creative ways to

improve

•fix root causes

•Find/Implement Preventive Fixes

•Deploy changes organization

•RESULT: Performance is more

predictable ; culture changing

•Plan and apply Improve Tools to fix

problems and reduce variation

•Implement improvement opportunities

•RESULT: Improve tools applied, changes

implemented and performance

improvement in place

DMAIC Improvement ModelA Road Map for guiding Improvement Projects

•Identify customer problems

•Identify performances standards

•Identify improvement objectives

•Link problem to the key performance metrics

•Find the right problems

•RESULT: Problem Statement & Project Charter

Define

Measure

AnalyzeImprove

Control

Results

Define

Measure

AnalyzeImprove

Control

Results

5

Define

Measure

AnalyzeImprove

Control

Results

Define

Measure

AnalyzeImprove

Control

Results

The “R-DMAIC –D” Model Recognize-DMAIC-Deploy: Extend Performance Improvements

Applied Learning

Theory Practice

Coached Projects

Training typically

occurs over a 4-6

month period using a

Learn Apply;

Learn Apply; …

format.

Projects that solve

real performance

problems in your

organization are part

of the certification

and learning process

Recognize:

•Find problems

•Link to organization's needs

•Form teams, define resources

•Understand program req’ts

Deploy:•Validate improved performance that stays in place•Spread improvements across the enterprise•Harvest the performance improvements•Change the culture

8

Defining Projects Starts with Problems

2) Clearly Define your Problem and the ImpactExample:Too many mistakes in purchase request specifications are causing rework rates of 34%, high costs and late deliveries (less than 50% on time) to our customers

1) Find

• Customer Issues

• Performance

Reviews

• Meetings

• New Projects

• Failures, Re-work

• Projects

• Continuous

improvement

activities

• Champions

• Stakeholders

• Employees

• SCOR Maps

• Research

• Brainstorm

• Wait until the

crisis hits you

• Value Stream

Maps

3) Clearly State the Initial Scope for your project

(Set targets and timelines!)Example:This first phase of this project – by 3Q 2004 - will identify problems and root causes in the purchase request process, increase first pass yield rates to 95%, reduce the cost of poor quality by 50% and increase on-time deliveries to our customers to greater than 90%

4) Link improvement targets to customer needs and

organizational objectives

5) Project Plan – Charter, Resources, Milestones

9

Process Flow

for Measure Phase

Data

Storage

and

Archiving

Foundations for Wisdom and good decision making start with Good Data

What data do we need?Can we use “old” historical data?

Is old data (still) usable? How was it collected?

Wisdom

Knowledge

Data

Information

Wisdom

Knowledge

Data

Information

Wisdom

Knowledge

Data

Information

Measure Phase Process FlowHow to Get Good Data

Data Selection

Data Integrity Analysis

Data Collection

Measurement System Analysis

Process Mapping

10

SIPOC Diagrams

X1 = _____ Y1 = _____

X2 = _____ Y2 = _____

X1 = _______ Y1 = _______

X2 = _______ Y2 = _______

X3 = _______ Y3 = _______

SIPOC Diagrams can be characterized as a 3-step, high-level (“30,000-foot)

Process Flow Diagram for a process

Critical to

Customer Quality

Requirements

(the “CTQ’s)

Key Inputs

(Materials &

Resources) and

Key Process Input

Variables

(KPIV’s)

Inputs Processes OutputsSupplier Customer

Secondary Metric

(e.g. Field Returns)

Primary Metric

(e.g. Scrap $/Month

11

SIPOC Process Improvement Model

Supplier provides inputs

Inputs: materials, equipment, information, people, money, environmental conditions

Process: activities & tasks that transform inputs

Outputs: product or service delivered to the customer

Customer receives outputs

PI OKPIV KPOV

CS

FEEDBACK

FEEDBACK

Measurement System Analysis (MSA)Is your error in the process or in the way you measure it?

Could it be that you actually are “good” but the error in the measurement system shows that you are not “good”?

Overall Variation

Occurrence-to-

Occurrence ( or Piece-

to-Piece) Variation

Measurement System Variation

Repeatability:

Variation due to gage

or measurement tool

Reproducibility:

Variation due to people or

operators who are measuring

TV= MSA + Process variation

13

Repeatability

Repeatability is the variation in measurements obtained with one measurement instrument used several times by one appraiser while measuring the identical characteristic on the same part.

For example:– Manufacturing: One person measures the purity of multiple samples of the same vial and

gets different purity measures.

– Transactional: One person evaluates a contract multiple times (over a period of time) and makes different determinations of errors.

Repeatability

Y

Source: iSixSigma

14

Reproducibility

Reproducibility is the variation in the average of the measurements made by different appraisers using the same measuring instrument when measuring the identical characteristic on the same part.

For example:– Manufacturing: Different people perform purity test on samples from the same vial and get

different results.

– Transactional: Different people evaluate the same contract and make different determinations.

Reproducibility

Operator A

Operator B

Y

Source: iSixSigma

15

Organize & Understand your data

Inferential Statistics

Descriptive Statistics

Sort, Collate, Investigate your data

Analyze Phase Process FlowTransitioning from Data to Information and Knowledge

Organize your data and put it into some sort of

perspective, concept, picture or visual

representation that is easier to understand

Use maps, graphs, charts, summaries, spread

sheets, etc., that organize the data

Wisdom

Knowledge

Data

Information

Wisdom

Knowledge

Data

Information

Wisdom

Knowledge

Data

Information

16

We know we must change X to create a change in Y…

But how do we know which Xs to change

and how to change them ?

Y =f (x1, x2, …)

process output key process input factors

affecting process outputfunction of

Identify Significant Factors (x’s)

17

P value< 0.5--Significant

18

Hypothesis Testing

Not different (Ho)

Reality

Test

Decis

ion

Different (H1)

Not different (Ho)

Different (H1)

Correct Conclusion

Correct Conclusiona Risk

Type I Error Producer Risk

Type II Error b risk Consumer Risk

Test Reality = Different

Decision Point

b Risk a Risk

19

Process Capability Cp

USLLSL

Voice of the

Customer

Voice of The Process

Voice of the Customer

Voice of the Process

Capability Ratio - compares the capability of a process (voice of the

process) to the specification limits (voice of the customer):

=USL - LSL

6s= Cp

Cp = 1: The process is

barely capable (Just fits into

the tolerance window).

Cp = 2: The process is a

six sigma process (The

tolerance window is twice

the process capability).

20

Process Capability Indices

Process capability relates actual product variability to the customer

specifications.

The Cp index estimates process capability if the process mean is

centered in the specification window (Process Entitlement).

The Cpk estimates the process capability if the process mean is off

center.

)Capability6(

Spec Lower -Spec UpperCp

3)

Z

Cap3(

Spec NearestMean=Cpk

21

Illustration of Cp

Cp = 0.5

Cp = 1.0

Cp = 1.5

Cp = 2.0

Black Belt

Training

22

7565554 535

66,800

ppm

Barely Capable Process: With Mean Shift

Lower

Specification Upper

Specification

Graph>Box plot

75%

50%

25%

Graph>Box plot

Without X values

DBP

Box plots help to see the data distribution

Day

DBP

10

9

10

4

99

94

109

104

99

94

Operator

DBP

10

9

10

4

99

94

Shift

DBP

10

9

10

4

99

94

24

Types of Data Charts

Box Plot

Analyze -jun07

25

0

5

10

15

20

25

30

35

40

45

50

Data Set#1

Brown Blue Green Yellow Red Orange

Box Plot Titles

Q1

Min

Median

Max

Q3

Other formats

Using Statistical

Software

26

Pareto Chart A method to display the “vital few” from the “trivial many.” These charts are

based on the Pareto Principle – 20% of the problems have 80% of the impact.

The 20% represents the “vital few.” The Pareto chart helps you to arrange data

in order of priority or importance.

90

30

20 10 5

Freq

uenc

y

Categories

Per

cent

age

75%

25%

50%

20

180

140

100

60

TITLE

n = sample size

(time period)

LEGEND

REFERENCE INFO

date

initials

source

100 %

33%

52%

66%

78%

90%

100%

0

100

200

300

400

500

600

700

800

brown red orange yellow blue green

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

27

Your Completed Pareto Chart

Statistical Analysis

0.0250.0200.0150.0100.0050.000

7

6

5

4

3

2

1

0

New Machine

Fre

que

ncy

0.0250.0200.0150.0100.0050.000

30

20

10

0

Machine 6 mthsF

req

ue

ncy

Is the factor really important?

Do we understand the impact for

the factor?

Has our improvement made an

impact

What is the true impact?

Hypothesis Testing

Regression Analysis

5545352515 5

60

50

40

30

20

10

0

X

Y

R-Sq = 86.0 %

Y = 2.19469 + 0.918549X

95% PI

Regression

Regression Plot

Apply statistics to validate actions & improvements

29

Improve Phase Process Flow(1) Select - (2) Apply - (3) Implement/Deploy

PROBLEM

FINDING

FACT

FINDING

PROBLEM

DEFINITION

IDEA

FINDING

EVAL. &

SELECT

PLAN

ACCEPT

-ANCE

ACTION

1

2

34

6

7

8

5

Select

Tool

Analyze Data & Information Apply tools & Make Changes

Six

Sigma

Projects5S PM/TPM TQM

Rapid Improve-

ment Event

Integrated

Process

Team (IPT)

Establish Baselines/Metrics

Other CI

ToolsKanban DFSS

Ap

ply

th

e T

oo

l

(1)

(2)

(3) Implement / Deploy Fixes

(0) Identify failure points, bottlenecks and improvement

opportunities from Define-Measure-Analyze Phases

PM = Preventive Maintenance

TPM = Total Productive Maintenance

CI = Continuous Improvement

DFSS = Design for Six Sigma

TQM = Total Quality Management

THE

PROCESS

PROCESS

OUTPUTS

CONTROLLED

VARIABLES

CUSTOMER

PROCESS

INPUTS

UNCONTROLLED

NOISE

VARIABLES

Design of Experiments (DOE)

P-value= 0.045

DESIGN OF EXPERIMENT Final Yield= 35.44-0.9xFFT

P-value= 0.045

FINAL Yield= 35.44-0.901xFFT

DOE RESULTS FINAL YIELD=35.44-0.901xFFT

FINAL YIELD=35.44-0.901xFFT

• FINAL FILTRATION TIME (FFT) SIGNFICANTLY impacts Final Yield (inversely)

• The greater Final Filtration Time the

lower the Final Yield

• If you control FFT you control Final Yield

• So what do we need to do to minimize final filtration time?

Virus Removal

FINAL FILTRATION

DOE RESULTS—Implement Final Filtration Controls

• Replace filters whenever FFT > 0.5 Hrs

• Monitor FFT SPC trends

Control Phase

Control Tools:

• Control Plan

• POKA YOKE

• Control Charts

• Statistical Process Control (SPC)

• Go/No-Go checks

• Preventative Maintenance (Spare Filters)

39

40

Control Phase Process FlowSustaining the Improvements

Maintaining and Improving Control: Managing Change

Quality Management Systems

What’s Left? What’s Next?

Control Tools:

• POKA YOKE

• Statistical Process Control (SPC)

• “Spec” Control

Document

Procedures and

Best Practices

Create Visual

Controls and

“Dashboards”

Update

Deployment Plan

Complete

Control

FMEA

Update and Deploy

Communication Plan

Execute

Deployment Plan

Document Other

Opportunities

Prepare and

Implement

Control Plan

Report out, Hand-

off and Closure

Prepare and

Implement Training

Plan

Process Flow Diagram for Control Phase:

41

In preparing a control plan, you should concentrate on:

Control Plan Elements

1) What needs to be monitored?2) Who is going to be keeping the process

functioning properly?3) How are they going to monitor?4) Where will the monitoring be conducted?5) Where should the plan reside?6) What will be done if the monitoring detects a

condition outside of the customer’s specifications?

42

Control Plan

Key Process

Input Variable

Key Process

Output Variable

Process

Specification

(LSL, USL,

Target)

Capability (Cp,

Cpk, DPMO) and

date

Measurement

TechniqueSample Size

Sample

FrequencyControl Element Contigency Remedial Audit

Phone Call

LSL - NA

TGT- 10 sec

USL - 30 secs

Cpk=.85

05 Dec 03

Call Center

Manager

Dashboard 30 1/2 hourly

CCM to adjust

seated agents

CCM to reprioritize

worload to meet

specification

Daily report to be

examined by

Supervisor N

New Customer

LSL - NA

TGT- 15 min

USL - 30 min

Cpk=. 48

05 Dec 03

New Acct Dept

Visual Display 100 Weekly

Nacct Mgr to

monitor and adjust

new Acct Reps

Nacct Mgr to

reprioritize

Monthly report to

NAM regional

manager Y

New Accounts 0 Field defects

DPMO = 66289

05 Dec 03

New Acct Dept

Visual Display 100 Weekly

Nacct Mgr to

monitor and adjust

new Acct Reps

Nacct Mgr to use

Pareto Analysis to

determine defect

category

Monthly report to

NAM regional

manager Y

Lead Time

LSL - NA

TGT- 5 Days

USL - 10 Days

Cpk = 1.43

05 Dec 03

Operations

Dashboard 30 Daily Ops Mgr Review

Identify trend root

cause

Director Review

Weekly N

Customer

Expedite Expedite Order 0 Expedited

DPMO = 24721

05 Dec 03

Operations

Dashboard 25 Weekly Ops Mgr Review

Identify trend root

cause

Director Review

Monthly Y

Key Process Input Variable

Key Process Output Variable

Process Specification or Target Output Value

Current Performance Value

Sample Size

Sample Frequency

Control Element

Contingency Plans

Remedial Plans

Is the

Performance

Audited

DPMO / Control Chart

20000

25000

30000

35000

40000

45000

50000

55000

60000

65000

70000

75000

80000

85000

90000

95000

100000

2/5/2001

2/19/2001

3/5/2001

3/19/2001

4/2/2001

4/16/2001

4/30/2001

5/14/2001

5/28/2001

6/11/2001

6/25/2001

7/9/2001

7/23/2001

8/6/2001

8/20/2001

9/3/2001

9/17/2001

10/1/2001

10/15/2001

10/29/2001

11/12/2001

11/26/2001

12/10/2001

12/24/2001

DPM

O

Baseline

DPMO

UCL

LCL

Mean

Goal

Process

Change

REGIONAL

CONTROL

BEGINS

PAST BKLOG PAST BKLOG P-DUE PAST

UNITS DUE PM DUE LUBE SVC DUE

REGION ASSIGN PM DAYS LUBE DAYS GOAL SVC

NETWORK 2327 1 0 107 4 93 77

CENTRAL 151 0 0 8 4.8 6 8

F.B.U. 74 0 0 0 0 3 2

MIDWEST 178 2 1 9 4.6 7 3

NORTHEAST 312 0 0 33 9.7 12 18

SOUTHERN 389 0 0 43 10.1 16 36

WESTERN 259 0 0 19 6.7 10 11

TOTAL 3689 3 0 219 5.4 150 155

OFFLINE 203 4 1.8 37 16.8 1 4

NEW ACCOUNT EXCEPTION SUMMARY 12/08/2003

Control “Dashboards”A Visual Display for the Key Control Elements and Metrics

Item

No. Problem/Opportunity Benefit

Person(s)

Responsible

Due

Date

%

Complete Comments

1

2

3

4

5

6

7

8

Kaizen Event Newspaper

Team: Kaizen Event:

Kaizen Newspaper

The Control Plan

explains how the

Dashboard

Works and what

to do if

performance

goes out of

controlImprovement Actions and Activities

Updated: __/__/__ By:________

Monitor FFT Filter integrity and performance

Proposed LFB SPARE PARTS LIST

Item # Equipment Description Qty

1 ANX bubble Trap 1

2 Methyl/Heparin bubble trap 1

3 UV light 2

4 Chromaflow Nozzle 1

5 Variable Column Seal Kit 2

6 Valve Seal 2

7 Chromaflow Frits 2

8 Digital Pressure Gauge 0-100psi 1

9 Digital Pressure Gauge -30-100psi 1

10 Seal Kit Nozzel 1

11 CIP Pump 1

12 ANX & 43-029 UF/DF Retentate Backup Flow Meter/Display 1

13 Heparin Backup Flow Meter/Display 1

14 Vaccuum Pump 1

15 Feed Pump Impeller Turbine 1

16 Citrate addition valve 2

17 360° Casters & hardware 1

18 Conductivity Sensor (Probe) 2

Control Charts

47

Deploy Phase Spreading change Across the Organization

Deploy:•Validate improved

performance that stays in

place

•Spread improvements

across the enterprise

•Harvest the performance

improvements

•Change the culture

•Where else can you tale the improvements

•Look further upstream

•Look further downstream

•What about other customers

•What about “re-fixing” processes you

already fixed

48

Kaizen Events

& TPM

SPC

Six Sigma (GB &BB)

D.O.E

Which Tools

to Apply

Intuitive/

Common

Sense

Data

Driven

Non-

Value

Added

Value

Added

Type of

Variation

Special

Cause

Variation

Common

Cause

Variation

Lean

vs

6

95%

vs

5%

CO

NT

INU

UM

5S

PIT

& B

rain

sto

rmin

g

Cre

ati

ve

Pro

ble

m S

olv

ing

Process

Content

Where do Visual Displays Fit In?

Visual Displays Identify and Communicate Where You Are

49

Kaizen Events

& TPM

SPC

Six Sigma (GB &BB)

D.O.E

Which Tools

to Apply

Intuitive/

Common

Sense

Data

Driven

Non-

Value

Added

Value

Added

Type of

Variation

Special

Cause

Variation

Common

Cause

Variation

Lean

vs

6

95%

vs

5%

CO

NT

INU

UM

5S

PIT

& B

rain

sto

rmin

g

Cre

ati

ve

Pro

ble

m S

olv

ing

Process

Content

Where do Kaizen Events Fit In?

© Ci2 2009

50

Kaizen Events

& TPM

SPC

Six Sigma (GB &BB)

D.O.E

Which Tools

to Apply

Intuitive/

Common

Sense

Data

Driven

Non-

Value

Added

Value

Added

Type of

Variation

Special

Cause

Variation

Common

Cause

Variation

Lean

vs

6

95%

vs

5%

CO

NT

INU

UM

5S

PIT

& B

rain

sto

rmin

g

Cre

ati

ve

Pro

ble

m S

olv

ing

Process

Content

Where do Green Belt/Black Belt Projects Fit In?

© Ci2 2009

Foundations for Wisdom and good decision finding the right problem

Numbers, words, quantities, values stored sitting in piles or queues waiting for future use

Charts, summaries, spread sheets, etc. that organize the data

Presentations, plans and tools that explain and communicate the information

Processes, organizations and team using the information to improve, manage, build systems and develop cultures

Push the data “up” to become Wise

Always loop back to check the data

Wisdom

Knowledge

Data

Information

Where are you looking from?

The WISDOM TOWER: Changing Perspectives