Quality Management

Heidi Maier Sagstad

 Agenda
Six Sigma tools can be effectively used in a Quality management program by
helping to identify CTQ’s, measure the project Y, and identify Key X’s:

Project Management
– Lean/Six Sigma
– DMAIC
Plan Quality
– Project Y = Identify Quality Measures
Quality Tools
– Cause and Effect Matrix
– Pareto
– Scatter Plot

2
 Introductions

• Name
• Company
• Experience with
– Project Management / Quality Management
– Lean / Six Sigma
• Expectations of today’s session

3
Project Management
Quality Management,
Lean / Six Sigma
DMAIC
 What is project quality management?

«Project Quality Management works to ensure that

the project requirements, including product
requirements, are met and validated.»

Minimize variation and deliver results that meet

defined requirements.

5
 Project Quality Management

Plan Quality Management

• Identify Quality Requirements

Perform Quality Assurance

• Auditing the quality requirements

Control Quality
• Monitoring and recording results to assess performance and make
changes

6
 Lean Six Sigma
• Total Quality Management System

• Total Employee involvement

Philosophy • Project Rigor

• Continuous Improvement

• Customer Focus

Method Tools • Deliver products to the customer:

• Time
• Cost
• Quality

7
 Lean Six Sigma - Philosophy
Six Sigma:
Do it Right the First Time
Understand the problem
Fix the right problem
Philosophy Sustain Results

Lean:
Method Tools Just In Time
GO SEE
Fix It NOW
Faster, Cheaper, Better

8
 Lean Six Sigma - Method

Six Sigma:
DMAIC
DIDOV Philosophy
DFSS

Method Tools
Lean:
DMAIC
KAIZEN
Lean Start Up

9
 Lean Six Sigma - Tools
Six Sigma:
Accuracy and Precision
Reduce variation
Philosophy
Identify critical X

Method Tools
Lean:
Improve flow
Reduce Cycle time

Identify Waste

10
11
 Project manager structure

Champion

Champion
Master Black Belt
Black Belt

Green Belt

Yellow Belt
12
 Lean

Lean Tools are used to:

• Eliminate Waste
• Improve Flow
• Reduce Cycle Time

You may use lean tools to address a quality issue!

13
Lean – 7 types of waste

Motion
Overproduction
Waiting
Transport
Inventory
Unnecessary Processing
Defects

(Under-utilized people)
 Lean - Improve Flow

15
 Lean - Reduce Cycle Time

Reduction in cycle time is achieved by making improvements which

ultimately improve overall quality performance.

16
 Six Sigma

You may use Six Sigma tools to measure and/or identify a quality issue!

17
18
19
 Six Sigma

Free Online Calculator

20
 Six Sigma 99,9% perfect

1 3 6
1106 107 <1

Lost your mail

470 45 <1

Your office coffee pot empty

4839 467 ,02

Dropped your phone call

21
 99.9% Perfect is not good enough

• 1 hour of unsafe drinking water every month

• 2 unsafe plane landings per day in Chicago
• 16,000 pieces of lost US mail every HOUR
• 20,000 incorrect drug prescriptions every year
• 500 incorrect operations each week
• 50 babies dropped at birth every day
• 22,000 bills drawn from the wrong bank account
every HOUR

22
 DMAIC
a problem solving approach
D Define
Clearly define the problem and form your team

Measure
M Measure what is key to the customer and confirm that your data is
reliable

A Analyze
Search for the root causes and identify the most significant factors

Improve
I Establish methods to improve the root causes and significant factors

C Control
Make sure the problem doesn’t come back

23
PDCA - Roots of the DMAIC Process

Change the Process

- Adopt the change or abandon it Plan the Improvement
- Standardize the change - Identify process to improve
- Gather additional data - Develop the test plan
Act
Plan

Check
Check the Results Do
- Identify the likely root causes
- Determine how to improve
Conduct the Test
- Identify potential causes
- Compare causes to facts

24
The DMAIC Roadmap

• Project charter • Y metric • Critical inputs and/or • Improvement plan • SOP revisions
Deliverables

• RACI matrix • Specification limits root causes • Improved process map • Control plan
• Communication plan • Detailed process map • High-level cost/benefit • Financial impact • Training
analysis
• Stakeholder analysis • Data collection plan • New process capability
• Preliminary
• Kick-off meeting • MSA • Commissioning meeting
improvement plan
• Normality of Y • Actual realized financial
• Stability of Y benefits
• Baseline capability
• Cost of non-quality

O
Phase

I
u
n

Define Measure
t
p
u
Process p

Improve Control
u
Analyze
t
t
s
s

• SIPOC Process Map • Swim-Lane Diagram • Brainstorming • Process Flowcharting • Risk Assessment of
• Bar Chart Value Stream Mapping • Fishbone Diagram • Value Stream Mapping Changes-FMEA
Activities/Tools

•
• Cause & Effect Matrix • Pareto Charts • Change Point Analysis • Mistake Proofing • Control/Action Plan
• Pareto Chart • Fishbone Diagram • Multi-Vari Charts • 5S/Kaizen • Standard Operating
• RACI Matrix • Concentration Diagrams • Scatter Diagrams • FMEA Procedures
• FMEA • Hypothesis Testing • Process Tolerancing • Project Commissioning
• Measurement System • ANOVA • Process Validation
Analysis/Gage R&R • Correlation & Regression • Process Control Charts
• Process Control Charts Analysis • Capability Studies
• Capability Studies • 7MP tools • Visual Controls

25
 Problem Solving Storyboard
1. Describe the Problem Goal Statement
Inputs Outputs 2. Identify Potential Causes Failure Modes & Effects Analysis

1. Achieve a defect M
D IS level less than Item Function Failure Mode Causes Effects O S D C/A

IS 500 dpm. START E

NOT 2. Improve throughput
E by 30 %. Input/Output Distinct Change
WHAT A EFFECT
F Diagram
Goal
WHEN S

Equipm’t
Adjusted
I

Process

Change
p=0.1%

p=8.6%
Failure Modes &
WHERE
U
N Effects Analysis
END R
E EXTENT Cause & Effect
Process or E Diagram (5M) Distinctions
2/96 7/98 8/98 4/99
Is / Is Not Diagram Usage Flow & Changes
Time Line
LSL USL
3. Collect & Organize Existing Data 4. Compare Causes to the Facts
CUSUM Chart of Defect Level A Facts From Step 1
18

100 xx
CUSUM

xx x N

Fact 4
Fact 2
Fact 3
Fact 1
8.5

M T WT F
Concentration
Cum Percent
Frequency

-1
A X Capability A

Causes from Step 2

3 6 9 12 15 18 21 24
Row
Diagram Cause 1 O
Change Point B X X X Study L
O O
Cause 2 O X
Analysis C X X Cause 3 O O
UCL D X X Y Cause 4 O O X
0
Defect Type Cause 5 O X
Checksheet Z
Pareto Analysis
LCL
Scatter Multi-Vari E
Diagram Contradiction Matrix
Control Chart Chart

5. Collect Additional Data to 6. Determine Best Solution(s) 7. Verify, Implement, Standardize

Identify Root Cause(s) Seal
Force Procedures LSL USL

I C
Seal Strength

THigh
TASK SCHEDULE
M Elimination O 1
ANOVA
TLow 2
Contingency P N
Output

Response Facilitation 3 Capability

Tables 4
Low High Surfaces R T Study
Pair t-Test Mitigation 5
Screening
F-test UDL
O R
Experiment Flagging
Implementation Plan O-C Curve
LDL Input
Other UDL V O UCL Pa

UDL
Statistical Robust Tolerance
Mistake Proofing
LDL
LDL E Analysis or VTA L
Tests % Defectives
Component LCL
Sampling
Analysis of Means Swapping Study Control Chart Plan
 Project Definition
Define

Buffer Preparation
Ing. Add Mixing Adjust/Release Transfer Cooling CIP

Process
Suppliers Inputs Output(s) Customer(s)
QMRI Assay Start:
Cpk LAFC Process
Initial Sample
Specs Sampling & Buffer First Pass Success LA Plant
Adjustments
Equipment Cycle Time Lessines
End:
Released Buffer
Paste your SIPOC here.

Have a clear, specific and

Goal: Improve yield by 10% attainable goal.

Est Completion Date for Measure: April 13, 2008

 Measure
Current Process Map
Attach a map of your current process (i.e. Current VSM,
Swim-lane, top-down chart).

Customer

Engineering

Start

Tooling

Production

End

Field Service
 Measure
Measurement System Analysis
How do you trust your data?

CYCLE COUNT ACTUAL INVENTORY LOGGED INVENTORY DELTA INVENTORY

1 $ 75,491.83 $ 75,512.20 $ 20.37
Gage R&R
2 $ 93,804.34 $ 98,686.00 $ 4,881.66 %Contribution
3 $ 49,105.00 $ 81,592.00 $ 32,487.00
4 $ 83,856.74 $ 83,810.61 $ (46.13) Source VarComp (of VarComp)
5 $ 121,657.00 $ 127,100.00 $ 5,443.00
6 $ 164,750.00 $ 187,314.00 $ 22,564.00 Total Gage R&R 332.21 20.74
7 $ 111,968.10 $ 111,931.30 $ (36.80)
8 $ 123,943.50 $ 122,800.40 $ (1,143.10)
Repeatability 154.55 9.65
9 $ 255,238.40 $ 247,571.30 $ (7,667.10)
10 $ Attach Measurement System Analysis forReproducibility
42,468.67 $ 42,468.67 $
177.66
your Y. Use Transactional
- MSA, 11.09
Gage
11 $ R&R, etc as applicable to your data.
54,409.91 $ 54,409.91 Part-To-Part
$ 1269.44
- 79.26
12 $ 224,343.00 $ 225,277.00 $ 934.00
13 $ 101,199.40 $ 101,199.10 $ (0.30) Total Variation 1601.65 100.00
14 $ 99,348.83 $ 96,632.61 $ (2,716.22)
15 $ 84,420.45 $ 85,984.91 $ 1,564.46
16 $ 123,372.00 $ 123,973.10 $ 601.10
17 $ 114,937.90 $ 114,938.20 $ 0.30
18 $ 87,678.17 $ 89,656.06 $ 1,977.89 Study Var %Study Var
19 $ 83,531.17 $ 83,610.20 $ 79.03
20 $ 136,368.30 $ 133,259.30 $ (3,109.00) Source StdDev (SD) (5.15 * SD) (%SV)
21 $ 80,436.30 $ 77,335.59 $ (3,100.71)
22 $ 100,647.30 $ 100,647.00 $ (0.30) Total Gage R&R 18.2267 93.867 45.54
23 $ 91,068.44 $ 91,068.44 $ -
24 $ 49,007.76 $ 53,499.08 $ 4,491.32
Repeatability 12.4318 64.024 31.06
Reproducibility 13.3290 68.644 33.31
Standard Deviation $ 50,503.24 $ 8,330.35
Variance $ 2,550,577,534.45 $ 69,394,668.73 Part-To-Part 35.6292 183.490 89.03
Error of Transaction = 2.72% Total Variation 40.0206 206.106 100.00
Number of Distinct Categories = 2
 Measure
Baseline Data - Normality
Normal or non-normal? Use “Graphical Summary”.

Summary for Reading at 405

A nderson-D arling N ormality Test
A -S quared 4.67 P < 0.05
P -V alue < 0.005

M ean 0.20541 Not

S tD ev 0.10428
V ariance 0.01088 Normal
S kew ness 0.874044
Kurtosis 0.563682
N 186

M inimum 0.04700
1st Q uartile 0.11600
M edian 0.17750
3rd Q uartile 0.28725
0.1 0.2 0.3 0.4 0.5 0.6
M aximum 0.63800
95% C onfidence Interv al for M ean
0.19033 0.22050
95% C onfidence Interv al for M edian
0.16215 0.19709
95% C onfidence Interv al for S tDev
95% Confidence Intervals
0.09465 0.11611
Mean

Median

0.16 0.17 0.18 0.19 0.20 0.21 0.22

 Measure
Baseline Data - Stability
Stable or unstable? Use appropriate Control Charts. Explain out-of-control points.

I-MR Chart of Reading at 405

Shut-down,
1
0.60 new equipment,
Individual V alue

1
1 1 etc
0.45 1 11
1
U C L=0.3771
0.30
_
X=0.2054
0.15

LC L=0.0337
0.00
1 19 38 57 76 95 114 133 152 171
O bser vation

1
0.48

0.36
M oving Range

1
1 1 1 1 1
0.24 11 1 1 1
U C L=0.2109

0.12 __
M R=0.0646
0.00 LC L=0
1 19 38 57 76 95 114 133 152 171
O bser vation
 Measure
Baseline Data - Capability
Capable or incapable?

Process Capability of Reading at 405

Calculations Based on Weibull Distribution Model

USL
P rocess D ata O v erall C apability
LS L * Pp *
Target * PPL *
USL 0.40000 PPU 0.55
S ample M ean 0.20541 P pk 0.55
S ample N 186
E xp. O v erall P erformance
S hape 2.11863
P P M < LS L *
S cale 0.23311
P P M > U S L 43316.5
O bserv ed P erformance P P M Total 43316.5
P P M < LS L *
P P M > U S L 37634.4
P P M Total 37634.4

-0.0 0.1 0.2 0.3 0.4 0.5 0.6

 Analyze
Identifying Significant Xs
Supporting data and/or charts that explain what led you to your significant Xs.
What impact do they have on your Y?

Possible tools: Pareto, VSM, Fishbone, CE matrix, ANOVA, Moods Median, Box&Whisker
charts, regression, etc. (Use only the tools that are appropriate.) See your coach for help.

Methods
Materials Man

Size of
Pareto Chart of Reason screen save (C)
1400 Write to database (X)
Number of clicks
1200 100 Burn overlay
until screen capture (X)
into pixel values (X) Disk Space (N)
1000 80 Percent
800
Count

60 Integration with Inheritance of DB Schema supports

600 current Display GUI (X) patient attributes (X)Secondary Captures (X)
Time to
40
400 generate
200 20 Secondary
Reformat Tools Capture
0 0 Orthogonal (X)
Reason r
Pet Error
he or es es es ou
s es e
he
r
at is su su su su ag Logging (N)
e e rv Is Is Is ne Is ut Ot Orthogonal WB (X) Patlak(X) Protocol Tools
p r la l O
W g te er l e er
Su tin pu ez is
ce ab
ed Hea m re M s/L Pow Rectilinear (X)
M o F FDG (X)
C rm MID (X) Status on GUI (X)
S/ Fo
DI Rectilinear WB (X)
Fusion (X)
Count 511.5 497.0 52.0 51.5 50.0 45.5 15.0 8.5 8.5 SID (X)
Polar Display (X) Stopwatch (C)
Percent 41.3 40.1 4.2 4.2 4.0 3.7 1.2 0.7 0.7
CBF (X)
Cum % 41.3 81.4 85.6 89.7 93.7 97.4 98.6 99.3 100.0 Oblique (X) Plot (X)
Normals(X)
Cardiac (X) Measurement
Machine
MIP (X)
 Improve
Short Term Process Capability

Process Capability of delta t 2

Calculations Based on Weibull Distribution Model

USL
P rocess D ata O v erall C apability
LS L * Pp *
Target * PPL *
USL 3 PPU 0,02
S ample M ean 4,34462 P pk 0,02
S ample N 13
E xp. O v erall P erformance
S hape 0,648483
P P M < LS L *
S cale 3,05148
P P M > U S L 371939
O bserv ed P erformance P P M Total 371939
P P M < LS L *
P P M > U S L 307692
P P M Total 307692

0 5 10 15 20 25 30

Baseline SQL to Improve SQL: 1.1 to 1.8

 New Process Capability
Control

Process Capability of delta t 2

Calculations Based on Weibull Distribution Model

USL
P rocess D ata O v erall C apability
LS L * Pp *
Target * PPL *
USL 3 PPU 0,02
S ample M ean 4,34462 P pk 0,02
S ample N 13
E xp. O v erall P erformance
S hape 0,648483
P P M < LS L *
S cale 3,05148
P P M > U S L 371939
O bserv ed P erformance P P M Total 371939
P P M < LS L *
P P M > U S L 307692
P P M Total 307692

0 5 10 15 20 25 30

Baseline SQL to Control SQL: 1.1 to 1.8

 Control
Confirm Statistical Difference
Two-Sample T-Test and CI: Before, After
Use any hypothesis test such as 1-sample or 2-sample t Tests,
Mood’s Median Test, 2 Variances Test, Chi-square Test Two-sample T for Before vs After
N Mean StDev SE Mean
Boxplot of Before, After Before 16 27.8 17.3 4.3
70
After 17 5.00 3.41 0.83
60 Difference = mu (Before) - mu (After)
50 Estimate for difference: 22.8125

40
95% CI for difference: (13.4767,
32.1483)
Data

30
T-Test of difference = 0 (vs not =): T-
Value = 5.18 P-Value = 0.000 DF = 16
20

10
If the P value is less than 0.05 then
0
Ho has to go;
Before After i.e. there is a statistical difference.

P value: 0.00

Is P value less yes There is a statistical

than 0.05? difference.
 Results

37
 Break time…

38
Identify Quality Measures
Cost of Quality
The 1-10-100 Rule!

The cost effect of defects when the are:

DISCOVERED DISCOVERED DISCOVERED

IN THE PROCESS INTERNALLY BY CUSTOMER
(AFTER PROCESS
COMPLETION)

1X 10X 100X

Prevention Detection Field Defect

40
What is Quality?
The degree of conformance to the customer requirements

41
 Who are the Customers?

• Internal Customer - Anyone inside the company who

receives the output of a process or the results of our work.

• External Customer - Anyone outside the company who

receives the output of a process or the results of our work.

• Process Owner - Anyone inside the company who controls

the inputs of a process and is responsible for the outputs.

• Project Sponsor - Anyone inside the company who requests

the project, and receives direct benefit from the project
results.
42
What are their Requirements?

• Requirements are related to measureable characteristics

in the outputs of our processes.

• These characteristics can be expressed in either

quantitative or qualitative ways.

• Internal requirements should to be connected to one or

more customer CTQs.

43
 How do we know what to measure?

Inputs Process Outputs

X1 Step 1 Dec 1 Step 2A

All X2
All
Y1
processes
Step 2B

processes
have have
X3 Y2
outputs
Step 3A Dec 2
inputs
X4 Step 3B

44
SIPOC
Title
Process 1 Process 2 Process 3 Process 4 Process 5 Process 6

Process
Suppliers Inputs Output(s) Customer(s)
Start:

End:

Supplier CTQ’s: Customer CTQ’s:

Example of a CTQ Tree
Need Drivers CTQ’s

Not acidic
Taste Rich)

Aromatic
>60 C
Customer Says: Temperature
70 C
“I want a coffee that is good.”
<80 C

Cost < 20 kr

Loyalty Stamp

General Specific

Hard to measure Easy to measure

46
 The VOC Flow Down Logic
Term

Voice of the Critical To Key Output Specification

Customer Quality Variable Limits
Abbreviation

VOC
VOC CTQ
CTQ YY USL, LSL
USL, LSL
Explanation

Customer
Specific Acceptance
Customer needs
measure that limits for the
concerns or translated into
reflects the measure
needs the process
CTQ collected
language

47
How do we define quality?

VOC: “I want my product as I agreed to when I want it.”

CTQ: “Product released in full, on time and at the potency demanded.”

Y: fulfillment = percentage of orders completed and shipped on time

Specification Limits: 98% shipped within 21 days

after issuance of the order

48
Specifications for CTQs
• In manufacturing,
100

Set specification
specification limits often 75 here
come from technical or

% Yield
50

mechanical requirements. 25

• Otherwise, base 0
0 5 10 15 20 25 30
Time (secs)
specification limits on data
about customer needs.

• Set specifications at the Set lower

specification here
Set upper
specification here
places where customer 100%

satisfaction starts to fall off

% satisfied

appreciably.
0%
66° 67° 68° 69° 70° 71° 72° 73° 74°
69° 75° 76°
Room temperature (°F)

49
Other Examples

Customer
Concern or Need Specification
(VOC) Customer CTQ Key Process Y Limits
“Our fill volumes are fill volume fill volume 50 ml as USL
not homogenous
across the lots, some
exceeding spec limits.”

“We need to find a release processing release processing cycle 28 days as USL
way to release our time time, which is the
products sooner.” difference between end of
production and final
issuance of approval sheet

“We need to get as process yield Yield = ratio of Yield greater

much saleable product than 85%
from each (weight/density*potency)in (LSL>=85)
manufactured batch as
possible.” (flow*time)*potency)out

50
The List of Terms

Abbreviation Description
DMAIC Define-Measure-Analyze-Improve-Control
Process Y Output of a process
Process X Input of a process
VOC Voice of the Customer
VOB Voice of the Business
CTQ Critical to Quality
Project Y Process Output Used by the Six Sigma Project
USL, LSL Upper and Lower Specification Limits

51
Data Collection Plans

• Select a project Y
– Can you measure it?
– How readily available are the data?
– Who will collect it?
– How often will it be collected? Sample size?
– Where will the data be stored?

Project Y

52
Data Collection Plan Features
Data Collection Plan Project ________________________
What questions do you want to answer?

Being clear about your question will help you make sure you collect the right data.

Data Operational Definition and Procedures

What Measure How Related Sampling How/where
type/ Data measured 1 conditions to notes recorded (attach
type record 2 form)

Recording what data NOTES

you are going to 1) Be sure to test and monitor
collect reminds you any measurement
what you want to An operational definition defines procedures/ instruments.
accomplish. exactly how you will go about 2) “Related factors” are
Noting the type of data collecting and recording the data. stratification factors or
helps you decide how potential causes you want to
you should analyze monitor as you collect data
the data.
53
Data Collection Plan Features, cont.

How will you insure consistency and What is your plan for starting data
stability? collection? (Attach details if necessary)

Just how will you go about collecting

the data?

What will you do to make sure the data

collected at one point in time is How will the data be displayed? (Sketch
comparable to data collected at other below)
times? That is, that no biases have
been introduced in the way the data is
collected? Thinking about how you will display
the data will help you make sure you’re
getting the right kind of data to answer
the question you have in mind.

54
Quality Tools
Fishbone
Cause and Effect matrix
Pareto
Scatter Plot
56
Cause & Effect Diagram

• Refines brainstormed ideas, and helps to identify

likely causes of an observed problem.

• Provides a discussion context that assists in

determining likely root causes.

• Sorts the factors that may affect a process when

little quantifiable data are available.

• To graphically displays the team's level of

problem understanding.

57
Cause & Effect Diagram

METHODS MATERIALS MAN

Cause #2 Cause #2
Cause #2 Why? Why?
Cause #1
Why? Why?
Why? Why?
Why? Why? Cause #1 Why?
Cause #1 Why? Why?
Why?
Why? Why?
Why?
Problem or
Why?
Effect
Cause #1

Cause #1 Cause #1 Why?

Why? Why?
Why? Cause #2
Why? Why?
Why? Cause #2
Why? Why?
Why?
ENVIRONMENT MEASURES MACHINE

Also known as “Fish Bone Diagrams”

Cause & Effect Diagram

PEOPLE MATERIAL MACHINE

Pesticides Crushed roots

Construction
Fertilizer

Trees
are
Dying

Acid rain Pruning

ENVIRONMENT METHOD MEASUREMENT

59
 Cause and Effect Matrix

60
Pareto Charts

• Definition
– A Pareto chart is a graphical tool used to rank items.
– It is a bar chart used to prioritize items in a process
and to identify the area where to focus the attention.
– 80% of the problem is related to 20% of the process

• Examples
– A process engineer wants to know which visual
defects are detected most frequently in order to
improve yield.
– A quality control administrator wants to identify the
analytical tests which contribute most to excessive
lead time.
61
 Pareto

62
Pareto Charts

• Pareto charts help us to prioritize observed causes of

problems so we can focus limited resources on areas
where we can get the biggest results.

• It can help us separate the “vital few" opportunities

from the “trivial many" observations.

• Often we use the “80/20 rule”, sometimes called

Pareto's rule, when we want to identify the small
fraction of the causes producing most of the defects or
problems.

63
 Fitted Line Plot (scatter plot)

• Determine correlation between an X and a Y

• How strong is the relationship
• Method to further evaluate vital X’s.

64
 Fitted Line Plot (scatter plot)

The higher the R-sq value, the stronger the relationship, but that does not
necessarily mean it is the cause.

Happiness vs. Income: is it the income that makes people happier, or that they
are better educated, work in better conditions, fields they enjoy?

65
 Summary

Six Sigma tools can be effectively used in a Quality

management program.

• Identify CTQ’s
• Measure the project Y
• Identify Key X’s

66