QUALITY MANAGEMENT

Complete Course Notes

Lean | Six Sigma | DMAIC

UNIT I: INTRODUCTION TO LEAN & SIX SIGMA

Background
Studying quality management is crucial for various reasons and plays a significant role in the success and
sustainability of organizations. The key reasons are:

• Customer Satisfaction: Quality management focuses on meeting or exceeding customer expectations.

Satisfied customers are more likely to be loyal, repeat buyers, and positive promoters of a business.
• Competitive Advantage: Organizations that consistently deliver high-quality products or services gain a
competitive edge in the market. Quality can be a key differentiator that sets a company apart from its
competitors.
• Cost Reduction: Proper implementation of quality management practices can lead to a reduction in
defects, errors, and waste. This, in turn, lowers operational costs and improves overall efficiency.
• Employee Engagement: Quality management involves engaging employees in the continuous
improvement process. This can boost morale, motivation, and a sense of ownership among employees,
leading to increased productivity.
• Risk Management: Quality management helps identify potential risks and issues early in the production or
service delivery process. Addressing these issues proactively can prevent larger problems and mitigate
risks.
• Regulatory Compliance: Many industries have specific quality standards and regulations that
organizations must adhere to. Studying quality management ensures compliance with these standards,
avoiding legal issues and reputational damage.
• Continuous Improvement: The core principle of quality management is continuous improvement.
Studying this discipline encourages organizations to continually assess and enhance their processes,
leading to ongoing innovation and adaptation.
• Enhanced Reputation: Consistently delivering high-quality products or services builds a positive
reputation for a company. A good reputation is essential for attracting new customers and retaining
existing ones.
• Decision-Making: Quality management provides data-driven insights and metrics that aid in informed
decision-making. This helps leaders make strategic choices based on actual performance and customer
feedback.
• Globalization: In an increasingly interconnected world, understanding and implementing quality
management principles can help organizations compete on a global scale. Consistent quality is a
prerequisite for entering and thriving in international markets.

By studying quality management, individuals and organizations can gain the knowledge and skills needed to
achieve these benefits and create a culture of excellence within their operations.

Course Objective
This course aims at providing a good understanding of the Six Sigma process which seeks to improve the quality
of the output of a process by identifying and removing the causes of defects and minimizing variability in
manufacturing and business processes.
Real-Life Example of Six Sigma: The Mumbai Dabbawallas
The Mumbai Dabbawallas (tiffin wallas) represent a remarkable real-life example of Six Sigma in the service
sector:
• Context: Mumbai dabbawallas
• Volumes: Every day 5,000 dabbawallas deliver 175,000 tiffin boxes
• Defects: One mistake every two months
• Performance: 1 error in 8 million opportunities (16 million including return trips)

Dabbawallas have their own unique community with a distinct culture, skills and lifestyle.
This translates to approximately Six Sigma performance — world class quality!

The Dabbawalla Process Steps (Flowchart)

• Step 1: Meals (dabbas) are picked from commuters’ homes in the suburbs and taken to the nearest
railway station
• Step 2: A second group sorts out the dabbas at the railway station and loads them onto trains
• Step 3: A third group travels with the dabbas to the train stations closest to the final destination
• Step 4: A fourth group picks up the dabbas and delivers them to recipients
• Step 5: The process is reversed in the evening to return empty tiffin boxes to the workers’ homes

Comments and Lessons Learned

• Dabbawallas rely on very low capital endowments
• Several groups work independently but have established networking practices to achieve their goals
• They hold improvement meetings once a month
• There is no retirement age
• Most dabbawallas are illiterate and use a simple information system made of colour codes
• They have their own credit society
• Dabbawallas receive social recognition and respect for the quality and importance of their work (and not
only at home)

It is not only about Technology!

Why is this Example Relevant in Six Sigma Terms?

The example comes from the service industry (Six Sigma is not limited to manufacturing processes).
Some of the main tenets of Six Sigma drive their daily work practice:
• Customer focus: Dabbawallas are very attentive to their customer needs (they know their CTQs (Critical to
Quality requirements))
• They use metrics and have a clear definition of defect
• Their process emphasises team work (as the only possible way to deliver a close to zero defects service)
• They have a strong culture of “process excellence”

This leads to the question: ‘What is an organisational culture and how does it influence the behaviour of the
organisation, its ability to manage change and strive for excellence?’

Quality Management — A Conceptual Framework

 • Oxford dictionary – Quality is defined as the degree of excellence
• Quality guru J. M. Juran – Fitness for purpose
• Quality guru Philip Crosby – Conformance to specifications
• Quality Control (QC) – is defined as maintaining requisite standards in products or services
• ISO 8402 defines Quality Control as – The operational techniques and activities that are used to fulfil
requirements of quality
• ISO 8402 defines Quality Assurance as – All those planned and systematic actions necessary to provide
adequate confidence that a product or service will satisfy the given requirements for quality

Quality guru A. V. Feigenbaum defines Total Quality Control (TQC) as:

Total Quality Control is an effective system for integrating the quality-development, quality-
maintenance, and quality-improvement efforts of the various groups in an organisation to
enable marketing, engineering, production, and service at the most economic levels which allow
for full customer satisfaction.

Various Dimensions of Quality

Quality should be perceived from the customers point of view. This is because it is the customer who decides to
buy or not to buy a product or service, according to his or her perceptions of quality. Thus it is important to
know the various dimensions of quality that are considered by a customer in assessing the quality of the
product.
Garvin (1988,1990), while focusing on the strategic potential of quality, recognized the eight dimensions of
quality as the basis for developing strategic options.

Following are the dimensions of quality:

• Performance: How well does the product perform with respect to its intended use? For example: Good
picture and sound effects of a TV.
• Features: What special features does the product have? For example: Watt of sound, Flat picture tube,
picture-in-picture feature, in a TV.
• Reliability: How much is the probability of breakdowns, need for adjustments, replacement of parts, etc.
in the product? For example: A TV performing well every time it is switched on.
• Serviceability: How easily, cheaply, and speedily can the product be repaired and serviced? For example: A
company providing on-spot repairs of TVs within 1 hour of customer complaint at nominal charges.
• Durability: How long can the product perform well without substantial repair or replacement of parts? For
example: Performance of a TV for 10 years without needing any repair.
• Appearance: How pleasant is the outward look, smell, taste, feel or sound of the product? For example:
Super-thin flat TV (which can be hung on a wall like a picture).
• Customer Service: How is the behaviour of the seller with the customer before, during, and after the sale
of the product? For example: Treatment given by the sales staff may prompt a customer to buy another TV
of the same brand from the same dealer.
• Safety: How much care has the company taken to make the product safe for users before, during or after
use? For example: A TV having features to protect the eyes of viewers from harmful radiations.

Companies must decide which subset of these dimensions differentiates their products or services from that of
its competitors.
Then he introduced the framework of strategic quality management, emphasizing that quality must be defined
from the customer’s point of view. Garvin further elaborated it by stating that quality should be linked to
profitability on both the market and the cost side. It should be linked to the strategic planning process which
requires organization-wide commitment. Also, quality should be viewed as a company weapon.

Reliability vs. Durability

• Reliability is the ability of an item to perform its function under specified conditions over a given period. It
is usually expressed in terms of mean time between failures (MTBF), which is the average length of time
that a device will operate without failure.
• Durability is the ability of an object to withstand physical damage from normal use over a long period. This
may include wear and tear, corrosion, impact, etc. In some cases, durability is also referred to as
“endurance.”

Bloom's Taxonomy
Bloom's Taxonomy is a hierarchical framework for categorizing educational learning objectives from lower-order
to higher-order thinking:
• Remember (lowest): Recall facts and basic concepts — define, duplicate, list, memorize, repeat, state
• Understand: Explain ideas or concepts — classify, describe, discuss, explain, identify, locate, recognize
• Apply: Use information in new situations — execute, implement, solve, use, demonstrate, interpret
• Analyze: Draw connections among ideas — differentiate, organize, relate, compare, contrast, distinguish
• Evaluate: Justify a stand or decision — appraise, argue, defend, judge, select, support, critique, weigh
• Create (highest): Produce new or original work — design, assemble, construct, develop, formulate,
investigate

Six Sigma Overview

Six Sigma Defined — The Multifaceted Nature
The multifaceted nature of Six Sigma:
• As a Benchmarking Method
• As a Metric (statistical measure of performance)
• As a Strategy
• As a Branding Tool
• As a Management Philosophy
• As Statistical Tools
• As a Symbol
• As a Methods

Six Sigma is a philosophy, a measure, and a methodology that provides businesses with a perspective and tools
to achieve new levels of performance in both services and products.

Six Sigma History — Timeline (1985–2002+)

Evolution of Six Sigma (Staircase Model)
• Stage 1: Measure
• Stage 2: Benchmark
 • Stage 3: Methodology
• Stage 4: Organisational Infrastructure
• Stage 5: Management Philosophy
• Stage 6: Branding Tool

All stages built on: Principles, Tools, Techniques & Best Practices

Six Sigma as a (Management) Philosophy

Six Sigma as a philosophy helps companies achieve very low defects per million opportunities over long-term
exposure.

• “This Six Sigma journey will change the paradigm from fixing products so they are perfect to fixing
processes so that they produce nothing but perfection, or close to it.” — Jack Welch (GE)
• “It is a business process that allows companies to drastically improve their bottom line by designing and
monitoring everyday business activities in ways that minimize waste and resources while increasing
customer satisfaction.” — Mikel Harry, Ph.D., and Richard Schroeder, 2000
• “First, what it is not. It is not a secret society, a slogan or a cliche. Six Sigma is a highly disciplined process
that helps us focus on developing and delivering near-perfect products and services. The central idea
behind Six Sigma is that if you can measure how many “defects” you have in a process, you can
systematically figure out how to eliminate them and get as close to “zero defects” as possible. Six Sigma
has changed the DNA of GE — it is now the way we work - in everything we do and in every product we
design.” — General Electric
• “Six Sigma is Powerful Stuff” — Michael Dell
• “Six Sigma refers to our overall strategy to improve growth and productivity as well as a measurement of
quality. As a strategy, Six Sigma is a way for us to achieve performance breakthroughs. It applies to every
function in our company, not just those on the factory floor. That means Marketing, Finance, Product
Development, Business Services, Engineering, and all the other functions in our businesses are included.”
— Honeywell

Six Sigma — Focus on Process Improvement

• Six Sigma is a process-focused approach to achieving new levels of performance throughout any business
or organisation
• In Six Sigma the focus is on process improvement to increase capability and reduce variation
• The process is considered as a system of inputs, activities, and outputs. It provides a holistic approach to
all factors and the way they interact to create value or waste
• The few vital inputs are chosen from the entire system of ‘controllable’ and ‘noise’ variables, and the focus
of improvement is controlling these inputs

Six Sigma — Voice of the Customer (VOC)

Six Sigma begins with VOC (Voice of the Customer). The VOC describes all stated and implicit customer needs
and requirements. In Six Sigma, it is not directly actionable, so it translates into measurable business
requirements:
• CTQ — Critical to Quality: Attributes that directly impact customer satisfaction and product/service
quality.
• CTD — Critical to Delivery: Factors affecting delivery performance such as time, reliability, and
responsiveness.
 • CTC — Critical to Cost: Elements influencing cost efficiency and financial performance.

These requirements exist at Product, System, and Subsystem levels. This ensures: Consistency and Traceability
from customer need → operational execution.
Six Sigma is defined as ‘the degree of confidence a customer has that his (or her) product and
service-related expectations will be met by the producer.’ — Michael J. Harry & J. Ronald
Lawson, 1992

Where does VOC come from?

• End Customers or Users
• Potential Future Customers
• Internal Customers
• Product Development teams
• Support Organisations
• Marketing Department
• State of the Art / Industry benchmarks

Six Sigma as a Measure — The Normal Distribution

Variation is the Enemy of Quality!
Six Sigma as a measure gives us a statistical scale (the normal distribution / bell curve) to measure progress and
benchmark companies, processes, or products. Key coverage percentages:
• ±1σ covers 68.26% of all values
• ±2σ covers 95.44% of all values
• ±3σ covers 99.73% of all values
• ±4σ covers 99.9937% of all values
• ±5σ covers 99.99943% of all values
• ±6σ covers 99.999998% of all values (3.4 DPMO)

Three milestones: Operations Excellence (±3σ), Team Excellence (±4σ), Business Excellence (±6σ).
Six Sigma implies a process where any value outside the specifications is an extremely rare occurrence.

Process Capability — Aligning VOC and VOP

• High Capability / Capable Process (green): Process distribution fits entirely within the specification limits
— all output meets requirements
• Low Capability / Not Capable Process (red): Process distribution extends beyond specification limits —
some output fails to meet requirements

Goal: Get the two Voices (VOC & VOP) into alignment — the process spread must fit within the specification
width.

Statistical Thinking
• All work is process
• Variation exists in any process
• Knowledge and management of variation are key to succeed in process improvement
 • Provides feedback loops for continuous improvement

Two types of variation:

• Common Cause Variation (Natural): Inherent to the process; random over time; process stays in control
• Special Cause Variation: Specific assignable cause; process goes out of control; must be identified and
eliminated

So, Six Sigma helps in taking right decisions

‘Six sigma is the best training an organization can give to its employees, it’s even better than
sending them to Harward Business School!’ — Jack Welch

Defining a Problem Statistically

In Six Sigma, every problem is expressed as: Y = F(X₁ ... Xₙ)
Process inputs (X’s — causes): People, Material, Equipment, Policies, Procedures, Methods, Environment,
Randomness → through Activities → to Outputs (Y’s — effects): Products, Services, Other results.
• Y is: Dependent, Output, Effect, Symptom — something we Monitor
• X₁...Xₙ are: Independent, Input-Process, Cause, Problem — something we Control

Goal: Find and control the critical X’s

Six Sigma as a Benchmark — DPMO Table

Process capability directly determines defects per million opportunities (DPMO), yield, and cost of quality (COQ):

Sigma Level DPMO Yield COQ Classification

6 3.4 99.9997% <10% World Class
5 233 99.976% 10–15%
4 6,210 99.4% 15–20% Industry Average
3 66,807 93% 20–30%
2 308,537 65% 30–40% Non Competitive
1 500,000 50% >40%

What Does 4 Sigma (99.9% Sure) Mean in Practice?

At the 4 Sigma level, the following problems would still occur:
• 9 hours/year of unsafe drinking water
• 107 incorrect medical procedures a day
• 200,000 incorrect drug prescriptions per year
• 18,322 pieces of mishandled mail an hour
• 2,000,000 documents lost by IRS a year
• Two short or long landings at any major airport each day
Organisational Infrastructure — Roles and Responsibilities
• Executives: Own vision, direction, integration, results; Lead change
• Champions: Project owner; Implement solutions; Black Belt managers
• Master Black Belts (MBBs): Full time; Train and coach; Statistical problem solving experts
• Black Belts (BBs): Devote 50%–100% of time on BB activities; Facilitate problem solving; Train and coach
Green Belts and project teams
• Green Belts (GBs): Part-time; Help Black Belts on projects
• Project Team Members: Part-time; Project-specific roles
• All Employees: Understand vision; Apply concepts

Six Sigma Methodology and Principles

• DEFINE: Identify the process, problem, and goals. Tools: SIPOC, VOC, CTQ Tree, Project Charter
• MEASURE: Quantify current performance (Y = f(X₁, X₂,...,Xₙ) + ε). Tools: Data Collection, Control Charts,
MSA
• ANALYSE: Determine root causes (identify the critical X’s). Tools: Fishbone, Pareto, Hypothesis Testing,
Regression
• IMPROVE: Implement and validate solutions. Tools: DOE, Poka-Yoke, Pilot Studies, Brainstorming
• CONTROL: Sustain improvements (monitor X₁ X₂ X₃ Xₙ). Tools: Control Plans, SPC, FMEA, SOPs

Six Sigma is a pragmatic approach to Empirical and Experimental Software and Systems Engineering!

Lean Thinking
In the late 1940s, TOYOTA wanted to produce quality cars at American mass-market prices, but the Japanese
market was not big enough for similar economies of scale. TOYOTA re-engineered the manufacturing process
following the fundamental Lean principle: eliminate waste. The Japanese car industry (Toyota, Honda) became
highly competitive by applying Lean principles.

Lean vs Six Sigma Comparison

Six Sigma Lean

Focus on Defects Target Waste in any process
Statistical and analytical tools Maps Value Stream
Systematic approach to solve problems House Keeping (5S)
Probabilistic approach Key for Process Improvement
Has its own Tools Defects are biggest category of waste

Housekeeping — 5S
• Seiri (Sort): Distinguishing between necessary and unnecessary things, and getting rid of what you do not
need
• Seiton (Straighten, Set): The practice of orderly storage so the right item can be picked efficiently (without
waste) at the right time, easy to access for everyone. A place for everything and everything in its place.
 • Seiso (Shine, Sweep): Create a clean worksite without garbage, dirt and dust, so problems can be more
easily identified (leaks, spills, excess, damage, etc)
• Seiketsu (Standardize): Setting up standards for a neat, clean, workplace
• Shitsuke (Sustain): Implementing behaviours and habits to maintain the established standards over the
long term, and making the workplace organization the key to managing the process for success

The Concept of Yield

First Time Yield (FTY): The number of good units produced divided by the number of total units going into the
process.

Example with 4 activities (starting with 100 units):

• Activity 1: 100 in → 10 defects → FTY1 = 0.9 (90/100)
• Activity 2: 90 in → 10 defects → FTY2 = 0.89 (80/90)
• Activity 3: 80 in → 5 defects → FTY3 = 0.94 (75/80)
• Activity 4: 75 in → 5 defects → FTY4 = 0.93 (70/75)

Total Process Yield = FTY1 × FTY2 × FTY3 × FTY4 = 0.9 × 0.89 × 0.94 × 0.93 = 0.70 (70/100)

Best Practices
• Do not re-invent the wheel
• Someone has already done what you are trying to do — find out what others have done (at the beginning
of your project!)
• Flatter them by reusing their results — as Robert Galvin of Motorola said: 'Thou shalt steal (non-
proprietary) ideas shamelessly' and give credit where it is due
• Use internal knowledge repositories, benchmarking databases, and industry publications

Six Sigma — Key Ingredients

The three key ingredients form a triangle:
• People: Change management + Leadership
• Process: Agile, ITIL, CMMI, Lean
• Product: Java, J2E, S.O.A. (and equivalent technologies)

At the centre is the Six Sigma Space, requiring: Analytical thinking, Statistical thinking, and System thinking.

DEFINE PHASE
"A well-defined problem is more than half-solved" — Shoji Shiba

DEFINE Phase Overview

Goal: To identify and validate the improvement opportunity, develop the business processes, define critical
customer requirements, and prepare to be an effective project team.
Key Activities:
• Define the Project Charter
• Define the Problem
 • Define the Scope
• Define Goals
• Define the Initial Business Case
• Understand the VOC
• Identify initial CTQs
• Define the AS-IS System Map
• Identify Stakeholders
• Identify the Organisation Change Plan
• Define Communication Plan
• Identify Initial Risks
• Tollgate Review
Potential Tools and Techniques:
• SIPOC (Supplier-Input-Process-Output-Customer)
• CTQ Tree
• Pareto Chart
• Project Charter
• Stakeholders Analysis
• Tracking sheet
Key Deliverables:
• Project Charter
• Business Case
• Project Objective
• Project Scope
• Project Plan
• Communication Plan

The DEFINE Process

• Step 1 — Define the Business Case: What's wrong? What do we want to achieve? → Problem & Goal
Statements, COPQ
• Step 2 — Understand the Customer: How does the problem impact our customer? → VOC Process,
Stakeholder Analysis
• Step 3 — Define the Process: Which process does the problem relate to? → VOP Analysis, SIPOC
• Step 4 — Plan/Manage the Project → Project Plan
• Step 5 — Secure Project Approval → Project Charter

Objectives of DEFINE: Understand the VOC, Understand the Process, Define the problem, Establish project goals,
Establish project scope, Define project plan, Secure sponsorship.

Defining the Problem

A problem is a situation that is judged as something that needs to be corrected.
• Most problems are initially identified by our customers
• Defining the problem improves focus
• To clearly define a problem is often a discovery-driven process
 • Problem definition/refinement continues throughout the project: We start with a conceptual and clear
definition and through MEASURE and ANALYSE refine and shape the problem (specific issues)

It is YOUR job to make sure that we are solving the right problem.
Problem Statement — Exercise
A typical problem statement for a Support Services improvement project:
“The current process for handling customer tickets is viewed as taking too long and could be
improved by introducing an automatic dispatching and monitoring system”
• Is the problem statement properly formulated?
• How would you improve the statement?

Issue Explanation
1. The problem is not Particularly for projects that are potentially wide in scope the problem statement
SPECIFIC! should clearly identify the boundaries of the project “The current process for
handling customer tickets in relation to the application areas XYZ, for customers
based in the EMEA region and with specific reference to messages with high
priority..”
2. The problem is not The problem statement should provide a high level description of the measurable
MEASURABLE impacts of the problem “....Does not allow to respect relevant, internal SLA’s
defined in relation to the processing time. The average resolution time is
approximately 14 days (against an SLA of 9 days) and we have also observed that
in 25% of the cases that messages with high priority are taking over 21 days to be
resolved”
3. Is the problem Relevance should be assessed from both the customer and the business
RELEVANT? perspective “As a consequence of this poor performance, our customer rating of
the support services we provide (see Survey May 2006) is, on average 15% lower
than those of our close competitors. There is a growing number of customers (7%
last year) who have decided to review their procurement decisions and moved to
our competitors. The impact for the business can be Estimated in lost revenues of
over 2 million $ per year”
4. The problem How do we know that an automatic dispatching of customer tickets to support
incorporates its own services staff would solve the problem? Do we have evidence or simply opinions?
SOLUTION!

Some problems are too generic to be solved just by refining them. Many problems are expressed as solutions!
Improved Problem Statement:
• The current process for handling customer tickets in relation to the application areas XYZ, for customers
based in the India region and with specific reference to messages with high priority does not allow to
respect relevant, Internal SLA’s (Service Level Agreement) defined in relation to the processing time.
• The average resolution is approximately 14 days (against an SLA of 9 days) and we have also observed that
in 25% of the cases that messages with high priority are taking over 21 days to be resolved.
• As a consequence of this poor performance, our customer rating of the support service we provide (see
survey May 2016) is, on average 15% lower than those of our close competitors. There is a growing
number of customers (7% last year) who have decided to review their procurement decisions and moved
to our competitors.
 • The impact for the business can be estimated in lost revenues of over 2 million $ per year.

Problem Statement Tips

What is our problem? When and how often does it occur? What is the financial impact?
• Describe briefly
• Avoid technical jargon
• Quantify the problem: use any available data and facts
• Explain the cost of the problem (see Cost of Poor Quality)
• Define the scope
• Think quantitative rather than qualitative: Numbers, metrics, and Money

Problem Statement — The Art of Questioning

• Formulating a problem is a teamwork exercise
• It should be done through communication and discussion
• The main tool used to formulate problems is Naive Questioning
• At this stage we are interested in answering to: the WHY’s and WHAT’s (rather than the HOW’s)
• Teamwork, brainstorming, and elicitation techniques can be used

Examples:
• During 2016, 20% of overseas customer payments took longer than the agreed invoice terms. This resulted
in an average outstanding debt of $357K, at a 5% cost of capital.
• From Oct 2016 to March 2017, 5% of Product A manufactured on production line 1 failed the final test.
This resulted in an extra inspection/rework process being implemented (at a cost of $25,000 per month),
and a scrap rate of 2% costing $20k per month in lost revenue.

What is good about these examples:

• Both of them provide dates and data that help baseline the problem
• The scope of the problems are defined by phrases such as product A, overseas and production line 1.
• Both statements explain how the problem affects the organization and provides an estimate of the cost.
• They do not refer to any solutions or root causes

Problem Statement: The 5 Why's

Question your problem down to the 'root need'. Perhaps problems are not located where you were told!
Ask 'Why' five times before coming to conclusions. Don't assume that the original problem, as stated, captures
the real problem. Ask questions until the person says 'I don't know'.

Example (Acceptance Testing process):

• Customer says testing takes too long compared to other phases
• Developer asks Why → testers spend more time writing test cases than testing
• Developer asks Why → testers don’t know what they are supposed to test
• Developer asks Why → there is no formal requirements document
• Developer asks Why → staff says that’s the BA’s job
• Developer suggests: 'Why don't you have a word with the BA's and the Project Manager?'
This iterative questioning reveals the real root problem rather than just the symptom.

Goal Statement — SMART Criteria

When your problem is properly formulated, you have to set a target, a goal for your project. The goal statement
is typically revised at the end of the MEASURE and ANALYSE phases to ensure it remain achievable. Define what
is within and out of scope.

SMART checklist to evaluate the Project Goal:

• S — Specific: Clearly identify the boundaries of the project
• M — Measurable: Provide a high-level description of the measurable impacts
• A — Achievable: The goal should be realistic
• R — Relevant: Linked to strategic objectives
• T — Time bound: Has a deadline

Example: 'Reduce the cost of defects by a factor of 3 before delivering the product to the end-customer.'
The Goal Statement will be a key reference to derive the KPIs for your Project.

IS / IS NOT Tool
To avoid misconceptions and misunderstanding about your project: Explain what your project IS and what it IS
NOT. This two-column table helps scope the project boundaries precisely.

IS IS NOT

Examples:
• Reduce the percentage of overseas customer payments that take longer than the agreed invoice terms
from 20% to 5%, by the end of quarter 4, 2017.
• Reduce the Product A final test failure rate on production line 1 from 5% to X% by the end of October
2017.

What is good about these examples?

• Both of them set a date for the improvement to be achieved by.
• Both of them are specific about the process measurement that will be improved.
• Neither of them contain a solution or root cause.
• They define success as improvement in a key measurement.
• They are brief.

Problem and Goal Statement: Key Conclusions

• Problem and Goal Statement are closely inter-related — they must be consistent!
• Improving the knowledge and developing a detailed description for one will improve understanding of the
other
 • Same tools applicable to both: Brainstorming, Questioning (5 Why's), Scope Definition (IS-IS NOT)
• Problem and Goal Statements follow a highly iterative cycle — refined through DEFINE → MEASURE →
ANALYSE → IMPROVE

Project Charter
• Problem Statement: detailing when the problem has been seen, what the problem is, the magnitude of
the problem, and the impact or consequence of the problem (such effect on customer Critical-to-Quality
expectations). Make sure the problem statement focuses on symptoms only (not on causes or solutions).
• Key Stakeholders: Who are they? How will they be involved in the project? How will progress be
communicated to them?
• Business Impact: reflecting expected financial benefits and assumptions
• Goal Statement: Clearly identifying the key output metric (Y) to be improved.
• Verification of Project Scope: broad enough to achieve the project objectives yet narrow enough to be
completed within the Project Plan timeframe.
• High-level Project Plan: showing the targeted completion date for the project and intermediate
milestones.
• List of team members: representing key stakeholders, appropriate mix of skills, and knowledge (especially
about the current process)

Stakeholders Analysis
Stakeholders Analysis is a process of systematically gathering relevant information to decide which interests
should be considered when implementing a change initiative. The analysis must identify:
• Stakeholders identity / groups of stakeholders
• Their interest (positive or negative) in the initiative to be undertaken
• Their influence and the level of support that can be expected from them
• The required level of support in order for the initiative to be successful
• The actions required to gain the necessary support from key stakeholders

Without support from relevant stakeholders, no change initiative will achieve its goals!
Identifying Stakeholders:
Stakeholders are all those who have a stake in the change initiative being considered: those who stand to gain
from it, and those who stand to lose.
• Any person, group or entity which can determine our system’s degree of success or failure by having an
“opinion” about system performance characteristics and system lifecycle constraints
• System stakeholders may come from a range of technical and nontechnical backgrounds and from
different disciplines
• Stakeholders determine requirements, and expectations!
• If you don’t identify all critical stakeholders → You probably cannot identify all critical requirements
• If you don’t identify all critical requirements → You will get delays and problems (and sooner or later
failures)
• The most common mistake in development efforts is to leave an essential player out of the process.
Categories of Stakeholders:
• Those responsible for design and development: project manager, designers, quality assurance, V&V
• Those with a financial interest, responsible for its sale or for its purchase: customer, buyer, business
analyst, marketing manager
 • Those responsible for its introduction and maintenance within an organisation: training and user support
staff, installation and maintenance engineers
• Those who have an interest in its use: user managers and all classes of users

Stakeholder Motivation Area of Expertise

Customer Introduce change with maximum Business and information system
benefit strategies, industry trends
User Representative Introduce change with minimum Business process, operating procedures
disruption
Project Manager Successfully complete the project Project and resources management
with the given resources
Analyst Specify requirements on time and RE methods and tools
within budget
Developer Produce technically excellent Latest techniques and creative design skills
system, use latest technologies
Quality & Compliance Ensure compliance to process and Methods, standards, compliance and
product standards regulations
Training & User Support To support existing accounts and to Knowledge of current user problems
generate future revenue

The Voice of the Customer (VOC)

The VOC describes all stated and implicit customer needs and requirements.
It can be captured using different techniques and tools: discussions, surveys, interviews, formal specifications,
field reports, customer complaints, focus groups, direct observation.

To capture the VOC a project team must:

• Determine how to identify target customers
• Decide which customers to contact
• Define what mechanisms to use to collect their needs
• Allocate schedule and estimate necessary resources (a plan)

Why listen to VOC?

• Decide what is the best product to deliver
• Identify the critical features
• Decide where to focus the efforts
• Get a baseline in measuring customer satisfaction
• Identify the key drivers of customer satisfaction

Basic Tools: Customer Complaints Analysis, Surveys, Focus Groups, Letters, Comment Cards, Direct Contact
Methods (Phone, Interviews), Become your own customer (phone your call centre, try to buy your own
product).
Often customers cannot express their needs: they ignore what the supplier can deliver, their creativity is focused
on doing his job rather than advising service providers, psychological factors.
Note: It's not enough to ask customers 'how do you like it' or 'how can it be improved'.

Critical to Quality (CTQ)

• CTQs (Critical to Quality) are key measurable characteristics of a product or process whose performance
standards must be met in order to satisfy the customer.
• CTQ’s align improvement or design efforts with customer requirements
• CTQ’s are those essential product of service features defined by the customer (internal or external)
• They may or may not include upper and lower specification limits
• CTQs are what a customer expects and demands from a particular product or service.

CTQ Flow Down / CTQ Tree

We move from Customer’s needs to Quantifiable goals through progressive decomposition. Needs are
understood through Problem and Goal statement.

Voice of the Process (VOP)

A process can be defined as the logical organization of people, materials, energy, equipment, and procedures
into work activities designed to produce a specified end result.

To improve (or study) any process, we need three basic ingredients:

• A Process Map to describe the AS-IS process in terms of activities, flow, input, output, organisation, etc.
Examples of mapping tools: SIPOC, flowchart, streamline diagrams
• A Maturity Grid, used as a benchmarking and evolutionary path for improvement. Examples of maturity
grid: Crosby Maturity Grid, CMM, CMMI, COBIT, ITIL
• An Assessment Method to continuously provide feedback for improvement. Examples: CBA-IPI, SCAMPI,
Interim Profile, ISO/IEC 15504

Processes: theory, reality, and best practice — Three models of a process:

• AS-IS Model: What you believe you have (formal, clean flowchart)
• Real Life Model: What you have in reality (complex with loops, workarounds, and unknowns)
• To-Be Model: What you would like to achieve (simplified, optimised)

Process Flowchart:
Flowchart is a simple step-by-step process of activities carried out in the process.

Process Mapping:
Process Mapping consists of various methods highlighting all aspects of a process:
• Activities and deliverables
• Organisation, structures and responsibilities
• Cause and effect
• Procedure and logic
Techniques and Formalism: SIPOC, activities and deliverables (Data Flow Diagram (DFD)), Procedure and logic
(Flow chart), Actors (Activity-Actors Matrix). Use a graphical representation and complete with textual
description for more details.

Benefits of Process Mapping:

• Helps understanding the process
• Provides clearer explanations
• Identifies customers, actors and stakeholders previously neglected
• Identifies responsibilities, owners and decision makers
• Identifies skills and knowledge needed
• Identifies opportunities for improvement
• Eases identification of boundaries

SIPOC Diagram
SIPOC = Supplier – Input – Process – Output – Customer. Key principle: Start at the end (with the Customer).

SIPOC Example — Change Request (CR) Process:

Problem Statement: Analyse the process for handling change requests for improvement purposes (defects and
time to deliver)
Process steps: Record CR → Analyse CR → Implement CR → Validate → Release New Version

Developing the Lean Six Sigma Business Case

• Business case: all the materials and arguments used to show decision makers that an idea is good and that
the numbers associated with it makes sense from a financial perspective.
• Business case development is a process and: It is not exclusively a matter of financial figures. It is iterative
in nature. Often becomes a mini project in itself.
• The content, timing and packaging of your proposal are the relevant factors
• Involves the financial department to assist you with numbers and estimates
UNIT II: MEASURE PHASE

Defining a Problem Statistically

In Six Sigma, every problem is expressed as: Y = F(X₁ ... Xₙ)
Where do the Y’s come from: Goals, symptom, issues, effect, specifications, and CTQs
Getting the X’s: how to derive the X’s, did we identify all the major X’s, do we need a process for measurement,
how to measure the X’s, how to be sure that we are measuring the same way, is their systematic process or
technique to identify the X’s, how are we sure that we are getting the right X’s, etc.

Measurement Framework — The 5 Questions

Question Explanation
Why measure? Link it to the problem, goal statement and CTQs. These are the drivers for
measurement
What to measure? What are the entities, their characteristics to measure and how do the link
to the goal?
Who measure? Who perform the measurement
When to measure? When to measure and how frequently to measure
Where? What are the sources of the data
How to measure? The procedures, tools and instruments used for measurement and data
collection

Why do we need metrics?

• To gather evidence that will support the decision making process
• Assumptions and feelings are not always a truthful picture of reality
• To baseline a process
• To measure improvement
• To characterise our process (how are inputs are output related)
• To gain a better understanding of complex phenomena
• To identify problems and establish casual relationships
• Assumptions can lead to major failures
• To evaluate customer satisfaction

Measure Phase Overview

Goal: To define a measurement system to quantify CTQs and critical measures. To develop a data collection
process. Define unambiguously what is a defect. To establish a baseline of the performance and/or capability of
the current system from a qualitative and quantitative perspective.
Key Activities:
• Establish Metrics Operational Definitions
• Establish Defects Definition
• Define the Measurement Framework
• Analyse Measurement System
• Define Data Collection Plan
• Define the Process Capability or the System Performance
• Tollgate Review
Potential Tools and Techniques:
Capability scorecard, Pareto, GQ(I)M
Key Deliverables:
Baselined measures | Process Capability | Measurement System | Data collection plans and forms | Operational
Definitions | Measurement System Analysis

What is a baseline?
• A “snapshot” of the state of a given system (process, product, artefact,..) in terms of Y’s and X’s, frozen at
a certain point in time.
• The baseline will be used as the basis (reference) for further comparison, analysis, studies, and actions.
• The baseline is defined qualitatively and quantitatively.
• Define a baseline for all CTQ.

Key Performance Indicators (KPI)

• A KPI is a type of performance measurement that helps you understand how your organization or
department is performing.
• A Key Performance Indicator is a measurable value that demonstrates how effectively a company is
achieving key business objectives. Organizations use KPIs at multiple levels to evaluate their success at
reaching targets.
• High-level KPIs may focus on the overall performance of the enterprise, while low-level KPIs may focus on
processes in departments such as sales, marketing or a call centre.

To be effective, a KPI must:

• Be well-defined and quantifiable.
• Be communicated throughout your organization and department.
• Be crucial to achieving your goal. (Hence, key performance indicators)
• Be applicable to your Line of Business (LOB) or department.

KPI Examples:
• Marketing and Sales examples: New customer acquisition, Status of existing customers, Customer
attrition, Outstanding balances held by segments of customers and terms of payment
• Manufacturing examples: Quality, Cycle time, Rejection rate
• IT Operations examples: Mean time between failures, Mean time to repair
• Education examples: Government of Ontario, Canada has been using KPIs since 1998 to measure the
performance of higher education institutions in the province. All post-secondary schools collect and report
 performance data in five areas: graduate satisfaction, student satisfaction, employer satisfaction,
graduation rate, employment rate.

Why Measure? — Human Factors

• “When you can measure what you are speaking about, and express it in numbers, you know something
about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge of it is
of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely, in your
thoughts, advanced it to the stage of science.” — Lord Kelvin (1824–1907)
• “You can't manage what you can't control, and you can't control what you don't measure.” — Tom
Demarco (1982)
• “Without a measurement process, managers do not have the information needed to identify and resolve
problems.” — James Rozum

Measurement is a Management Tool.

Measurements Drive Behaviours!

Measurements drive behaviours, and behaviours, in turn, create culture.
The software process is performed by people, not machines.
The very act of measuring (observing) a software process may change the human behaviour in dysfunctional
ways.
Any observation is a form of interpretation and is not neutral.
In making only relevant observations, we are already making choices and selecting data - unknowingly or
willingly.

WYMIWYG — What You Measure Is What You Get

‘Effort moves toward what is measured’ — Tom Demarco
People respond to whatever they are asked to do:
• If you measure Lines of Code — you get Lines of Code
• If you measure how many Six Sigma Projects you complete — you will get Six Sigma projects (not
necessarily results)
• If you measure how many projects are completed on schedule — you are likely to alter estimating and
reporting practices more than improve performance

Measures of “productivity” are inherently dangerous because they will drive behaviour, even if it is counter-
productive. They will lead to optimising various steps in the process while lengthening the overall process.
‘Measures of productivity do not lead to improvement in productivity’ — Deming
Measurement is only a mean to an end. Measure quality and let productivity take care of itself (because it will).

Precepts of Measurements
• Measure what the customer cares about
• Measure the process, not the people
• Know what are the impacts and results of measurements before you measure
• Know what to do with the results you obtain
• Anticipate the results of actions and decisions taken from measurements
• Measurements should be natural and the process itself should yield measures
 • Measurements should be numerical and objective (to enable comparability)
• Publish and publicize measurement results
• Be careful about the WYMIWYG syndrome

Measurement Framework
Software and IS Measurement examples include: lines of code, cycle time, response times, number of defects,
service availability, cost of defect, volatility of requirements, employee productivity, and many others.
Measurement is not just the collection of data and measures.

Measurement Definition
Measurement is the process by which numbers or symbols are assigned to attributes of entities in the real world
in such a way as to describe them according to clearly defined rules.
Measurement requires: entities (objects or events of interest), attributes (features or properties of an entity),
rules (and scales) for assigning values to the attributes.
Before being able to measure, we must clearly identify the entities and attributes we will address and the rules
we will use to assign values.
‘Measurements are important, but what is measured is more important.’ — Francis S. Patrick

Examples of Entities, Attributes, and Measures

Entity Attributes Measures

Report Length Defect density Complexity Number of pages Number of customer-
relevant defects Number of data sources
Number of steps in composition
Defect Type Origin Severity Effort to fix Age (of Type names Name of activity where
open defects) introduced Ordered set of severity classes
Staff-hours Elapsed time (days) since
receipt of defect report
Support Services Effectiveness Efficiency Cost Total number of queries resolved Average
resolution time Total cost per Region

Measures Operational Definition

An operational definition is a precise description of: (1) What is the characteristic or feature to be measured, and
(2) How specifically data collectors are to measure the characteristic or feature.
Purpose of Operational Definitions:
• Remove ambiguity and ensure all data collectors have the same understanding
• Reduce chances of disparate results between collectors
• Play a key role in the reliability of the Measurement System
A Precise Definition Must Consider:
• Communication — the ability of others to understand what to measure, how, what is included and
excluded
• Repeatability — the ability of others to repeat the process using these definitions and get the same results
 • Traceability — tracing the data to its origins in terms of time, source, sequence, activity, product status,
environment, tools used, and collector

Traceability and contextual data are very important aspects for process improvement.
Operational definitions are developed to provide clear and an ambiguous descriptions of each measure.
An Operational Definition Must Include:
• Name of the measurement
• What is the measure supposed to represent
• Process or system diagram
• Detailed definition
• Measurement Scope
• Questions that the measure helps to answer

Without an Operational Definition of the measurements: Unreliable data will be collected in different ways, and
different standards will be applied in different areas.

Introduction to Lean
Definition:
• Basically waste control technique
• All about eliminating waste from a process and making the product or service flow
• Lean focuses on waste elimination
• Lean thinking focuses on identifying and eliminating waste (non-value-adding activities) from processes.

The 7 Wastes in Lean

• Waiting: Idle time
• Over production: Produce more than required.
• Re work (defects): Doing the same job more than once.
• Motion: Excess human movements at work place.
• Process: Over processing, processing more than what is required.
• Inventory: Excess storage.
• Transportation: Excess movement of material or information.

The 7 wastes are at the root of all unprofitable activity, and all tools of lean production should focus on getting
rid of waste.

The 8 Wastes — DOWNTIME Acronym

• D — Defects: Efforts caused by rework, scrap, and incorrect information.
• O — Overproduction: Production that is more than needed or before it is needed.
• W — Waiting: Wasted time waiting for the next step in a process.
• N — Non-Utilized Talent: Underutilizing people’s talents, skills, and knowledge.
• T — Transportation: Unnecessary movements of products & materials.
• I — Inventory: Excess products and materials not being processed.
• M — Motion: Unnecessary movements by people (e.g., walking).
• E — Extra-Processing: More work or higher quality than is required by the customer.
Five Lean Principles
• 1. Define Value: Value is always defined by the customer's needs for a specific product. For example, what
is the timeline for manufacturing and delivery? What is the price point? What are other important
requirements or expectations that must be met? This information is vital for defining value.
• 2. Map the Value Stream: Once the value (end goal) has been determined, the next step is mapping the
“value stream,” or all the steps and processes involved in taking a specific product from raw materials and
delivering the final product to the customer. In this step, the goal is to use the customer’s value as a
reference point and identify all the activities that contribute to these values. Activities that do not add
value to the end customer are considered waste. The waste can be broken into two categories: Business
non-value added and non-value added.
• 3. Create Flow: After removing the wastes from the value stream, the following action is to ensure that
the flow of the remaining steps run smoothly without interruptions or delays. Some strategies for ensuring
that value-adding activities flow smoothly.
• 4. Establish Pull: Inventory is considered one of the biggest wastes in any production system. The goal of a
pull-based system is to limit inventory and work in process (WIP) items while ensuring that the requisite
materials and information are available for a smooth flow of work. In other words, a pull-based system
allows for Just-in-time delivery and manufacturing where products are created at the time that they are
needed and in just the quantities needed.
• 5. Pursuing (Seek) Perfection: Wastes are prevented through the achievement of the first four steps: 1)
identifying value, 2) mapping value stream, 3) creating flow, and 4) adopting a pull system. However, the
fifth step of pursuing perfection is the most important among them all. It makes Lean thinking and
continuous process improvement a part of the organizational culture. Every employee should strive
towards perfection while delivering products based on the customer needs. The company should be a
learning organization and always find ways to get a little better each and every day.

Data Types
The two main types of data: Quantitative and Qualitative

Data Collection & Summarization

Before data collections start, classify the data into different types:

Continuous or variable Discrete, categorical or attribute

Measured on a continuum Count or categories
Objective: Time, Money, Length, Weight Objective: Count defects, # approved, # of errors,
Type of document Subjective: Yes/No, Categories,
Performance rating, Satisfaction, Agreement, Extent

• Sometimes it is very costly or time consuming to collect all the data that is available.
• In these cases, we resort to sampling
• Sampling refers to collecting only a subset of the data and still be able to make good decisions from this
subset of data instead of having the collect the data for the entire population.
• Is the sample representative of the process or population?
• Is the process stable?
• Is the sample random?
• Is there an equal probability of selecting any data point within a homogenous group?
Probability Sampling
• Simple Random Sampling: Simple random sampling, which involves the random selection of data from an
entire population, so each possible sample is equally likely to occur.
• Stratified Sampling: Stratified random sampling is a method of sampling that involves the division of a
population into smaller sub-groups known as strata. In stratified random sampling, or stratification, the
strata are formed based on members’ shared attributes or characteristics such as income or educational
attainment and are homogeneous.
• Systematic Sampling: Systematic sampling is a probability sampling method in which a random sample,
with a fixed periodic interval, is selected from a larger population. The fixed periodic interval, called the
sampling interval, is calculated by dividing the population size by the desired sample size.
• Multistage sampling: Multistage sampling divides large populations into stages to make the sampling
process more practical. A combination of stratified sampling or cluster sampling and simple random
sampling is usually used.

Non-Probability Sampling
• Judgment Sampling: Selection of sample based on existing knowledge / professional judgement
• Quota Sampling: Sample has the same proportions of individuals as the entire population with respect to
known characteristics, traits or focused phenomenon.
• Cluster or Block sampling: Here, only the selected clusters are sampled. A common motivation of cluster
sampling is to reduce costs by increasing sampling efficiency.
• Snowball sampling: Snowball sampling (or chain sampling, chain-referral sampling, referral sampling) is a
nonprobability sampling technique where existing study subjects recruit future subjects from among their
acquaintances. Thus the sample group is said to grow like a rolling snowball.

Determination of a Proper Sample Size

• Sample size for estimating a population mean: n = Z × σ / ε² — Where Z: is set by the level of the
confidence interval, σ: the value of the population s.d., E: (x - μ) is error term
• Sample size for estimating a population proportion: n = Z² × P × Q / ε² — Where p: sample proportion, P:
population proportion, Z: is set by the level of the confidence interval, E = p – P (Error term)

Tools For Data Collections

• Check Sheets: Are probably the most common type of data collection forms used, but there are a variety
of other data collection forms. Simple data collection form that helps to determine how often something
occurs.
• Concentration Diagrams: Are most commonly physical representations of a product that has marks on it
to show what problems occurred and where on the product they occurred. By using a concentration
diagram, the company can see what errors are being made and where on the object the errors occur most
frequently.

Six Sigma Statistics

Descriptive Statistics: Averages, Mean, Median, Mode
Measure of spread (Dispersion): Standard Deviation and Variance.

Definition: Average is the single value which is most representative or typical value of the data.
Average is the value which is close to all observations.
Average is central value or central part of the data.
Therefore it is Known as a measure of central tendency or measure of location.

Dispersion: It is the method of measuring the amount of variation within the data. In other words dispersion is
the method of measuring the spread or scatterness of data.

Important:
• 1. Arithmetic Mean: This measure is arithmetic or algebraic average. Since it is based on all observations it
is affected by extreme values or outliers. i.e. If there is an outlier A.M. is not appropriate measure, median
is better.
• 2. Median: Median does not depend upon all values, it is positional or locational average it is not affected
by outliers, hence it is a better indicator of measure of central value when one or more of the lowest or
the highest observations are wide apart or not so evenly distributed (i.e. in case of many outliers in the
data).

Normal Distribution
• A continuous random variable X is said to follow normal distribution with the random variable values -∞ <
x < ∞ and its probability density function is given by: f(x) = (1/σ√2π) × e^(-½ × ((x-μ)/σ)²)
• It is a very important statistical distribution pattern occurring in many natural phenomena, such as height,
blood pressure, lengths of objects produced by machines, etc.
• Certain data, when graphed as a histogram creates a bell-shaped curve known as normal or normal
distribution.
• Normal distribution is symmetrical with a single central peak at the mean(average) of the data.
• Fifty percent of the area lies to the left of the mean and fifty percent lies to the right of the mean of the
distribution.
• The spread of a normal distribution is controlled by the standard deviation.
• The smaller the standard deviation the more concentrated the data.
• The mean, median and mode are the same in normal distribution.

Descriptive Statistics
Descriptive Statistics describe and summarize data. The two main branches are:
Graphical Methods:
DotPlot, Histogram, Stem-and-Leaf, Scatter Diagram, BoxPlot, Time Plot, Bar Chart, Pie Chart
Numerical Methods:
• Location (Central Tendency): Mean, Median, Mode, Percentiles, Quartiles
• Dispersion (Spread): Range, Variance, Standard Deviation, Interquartiles
• Shape: Skewness, Kurtosis
• Correlation: Correlation measures

Pareto Diagram
A Pareto diagram is a variation of a histogram (or more exactly a bar chart) for categorical data. Each category
represents a different type or category. The categories are ordered so that the one with the largest frequency
appears on the far left, then the category with the second largest frequency and so on.
The Pareto diagram is a key diagram in quality control and Six Sigma. It allows us to focus on the most critical
issues and identify the vital few.
Measurement System Analysis (MSA)
What is measurement system?
A measurement system is not just a device, such as a ruler or timer etc, but it includes the people, standards,
and procedures that surround the measurement process itself.

Measurement System Analysis refers methods to evaluate our measurement systems and determine whether
we can trust our data.
The evaluation of measurement system should be done prior to capability analysis, control charting or any other
analysis to prove that the measurement system is accurate and precise, and the data are trustworthy.

Types of Measurement Errors

• Accuracy: (Average is good but variation is more) Difference between the average of observed values and
the standard.
• Precision: (Variation is less) Variation observed between repeated observations of the same unit- either
due to operator or due to equipment.
• Stability: (Time based issue) Variation which occurs with time. It is observed if the same operator
measures the same unit with the same measuring equipment over an extended period.
• Linearity: (Range based issue) Consistency of the measurement system across its entire range.

MSA Formula
Observed Variation = Process Variation + Measurement System Variation
Formula: σT² = σp² + σm²

Components of Measurement System Variation (σm²):

• Bias: Difference between the observed value and the reference value
• Linearity: A measure of the bias values through the expected range of measurements
• Stability: Consistency of measurements over time
• Repeatability (Equipment variation): Variation in measurements under exact conditions with the same
instrument
• Reproducibility (Appraiser variation): Variation in the average of measurements when different operators
measure the same part

Repeatability + Reproducibility = Gauge R&R

Gauge R&R
Repeatability: Variation observed if the same operator measures the same unit repeatedly with the same
instrument.
Reproducibility: Variation observed if two or more operators measure the same unit with the same measuring
instrument.

Gauge R&R involves evaluating the repeatability and reproducibility of a measurement system.

Repeatability refers to the inherent variability of the measurement system. Conditions:

• Same person
 • Same thing being measured (part)
• Same characteristic
• Same instrument
• Same set-up
• Same environmental conditions

Reproducibility is the variation in the average of the measurements made by different operators. Conditions:
• Different person
• Same part
• Same characteristic
• Same instrument
• Same set-up
• Same environmental conditions
To use Gauge R&R:
• Identify the elements of your measurement system (equipment, operators or data collectors,
parts/materials/process, and other factors).
• Check that any measuring instruments have a discrimination that is equal to or less than 1/10 of the
expected process variation / specification range
• Select the items to include in the Gage R&R test. Be sure to represent the entire range of process
variation. (Good and Bad over the entire specification plus slightly out of spec on both the high and low
sides)
• Select 2 or 3 operators to participate in the study
• Identify 5 to 10 items to be measured.
• Make sure the items are marked for ease of data collection, but remain “blind” (unidentifiable) to the
operators
• Have each operator measure each item 2 to 3 times in random sequence
• Gather data and analyse. See following example for interpretation of typical plots generated by statistical
software.

Interpreting Gauge R&R Results

Components Of Variation Chart:
• You want the part-to-part bars to be much taller than the others because that means most of the variation
is from true differences in the items being measured.
• If the Gage R&R, Repeat, and Reprod bars are tall, that means the measurement system is unreliable.
(Repeat + Reprod = Gage R&R)
• Focus on the % Study Var Bars – This is the amount of variation (expressed as a %) attributed to
measurement error. Specifically, the calculation divides the standard deviation of the gage component by
the total observed standard deviation then multiplies by 100.
• Common standards (such as AIAG - Automotive Industry Action Group) for % Study Var are:
• Less than 10% is good & acceptable – it means little variation is due to your measurement system; most of
it is true variation. If number of distinct categories is >= 4 - acceptable
• 10% - 30% may be acceptable depending on the application (30% is maximum acceptable for any process
improvement effort)
• More than 30% unacceptable (your measurement system is too unpredictable)
Repeatability:
Repeatability is checked by using Range Graph of the Gage R&R control charts. This chart shows the variation in
the measurements made by each operator on each part. If chart is “in control” the gage and operator are
“repeatable”. Is the range chart in control? Any points that fall above the UCL need to be investigated. If the
difference between the largest value and the smallest value of the same part does not exceed the UCL, then that
gage and operator may be considered Repeatable (depends on the application).

Reproducibility:
Reproducibility is graphically represented by looking for significant differences between the patterns of data
generated by each operator measuring the same items. Want chart to be “out of control” – that means the gage
variance is much smaller than the differences in the items being measured. This is one instance where you want
the points to consistently go outside the upper and lower control limits (LCL,UCL). The control limits are
determined by gage variance and these plots should show that gage variance is much smaller than variability
within the parts.

By Part Chart:
The By Part graph shows the data for the parts for all operators plotted together. It displays the raw data and
highlights the average of those measurements. This chart shows the measurements (taken by 3 different
operators) for each of 3 parts. Want the range of readings for each part to be consistent with the range for other
parts. The chart should show a consistent range of variation (smallest to the largest dimensions) for the same
parts.

By Operator Chart:
The By Operator graph groups data by who was collecting the data (“running the process”) rather than by part,
so it will help you identify operator issues. In this example each of 3 operators measured the same 3 parts twice.
If line connecting the means is flat or almost flat, there are no significant differences between operators. Any
significant slope indicates that at least one operator has a bias to measure larger or smaller than the other
operators.

By Operator*Part Chart:
This graph shows the data for each operator involved in the study. It is the best chart for exposing operator-and-
part interaction (meaning differences in how different people measure different parts). Look at how closely the
lines parallel each other. If they diverge, then some operators are consistently measuring some parts differently.
These lines should be parallel.

Process Capability Index

Process capability measures how well a process performs relative to specification limits.

Process Capability Analysis:

An important technique used to determine how well a process meets a set of specification limits is called a
process capability analysis. A capability analysis is based on a sample of data taken from a process.

Process Capability Index:

• A metric used to quantify and evaluate process capability.
• It shows how capable process is in terms of meeting (internal or external) customer requirements.
Capability Formulas
• Cpl: Measures how close the process mean is running to the lower specification limit. Cpl = (Mean – LSL) /
3Sst — where Sst: Estimated Standard deviation (short term)
• Cpu: Measures how close the process mean is running to the upper specification limit. Cpu = (USL – Mean)
/ 3Sst

Specification Width vs Process Width

• Specification Width (SW): The distance from LSL to USL — determined by the customer
• Process Width (PW): The natural spread of the process = 6σ (±3σ from the mean) — determined by the
process

For the process to be capable: Specification Width / Process Spread > 1.00
Capability Indices:
• Cp: Indicates the inherent capability of the process. A high value of Cp is indicative of good process control
and repeatability as the data points are well clustered about the mean. Cp = (USL – LSL) / 6Sst = SW / PW
• Cpk: Indicates the resultant process capability for a non-centred process. Shift = Absolute value of
difference between mean and target (always positive). K = (Shift)/(0.5 SW). Cpk = Cp (1 – K) OR Cpk =
Minimum of [(X̅ - LSL)/3S and (USL – X̅)/3S] = Minimum of (Cpl , Cpu)
• Pp and Ppk: Similar to Cp (Process Capability) and Cpk (Capability with centering) but use overall long-
term variation
Interpretation of Capability Values:
• Cp / Cpk > 1.33: Capable process (4 sigma equivalent)
• Cp / Cpk > 1.67: Highly capable process (5 sigma equivalent)
• Cp / Cpk > 2.00: Six Sigma capable process
Process Capability Goals:
• Cpl = Indicates capability when only a lower specification limit exists (ex: amt. of chicken in frozen pot pie
cannot be less than 1 oz)
• Cpu = Indicates capability when only an upper specification limit exists (ex: delivery time cannot exceed 24
hours)
• Typical goals for capability indices are greater than 1.33 (or 1.67 if safety related)

Process Stability
Process stability is the ability of the process to perform in a predictable manner over time. A Run Chart gives a
good picture of stability.
ANALYSE Phase
Purpose: Analyze the data to investigate and verify cause-and-effect relationships. Determine what the
relationships are, and attempt to ensure that all factors have been considered. Seek out root cause of the defect
under investigation. To verify causes affecting the key input and output variables tied to project goals. (“Finding
the critical Xs”)

Key deliverables:
• Documentation of potential causes considered in the analysis
• Data charts and other analysis that show the link between the targeted input and process (Xs) variables
and critical output(Y)
• Identification of value–added and non–value–added work
• Calculation of process cycle efficiency

In summary: we analyse data to identify key factors driving performance and to understand the root causes of
problems.

The Process Door and Data Door

Six Sigma provides a large range of tools and techniques that can be used within the Analyze phase. The
challenge therefore, becomes knowing which tool to use when!
The phrases Process Door and Data Door are a useful way of dividing the Analyze phase up into two manageable
sections, which is particularly useful in a training environment.

The Process Door refers to a selection of tools and techniques that help you to understand and gain clues
directly from the process itself. It includes tools such as:
• Process mapping techniques (VSP etc.)
• Seven Wastes
• Root cause analysis tools (brainstorming, fishbone diagrams, 5 whys)
• FMEA (Failure Modes and Effects Analysis)

The Data Door refers to a selection of tools and techniques that help you to gain an understanding of the
process from the data itself. It includes tools such as:
• Graphical Analysis tools (e.g histograms, Run Charts, Box plots, Scatter plots, Dot plots)
• Statistical Analysis tools (e.g confidence intervals, hypothesis testing, correlation and regression, Normality
testing)

Root Cause Analysis

Root Cause Analysis is a structured evaluation method that identifies the root causes for an undesired outcome
and the actions adequate to prevent recurrence. It helps answering:
• What happened?
• How did it happen?
• Why did it happen?
RCA allows us to learn from past problems, failures, and incidents, and may be used for Corrective and
Preventive Actions.

Finding Root Causes

The process for the identification and validation of root causes is an iterative process. During the analyse phase,
concepts and models are integrated with empirical evidence and refined. This is quite likely to lead to reviewing
initial statements and re-formulating hypotheses and theories. The four-step cycle:
• 1. Brainstorm on Possible Causes
• 2. Analyse Data and Processes
• 3. Develop Hypotheses
• 4. Test and Validate

Cycle repeats until root cause is confirmed.

Cause and Effect Analysis

Decomposition is a type of analysis seeking to isolate different influencing factors in order to assess their relative
importance and identify root causes of a phenomenon.

The Fishbone/Ishikawa Diagram structure:

• The 'effect' (problem) is placed at the fish's head (right side)
• Major cause categories become the fish 'bones' (e.g., Equipment, Outage, Infrastructure, People,
Application)
• Sub-causes branch off each bone (e.g., for Equipment: Lack of Capacity Data, Lack of Configuration
Information, Inaccurate Asset Inventory)
• The 6M categories commonly used: Man, Machine, Method, Material, Measurement, Mother Nature

Example (Poor Availability problem):

Equipment bone includes Lack of Capacity Data, Lack of Configuration Information, Inaccurate Asset Inventory;
People bone includes Poor Training, Wrong Skills, Wrong Training; Application bone includes Poor Design, Poor
Testing; etc.

Root Cause Analysis — The 5 Why's (in ANALYSE context)

• The 5 Why's approach: a simple method to analyse a negative situation and identify the root cause (or
causes) of the problem. It is particularly useful during the Analyse phase of a DMAIC/DMADV project to
help isolate symptoms of a problem from its real origin (its root cause).
• A Root Cause is the most basic explanation of a problem origin and it’s the entity which, if eliminated,
would prevent recurrence.
‘If you don’t ask the right questions, you don’t get the right answers. A question asked in the
right way often points to its own answer. Asking questions is the ABC of diagnosis. Only the
inquiring mind solves problems.’ — Edward Hodnett
• Keep asking questions until the answers are: 'I don't know!' or 'I don't care!'
• The 'ask why five times' approach is credited to Taiichi Ohno, creator of the Toyota Production System.
Ohno's idea is that we have to make sure that we really do get to the root problem or opportunity, the
bottom of the stair, to make sure we deliver the customer needs.
• The 5 WHY's approach is one of the Six Sigma tools for identifying root causes and can be combined with
other methods such as FMEA analysis.
The Data Door — Analyse Phase
• Gain clues and understanding from the data itself
• Verify Root Causes with data to confirm they are real causes
• Verify data only on most likely root causes

Data → Analysis → Interpretation

• Data are generally collected as a basis for action
• We analyse data to know when a change has occurred in our processes or systems
• We analyse data to know about such a change in a timely manner and respond appropriately

Without meaningful data there can be no meaningful analysis. Voice of the Data: Listen to the data.

Correlation
Correlation is the logical relationship that exists between two or more variables. If two or more variables are
related each other in such a way that change increases or decreases the corresponding change in other, then the
variables are said to be correlated.

Examples:
• Relation between heights and weights
• Relation between dose of insulin and blood sugar
• Income and expenditure
• Cigarette smoking and lung cancer

Types of correlation:
• Positive correlation: If both the variables are varying in the same direction then the correlation is said to
be positive.
• Negative correlation: If both the variables are varying in opposite direction then the correlation is said to
be negative.
• Zero correlation: If there is no relationship between two variables then the correlation is said to be zero.

Regression Analysis
Regression takes its name from studies made by Sir Francis Galton. He compared the heights of persons to the
heights of their parents. His major conclusion was that the off-spring of unusually tall persons tend to be shorter
than their parents while children of usually short parents tend to be taller than their parents.
In a sense, the successive generations of off-springs from tall persons “regress” downward towards the height of
the population, while the reverse is true originally for short families. But the distribution of height for the total
population continues to have the same variability from generation to generation.
• By regression (linear) we mean average relationship between two variables, and this relationship is used
to estimate or predict the most likely values of one variable for specified values of the other variable.
• Regression is cause and effect relation between variables.
• Eg. Demand (cause) and supply (effect), Age (cause) and height (effect).
• One of the variables is called independent or the explained variable and the other is called dependent or
the explaining variable.
Regression model:
y = β0 + β1x + e — Where y is dependent variable, x is independent variable and e is residual (or error term)
which follows normal distribution with mean E(e) = 0 and variance V(e) = σi²

Some Statistical Terms

• Population: A population is any entire collection of people, animals, plants or things from which we may
collect data. It is the entire group we are interested in, which we wish to describe or draw conclusions
about the population characters.
• Sample: A sample is a subset or portion or a part of population.
• Parameter: Parameter is a statistical quantity will be based on (computed from) population units. It is a
function of population units.
• Statistic: Statistic is a statistical quantity will be based on(computed from) sample observations only.
Statistic is a function of sample observations. Any function of the random sample x1, x2, x3, ------ xn that
are being observed say Tn (x1, x2, x3, --------- xn) is called a “statistic”.

Testing of Hypothesis
In many situations, we need to make decisions about population on the basis of only sample information.

Examples:
• Quality control manager is to determine whether process is working properly
• A drug company is to decide whether a new drug is really effective in curing a disease
• A statistician has to decide whether a given coin is biased

Statistical Hypothesis:
In attempting to arrive at decisions about the populations on the basis of sample information, it is necessary to
make assumptions or guesses or statements about the population parameters involved. Such an assumption or
statement is called a Statistical Hypothesis which may or may not be true. The procedure which enables us to
decide on the basis of sample results whether a hypothesis is true or not, is called Test of Hypotheses or Test of
significance.
Null & Alternative Hypothesis:
• Null Hypothesis: The null hypothesis asserts that there is no significant difference between the statistic
and the parameter and whether observed difference is there, is merely due to fluctuations in sampling
from the same population. The hypothesis which is tested for possible rejection under the assumption
that it is true is known as Null hypothesis. It is usually denoted by H0.
• Alternative hypothesis: Any hypothesis which contradicts the null hypothesis is called an alternative
hypothesis and it is denoted by H1.

The two hypotheses H0 and H1 are such that, if one is true, the other is false and vice versa.
Type I And Type II Errors:
• The errors made in rejecting H0 when it is true is called Type I error.
• The error committed in acceptance of H0 when it is false is called Type II error.
• α = Probability of Type I error = Probability of rejecting H0 when H0 is true.
• β = Probability of Type II error = Probability of accepting H0 when H1 is true.
 Decision Accept H0 Reject H0
H0 is true Correct Decision P (1 – α) Type I error P α
H0 is false Type II error P β Correct Decision P (1 – β) Power

Example in Job Placement Interview:

Candidate Qualified Not qualified

Accepted No error Error β
Rejected Error α No error

Level of Significance:
In testing a hypothesis, the maximum probability with which we are willing to risk a Type I error is called the
Level of significance of the test.
• In practice we take either 5% (i.e. 0.05) or 1% (0.01) level of significance, although other levels such as
10%, 2% etc. may also be used.
• When we choose 5% level of significance in a test procedure, there are about five cases in hundred that
we would reject the hypotheses when it should be accepted, i.e. we are about 95% confident that we have
made the right decision.
• Similarly, if we choose 1% level of significance in testing a hypothesis, then there is only 1 case in 100 that
we would reject the hypotheses when it should be accepted i.e., we are about 99% confident that we have
made the right decision.
P – Value (Probability Value):
• It is the probability of getting the observed difference or greater between the sample means when H0 is
true.
• If P – value > α, Accept H0
• If P – value ≤ α, Reject H0
• A small p-value (typically ≤ 0.05) indicates strong evidence against the null hypothesis, so you reject the
null hypothesis.
• A large p-value (> 0.05) indicates weak evidence against the null hypothesis, so you fail to reject the null
hypothesis.

Hypothesis Testing with Normal Data

Large sample tests when n > 30 — Z test:
• Single sample mean testing, H0: μ = μ0
• Double sample means testing, H0: μ1 = μ2
• Single sample proportion testing, H0: P = P0
• Double sample proportion testing, H0: P1 = P2

Small sample tests when n ≤ 30 — t-test:

• Single mean testing, H0: μ = μ0
• Double mean testing, H0: μ1 = μ2
• Paired t test, H0: μ1 = μ2
ANOVA: Analysis of Variance (Three or More Samples)
• Purpose: To compare three or more samples to each other to see if any of the sample means is statistically
different from the others.
• An ANOVA is used to analyse the relationships between several categorical inputs (KPIVs) and one
continuous output (KPOV)

When to use ANOVA:

• Use in Analyse to confirm the impact of variables
• Use in Improve to help select the best option from several alternatives

In the statistical world, inputs are sometimes referred to as factors. The samples may be drawn from several
different sources or under several different circumstances. These are referred to as levels.

To tell whether the three or more options are statistically different, ANOVA looks at three sources of
variability:
• Total – Total variability among all observations
• Between – variation between subgroup means (factor)
• Within – Random (chance) variation within each subgroup (noise, or statistical error)

• In One-Way ANOVA, we look at how different levels of a single factor affect a response variable
• In Two-Way ANOVA, we examine how different levels of two factors and the interaction between those
factors affect a response variable

One-Way ANOVA
• One-way ANOVA focuses on a comparison of more than two population or treatment means.
• I ≥ 2 is the number of populations or treatments being compared.

Hypotheses for One-Way ANOVA:

• H0: The means of all groups are equal, i.e. μ1 = μ2 = ……… = μi
• Ha: Not all the means of the groups are equal, i.e. at least two of the μi are different.

For One-Way ANOVA, residuals are the differences between the observed values and the group means. Viewing
the residuals allows us to get more information and to verify some of the assumptions (normal distribution of
the residuals), which we cannot get from the ANOVA table.
(3-4 Examples are there is the ppt)

Two-Way ANOVA
Two-way ANOVA uses the same principles as one-way ANOVA, and produces similar Minitab output. Key
characteristics:
• The factors can take on many levels; you are not limited to two levels for each.
• Total variability is represented as: SST = SSA + SSB + SSAB + SSe

Where:
• SST = Total sum of squares
 • SSA = Sum of squares for factor A
• SSB = Sum of squares for factor B
• SSAB = Sum of squares due to the interaction between factor A and factor B
• SSe = Sum of squares from error (residuals)

Source DF SS MS F P
Order Type 1 3.968 3.968 4.34 0.048
Location 2 13.404 6.702 7.34 0.003
Interaction 2 0.364 0.182 0.20 0.821
Error 24 21.929 0.914
Total 29 39.665

As with other hypothesis tests, look at the p-values to make a judgement based on your chosen alpha level
(typically 0.05 or 0.10) as to whether the levels of the factors make a significant difference.

Main Effects Plots:

These plots show the average or mean values for the individual factors being compared. You will have one plot
for every factor. Differences between the factor levels will show up in ‘non-flat’ lines: slopes going up or down,
or zig-zagging up and down.
Example: Main Effects Plot – Data Means for Order Processing (Order Type and Location):
• Left panel – Order Type: The left side of the chart shows that consumer orders process faster than
commercial orders (consumer mean ≈ 4.8 days vs. commercial ≈ 5.5 days).
• Right panel – Location: The right side shows a clear difference in processing times between the three
locations: California (CA), New York (NY), and Texas (TX). NY and TX have higher mean processing times
than CA, and the line is clearly non-flat, indicating that location is a significant factor.
From the ANOVA table: Order Type p = 0.048 (significant at α = 0.05), Location p = 0.003 (highly significant),
Interaction p = 0.821 (not significant).

Interaction Plots:
The example below shows a different pattern for each region, meaning the factors ‘act differently’ at different
locations:
• Region 1: Color and plain packaging drive higher sales than point-of-sale displays.
• Region 2: Color and point-of-sale promotions have higher sales than color.
• Region 3: Has lower overall sales; unlike in Region 1 and Region 2, color alone does not improve sales.
The interaction plot shows crossing lines for the three regions, which is a classic indicator of a significant
interaction effect between packaging type and region.

General Linear Model (GLM) in ANOVA

Continuous Y, Discrete Xs (more than 2 factors)
• While using two-way ANOVA, Minitab demands a balanced design between X’s (i.e., equal sample sizes
across all factor level combinations).
• If the design is unbalanced, teams can use the General Linear Model (GLM) instead.
• GLM can also be used if there are more than two X’s (i.e., more than two factors).
ANOVA: General Linear Model – Example:
Suppose the Sales figure of a company depends on 3 factors: Number of Advertisements, Free service camps (Y
or N), Gender of Sales Representative (M or F). We can use the General Linear Model (GLM) to study more than
2 factors at a time.

Sales Figure Advertisement Free Service Sales Rep

10 10 Y M
12 20 Y M
14 10 N M
15 20 N M
11 10 Y M
13 20 Y M
9 10 N M
12 20 N M
17 10 Y F
22 20 Y F
24 10 N F
28 20 N F
18 10 Y F
25 20 Y F
13 20 Y M
16 20 Y M
8 10 N M
12 20 N M
18 10 Y F
23 20 Y F
19 10 N F
22 20 N F
18 10 Y F
20 20 Y F

Points to Remember in ANOVA:

When you use multiple X’s in ANOVA, they have to be mutually independent. This is a critical assumption: if the
independent variables (factors) in an ANOVA model are correlated with each other, the test results may be
unreliable. Each factor must be independently varied to isolate its effect on the response variable Y.
UNIT III: IMPROVE PHASE

Improve Phase
To evaluate, select and develop the right improvement solutions that meet the required performance. It
includes determining the best combination of inputs and controls and refining the solutions as necessary. The
Improve phase must develop the corresponding implementation and change approach to enable broad buy-in
from the stakeholders and the organisation.
Purpose: Improve or optimize the current process based upon data analysis using techniques. This phase helps
determine the relationships of key variables to the project ‘Y’ and leads to improvement ideas. All team
members work together for execution of the agreed plan of action from all stakeholders.

Key Deliverables:
• Screen the contributing X’s
• Build action plan for improvement
• Implement the action plan
• Collect data on improved Y
• Statistically validate if Y has reached its desired level
• Monitor before and after for X and Y

Brainstorming
Brainstorming is a group creativity technique by which efforts are made to find a conclusion for a specific
problem by gathering a list of ideas spontaneously contributed by its members.
In other words brainstorming is a situation where a group of people meet to generate new ideas and solutions
around a specific domain of interest by removing inhibitions. People are able to think more freely and they
suggest many spontaneous new ideas as possible. All the ideas are noted down and are not criticized and after
brainstorming session the ideas are evaluated.

Rules of Brainstorming:
• A group problem-solving technique that involves the spontaneous contribution of ideas from all members
of the group.
• Generate as many ideas as possible in short time
• Have lateral as well as vertical thinking
• Do not criticize any idea
• Encourage all participants. Avoid the dominance of some persons

Negative Brainstorming
A potentially better approach to brainstorming is negative brainstorming. This technique uses the human
tendency to see the negative side of things. When an topic or problem is put forward, it is often easier to come
up with reasons why it won’t work or why the problem cannot be solved rather than looking for what’s good
about the idea or solutions to the problem.
So – if we’re good at coming up with negatives, we might as well use this skill!
Example — Credit Management Negative Brainstorming:
• Get angry about credit policy
• Don’t communicate with credit
• Don’t involve credit in strategic marketing/sales initiatives
• Inflexible credit management processes
• Don’t communicate credit policies to customers
• Don’t build a logical business case backed by data to change policy

Assumption Busting
An advanced brainstorming technique that works by challenging traditional assumptions that have gone un-
debated for years or even decades. Ideas and assumptions that worked historically may not be valid in today’s
high paced, high tech environment. Key assumptions are identified and their accuracy mercilessly questioned.
Alternative new assumptions are formed and then tested by asking ‘what if’ questions.

When to use:
• Assumption-busting can be used in most creative situations
• It is particularly effective when you are stuck in current thinking or have run out of ideas

Example:
• Situation: We cannot sell SENSES over the internet.
• Assumption: We do not have the technology. Busting: We can learn. We can buy it in.
• Assumption: Buying it in would be expensive. Busting: You don’t know until you ask. Get my teenage son
to do it.
• Assumption: Your teenage son would not do it professionally. Busting: Look at these sites he did. He can at
least do a trial site to prove the principle.
• Assumption: It won’t work. Busting: A trial site would help prove this.

Assumption Busting — Real Example:

• In a personal loan approval process the problem is lengthy cycle time for approvals which is leading to
dissatisfaction.
• Investigation of the process reveals that there is a rule existing in the process which makes every vital task
in processing the loan pass through a specialist (legal expert, financial expert, credit expert) after it has
been processed by an agent.
• The reason for rule existing in the process is an assumption which says that all loan deals are complex.
• On investigation and data-collection it is found that only 5% of deals are complex and thus the process
assumption is wrong.

Error Proofing / Poka-Yoke

The word 'poka-yoke' is of Japanese origin. 'Poka' in Japanese means 'mistake' and 'yoke' means 'prevent'. This
roughly translates to 'preventing inadvertent mistakes'.
The term was coined by Shigeo Shingo, in the 1960s. Shigeo Shingo was in charge of the statistical process
control of a leading Japanese manufacturing plant in the 1950s, but was always dissatisfied with the statistical
approach regarding untested devices.
In the early 1960s, while visiting an electrical plant, Shigeo Shingo designed what is believed to be the first poka-
yoke device, in order to manage the defects regarding the push buttons of a small switch. He realized the
importance of mistake-proofing, and consequently, developed and worked on this concept for the next few
years. As of today, the concept is not only popular in Japan, but all over the world.
• Poka-yoke is a structured methodology for mistake-proofing operations.
• It is any device or mechanism that either prevents a mistake from being made or ensures that the
mistakes don’t get translated into errors that the customers see or experience.
The goal of poka-yoke is both prevention and detection: “errors will not turn into defects if
feedback and action take place at the error stage.” — Shigeo Shingo

Three Approaches to Poka-Yoke

• Warning: let the user know that there is a potential problem – like door ajar warning in a car
• Auto-correction: automatically change the process if there is a problem – like turn on windshield wipers in
case of rain in some advanced cars
• Shutdown: close down the process so it does not cause damage – like deny access to ATM machines if pin
entered is wrong 3 times in a row

The best operation is one that both produces and inspects at the same time.

Real Examples of Poka-Yoke

• Computer Monitor Cable: You must have dismantled your CPU a number of times. Have you noticed how
all the cables and connectors fit into their respective slots with the right precision and accuracy? This is
because of mistake-proofing. While manufacturing these cables, the operator needs to take care the right
pins are secured in the insulator, and the exact number of slots are drilled for perfect placement. If this is
not done, the device will definitely not work.
• Alarm Systems: They are activated with alarms for security purposes. If an unauthorized person enters the
bank, or the electric circuit of the doorbell catches fire, a warning alarm rings, which ensures that you can
take care of the problem before it gets out of hand.
• Fingerprint identification: This is used extensively in biometrics, forensics, detective agencies, etc. Every
fingerprint in the world is unique, no two are ever the same. The device is such that only the fingerprint of
that particular person will be required as input. Trying to replicate or use some one’s fingerprint forcibly
will not work either.
• Microwave Oven: You might have observed that you cannot operate the microwave until the door is
closed. Also, if you do open it, then the microwave stops immediately. Sometimes, you place a dish in the
microwave and forget all about it, the microwave switches off on its own. This is called ‘auto cooking
mode’.
• Washing Machine: A control mechanism is present in washing machines to ensure that they start
functioning only after the door is closed. Also, once started, the door cannot be opened until one full cycle
is complete.

Benchmarking
Benchmarking is comparing ones business processes and performance metrics to industry bests and best
practices from other companies.
A measurement of the quality of an organization’s policies, products, programs, strategies, etc., and their
comparison with standard measurements, or similar measurements of its peers.
The objectives of benchmarking are:
• To determine what and where improvements are called for
• To analyze how other organizations achieve their high performance levels
• To use this information to improve performance

Chain Letters
• Chain letters are messages written to a person, that attempts to get the person to copy the message and
send it to a specified number of people, word for word.
• A chain letter is a message that attempts to convince the recipient to make a number of copies of the
letter and then pass them on to a certain number of recipients (either a predefined number or as many as
possible).
• Chain letters are information sent around to a group of specific people (experts) and each participant is
requested to review previous statements and add comments, alternatives, and modifications (but not to
delete any information). It allows adding information in a more private atmosphere or also across big
distances (useful for global organizations).
• E-mail is a good medium to support chain letters to reach a significant amount of people in any location.
By copying the letters to original sender it has to be assured that none of the information gets lost. All
information at the end needs to get categorized and analyzed.
• The chain letter should provide an introduction, clear instructions on the subjects as well as handling
guidelines (forwarding, feedback). Also, timeframe and feedback to participants should be considered.

Billboards
• Billboards work similar as Chain letters; participants are invited to add comments or ideas on the Billboard.
It only requires that the target groups are locally available and Billboards are placed at a location where
they are noticed.
• Also no information get criticized or deleted!!!
• The comment can be given anonymously!!!

It should have an introduction to specify the objective and asking to add ideas and comments. Assure Billboard is
used in timely manner and for a limited time period only. Also feedback should be provided to the participants.

Pugh Matrix
• Pug matrix refers to a matrix that helps to determine a solution from a list of potential solutions.
• It is a scoring matrix used for concept selection, in which options are assigned scores relative to some pre–
defined criteria.
• The selection is made based on the consolidated score.
• The pug matrix is a tool to facilitate a methodical team based approach for the selection of the best
solution.
• It combines the strengths of different solutions and eliminates the weaknesses.
• This solution then becomes the datum of the base solution against which other solutions are compared.
• Results shown as Same (S), Better (+), or Worse (-) than the datum. The process is iterated until the best
solution emerges.

Evaluation Criteria Imp. A (Datum) B C D

Ease of guests finding lobby 3 S S S -
 Evaluation Criteria Imp. A (Datum) B C D
Min. waiting times for Check-in 5 S - - -
Min. errors in room assignment 5 S - - +
Appearance of lobby/cleanliness 4 S - - S
Sum of Same (S) 4 1 2 1
Sum of Positives (+) 0 0 0 1
Sum of Negatives (-) 0 3 2 2
Weighted Sum of Positives 0 0 0 5
Weighted Sum of Negatives 0 14 9 7

Table shows an example of Pug matrix used to compare current situation with alternative solutions. In this
example current venue is being compared with alternative solutions. From this example, we can conclude that
options B, C, and D have more of negatives compared to option A. Hence, we would pick option – A (current
situation) as the best option and would not consider any change in venue.

Prioritization Matrix
• The prioritization matrix is a tool used to prioritize items and describe them in terms of weighted criteria.
• Prioritization Matrix is a visual prioritization tool to compare the relative merits of alternative actions.
• A prioritization matrix can help an organization make decisions by narrowing options down by
systematically comparing choices through the selection, weighing, and application of criteria.

Benefits of Prioritization Matrices:

• Quickly surface basic disagreements, so disagreements can be resolved openly
• Force a team to narrow down all solutions from all solutions to the best solutions, which are more likely to
increase chances for successful program implementation
• Limits “hidden agendas” by bringing decision criteria to the forefront of a choice
• Increases follow-through by asking for consensus after

Method Cost Effective (0.4) Response Time Quantity (0.3) Weighted Ra

(0.2) Score nk
Questionnaire 4 1 5 3.3 1
Interview 1 4 1 1.5 4
Observation 3 2 2 2.2 2
Focus group 1 4 2 1.8 3

Result: Questionnaire scores highest (3.3) and is ranked first as the most efficient data collection method.
SCAMPER Technique
SCAMPER is a creative thinking framework created by Bob Eberle (originated from Alex Osborne’s brainstorming
questions). It is used as an excellent creative thinking exercise to generate new ideas. It is a useful list of words
to help us think differently about the problem, idea, concept or issue. It is an acronym of seven idea-triggering
questions:
• S — Substitute: What can you substitute? What or who could you use instead? The substitute technique
focuses on the parts in the product, service or solution that can be replaced with another.
• C — Combine: What can you combine or bring together? The combine technique tends to analyse the
possibility of merging two ideas, stages of the process or product in one single more efficient output.
• A — Adapt: What can you adapt for use as a solution? Adapt refers to a brainstorming discussion that
aims to adjust or tweak product or service for a better output. This adjustment can range between minor
changes to radical changes in the whole project.
• M — Modify/Minify/Magnify: Can you modify or change the item in some way? The modify technique
refers to changing the process in a way that unleashes more innovative capabilities or solves problems.
This change is more that just adjustment as it focuses on the overall process.
• P — Put to Another Use: What other or new ways could this be used? This technique concerns how to put
the current product or process in another purpose or how to use the existing product to solve problems.
• E — Eliminate/Elaborate: What can you get rid of? This technique aims to identify the parts of the process
that can be eliminated to improve the process product or service. It also helps to explore the unnecessary
parts of the project.
• R — Reverse/Rearrange: What could you rearrange or change around? The reverse or rearrange
technique aims to explore the innovative potential when changing the order of the process in the
production line. Reversing the process or part of it can help solving problems or produce more innovative
output.

Paired Comparison
LL Thurstone first established the scientific approach for using this approach for measurement.
Paired comparison involves pairwise comparison – i.e., comparing entities in pairs to judge which is preferable
or has a certain level of some property.

Employee Evaluation Example:

• If there are 5 employees A-E, A will be compared individually to B, to C and similarly to the remaining
employees.
• If A is better than B a “+” will be marked against his name, and if he is not as good as C, a “-”will be
marked.
• The total number of decisions in this case will be (5×4/2 =) 10.
• The number of decisions can be calculated by the formula N(N-1)/2, where N represents the total number
of employees being evaluated.
• In the diagram below employee C has the most “+” and hence will receive more incentives.

Job Evaluation:
• Job evaluation is done by comparing the worth of one job against that of another.
• Ranks and grades can be decided depending on the number of points scored by each job.
• If Job A is worth more than B, it will be rated with a “1”, and if it is not worth as much as Job C it will be
rated with a “0”.
 • In the diagram below job E has scored the most number of points and hence will be at a higher rank when
compared to the other jobs.

Solution Screening and Pilot Studies

In the ideal situation, the team would immediately know which factors are critical and which are not, so they
could ignore all the non-critical factors and simply pursue the critical ones.
When Six Sigma teams use the screening tools to approach a problem, they start with these correct conditions,
and then discover which factors are critical and vary excessively whenever defects occur. The solution they
develop is simply tighter control of these critical factors.
What is a Pilot Study?
An experimental or preliminary trial or test — process improvement teams will pilot the solution within the
Improve phase of the project.

Why Pilot?
• Confirm expected results and relationships
• Understand expected variation in the process and its possible impact to the customer
• Improve a solution
• Improve the implementation
• Lower the risk of failure
• Improve the ability to predict monetary savings from a proposed solution
• Increase opportunities for feedback
• Increase buy-in
• Quickly deliver a version of the solution to a particular segment
• Validate the measurement system

When to Pilot:
In general, there is a way to pilot all or some part of every solution or new design that an organization wants to
implement. It is almost always worth the extra effort to pilot. Consider doing a pilot when:
• The scope of the design is large
• The new product/service could cause far-reaching, unintended consequences
• Implementing the design will be a costly process
• The design and/or solutions would be difficult to reverse
• If conducting a DFSS (Design for Six Sigma) project, as opposed to a DMAIC (Define, Measure, Analyze,
Improve, Control) project, piloting is almost always a must.

Example:
• The first project had a productivity goal of 85 percent. The Green Belt conducted a pilot and found that
productivity would reach 96 percent. Although this clearly surpassed the goal of the project, the Green
Belt used information from the pilot to modify the solutions. After full-scale implementation, the project
was performing at 100 percent.
• The second project’s pilot was conducted during the improve phase of the project. It was a full-scale pilot.
The results not only verified the solutions, but also accelerated the completion of the project through the
control phase because there was nothing left to implement. The pilot was the implementation.
Assessment Criteria
Assessment criteria relates to the third part of the objective, the standard of performance. Criteria are
developed by analyzing the learning outcomes and identifying the specific characteristics that contribute to the
overall assignment.
• These are the standards by which learning is judged.
• Evaluation that attempts to uncover the strengths and weakness of a student or trainee in terms of what
he or she knows or doesn’t know, understands or doesn’t understand, or can do or cannot do, as
measured against a benchmark or standard. Also called content referenced assessment or criterion
referenced test.

One-Piece Flow / Continuous Flow

One-Piece Flow comes from the Toyota Production System, famous in the United States as Lean Manufacturing.
One-Piece Flow is also known by various other names such as the following: One-piece Flow; Make-one, Move-
one; Single-piece Flow; Continuous Flow; Flow Manufacturing.
One Piece Flow refers to the concept of moving one workpiece at a time between operations within a workcell.
At the opposite extreme, we might process an entire batch or lot at each operation before moving it to the next
operation.

Benefits:
• Improves safety
• Builds in Quality
• Reduces inventory
• Improves productivity
• Frees up floor space
• Reduces waiting time

One-piece flow means that parts are moved through operations from step-to-step with no work in process in
between; either one piece at a time or a small batch at a time. Batching and queuing tends to drive up inventory
and lead time, and creates inefficiency in an operation.
One Piece Flow manufacturing refers to the method of production that moves product along the stages of
manufacturing once a single piece is complete, regardless of that piece’s relation to a batch or not.
For example, when a glass bottle is produced for a six pack of soda, it does not matter if the other five bottles
are completely labelled before moving on to its bottle being capped, the other five will move along when they
have individually completed the steps required to finish that stage of production. The six pack, or batch, is
bundled at the end of production, not at every step.
CONTROL PHASE

Control Phase Overview

In summary: For DMAIC Projects, develop and implement a control plan and ongoing measurement, transition
and hand over to system owner, close the project. The CONTROL phase uses Descriptive Statistics to monitor the
ongoing performance of the improved process.
Purpose: The primary objective of the Control phase is to ensure that the gains obtained during Improve are
maintained long after the project has ended. To that end, it is necessary to standardize and document
procedures, make sure all employees are trained and communicate the project’s results.
In addition, the project team needs to create a plan for ongoing monitoring of the process and for reacting to
any problems that arise.

Key Deliverables of Control Phase:

• Documented and implemented monitoring plan
• Standardized process
• Documented procedures
• Response plan established and deployed
• Transfer of ownership (project closure)

Control Plans
Most organizations are good at rallying around chronic problems and finding solutions. But long term discipline
in sustaining those solutions is a different story – all too often a process will slowly return to its original state as
the path of least resistance takes control in the manufacturing environment.
This is where control plans come into play: control plans play a vital role in sustaining process improvements
over the long run.
Control plans are typically Excel documents that list the product and process characteristics that must be
monitored during the manufacturing process, including measurement methods and necessary reaction plans for
deviant conditions.
Without control plans, most processes will gradually drift into a state of disrepair due to employee turnover,
knowledge loss, and short-term production priorities.

What Makes an Effective Control Plan?

• 1. Product Specifications are clearly defined and linked to customer requirements.
• 2. PFMEA: A PFMEA (Process Failure Modes and Effects Analysis) has been conducted on all process steps,
and the team fully understands the quality risks associated with each step.
• 3. Effective Measurement Methods and Reaction Plans: Process and product measurement methods
must be clearly defined and repeatable.
• 4. Internal Audit: Internal audit is essential for maintaining control plan discipline. Effective internal audit
programs include a few key elements:
• An underlying SOP (standard operating procedure) that states control plan responsibilities such as
maintaining and following control plans.
 • Process engineers typically publish and maintain control plans.
• An internal audit calendar that shows which production areas will be audited at specific times throughout
the year.
• Management review of audit findings.
• Management must review audit findings and address them. This activity is extremely important, and will
strengthen control plan discipline in the facility.

FMEA (Failure Mode and Effects Analysis)

FMEA can be used during Analyse, Improve, or Control to highlight aspects of a product or process that should
be targeted for improvement (without any failures having actually occurred).
An FMEA is essentially a risk analysis tool that can be useful in environments where you have to prevent an
event ever happening (e.g. safety) or where the failure rate for a process is so low that there is little opportunity
to learn from past failures.
• Product FMEAs: Analyse the function, design, and potential failure of each component of a product
• Process FMEAs: Analyse the key outputs and potential failure of each process step of a process, and
consider the effect of process failure on the product or service concerned.

The FMEA Process

• Step 1: Identify process steps or components of the product
• Step 2: For each process step (or component), list the different failure modes that might occur, and rate
their severity (SEV)
• Step 3: For each failure mode, consider the different potential causes that might cause the failure and rate
their occurrence (OCC)
• Step 4: For each potential cause, consider the controls that are in place to prevent it happening and/or to
detect the failure if the cause does occur. Then rate the likelihood of detection (DET).
• Step 5: Calculate a Risk Priority Number (RPN) for each potential failure by multiplying the severity by
occurrence by detection.
• Step 6: Act on the results! Assign actions to tackle the highest RPNs, and set a date to review progress.

The FMEA structure above shows a typical format. The left hand side (first nine columns) are completed during
the first draft of the FMEA, and the right hand side (remaining seven columns) are used to track subsequent
actions and document reduced RPNs. FMEAs should be seen as living documents, not something that you do and
store away in a filing cabinet somewhere. Having identified a risk, you must do something about it!

Statistical Process Control (SPC)

Developed by Shewart in 1924. SPC is a method to operate process efficiently using statistical methods (control
charts). It uses statistical analysis tools to identify sources of variation and controls the process to meet the
goals. Applicable when mistake–proofing cannot be done. SPC may be eliminated over time when the process
inputs are found in place after regular monitoring. SPC uses control charts to monitor the process.

Types of Variation
• Random Variation (Natural Cause): also called as common cause variation; natural for the process; occurs
randomly; generally process remains in control
• Non-Random Variation (Special Cause): also called as special cause variation; result of some specific
assignable cause; process goes out of control. Typical sources: Operator, Shift, Raw materials
Advantages of SPC
• To identify assignable causes and eliminate them from the process living process alone with chance causes
• There will be guarantee of the product to attract the consumer
• If the testing is of destructive nature a fixed quantity can be assured of the product which is not tested
• QC helps to maintain a better quality with less cost
• There is a quality consciousness in the industry due to the proper quality control
• It reduces waste of time and material

Control Charts for Variables

The three main types are:
• I–MR Chart (Individuals – Moving Range Chart) – used when subgroup size = 1, n = 1
• X-bar Chart (Mean Chart) – used when subgroup size > 1
• R Chart (Range Chart) – used in conjunction with the X-bar Chart

I–MR Chart
• I-MR Chart: An I-MR chart is a combination of control charts used to monitor the process variability (as the
moving range between successive observations) and the process average (as the mean) when measuring
individuals at regular intervals from a process.
• I Chart (Individuals Chart): An I-chart is a type of control chart used to monitor the process mean when
measuring individuals at regular intervals from a process.
• MR Chart (Moving Range Chart): An MR-chart is a type of control chart used to monitor process variability
(as the moving range of successive observations) when measuring individuals at regular intervals from a
process.
I-MR Chart – Visual Example:
The I-MR chart displays two panels stacked vertically:
• Upper panel (I Chart): plots each individual measurement. The centre line is X-bar (mean of all
observations). The Upper Control Limit (UCL) and Lower Control Limit (LCL) are the natural process limits.
• Lower panel (MR Chart): plots the moving range between consecutive observations. The centre line is MR-
bar. The UCL is calculated from D4 × MR-bar.
I-MR Chart – Formulas for Control Limits:
The formulas to calculate the Control Limits for both individual values and the average moving range are given
below. Here m is the total number of data points.

Term Formula
kth (two-point) moving range mRk = |Xk+1 − Xk|
Average moving range (MR-bar) mR̅ = (1/(m−1)) × Σ mRk
Center line (the average of individual values) CLX = X̅ = (1/m) × Σ Xk
Upper Natural Process Limit UNPLX = X̅ + 2.660 mR̅
Lower Natural Process Limit LNPLX = X̅ − 2.660 mR̅
Center line (avg moving range) CLR = mR̅
Upper Control Limit for MR UCLR = D4 × mR̅ = 3.268 mR̅
 Term Formula
Lower Control Limit for MR LCLR: Does not exist for n = 2 to 6; LCLR = D3 × mR̅ for
n>6

Where:
• m = total number of data points
• mRk = kth two-point moving range = |Xk+1 − Xk|
• D4 = 3.268 (control chart constant for n = 2)
• D3 = not applicable for n = 2 to 6 (LCL does not exist)

X-Bar and R Charts

X-bar and R charts are typically used when multiple measurements of process behaviour are gathered in a short
time frame and grouped into sub-sets (subgroups) that can reasonably be expected to contain only common
cause variation.
X-bar charts answer questions like: “What is the central tendency of the data?” and “How much variation occurs
from sub-group to sub-group over time?”
The corresponding R charts indicate the variation within the subgroups.

X-bar Chart:
If we want to control the process mean in any production process, we use the control chart for mean — i.e., the
X-bar chart. With the X-bar chart, we monitor the variation in the mean of the samples that have been drawn
from time to time from the process. We plot the sample means instead of individual measurements on the
control chart because sample means are calculated from n individuals and give additional information. The X-bar
chart is more sensitive than the individual chart for detecting changes in the process mean.

R Chart:
The underlying logic and basic form of an R-chart are similar to the X-bar chart. The primary difference between
the X-bar chart and the R-chart is that instead of plotting the sample means and monitoring their variation, we
plot sample range and monitor the variation in sample ranges. To get the control limits for the R-chart, we need
to know the sampling distribution of the range.
Procedure for Computing Control Limits for X-bar and R Charts:
• Compute the average X and range R for each subgroup of size n, for each of the k subgroups: XK = (X1+X2+
……+Xk)/k and Rk = |Xmax − Xmin|
• Compute the grand average X̅ by averaging each of the k subgroups averages: X̅ = (X̅ 1+X̅ 2+…+X̅ k)/k
• Compute the grand range R̅ by averaging each of the k subgroups ranges: R̅ = (R1+R2+…+Rk)/k

Chart Centreline UCL LCL

X-bar Chart CLX = X̅ UCLX = X̅ + A2R̅ LCLX = X̅ − A2R̅
R Chart CLR = R̅ UCLR = D4R̅ LCLR = D3R̅

Constants Table for X-bar and R Charts (A2, D3, D4):

 Sample Size (n) A2 D3 D4
2 1.880 0 3.267
3 1.023 0 2.574
4 0.729 0 2.282
5 0.577 0 2.114
6 0.483 0 2.004
7 0.419 0.076 1.924
8 0.373 0.136 1.864
9 0.337 0.184 1.816
10 0.308 0.223 1.777

Control Charts for Attributes

Attribute control charts are similar to variables control charts except they plot proportion or count data rather
than variable measurements.
Attribute control charts have only one chart which tracks proportions or counts over time (there is no range
chart or standard deviation chart like there is with continuous data).

Binomial Data
When data points can have only one of two values — such as when comparing a product or service to a standard
and classifying it as being acceptable or not (pass/fail) — it is called binomial data.
• p-chart: Charts the proportion of defectives in each subgroup
• np-chart: Charts the number of defectives in each subgroup (must have same sample size each time)

P-charts are often used in transactional situations: billing errors, defective loan applications, proportion of
invoices with errors, defective room service orders, sales order data, etc.

p-Chart — Formulas (Standards given):

• CL = p̅
• UCL = p̅ + (3/√n) × √[ p̅ × (1 − p̅) ] = p̅ + A × √[ p̅ × (1 − p̅) ]
• LCL = p̅ − (3/√n) × √[ p̅ × (1 − p̅) ] = p̅ − A × √[ p̅ × (1 − p̅) ]

np-Chart — Formulas:
• UCL = n × p̅ + 3 × √( n × p̅ × q̅ )
• LCL = n × p̅ − 3 × √( n × p̅ × q̅ )
• where q̅ = (1 − p̅)

Poisson Data
A Poisson distribution describes count data where you can easily count the number of occurrences (e.g., errors
on a form, dents on a car), but not the number of non-occurrences.
• c-chart: Charts the defect count per sample (must have the same sample size each time)
 • u-chart: Charts the number of defects per unit sampled in each subgroup (uses a proportion, so its OK if
sample size varies)

Rule: If the sample size is always the same (up to 10% variation in sample size is acceptable), use c-charts. If the
sample size varies, use the u-chart.

c-Chart — Formulas (Standards given):

• CL = λ (lambda, the process average number of defects)
• UCL = λ + 3 × √(λ)
• LCL = λ − 3 × √(λ)

Control Limit Formulas for Attribute Data – Summary Table

Chart Type Centerline Upper Control Limit (UCL) Lower Control Limit (LCL)
p p̅ UCL = p̅ + 3√(p̅(1−p̅)/n) LCL = p̅ − 3√(p̅(1−p̅)/n)
np np̅ UCL = np̅ + 3√(np̅(1−p̅)) LCL = np̅ − 3√(np̅(1−p̅))
c c̅ UCL = c̅ + 3√c̅ LCL = c̅ − 3√c̅
u u̅ (u̅ = c̅/n) UCL = u̅ + 3√(u̅/n) LCL = u̅ − 3√(u̅/n)

Where:
• p̅ = average proportion defective
• n = sample size
• c̅ = average number of defects per sample
• u̅ = average number of defects per unit (u̅ = c / n)
• np̅ = average number of defectives (n × p̅)
Complete DMAIC Summary

Phase Key Question Key Tools Key Output

DEFINE What is the problem? SIPOC, VOC, CTQ Tree, Project Project Charter, Problem
Charter, Stakeholders Analysis Statement, Goal Statement,
Business Case
MEASURE How big is the Control Charts, MSA, Gauge R&R, Baseline Performance, Process
problem? Process Capability (Cp/Cpk), Data Capability, Measurement Plan,
Collection, Operational Data Collection Plan
Definitions
ANALYSE What are the root Fishbone/Ishikawa, Pareto, 5 Root Causes Identified and
causes? Why’s, Hypothesis Testing, Validated, Data Door Analysis
Regression, Correlation, DOE
IMPROVE How do we fix it? Brainstorming, SCAMPER, Poka- Validated Solutions, Pilot
Yoke, Assumption Busting, Pilot Results, Optimised Process
Studies, Pugh Matrix,
Prioritization Matrix, Paired
Comparison, One-Piece Flow
CONTROL How do we sustain Control Plans, SPC, FMEA, Control Plan, Standardized
the gains? Statistical Charts (I-MR, X-bar R, p, Process, Ongoing Monitoring
np, c, u), Standard Operating System, Project Closure
Procedures