VARSIGMA

Total quality management

Total productive maintenance

Advanced product quality planning

Toyota production system

Most of the lean six sigma tools have been around for more than 100 years.

Motorola:

- Correlate $$$$ with performance indicators

i) Quality as you know it was a by product

Reactive

- Customer service
- - Warranty
- Quality check
- Supervision
- Re-training

$$$$ - How much money are we spending/wasting on this ?

- While process improvement look for the feedback from bottom of the pyramid to improve
process

“Professor Temple grandin on how to psychology train cows”

Metrics – Performance Indicators – These performance indicators are reviewed periodically. If we

observe an underperforming indicator, we initiate correction, corrective action and preventive
action

Methodology – If an indicator continues to underperform even after repetitive effort, then we

consider it a chronic pain area.

We initiate a structured improvement initiative

- Define, Measure, Analyse, Improve, Control

Management System – Whenever we consider implementing new process, we may want to design
the new process with the same rigor of designing a product/service

- Define, Measure, Analyse, Design, Verify/Validate

Can we sustain the spirit of data driven thinking (decision making) and innovation?

- How do we engage all levels of the pyramid in this journey?

- How do we get all team to participate in collecting customer feedback and understanding
business rationale?
 Lean – Was coined by James Womack – Lean Enterprise institute

- PHD in industrial policy in Japan, America and Germany

- He observed – Japanese can do MORE with less

Six Sigma – Introduced and term coined by Motorola

Tools and techniques

- Agricultural Statistics
- Industrial Engineering
- Operations Research
- Econometrics
- Cost Finance

Operations Management

DMAIC

- Define - What should we improve?

i) Identify measurable parameters
ii) Select a measurable parameter which is a chronic pain area
iii) Prepare proposal for Senior management for a project to alleviate a chronic pain
area – project Charter
iv) Develop a high level process map – explain ‘As is’ Process to your team members
- Measure - What is our current performance and gap?
i) Measurable parameter – Measurement definition
ii) Validate Measurement system
iii) Evaluate / assess current performance
- Analyse - What are the factors responsible for gap?
i) Identify factors for waste, risks and variation
ii) Validate factors using data
iii) Correlate critical factors with business impact ($$$)
- Improve - How do we bridge these gaps?
i) Generate solution
ii) Evaluate solution
 Paper
 Mock
 Pilot
iii) Full scale implementation
- Control - How do we sustain improved performance?
i) Implement control system for critical X, Y
ii) Process Hanover to process owner
iii) Lessons learnt documentation
iv) Executive summary and projected financial benefits

 Look for the term “Autonomation”

  Read about THERP on human error – rate prediction

DMAIC
Define:

- Generate project Ideas

 VOC,VOB, COPQ should be translated to measurable parameters
 I waited for an hour to purchase a movie ticket
o Y – Waiting time
o Y is a measurable parameter which reflects the voice of business, customer,
financials
 There are too many incorrect entries in a form. We have to re-process it
o Y – Error %
 List all Y measures for your process
 Any Y measure that is under – performing is a potential project
 Product – Football – What should we measure in a football to be able to state
that it does what it should do?
 Y1 – Circumference
 Y2 – Weight
 Y3 – Rebound (20 and 5 degree Celsius)
 Y4 – Sphericity (45000 roller point check)
 Y5 – Pressure loss rate (72 hour)
 Y6 - Water Absorption (250 compressions)
 Y7 – Stress test
o Stress
2000 propulsions (Kicks)
50KMPH
2.5 metre travel distance
Impact: Steel plate
o Y1, Y4, Y5
o Defects- Seam, Valve

- Citrus
 5 fly eggs per 250 ml
 1 maggot per 250 ml

Y – Average Handling
Opportunity/Problem statement – Why do we need to work on this initiative?

Business case

- Financial Benefits - $$$$$ (Opportunity)

 Not to the penny accuracy, you will be able to provide a definitive estimate only in
improve phase
- Customer Impact – Neural to positive

Scope

- What are the boundaries of the initiative

- In/ Out scope

Y measure and goal

- How do we measure success?

Team members

- Sponsor
- Process Owner, Subject Matter Expert
- Engage cross functional representatives

Milestones

- Projected DMAIC tollgate review schedule

 Week ending
-

Measure Phase:

Performance Standards

Operational Definition

- Data collection definition

- Clear, concise, unambiguous definition
- Ensures uniformity in data collection across the organization

Specification limits

- Customer specified conformance boundary for a desired characteristic

- Limits within which our product/ services should operate

Step 4 – Performance Standards for Y (Pg 58)

Measure:

Objective: Assess current performance, establish baseline and determine gap from desired
performance.
What do we need? – We need data

Where would I get the data frm? – Measurement system

Can we trust the data we have collected? – ohh, we should validate the measurement system

Have we define the measurement? ---ohh, we should document the definition of the measurement
system

Step 4 - Y performance standards

- Operational definition
o Clear, concise, unambiguous definition recorded for collecting data
o Appraiser/Data collector, Prodedure to record measurement, Tools/Gauge to
record measurement
o Please not that a measurement system is
 Appraiser/ Human/ Machine
 Tools/Gauge to record the measurement
 Procedure/ process to record the measurement
- Specification limits, Tolerance
o Specified/ stated by a customer/ client
o Specified by a third party body protecting the interest of a customer (FDA, AIAG,
NIST)
o Competitive Benchmarking

Step 5 – Validate the measurement system

- Objective – Assess the reliablitiy of the data collected using the existing measurement
system.
- Types of measurement
o Automated measurement
 Record/ Capture, Storage and retrieval is by a machined
 Check for – Periodic calibration, Periodic maintenance (schedule from
manufacturer)
o Manual measurement system – You have to perform a study to establish the
reliability
 Continuous data – Gage R&R study
 Discrete data – Attribute Agreement Analysis

Step 6 – Evaluate current performance

- Continuous
o Probability of Failure
o Process Capability indices – Cp, CP<
- Discrete
o Defect%
o Defects per unknit
 o Defective%

Represented in Z score (Standard Score) – Do you have to?

IS CP > 1.66? (at least > 1.33)

- If yes, Process is consistent “good”

- If no, Process inconsistent
o Reduce variation to improve

IS Cpk =~ Cp?

- If yes, Mean =~ Target

- If no, Mean is AWAY from target
o Move mean closer to target to improve process

At the conclusion of Measure phase, we have established a baseline and compared to expected
performance – GAP

There must be factors responsible for GAP in performance

These factors are generally from:

- Sources of waste – VA/ NVA analysis

- Sources of Risks – Failure modes and effect analysis
- Sources of Variation – Qualitative screening and Exploratory data analysis

Sources of Waste – VA/ NVA analysis

- List all activities in your process

- Determine time and cost of each activity
- Classify each activity as VA/NVA
o Value add:
 The activity transform the product/service – Core activity
 Customer is willing to pay for it – non – core activity
 It shouldn’t be WORMPIT
 Waiting
 Over-production
 Rework
 Motion
 Over-processing
 Inventory
 Transportation
 To make you happy – underutilization of talent of skills
o Recognize the NVAs
 Required -
 Statutory obligation – Audit, Filing
 Payroll
 Accounting
 o WASTE – WROMPIT
 Immediate action - $$$$$ - Please take a look at it

Customer’s perspective

For each activity – What can go wrong? – Failure modes and effect analysis

Balanced review of the process

- Activity
o Is it VA/ NVA
o What can go wrong? – FMEA

There are 3 sources of factors responsible for gap

- Waste
o Identify NVA
o Classify NVA
 Generate solutions to eliminate WASTE NVA
 Generate solutions to replace required NVAs with robust processes
- Risks
o Risk Identification
o Risk Prioritization
o Response plan
o Risk Re-evaluation
- Variation
o Identify potential factors
o Group potential factors
o Drill down each potential factor to actionable root cause
o Prioritize actionable root causes

FMEA – Failure modes and effect analysis

Background: Introduced by the US Army in 1940s

Objective: Proactively manage risk – What can go wrong?

Types:

- Design FMEA
- Process FMEA

Process FMEA

Input – Detailed Process map with a list of activities.

Risk Identification:

- List all activities and requirement

o Requirement – the output expected from each activity
- What can go wrong? – Potential failure mode
- What is the impact? Potential failure effect
- How sever is the impact? Severity rank
 o 10,9 – Hazardous Outcome/Statutory obligation
o 8 – Product/ Service inoperable
o 3-7 – your organization can standardize
o 1,2 – Defects with no impact on primary, secondary features of product/ service
(Only MiL can find these defects)
- What caused the failure? – Potential causes of failure
- How often do these causes result in failure? Occurrence rank
- How do we protect the customer? – Process controls
- How effective are these controls? Detection rank
- Risk Priority Number = * Occurrence * Detection

Risk Prioritization

- In general, higher the RPN, greater is the risk associated with the failure mode
- RPN > Threshold set by your enterprise should e logged in a risk register for review.
(Commonly observed – 150)
- RPN > 1000 with Severity 10,9 – Immediate review recommended
- You can use a pareto plot on the RPNs, and identify vital few failures that contribute to
majority of risks
- An alternative to RPN is the S*O number, in recent year, risk managers prefer
- S*O over RPN at design stage (Detection is a Non-Value add)
- All prioritized risks are logged in a risk register for review by risk review board

Response plan

- Risk Review Board:

o Senior management
o Functional Expert
- Risk Review Board Tasks
o Recommend Response strategy (negative risk)
 Accept – Live with it
 Transfer – Share liability
 Contractor payroll
 Insurance
 Mitigate – Reduce probability
 Avoid – Eliminate possibility of the risk (Hazardous outcome)
o Assign risk owner
- Risk Owner
o Root cause analysis
o Generate and Evaluate solution
o Mock trials and Pilot
o Implement solution

Risk Re-evaluation:

- Re-evaluate Severity, Occurrence, Detection

o Severity change usually required design change
- RPN should subside below actionable threshold for the risk to be closed.
What happens in most brainstorming sessions?

- You will list all the common generic causes

o Lack of manpower, Lack of Training, and other generic terms

How will Japanese look at it:

Qualitative Screening

- Identify
o Collate all incidents, adverse events and near misses that have occurred over a
period (12 months
o Identify a primary cause for it
- Group
o Group these causes in relevant categories – Affinity Diagram
 Human/Man
 Machine/Hardware/Software
 Measurement/Metric
 Method/Procedure
 Material/Input
 Mother nature/ Environment
 External
- Drill down
o Drill down each cause to an actionable root cause
 5 Whys
 Ishikawa Diagram/Mind map
o Please expect this list to be extremely long
- Prioritize
o Data (Frequency of each actionable root cause) – Pareto Plot
o No data
 Voting – Multi – Voting
 Ranking – Nominal Group technique

Whenever you are thinking of actionable root causes, tell yourself – training, supervision and
inspection aren’t primary solutions.

List of factors:

- Data driven – You man want a team to start generating solutions

- Opinion – You must validate using data

How do we validate our opinion?

Population data – We have data for all observations/ nearly all observations

- Slice, Dice and Visualise

- Tableau, Qlikview, PoweBI(Excel friendly than this can be used)

Sample data

- Hypothesis test
No data

- Experimental Trials and Hypothesis Test

- Black Belt – Design of Experiments

Hypothesis test

- Objective – Draw inferences for population parameter using sample data

o Data collected: Sample
o Conclusion drawn: Population
- Y measure and data type
o Y – Cutlet diameter, continuous (any unit of measurement which includes
measurement, km, centimetre etc.)
o Y – Turnaround time, continuous
o Y- Transaction time, continuous
o Y – Buy/didn’t buy%, discrete
o Y-
- X factor and data type
o X – Production unit, discrete – 2 categories
o X – Laboratory, discrete – 4 categories
o X – Vendor, discrete – 3 categories
o X – Gender, discrete – 2 categories
o X – Region, discrete – 4 categories
- Parameter for comparison
o Y continuous, X discrete
 Variance/ Stdev
 Mean
o Y discrete, X discrete
 Proportions
- Test of comparison
- Ho, Ha
o Ho: Null
 =, statement of no difference
 Ho: Sweety – Good
 Ho: Accused = Innocent
o Ha: Alternate
 Not equal to, statement of difference
 Ha: Sweety not equal to Good
 Ha: Accused not equal to Innocent
- Acceptable error rate
o Type 1 error
 You reject Null though Null is true
 Sweety is Good however based on the evidence – you thinks she is not good
-----your Loss
 Probability of Type I error – alpha
 1-alpha = confidence level
o Type II error
 You stay with Null though Alternate is true
  Sweety is Not good however based on the evidence – you think she is good
- ----- ;☹☹☹ - Life ☹☹☹
 Probability of Type II error – beta
 1-beta = Power
- Minimum required sample (BB)
- Sampling Plan (BB)
- Collect data and perform the test
- Observe p-value and draw statistical inferences
o P value – actual probability of Type I error
o Rule of thumb
 P> significance level, Stay with NULL, Ho
 P>0.05, p high, null fly, Ho
 P< significance level, Go with Alternate, Ha
 P<0.05, p low, null go go, ha
- Correlate statistical inferences with business impact
-

Y – Weight Gained (Continuous)

X – Calories consumed (Continuous)

- Scatter plot – visually explore the relationship between X and Y

- Correlation – Determine the degree of linear association between 2 variables
- Regression – Determine an equation to predict Y for a given value of X