Six Sigma Tutorial

Pursuit of Perfection

Six Sigma is used by virtually all industries to improve quality and overall results. It is a structured and disciplined approach of reducing variability in products and processes, based on principles from business, statistics, and engineering. The ultimate goal of Six Sigma is achieving 'Perfect Quality' below 3.4 defects per million opportunities. Results are more reliable products (higher MTBF) and more stable and predictable processes (higher Cp, Cpk). Six Sigma methodology is applicable to manufacturing and service, but also government and infrastructure. Leading organization makes 'Quality' a central element of corporate plan and strategy.

Reducing Variability is Key

Six Sigma stands for six standard deviations from process average, represented by the Greek letter 'Sigma'. It means that the typical variation of a product or process is six times smaller than the tolerance limit. Less variability leads to higher capability, profitability, employee-motivation, and customer satisfaction. Six Sigma is being trained and implemented by Green-Belts and Black-Belts. Green Belts are quality specialists, improving product reliability and process capability. Black Belts are quality executives, leading corporate-wide Six Sigma initiatives and coaching Green Belts.

Benefits from less Variability

• Direct benefits: lower operating cost, controlled processes, predictable results
• Indirect benefits: trust from stable and reliable output, creates a competitive edge

Defects are Expensive!

• Non-confirming output must be corrected or scrapped, market-potential is lost
• Cost of rectification and lost potential make up 'Cost of Poor Quality' (COPQ)
• COPQ includes costs for inspection, rework, overhead, lost future orders
• Defects cost 5% of Sales at 6 Sigma and 20% of Sales at 3 Sigma level
• YIELD is the process pass-rate with 'single touch' from start to finish
• DPMO Defects Per Million Opportunities is the defect-rate for each step

The table below shows quality versus yield, defect level, and cost impact. Long-term effects have been accounted for, assuming +1.5…-1.5 Sigma drift around the mean to reflect a real operating environment.

Quality	Yield	DPMO	COPQ
1 Sigma	32%	700.000	45%
2 Sigma	69%	310,000	30%
3 Sigma	93%	67,000	20%
4 Sigma	99.4%	6,200	15%
5 Sigma	99.97%	240	10%
6 Sigma	99.999%	3.4	5%

Short-term versus Long-term Quality

Short-term variability (quality right after adjustment) is significantly higher than long-term as variability increases over time. Per empirical rule, the quality level can be translated into yield: 1 Sigma = 68%, 2 Sigma = 95%, 3 Sigma = 99.7%.

Six Sigma Techniques

Virtually all businesses can benefit from Six Sigma by using two proven techniques:

• DMAIC – Design-Measure-Analyze-Improve-Control – improvement-process
• DFSS – Design for Six Sigma – reliable and high-yield products and services

Six Sigma Process (DMAIC)

Six Sigma emphasizes the structured, scientific experimentation and uses statistical tools in a systematic project-oriented manner through the DMAIC-improvement-cycle:

1. DEFINE opportunities for improvement
2. MEASURE baseline performance
3. ANALYZE causes of non-conformance
4. IMPROVE by eliminating root-causes
5. CONTROL by locking improvements

Six Sigma Design (DFSS)

Design for Six Sigma (DFSS) creates reliable products with very little variability at near perfect quality. Instead of correcting deviations, DFSS focuses to building-in quality, to 'get it right the first time'. Design for Six Sigma improves margin and creates a competitive advantage in the long run:

• Short-term: savings are directly proportional to the reduction in scrap and rework
• Long-term: higher market potential from reliable product and stable deliveries

Operating at high Sigma level allows repositioning the product or serve to a segment where customer are willing to pay a premium for quality, and where the firm is less likely to be attacked from competitors.

Limits – where Process Management does not work

Correctly applied, process management (like Six Sigma, TQM, ISO9000) improves quality and efficiency but is not effective everywhere. Process management focuses on continuous improvement (Kaizen) which limits innovation by restricting thinking on what exists already. Especially in areas that are exploratory and where radical thinking (Kaikaku) is necessary to find new ways and break through traditional layers and thinking.

Guidelines for successful Process Management

• Apply where processes are data-driven – do not apply to creative processes
• Some variations do not cause problems – avoid 'Quality Overkill'
• Avoid where processes 'touch the customer' – focus on relationships, not data
• Lower Sigma level may be acceptable – consider benefit versus cost

Six Sigma Example – filling Bottles at 3 and 6 Sigma Quality

A wine-bottling company wants to improve the accuracy of their filling process. The target is 1000ml with 6ml tolerance to each side (plus/minus). The following calculation shows the impact of reducing variability from 3 to 6 sigma.

We assume the process is normal, stable, and repeatable (no strike, no power-outage). Short-term performance shows variability right after the machine was adjusted. Long-term performance considers drift over time, the process-average shifts around the mean, caused by people (sickness, mistakes), materials (aging), machines (breakdowns), environment (power outages). The typical value for long-term drift is a 1.5 sigma shift (difference between newly adjusted and long-term).

• Requirements and Specifications
  Filling bottles to target: T = 1000ml
  Acceptable tolerance: dT = ± 6ml
  Lower Specification Limit LSL = 994ml
  Upper Specification Limit USL = 1006ml
• 3 Sigma Process short-term
  Machine just calibrated, average on target
  Process is centered, no shift: μ = 0
  Process variation: μ–3σ…μ+3σ = 6σ
  Typical variation: (1006–994)/6 = 2ml
  Defectives: 2,700ppm = 1 of 370 fail
• 3 Sigma Process long-term
  Shift due to operator, material, temp etc.
  Average deviates from target by ±1.5σ
  Total process-shift μ = 3σ
  Defectives: 66,807ppm = 1 of 15 fail
• 6 Sigma Process short-term
  Machine just calibrated, average on target
  Process is centered, no shift: μ = 0
  Process variation: μ–6σ…μ+6σ = 12σ
  Typical variation: (1006–994)/12 = 1ml
  Defectives: 0.1ppm = 1 of 10,000,000 fail
• 6 Sigma Process long-term
  Shift due to aging, operator, material, temp etc.
  Average deviates from target by ±1.5σ
  Defectives: 3.4ppm = 1 of 294,000 fail

Conclusion: the 6 Sigma process produces 794 times less defects compared to 3 Sigma. Defects from non-conformance are costly due to (a) higher expenses from rework and scrap and (b) lower margin from selling a lower quality product at a discount.

Cost of Quality

The Economical Quality Level (EQL) is the point where quality-cost (for reaching and maintaining a high level of quality) divided by the 'Cost of Poor Quality' (COPQ) reaches a maximum.

Cost Of Poor Quality

COPQ includes expenses for identifying the defect, rework, repair, replacement, communicating to customer, spending additional time and money to smoothen the situation, additional administrative expenses, lost image and market-potential etc. Putting it into perspective:

• If all orders would have been fulfilled perfectly, how much MORE money could you have made?
• How much money did you actually spend on rectifying problems and handling emergencies?

Six Sigma Project Basics

1. ANALYSIS – what is the situation and condition?
   Reviewing current performance and environment
2. TREND – what is going to happen?
   Projecting the future based on baseline and trend
3. REASON – what can be done?
   Translating findings into one compelling reason to act
4. PLAN – how to make it happen?
   Planning implementation, resources, tasks, expectations

Six Sigma Project Phases

1. DEFINE – Outline and Charter
   Tasks: understand how things work and framing the project
   Deliverables: current-state-map and project-charter
2. MEASURE – Baseline, Trend, Impact, Benchmarks
   Tasks: measuring performance against target
   Deliverables: metrics, logs, charts
3. ANALYZE – Root-Causes
   Tasks: analyzing deviations and identifying root-causes
   Deliverable: Failure Mode and Effect Analysis (FMEA)
4. IMPROVE – Eliminating Causes
   Tasks: implementing improvements and coaching people
   Deliverable: rollout and skill development plan
5. CONTROL – sustaining improvements
   Tasks: standardize processes, transfer knowledge
   Deliverable: results are validated, processes are stable

Impact from Six Sigma Rollout

A successful implementation leads to higher capability short-term and improved competitiveness long-term. Resources occupied with fire-fighting and rework can now be reassigned to advance and grow the business:

• Increasing RANGE by adding more products or services
• Increasing VALUE by adding features or lowering cost
• Increasing QUALITY by serving the premium segment