2.3 Additional Quality Pioneers: Ishikawa, Shewhart, Feigenbaum & Taguchi

2.3 Additional Quality Pioneers: Ishikawa, Shewhart, Feigenbaum & Taguchi

While Deming, Juran, and Crosby provided high-level philosophical and managerial frameworks, the technical, statistical, and operational toolkit of modern quality was forged by four other seminal masters: Walter A. Shewhart, Kaoru Ishikawa, Armand V. Feigenbaum, and Genichi Taguchi. Mastery of their specific tools, models, and formulas is essential for the ASQ CQIA examination.

---

1. Walter A. Shewhart: The Father of Statistical Process Control

Dr. Walter A. Shewhart (1891–1967), a physicist and statistician at Bell Telephone Laboratories and Western Electric's Hawthorne Works, is universally recognized as the Father of Statistical Process Control (SPC).

The Invention of the Control Chart (1924)

On May 16, 1924, Shewhart authored a historic one-page memorandum introducing the world's first statistical control chart. Shewhart realized that while all processes exhibit variation, that variation falls into two fundamentally distinct categories:

• Common Cause Variation (Chance Causes): Inherent, random background variation natural to a stable system (e.g., slight ambient temperature shifts, normal bearing vibration). Common causes can only be reduced by management changing the underlying process design or equipment.
• Special Cause Variation (Assignable Causes): Specific, identifiable, intermittent disruptions external to the normal process (e.g., a broken cutting tool, an untrained operator, a contaminated batch of raw material). Special causes must be detected, isolated, and eliminated by frontline operational personnel.

                     SHEWHART CONTROL CHART
      ▲
Value │
      │  Upper Control Limit (UCL = μ + 3σ)
──────┼───────────────────────────────────────────── [Statistical Limit]
      │       *       *               *
      │   *       *       *       *       *         [Common Cause: Stable]
      │                       *               *
──────┼── Process Centerline (μ) ───────────────────
      │               *               *
      │   *       *       *       *       *    * <-- Out of Control (Special Cause)
──────┼─────────────────────────────────────────────
      │  Lower Control Limit (LCL = μ - 3σ)
      └─────────────────────────────────────────────► Subgroup / Time

The $3\sigma$ Economic Limits

Shewhart established control limits placed at $\pm 3$ standard deviations ($3\sigma$) from the process centerline ($\mu$). He chose $3\sigma$ because it struck an optimal economic balance between two costly errors:

1. Type I Error (Alpha Risk): Looking for a special cause when only common cause variation is present (false alarm).
2. Type II Error (Beta Risk): Failing to detect a true special cause when one has occurred (missed signal).

The Shewhart Cycle: Plan-Do-Check-Act (PDCA)

Shewhart introduced the iterative four-stage cycle for continuous process improvement and scientific problem-solving. Later popularized by Deming (who modified it to Plan-Do-Study-Act, or PDSA), the cycle consists of:

1. Plan: Formulate an objective, establish baseline hypotheses, and design the experiment or change.
2. Do: Execute the plan on a small, controlled scale.
3. Check / Study: Analyze experimental data, observe results, and verify against predictions.
4. Act: Standardize successful changes across operations or revise hypotheses and re-cycle.

---

2. Kaoru Ishikawa: Democratizing Quality & Total Participation

Dr. Kaoru Ishikawa (1915–1989), an engineering professor at the University of Tokyo and leader in JUSE, believed that quality should not be restricted to statisticians and specialists. He dedicated his career to democratizing quality tools so that frontline workers could solve operational problems independently.

The Cause-and-Effect Diagram (Fishbone / Ishikawa Diagram)

In 1943, Ishikawa developed the Cause-and-Effect Diagram (often called the Fishbone Diagram due to its skeletal appearance). The diagram visually organizes potential root causes into structured categories pointing toward a specific effect or problem statement.

In manufacturing, causes are typically categorized using the 6 Ms:

• Manpower (People): Training, operator fatigue, ergonomics, skill level.
• Methods: Standard operating procedures, work instructions, process sequence.
• Machines: Tool wear, calibration, preventive maintenance, equipment speed.
• Materials: Raw material composition, supplier variation, moisture content.
• Measurement: Gauge repeatability and reproducibility (GR&R), calibration drift, visual bias.
• Mother Nature (Environment): Ambient temperature, humidity, lighting, dust, vibration.

                  CAUSE-AND-EFFECT (ISHIKAWA) DIAGRAM

   Methods         Machines         Materials
       \               \               \
        \               \               \
─────────┴───────────────┴───────────────┴─────────► [ PROBLEM / DEFECT ]
        /               /               /
       /               /               /
   Manpower       Measurement      Environment

Quality Circles (1962)

Ishikawa founded the Quality Control Circle (QC Circle) movement in Japan. Quality Circles are small groups of frontline workers (typically 4 to 10 employees from the same work area) who meet voluntarily on a regular basis to identify, analyze, and solve work-related quality problems using basic analytical tools.

Company-Wide Quality Control (CWQC) & Internal Customers

Ishikawa expanded quality management across all non-manufacturing business functions (finance, marketing, human resources, field service). He coined the famous principle: "The next process is your customer", emphasizing that every employee serves an internal customer whose requirements must be fully satisfied.

The 7 Basic Quality Tools

Ishikawa asserted that 95% of quality problems in an organization can be solved using the 7 Basic Quality Tools:

1. Cause-and-Effect Diagram
2. Check Sheet
3. Control Chart
4. Histogram
5. Pareto Chart
6. Scatter Plot
7. Flowchart / Stratification

---

3. Armand V. Feigenbaum: Total Quality Control & Cost of Quality

Dr. Armand V. Feigenbaum (1922–2014), Director of Manufacturing Operations at General Electric, published the seminal book Total Quality Control in 1951. Feigenbaum was among the first to view quality as an end-to-end strategic system.

Total Quality Control (TQC)

Feigenbaum defined Total Quality Control as:

"An effective system for integrating the quality-development, quality-maintenance, and quality-improvement efforts of the various groups in an organization so as to enable marketing, engineering, production, and service at the most economical levels which allow for full customer satisfaction."

He famously observed: "Quality is everybody's job, but because it is everybody's job, it can become nobody's job" unless executive leadership establishes a clear, cross-functional organizational structure with designated responsibilities.

Originator of the Cost of Quality (COQ) PAF Model

Feigenbaum was the first to systematically classify the Cost of Quality (COQ) into the classic PAF Model:

• Prevention Costs: Investments made to prevent defects from occurring (e.g., training, design reviews, supplier qualification).
• Appraisal Costs: Expenses associated with inspecting, testing, and auditing products and processes (e.g., in-process testing, gauge calibration).
• Internal Failure Costs: Costs resulting from defects identified before product delivery to the customer (e.g., scrap, rework, reinspection).
• External Failure Costs: Costs resulting from defects identified after product delivery to the customer (e.g., warranty claims, recalls, customer returns, litigation).

The "Hidden Plant"

Feigenbaum introduced the concept of the Hidden Plant (or Hidden Factory)—the proportion of an organization's manufacturing and operational capacity (often estimated at 20% to 40%) that is wasted producing scrap, performing rework, investigating defects, and dealing with customer complaints. Eliminating this hidden waste releases enormous untapped capacity without requiring new capital expenditures.

---

4. Genichi Taguchi: Robust Design & The Taguchi Loss Function

Scope note. Body of Knowledge entry I.C names exactly six foundational thought leaders: Shewhart, Deming, Juran, Ishikawa, Crosby, and Feigenbaum. Taguchi is not among them. He is included here because the loss function is the clearest available argument against "conformance to specification is good enough," a concept that does appear on the exam through variation and process capability. Learn Taguchi for the idea; do not expect a question that asks you to name him as one of the BoK's foundational leaders.

Dr. Genichi Taguchi (1924–2012) revolutionized modern quality engineering by integrating statistical methods into product and process design.

The Taguchi Loss Function

Traditional manufacturing operates on a "Goalpost Mentality" (binary pass/fail compliance). Under this old view, any part produced within tolerance limits (Upper Specification Limit $USL$ and Lower Specification Limit $LSL$) incurs zero financial loss, while any part outside incurs a step-function loss.

Taguchi demonstrated that this assumption is fundamentally flawed. Quality loss does not suddenly jump at the specification limits; rather, financial loss to society begins the moment a product parameter deviates from the target (nominal) value ($m$) and increases quadratically:

$L(y) = k(y - m)^2$

Where:

• $L(y)$ = Financial loss incurred when the quality characteristic equals $y$
• $y$ = Actual measured value of the quality characteristic
• $m$ = Target (nominal) optimal value
• $k$ = Cost coefficient constant determined by the cost of replacement or repair at the specification limit ($k = \frac{\text{Cost}}{(\text{Tolerance})^2}$)

Loss ($)
  ▲                 TAGUCHI QUADRATIC LOSS FUNCTION
  │                            L(y) = k(y - m)^2
  │       \                                         /
  │        \                                       /
  │         \                                     /
  │          \                 Target            /
  │           \                  │              /
  │────────────\─────────────────┼─────────────/──────────── [Traditional Step Loss]
  │             \                │            /
  │              \_              │          _/
  │                \_____________│_________/
  └──────────────────────┬───────┴───────┬──────────────────► Characteristic (y)
                        LSL      m      USL

Robust Parameter Design

Taguchi divided product and process design into three progressive stages:

1. System Design: Selecting basic technologies, architectures, and materials.
2. Parameter Design (Robust Design): Setting nominal factor levels to make the system insensitive (robust) to noise factors (environmental temperature, component aging, manufacturing variation) without eliminating the noise sources themselves.
3. Tolerance Design: Tightening tolerances and buying higher-grade components only when parameter design fails to achieve required robustness.

Orthogonal Arrays & S/N Ratios

Taguchi developed streamlined orthogonal arrays for Design of Experiments (DOE), allowing engineers to test multiple factors simultaneously with significantly fewer experimental runs than full factorial designs, measuring stability using Signal-to-Noise (S/N) Ratios.

---

5. Master Comparison: Additional Quality Pioneers