Intro to Business › Section J · Operations management › Topic J.5

Quality and continuous improvement

A defect caught by a customer costs far more than one prevented on the shop floor: it means refunds, repairs, lost sales and a damaged reputation. This topic follows how firms moved from inspecting quality at the end of the line to building it into every step, and to cutting every activity customers do not value.

What quality means

Customers judge quality by fitness for use: whether a product performs reliably, lasts, has the features they want and is backed by good service. Producers translate this into conformance: the degree to which output meets set specifications, such as a bolt's diameter or a call centre's maximum waiting time. Poor quality is expensive. There are the costs of scrap and rework inside the plant, of warranty claims and returns outside it, and of customers who never come back. Spending on prevention, through better design, training and process control, is usually far cheaper than paying for failures. Quality and productivity therefore go together: every defect wastes the materials and time that went into it.

From inspection to assurance

Quality control traditionally meant inspecting finished products and removing the defective ones. Inspection finds problems but does not stop them happening, and it is costly when defects are common. Quality assurance aims to prevent defects by building quality into every stage, through clear standards, capable processes, trained staff and checks at each step. A key tool is statistical process control (SPC), developed by Walter Shewhart, who drew the first control chart in 1924 as an engineer at Western Electric, whose research arm became Bell Telephone Laboratories the next year. A control chart plots a measurement from regular samples against limits, usually set three standard deviations either side of the process average. Points outside the limits signal a special cause that someone should investigate.

Worked example: a filling line's control chart

A drinks line fills bottles to a target of 500 grams. When the process is stable, individual fills have a standard deviation of 6 grams. Each hour an inspector weighs a sample of 4 bottles and plots the sample mean. 1. Standard deviation of a sample mean: \(6 \div \sqrt{4} = 3\) grams. 2. Control limits: \(500 \pm 3 \times 3\), so 491 to 509 grams. 3. Four hourly means: 498, 503, 506 and 511 grams. 4. The first three lie inside the limits, so their variation is normal. 5. The fourth, 511, lies above 509: stop and look for a special cause, such as a worn valve. 6. Check: the limits are for sample means, which vary less than single bottles, which is why the limits are narrower than \(500 \pm 18\).

Deming and total quality management

W. Edwards Deming (1900-1993), an American statistician, lectured to Japanese engineers and managers from 1950, and the Union of Japanese Scientists and Engineers named its Deming Prize after him in 1951. Deming argued that most defects come from the system, not from careless workers, so management must improve the system. His 14 points included ending dependence on mass inspection, driving out fear and improving constantly. He also popularised the plan-do-check-act cycle, which he credited to Shewhart. Total quality management (TQM) applies these ideas across the firm: quality is everyone's job, from purchasing to shipping, decisions rest on data, and the goal is continuous improvement rather than one-off fixes.

Six Sigma

Six Sigma, launched at Motorola in 1986 and adopted company-wide by General Electric in the mid-1990s, aims for at most 3.4 defects per million opportunities (DPMO). An opportunity is each chance for a defect, so an invoice with five fields that could be wrong has five.

\text{DPMO} = \frac{\text{Defects}}{\text{Units} \times \text{Opportunities per unit}} \times 1{,}000{,}000

Projects follow the DMAIC cycle: define the problem and what customers need, measure current performance, analyse the causes of defects, improve the process and control it so gains last. Staff trained in the method are often given martial-arts ranks such as Green Belt and Black Belt. Service firms and governments use Six Sigma too.

Worked example: invoice errors

A billing team issues 50,000 invoices a year. Each invoice has 5 fields that could contain an error, and an audit finds 300 errors. 1. Opportunities: \(50{,}000 \times 5 = 250{,}000\). 2. DPMO: \(300 \div 250{,}000 \times 1{,}000{,}000 = 1{,}200\). 3. Compare with the Six Sigma target: \(1{,}200 \div 3.4 \approx 353\), so the error rate is about 350 times the target. 4. DMAIC next step: analyse which fields cause most errors before changing anything. 5. Check: had the team divided by invoices instead of opportunities, it would report 6,000 defects per million and overstate the problem fivefold.

Awards and standards

The Malcolm Baldrige National Quality Award, created by Congress in 1987 and first presented in 1988, is run by the National Institute of Standards and Technology (NIST). It judges organisations on leadership, customer focus, use of data and results, and is open to manufacturers, service firms, schools, health-care providers and non-profits. Internationally, the International Organization for Standardization (ISO) in Geneva publishes the ISO 9000 family of quality-management standards, first issued in 1987. ISO 9001 is the standard a firm can be certified against after an audit of its processes. ISO 14001, first published in 1996, does the same for environmental management systems.

Lean production

Lean production grew from the Toyota Production System, developed from the late 1940s by Taiichi Ohno and colleagues. The term lean was introduced in 1988 by John Krafcik and spread through the 1990 book The Machine That Changed the World. Lean means removing every activity that does not add value for the customer. Toyota identified seven wastes: overproduction, waiting, unnecessary transport, over-processing, excess inventory, unnecessary motion and defects. Key tools include kaizen, continuous improvement through many small changes suggested by workers; jidoka, stopping the line when a fault appears so it is fixed at once; and the andon cord or button that any worker can use to signal a problem.

Just-in-time and its risks

Just-in-time (JIT) delivers materials exactly when they are needed rather than storing them. Production is pulled by customer demand: a kanban, a card or electronic signal, tells the previous step to make or send more only when parts are used. JIT cuts holding costs, frees space and exposes problems that big stocks would hide. But it depends on reliable suppliers and transport. Natural disasters, such as the 2011 earthquake and tsunami in Japan, and the COVID-19 pandemic and the semiconductor shortage of 2020 to 2022, halted many plants that held little stock. Many firms now keep extra safety stock of critical parts and use more than one supplier, trading some efficiency for resilience.

Common mistakes to avoid

First, equating quality control with end-of-line inspection; assurance prevents defects at every step. Second, setting control limits for sample means with the standard deviation of single items, which makes them too wide. Third, treating every point inside the control limits as needing action; that is normal variation. Fourth, calculating DPMO per unit instead of per opportunity. Fifth, crediting Deming with inventing TQM single-handedly; many people shaped it, and Deming himself disliked the label. Sixth, assuming lean or JIT means no inventory at all; even lean firms keep buffers for critical parts.

Key takeaways

Quality means fitness for use to customers and conformance to producers, and prevention is cheaper than failure. Shewhart's control charts (1924) set limits at three standard deviations around the average. Deming taught that systems, not workers, cause most defects; TQM makes quality everyone's job. Six Sigma targets 3.4 DPMO using DMAIC. The Baldrige Award (first given 1988), ISO 9001 and ISO 14001 recognise and certify quality and environmental systems. Lean removes the seven wastes; JIT cuts stock but needs reliable supply.

Flashcards

Flashcard

Quality (customer view)

Fitness for use: reliability, durability, features, service

Producers translate it into conformance.

Test your understanding

Question 1 of 5

A process averages 250 ml with a standard deviation of 10 ml for single items. Samples of 25 are taken and their means plotted. What are the control limits?

Question 2 of 5

A factory makes 20,000 units, each with 4 opportunities for a defect, and finds 160 defects. What is its DPMO?

Question 3 of 5

Which statement best reflects Deming's view of quality?

Question 4 of 5

A worker sees a fault and pulls a cord that stops the line so the problem can be fixed at once. Which lean idea is this?

Question 5 of 5

What is the main risk of a just-in-time inventory system?

← Previous topicNext topic →