Book summary “Out of the crisis”: summary

Introduction

You face a crisis of industrial decline when your company chases short-term profits and quarterly dividends at the expense of long-term survival. Traditional management relies on mass inspection, annual performance ratings, and lowest-bid purchasing, which destroys productivity and morale. You watch foreign competitors capture markets because American manufacturing clings to obsolete practices that treat workers as the primary source of defects.

W. Edwards Deming argues in Out of the Crisis that quality problems stem from stable management systems rather than careless employees. You must abandon the hope for instant pudding and commit to a profound transformation of Western management. The central idea requires you to adopt statistical quality control and a total overhaul of administrative and production processes. Improving quality initiates a chain reaction that lowers costs, captures markets, and preserves jobs through constancy of purpose.

The book provides a comprehensive framework across eleven detailed sections that guide you through this industrial revolution. You will learn how to implement the fourteen points for management, eradicate deadly diseases and obstacles, and distinguish between common and special causes of variation. The text explains operational definitions, the economics of testing, and the reorganization needed to foster leadership and pride of workmanship across both manufacturing and service organizations.

You must consider several critical questions as you read the book.

• How does improving quality reduce total costs while increasing productivity?
• Why do annual ratings and management by objective destroy teamwork and intrinsic motivation?
• What role does statistical theory play in separating system faults from worker errors?
• How can service industries eliminate hidden administrative waste and excessive verification?

1. Chain Reaction and Management Responsibility

Improving quality initiates a chain reaction that lowers costs and captures the market, while quality problems stem from stable systems rather than workers.

When you improve quality, you reduce rework and the waste of man-hours and machine-time. This transforms wasteful processes into good products, yielding lower costs, a stronger competitive position, and happier workers. Starting in 1950, Japanese management and workers embraced a common aim of quality based on statistical methods. They turned shoddy goods into globally sought-after automobiles and electronics.

The Chain Reaction of Quality

Poor quality embeds massive hidden waste into American products. You consume a substantial portion of payroll and burden just to manufacture and repair defects.

In one manufacturing plant, 21 percent of the line workforce was dedicated entirely to reworking defects made in previous operations. This meant 42 percent of payroll and burden was spent simply making and fixing errors. According to A. V. Feigenbaum, manufacturer's costs spend 15 to 40 percent on embedded waste. Many factory settings also experience handling damage costs of 5 to 8 percent.

Management and Systems

Quality problems are primarily built into the stable system of production. Their improvement is the direct responsibility of management rather than the fault of workers.

When a production line operates as a stable system of random variation, pleading with workers to try harder is futile. Substantial improvements require you to change the system itself, such as by providing clear operational definitions.

In a manufacturing plant with 24 people on the work force, plotting daily defects revealed a stable system averaging 11 percent defective. By establishing operational definitions for acceptable work over 7 weeks, the manager and supervisors dropped the defect rate immediately to 5 percent with zero financial cost and no new machinery.

Reduction of Costs and Hidden Waste

Achieving statistical control over a manufacturing process enables engineers to innovate, reduce material usage, and generate massive cost savings. Instead of purchasing expensive new machinery, you can analyze your existing process using statistical control to eliminate variation.

At the Nashua Corporation, engineers analyzed the carbonless paper coating process and brought it into statistical control. They considered purchasing a new coating head for $700,000. Instead, they reduced the dry coating applied from 3.6 down to 2.8 pounds per 3000 square feet. This saved $800,000 per year at present volume levels, achieving an annual reduction in coating material cost of $100,000 per tenth of a pound saved while paving the way for further chemical innovations.

2. Principles for Transformation

When management responds to quality issues by simply asking everyone to do their best, it results in random actions and dissipation of energy. Best efforts without guidance, consistency of effort, and proper management principles lead to chaos and wasted knowledge rather than improved quality.

The fourteen points provide a comprehensive framework to replace mass inspection and lowest-bid purchasing with constancy of purpose and statistical methods.

Adopting the 14 points signals your intent to stay in business, protect jobs, and safeguard investors. These points demand a complete departure from traditional practices:

• create constancy of purpose toward improvement of product and service;
• adopt the new philosophy;
• cease dependence on mass inspection;
• end the practice of awarding business on the price tag alone;
• institute training on the job;
• institute leadership;
• drive out fear;
• break down barriers between departments.

Routine 100 percent mass inspection is costly, ineffective, and equivalent to planning for defects because you cannot inspect quality into a product after it is made. Inspection happens too late to fix the process. Inspectors suffer from fatigue and boredom, frequently disagree with one another, and often rely on faulty testing instruments.

A printing company called David S. Chambers for help because they proofread everything 11 times and were still plagued by customer complaints because each proofreader depended on the others.

Awarding business on the price tag alone drives companies toward low quality, while long-term relationships with a single supplier minimize total cost and foster trust. Purchasing managers must focus on total net cost over time rather than initial purchase price. Working with a single supplier for an item avoids the variation introduced by multiple sources and allows the vendor to innovate and improve processes.

General Motors required steelmakers to bid annually for its business, but later considered negotiating long-term price and supply contracts to improve efficiency and cut costs.

Numerical quotas, work standards, and piecework for hourly workers act as fortresses against improvement and destroy pride of workmanship. Quotas set for average workers cause peer pressure to limit output, while forcing struggling workers to rush and turn out defects.

An airline reservationist was required to make 25 calls per hour while also being told to provide courteous service without rushing callers, creating an impossible contradiction.

Fear causes impaired performance, padded figures, and resistance to knowledge, preventing employees from working effectively for the company. Employees hide mistakes, avoid asking questions, and refuse to stop production for necessary maintenance because they fear negative annual ratings, penalties, or losing their jobs. True improvement requires a secure environment where people can speak up.

Management must remove the everyday barriers, such as poor maintenance, erratic material quality, and unknowledgeable supervisors, that rob employees of pride of workmanship. Hourly and salaried workers face constant handicaps including broken machinery, ambiguous definitions of acceptable work, and supervisors focused solely on numbers.

3. Diseases and Obstacles

Western management is crippled by deadly diseases like lack of constancy of purpose, annual performance ratings, and the hope for instant pudding.

Corporate leaders focus heavily on financial outcomes rather than building efficient systems for long-term product and service quality. This is exacerbated by the fear of unfriendly takeovers and leveraged buyouts, which force executives into paper entrepreneurialism and short-term tactics like cutting research and training. William E. Hoglund noted that Blue Cross became the second largest supplier for the Pontiac Motor Division, with direct medical costs reaching $400 per automobile.

Evaluation of Performance and Fear

Annual performance appraisals and merit ratings destroy teamwork, nourish fear, and penalize employees for variation that is actually caused by the system itself. Performance reviews focus on end products rather than leadership, forcing managers into counting numerical outputs and discouraging employees from taking risks or helping colleagues. Statistical experiments demonstrate that apparent differences in worker performance often fall entirely within the expected chance variation of a stable system.

William W. Scherkenbach contributed an example where nine employees recorded varying numbers of mistakes, but statistical calculations proved none fell outside the control limits of the system.

Running on Visible Figures Alone

Managing solely on visible financial figures while ignoring unknown and unknowable figures will eventually destroy a company. While basic financial metrics like payroll and taxes are necessary, relying exclusively on them blinds management to critical qualitative factors.

You confront invisible losses from:

• the multiplying effects of happy or unhappy customers;
• the hidden losses caused by poor workplace morale.

A Technical Assistance Research Program study found that a car owner satisfied with their vehicle is likely to buy four more cars of the same make over twelve years and tell eight people. An angry buyer tells an average of sixteen people about their troubles.

The hope for instant pudding—the belief that quality and productivity can be achieved overnight through affirmation of faith or consultants without deep study—is a major obstacle. Many managers seek quick fixes, simple formulas, or cookbook procedures without understanding the underlying management theory. Simply appointing someone to a quality role or expecting automated machinery to solve systemic issues without proper education leads to frustration and failure.

A furniture manufacturing company took apart a Steinway piano to copy it exactly, only to produce thuds. They found that even reassembling the Steinway left it producing thuds.

Blindly copying Japanese management techniques like QC-Circles without understanding management's role in the broader corporate culture results in failure. American firms often establish formal staff organizations or use suggestion boxes as superficial tools to boost production or reduce costs, while ignoring management's primary responsibility to improve the system and protect workers. Akira Ishikawa noted that in the United States QC-Circles are organized as formal staff organizations led by managers, whereas in Japan they are informal voluntary worker groups advised by managers.

4. When and How Long for Transformation

Transformation Timeframe

Reforming American industry and establishing a competitive position will require decades of profound transformation rather than quick fixes. American management is paralyzed by a focus on quarterly dividends and short-term profits. Achieving real change requires removing deep-seated management obstacles and deadly diseases, a process that will take from 10 to 30 years.

Reforming American industry and establishing a competitive position will require decades of profound transformation rather than quick fixes.

A man hired by a large company to lead future plans was dismissed because fourth-quarter dividends dipped in 1981, as reported in Business Week. Shifting the purchasing department from lowest bidder to quality-based purchasing takes 5 years.

Government Regulations and Obstacles

Well-intentioned government regulations and antitrust enforcement often hinder productivity and stifle competition in quality and service. Regulatory bodies often focus narrowly on price competition while ignoring service and quality. Antitrust enforcement and excessive red tape force companies to waste massive resources on legal compliance rather than productivity.

The Justice Department's Antitrust Division wrecked telephone communication and transportation systems under the belief that price competition benefits the public. Government regulation cost American business $30 billion in 1976.

Multiple Faces of Quality

Quality has multiple dimensions and scales, evaluated differently by production workers, plant managers, and ultimate consumers. A product can score high on one quality scale and low on another. Translating the future needs of the user into measurable characteristics is difficult because consumer needs and competitive materials constantly change.

Textbook quality is viewed differently by the printer who cares about typography, the author who wants message clarity, and the publisher who needs sales. The Washington Metropolitan Transit Authority management set a goal of 5.7 percent for escalators out of order.

Dissatisfied customers rarely complain; they simply switch to competitors, making repeat customers the true source of profit. Relying on customer complaints gives a biased picture of performance because unhappy customers just take their business elsewhere.

Manufacturers must add a fourth step of testing products in service to create a continuous helix of quality improvement. Traditional manufacturing relied on three independent steps of design, production, and marketing based on guesswork. The new way introduces consumer research and service testing as a fourth step, creating a continuous feedback loop that improves quality while lowering cost.

Oliver Beckwith noted that a dissatisfied customer does not complain but just switches, and Robert W. Peach remarked for Sears, Roebuck & Co. that goods come back but not the customer. A grocer used to work with small cheese factories to hand-prepare cheese tailored precisely to local patrons before industrial expansion created barriers.

5. Quality and Productivity in Service Organizations

Census figures show that 75 people out of 100 are employed in service organizations. When you add service workers embedded within manufacturing firms, 86 people in 100 are engaged in service work. Only 14 out of 100 make tangible items or work in agriculture.

National productivity and your standard of living heavily depend on improving service quality.

Traditional management views administrative and service work as immune to quality methods. You must recognize that the same statistical principles applied to factories work equally well in offices and hospitals.

Service Sector Dominance

Service Sector Dominance

Employees who have direct contact with customers serve as your marketing department. These contact men hold the primary responsibility for shaping customer perception. A bus driver in Washington considered passengers to be a bother, failing to realize that helping them with directions builds future business. Management must screen and train personnel for their ability to please customers.

Four Costs of Quality in Banking

Bank experts note that between 40 and 60 percent of any bank's staff is involved in verifying the work of other staff members. Layers of checkers hide massive operational expenses under four distinct categories:

• cost of appraisal;
• cost of internal failure;
• cost of external failure;
• cost of prevention.

Traditional banking relies on endless verification under the false assumption that catching errors solves quality issues. You must shift from appraisal to prevention.

Municipal and Transit Improvements

Studying the statistical control of transit time allows companies to tighten delivery limits. Automobile parts shipped from Buffalo to Kansas City required a 5-day inventory previously. Observing transit time limits revealed a calculated upper limit of 4.2 days. That 0.8 day difference yielded $500,000 per year in savings for those parts, contributing to $25 million in total savings across routes.

Applying statistical thinking to municipal vehicle fleets shifts focus from expensive repairs to cost-effective upkeep. In the Motor Equipment Division of the City of Madison, led by Mayor Joseph Sensenbrenner, mechanics demonstrated the power of preventive maintenance. Mechanics showed that spending $164 on corrosion prevention could save a jeep that later cost $4,200 to rebuild. They presented the Mayor with a broken aluminum piston jammed with a steel valve, representing a $3,200 failure that a $1.50 spring replacement would have prevented.

6. Training, Leadership, and Statistical Control

Leadership must foster pride of workmanship and recognize that training is ineffective once workers reach statistical control.

The primary aim of leadership is to improve the performance of people and machines. You must enhance quality, increase output, and foster pride of workmanship by removing causes of failure rather than simply recording mistakes. A leader must calculate or judge who lies outside the system and needs individual help. You are responsible for continually improving the system so everyone can work with greater satisfaction.

Further training is ineffective for workers whose work has reached a state of statistical control. Workers in chaos or out of control can develop their potential ability through proper training, but continued training once a learning curve levels off accomplishes nothing because the worker is in a rut.

A beginner in golf showed scores wildly out of control before taking lessons. After lessons, his scores achieved statistical control and dropped significantly. An experienced golfer, already in statistical control, took lessons that accomplished nothing because his techniques were already ingrained.

Workers achieving statistical control should not be shown defective items or blamed for performance they cannot govern. Pointing out defectives is counterproductive unless a special cause is detected via a control chart. Management must investigate the system, tools, training, or materials when performance is poor.

A corporate accounting supervisor noted that an employee doing poor work was likely in statistical control. Transferring him to a new job with proper training turned him into an extra asset for the department.

Fear of management reprisal or harsh penalties leads employees and inspectors to distort data, hide defects, or make decisions that harm company interests. When management relies strictly on numbers and punishes workers based on arbitrary thresholds, people become insecure and manipulate records to protect themselves and their colleagues.

An inspector on final audit reported defect rates tightly clustered just under 10 percent because rumor stated the manager would shut the plant down at 10 percent. In another case, a foreman refused to fix a failing bearing in a blast machine because he feared missing his production quota.

Reliable inspection requires operational definitions and statistical control of test methods to prevent inspectors, testing machines, and vendors from disagreeing with one another. A recorded measurement is the end product of a long series of operations. Without operational definitions consisting of a test method, test, and criterion, inspection results become arbitrary.

A study across 4 inspectors examining 1,590 total pieces from 17 operators revealed large discrepancies between pairs of inspectors due to a lack of operational definitions. The average proportion defective across the study was 0.096.

7. Operational Definitions and Standards

Words and specifications require operational definitions and voluntary standards to prevent administrative chaos and legal penalties.

Operational Definition Necessity

Concepts like good, reliable, round, or safe are ineffable in pure mental form. Business transactions, legal regulations, and engineering specifications break down when parties fail to state in advance the exact sampling, measurement procedures, and pass-fail criteria.

A plant manager had an unacceptable defect rate of 35 to 50 percent for automobile instrument panel wrinkles because inspectors disagreed on what constituted a wrinkle. Establishing clear operational definitions dropped defect rates to 10 percent within a week and increased production by 50 percent at no cost.

The Absence of True Value

Applied science and commerce deal with deviations from an ideal rather than exact values. There is no true value for physical measurements, census counts, or defect rates, because every result depends entirely on the specific procedure, method of sampling, and test used.

Mayors of various cities protested the 1980 Census count, assuming an absolute true number of inhabitants existed if officials worked hard enough. Extra search efforts cost 100 dollars to 200 dollars for each further addition to the census count.

Voluntary Standards and Regulations

Voluntary industrial standards created cooperatively by producers, consumers, and experts prevent the proliferation of inflexible, mandatory government regulations. Standardization enables a free national market, mass production, lower prices, and better quality by ensuring interchangeability.

• Standardization builds an efficient national market and lowers prices.
• Cooperative agreements involve 4000 executives serving on standardization committees.
• Government intervention often follows industrial failure to regulate quality.

A U.S. importer of wool-blend fabrics agreed to pay a 25,000 dollars fine paid by an importer for incorrect wool labels to settle charges of selling fabrics with less wool than labels claimed. You must recognize that poorly defined specifications invite costly legal disputes.

8. Common and Special Causes of Variation

Most variation stems from common causes built into the system, making tampering through overadjustment disastrous for stable processes.

Management often mistakenly blames workers or fleeting events for defects, mistakes, and accidents. You must realize that fundamental system faults cause most variations. Treating system faults as special causes leads to higher variability, frustration, and increased costs.

• 94 percent of troubles belong to the system and are the responsibility of management;
• 6 percent of troubles are special causes.

A motor freight company manager blamed all shipping shortages and driver damage entirely on the drivers. This guaranteed that losses would continue until management addressed the underlying system.

Common Versus Special Causes

Adjusting a stable process to compensate for a specific result that is either undesirable or exceptionally good always makes the output worse than leaving it alone. Tampering with a stable system by overadjusting after every individual event ignores rational statistical limits and treats random variation as a special cause. This practice doubles variance or leads to an exploding system where results drift infinitely away from the target.

The funnel experiment demonstrates this principle using 3750 total beads in the box. The box contains 3000 white beads and 750 red beads, distributed across 50 beveled depressions on the paddle. Rules requiring the adjustment of a funnel's position based on the previous drop's error either double the variance or cause the system to explode completely.

The Futility of Tampering

Workers are completely helpless to improve quality as long as defects are built into the raw materials and procedures governed by the system. Management frequently evaluates employees based on daily output quotas and penalizes or rewards them for variations generated entirely by random draws. Such evaluations of individual performance are utterly useless for prediction.

During the red bead experiment, ten volunteers acted as workers and inspectors under strict work standards. Despite their best efforts and exhortations from management, they continued producing red beads because the system itself contained red beads.

Shewhart Control Charts and Process Capability

Shewhart control limits provide a rational and economic guide to distinguish between common causes and special causes, minimizing net economic loss from errors. Using the naked eye or arbitrary judgment to spot anomalies leads to costly mistakes. Control charts use statistical signals to indicate when variation is due to chance within a stable system versus when an assignable special cause requires intervention.

An insurance company with 10,500 employees faced canceled coverage due to frequent fires. A control chart showed a stable system averaging 1.2 fires per month with an upper limit of five. This proved the president's letter urging employees to set fewer fires was pointless without systemic change. A process has a definable capability and predictable performance only when it has been brought into a state of statistical control.

9. Systemic Improvements and Disappointments

Management is responsible for the entire system of production, and treating systemic errors with individual punishments fails.

You cannot blame workers for errors inherent in a flawed system. A London manager blamed his billing staff for massive errors and delays, awaiting new data-processing machinery. Simplifying charges and training workers solved the crisis instead of new hardware. The most important mistake rate dropped from 39 to 6 percent, and the second from 27 to 4 percent.

Management System Responsibility

External consultants do not need to know the specific technical details of a process. Workers already know how the work is done. External knowledge brings statistical methods to improve the system. Top management must undergo deep indoctrination rather than delegating quality to subordinates.

A company president mistakenly believed a statistician needed to be raised in the manufacturing business. Meanwhile, the plant manager was harassed by daily problems and suspicious of outsiders speaking a new language. Management must attend a four-day seminar for indoctrination and take personal responsibility.

Stocking Production Improvement

Involving supervisors in statistical training and tracking individual operator charts can dramatically reduce defect rates and scrap without purchasing new equipment. At a stocking plant, 20 supervisors were sent to a 10-week course. Individual charts for 47 loopers helped discover that Operator 22 was nearly blind in one eye. Corrective medical treatment and showing charts to workers reduced defective output from 4.8 percent to 0.8 percent in seven months.

A large motor freight company with 20 terminals suffered $31,000 in claims alone from 617 mistakes across 150 drivers, averaging a loss of $50 per mistake. Management had been sending identical warning letters for any single mistake, demoralizing good drivers:

• 150 drivers worked all year long;
• 617 total mistakes accumulated across seven categories;
• $620,000 total loss occurred across 20 terminals;
• 11 mistakes marked the upper limit for normal variation.

Above-Average Punishment Trap

Punishing or counseling employees who simply fall above the average of a stable system is a statistical fallacy. The president of a company had 60 women punching cards with an average error rate of 3 per 100 cards. Every Wednesday he reprimanded the roughly 30 women who fell above the average during the previous week. Because half a group must always fall above average in any stable system, his intervention achieved no improvement and caused frustration.

10. Economics of Testing and Incoming Materials

Managers should use break-even quality points and all-or-none rules instead of obsolete standard sampling plans to minimize total cost.

All-or-None Inspection Rules

To achieve minimum average total cost for incoming materials, you should apply simple all-or-none rules based on the break-even quality point. You define this point as the ratio of initial test cost to downstream repair cost.

If the worst lot's fraction defective falls below this break-even point, you require no inspection. If the best lot stays above it, you mandate 100 percent inspection.

A television manufacturer was performing 100 percent inspection on integrated circuits when the fraction defective was extremely low. This practice maximized total cost and wasted substantial money. The initial test cost was 30¢, the repair cost was 2000¢, and the average fraction defective was 0.00015.

Service Organization Verification

The economic theory of inspection applies equally to service organizations. Downstream correction costs for administrative mistakes can be thousands of times higher than initial verification costs.

As work moves through stages in banks, department stores, or payroll departments, errors compound and make early verification critical. Parallel computations by independent workers matched via machine detection provide the most reliable verification.

At Irving Trust Company, William J. Latzko noted that the cost of review was 25¢ while downstream correction of a mistake averaged $500. The cost to correct a mistake was 2000 times the cost of review.

Rejecting Obsolete Sampling Plans

Standard acceptance sampling plans like Military Standard 105D and Dodge-Romig tables should be discarded. These standard acceptance plans ignore economic costs and often double total expenses.

They focus on arbitrary quality limits or AOQL without considering actual inspection and repair costs. As the number of parts or operational steps in an assembly increases, the cumulative probability of failure multiplies. Complex assemblies demand near-zero defect rates for components because failure probabilities rise sharply. An automobile trunk with 70 seams had 1 fault in 2100 welds, which resulted in 1/30 of the trunks leaking and requiring rework for a lot size of 1500 assemblies, a $24 test cost per assembly, and a $780 replacement cost.

11. Organization for Improvement and Japan Transformation

Companies must appoint a statistical leader reporting to top management, following the Japanese model where management took personal responsibility.

Statistical Leader Requirements

A company must appoint a leader of statistical methodology who reports directly to top management. This person possesses advanced academic and practical qualifications and holds companywide authority. You need someone with the equivalent of a master's degree in statistical theory, industrial experience, published papers, and the ability to lead top management. This leader operates by pedagogy and guidance, participating in major staff meetings and having the right to investigate any activity. Morris H. Hansen instituted this type of organization for statistical work in the U.S. Bureau of the Census around 1940, bringing global preeminence by 1945.

Accidents and the System

The vast majority of accidents and system failures stem from the system itself rather than human carelessness. Computing figures on accidents does not reduce their frequency unless statistical methods are used to determine whether causes originate from the system or specific human errors. As precision improves, the proportion of outliers or failures often remains constant relative to increasingly severe requirements. In the United States, two billion medical interventions occur annually, and one hundred thousand unfavorable results represent a high reliability of one part in 20,000.

Medical interventions and malpractice can only be properly understood through statistical theory by distinguishing system failures from special causes like carelessness. The vast majority of unfavorable medical results belong to the underlying system of care rather than individual incompetence, and operational definitions of special causes are necessary to protect practitioners from unjustified accusations.

Japan Quality Transformation

Japan's economic miracle in quality began in 1950 when top management took personal responsibility for quality, embraced statistical methods companywide, and partnered with workers. Through the Union of Japanese Science and Engineering, known as JUSE, and expert guidance, Japanese management learned that the consumer is the most important part of the production line and focused on process improvement rather than isolated fixes. By 1954, Japanese quality captured global markets. Between 1950 and 1970, JUSE taught statistical methods to 14,700 engineers.

Key achievements show the power of this transformation:

• Keizo Nishimura of Furukawa Electric Company reduced rework in insulated wire to 10 percent of its former level.
• Gohei Tanabe reported producing three times as much PAS pharmaceutical product with the same resources.
• Fuji Steel Company achieved a 29 percent reduction in fuel required to produce a ton of steel.

Conclusion

You face a stark choice between the immediate pursuit of short-term profit and the long-term preservation of your company through profound transformation and statistical thinking.

The ideas in this book fit together as a unified system rather than a collection of separate tools. You cannot cherry-pick individual techniques while leaving the underlying management philosophy intact. Constancy of purpose demands that you abandon mass inspection, numerical quotas, and fear. You must instead treat quality as a chain reaction that lowers costs, captures markets, and preserves jobs. Statistical methods reveal that workers rarely cause defects; instead, faults stem from stable systems that only management has the power to change. By breaking down barriers between departments and instituting rigorous training, you create an environment where pride of workmanship can flourish. You must take personal responsibility for leading this transformation and commit to decades of continuous improvement.

To put these principles into practice, you can follow a clear set of operational rules.

1. Adopt constancy of purpose for the continuous improvement of products and services, aiming to become competitive and stay in business.
2. Cease dependence on mass inspection by building quality into the product in the first place through statistical control.
3. End the practice of awarding business on the price tag alone, and instead minimize total cost through meaningful relationships with single suppliers.
4. Drive out fear so that everyone may work effectively for the company and report problems without dreading retaliation.
5. Remove the barriers that rob people of their right to pride of workmanship, including annual ratings, merit systems, and arbitrary numerical quotas.
6. Institute a vigorous program of education and self-improvement for every employee, ensuring that training utilizes operational definitions and statistical methods.
7. Put everybody in the company to work to accomplish the transformation, recognizing that top management must take personal responsibility for the system.

10 Key Ideas

1. Improving quality initiates a chain reaction of lower costs and captured markets, requiring management to adopt constancy of purpose rather than chasing short-term profits.
2. Quality problems stem primarily from stable management systems rather than careless employees, making system improvement the direct responsibility of management.
3. The fourteen points provide a comprehensive framework to replace mass inspection and lowest-bid purchasing with constancy of purpose and statistical methods.
4. Western management is crippled by deadly diseases like the hope for instant pudding and annual performance ratings that destroy teamwork and nourish fear.
5. National productivity and the standard of living heavily depend on improving service quality by applying statistical principles to administrative work.
6. Leadership must foster pride of workmanship and recognize that further training is ineffective once workers reach a state of statistical control.
7. Words, specifications, and commerce require operational definitions and voluntary standards to prevent administrative chaos, legal penalties, and misunderstandings.
8. Most variation stems from common causes built into the system, making tampering through overadjustment disastrous for stable processes.
9. Managers should use break-even quality points and all-or-none rules instead of obsolete standard sampling plans to minimize total cost for incoming materials.
10. Companies must appoint a statistical leader reporting directly to top management, following the Japanese model where management took personal responsibility for quality.