Understanding A3 Thinking

A Critical Component of Toyota's PDCA Management System

Durward K. Sobek II., Art Smalley

• Publisher: Productivity Press
• Year: 2008
• ISBN:
  9781439885871
• Edition: 1
• Topic: Business
• Subtopic: Operazioni
• Index: Business

Winner of a 2009 Shingo Research and Professional Publication Prize. Notably flexible and brief, the A3 report has proven to be a key tool In Toyota's successful move toward organizational efficiency, effectiveness, and improvement, especially within its engineering and R&D organizations. The power of the A3 report, however, derives not from th

Chapter 1

A Basis for Managerial Effectiveness

Toyota Motor Corporation is arguably the most studied company of the modern era. More than a dozen books¹ have been written about the company, its management system and philosophy, and its approaches to various business and operational problems. One of the latest books, The Toyota Way by Jeffrey Liker², became an immediate best-seller, indicating the instant draw the Toyota name has within the business community. A search on Business Source Premier (http://www.uwe.ac.uk) turned up over 3,000 articles published over a ten-year period with “Toyota” in the title. And this number does not include the hundreds of volumes and countless articles on lean manufacturing or various aspects thereof (for example, 5S, kanban, poka-yoke) that are largely based on tools and practices developed by Toyota.

Such attention is well deserved. As of the writing of this book, Toyota had just surpassed Ford Motor Company in number of vehicles sold annually in the United States, having already beat out Ford in global sales several years prior, and it is poised to topple General Motors to become the largest auto manufacturer in the world. In 2005, Toyota produced one vehicle approximately every four seconds somewhere in the world while at the same time setting the benchmark for product quality. Toyota perennially wins national and international acclaim in all of the major automobile-quality ratings. For instance, Toyota’s flagship Lexus nameplate has earned the top spot in J. D. Power’s Initial Quality Survey for over ten years running. On top of all this, Toyota is profitable—in fact, very profitable. Toyota set record profits in 2003, 2004, and 2005, earning over $10 billion annually, even while its North American competitors saw significant drops in earnings, and even losses.

But other companies have also been successful. What makes Toyota intriguing is that its success has been sustained over an extremely long time period, by most business standards. From the ashes of World War II, Toyota initially struggled to maintain solvency but rose over the following decades to become Japan’s leading manufacturer. As it grew, Toyota began seeking markets outside of Japan, and by the early 1980s, Toyota was well established in the U.S. market. Toyota has grown each year for the past fifty years and has not experienced a loss in net earnings since the early 1950s. This is a standout performance in an industry characterized by cyclical ups and downs.

Toyota is also intriguing because its business and management philosophy is unique, its approach to manufacturing is exceptional and counterintuitive, and its collective understanding of operational dynamics is breathtakingly insightful. Toyota is perhaps most well known for its production system, first documented in a detailed eighty-page handbook published internally in Japanese in 1973. The first English publication on it appeared in 1977 by Sugimori et al.³ as a high-level summary. However, it was not until the early 1990s that the uniqueness of Toyota’s system became well known, with the publication of the book The Machine That Changed the World.⁴ In it, the MIT professors detail the strikingly robust, flexible, and efficient systems they observed in Japan, and dubbed it “lean manufacturing” for their ability to design, produce, and deliver higher-quality products in volume with a fraction of the resources used by their North American and European competitors. The manufacturing community learned later that the model of lean manufacturing was the Toyota Production System (TPS).

Toyota has been remarkably open in sharing its system with others, even establishing the Toyota Supplier Support Center to provide consulting assistance to U.S. companies wanting to operate more efficiently, at no cost to the client. More recently, we have come to understand that Toyota’s uniqueness extends into many other areas as well, including product development and logistics. Cottage industries are sprouting in many arenas to provide training in lean tools and concepts and assistance in implementation. Lean applications that were once targeted primarily at high-volume manufacturing plants are rapidly finding their way into other sectors of the economy, including engineering, financial services, transportation and logistics, health care, food and beverage services, and government (including military operations). Toyota’s impact is being felt well beyond the automotive industry.

PDCA: Heart of the Toyota Way

The lean model is dramatically altering the face of manufacturing in the developed world. Inventories are dropping, lead times are shortening, quality is holding steady or increasing, and prices are falling. We expect this trend will follow in other sectors, just as it has in manufacturing. Yet, to our knowledge, with all that we know, with all that has been published, with all the resources that are available, no American companies have reached Toyota’s level of efficiency and effectiveness. In fact, Toyota is building factories in the United States even as most of U.S. manufacturing is trying to move operations overseas or outsource them altogether. Why do we not see more companies emulating Toyota’s success?

While there may be many explanations, perhaps the most crucial one is that most of us do not understand—or if we understand, do not appreciate—what is at the heart of the Toyota business, management, and manufacturing approach. We tend to see the intricate set of tools as the system. But although they are important to the system as currently enacted, they are at the surface, not the heart. In fact, Taiichi Ohno, the father of the modern Toyota production system, said that the tools are just countermeasures to business problems that Toyota has faced and that they will be used only until better countermeasures are found.⁵ In other words, the interconnected web of tools and practices we know as lean manufacturing is the outcome of a deeper set of processes. These deeper processes, we argue, are at the heart of Toyota’s system.

So how did the Toyota system emerge? In simple terms, it emerged as Toyota’s people saw problems or opportunities, “solved” the problems aggressively and systematically to find a better way to do things, and then rigorously verified that the better way was indeed better. If the remedy did indeed improve the system, the new way became the standard way to do the work; if not, problem solving and verification continue until the problem is satisfactorily addressed.

The roots of this process are grounded in the scientific method of inquiry. The founders of the original Toyota companies were Sakichi Toyoda and his son, Kiichiro. Sakichi is recognized in Japanese textbooks as one of the great inventors of his time. His primary accomplishment was the creation of an automatic loom in 1924 that far surpassed the productivity and quality of any similar product on the market at that time. Over 24 patents were granted on this machine alone. Most of the patents were filed by Sakichi; however, quite a few were also developed by his son, Kiichiro, a graduate of the mechanical engineering department of Tokyo University, the most prestigious university in Japan. Like his father, he had a knack for invention and liked to create and tinker by nature. The machines they developed emerged out of repeated experimentation. This inclination toward actions, and of trying out ideas through experimentation, continues to this day.⁶

The problem-solving approaches in use at Toyota today are also deeply influenced by a high-level methodology initially developed by Walter Shewhart of Bell Labs in the 1930s, and later adopted by W. Edwards Deming who became its biggest proselytizer.⁷ The methodology is the Plan-Do-Check-Act (PDCA) cycle, also called the Deming Cycle. Additional education came from representatives of the Japan Union of Scientists and Engineers (JUSE), who put on lectures at Toyota and other manufacturing companies in Japan after World War II, teaching scientific principles for quality control and improvement.⁸ The heart of these teachings is the PDCA cycle.

The PDCA cycle begins with the Plan step, in which the problem-solver thoroughly studies a problem or opportunity to understand it from as many viewpoints as possible, analyzes it (quantitatively, if possible) to find the root causes, develops one or more ideas to remedy the problem or seize the opportunity, and devises a plan for implementation. In the Do step, the plan is put into action as immediately as is possible and prudent. The Check step involves measuring the effects of implementation and comparing them to the target or prediction. Act refers to establishing the new process, solution, or system as the standard if the results are satisfactory, or taking remedial action if they are not. The PDCA cycle simply follows the steps of the scientific method: Plan is developing a hypothesis and experimental design; Do is conducting the experiment; Check is collecting measurements; and Act is interpreting the results and taking appropriate action.

Over the years, Toyota has honed a set of norms and practices on how to most effectively conduct each step of the cycle. At each juncture, the problem- solver tries to confront his or her own assumptions and preconceptions in order to gain insight into a situation or phenomenon, or to validate that his or her understanding is accurate. If one gains insight, that new learning must be confirmed through experimentation. If one’s current understanding is found deficient, immediate remedial action is required.

What Don’t We Get?

None of this should sound new or exotic to Western managers. Indeed, the concepts of the scientific method or PDCA are ironically Western and not original to Japanese culture. So why have so many organizations not embraced this process or style of thinking to the extent that Toyota has? We believe multiple reasons exist.

Perhaps most significantly, in the West, we tend to be oriented toward short- term results. We want to get the problem taken care of and move on. At Toyota, however, the process by which the results are achieved is equally—if not more— important, and the ultimate goal is not just a problem resolved in the immediate term, but also that 1) the problem is less likely to occur in the future because the overall system is improved, and 2) the problem-solver has enhanced his or her problem-solving skills and is prepared to tackle even more challenging tasks in the future. This difference in perspectives fundamentally alters the way we see PDCA.

It has been said that the typical U.S. firm, when facing a vexing problem in which it has one year to solve, would spend three months planning, three months implementing, and six months tweaking and picking up loose ends. Toyota, facing a similar situation, would spend eleven months planning and one month implementing (with no loose ends to clean up!). This comparison is an exaggeration, of course, but it nonetheless contains an important element of truth. The reason Toyota spends so much time and effort on the planning phase is because that phase is so critical to learning. Toyota managers want to make sure they deeply understand the background and facts of the current situation before moving forward. After the current state is thoroughly probed, they want to establish a high degree of certainty that they have accurately identified the root cause of the problem. This includes understanding the situation from multiple perspectives, not just one’s own, and gathering and analyzing system performance quantitatively.

Toyota’s problem-solvers also plan out the change in detail (including which steps will be taken, by whom, and when) and get consensus from each individual involved and, if appropriate, their supervisors. This level of implementation planning is important to learning because if the expected improvement is not achieved, the team will want to know whether it was lack of understanding of the situation or a faulty implementation that produced the disappointing results.

Furthermore, thorough planning includes a reasonable prediction of the change in performance, along with a plan to follow up—again, who will do what and when. The prediction of future performance is actually an informal statement of hypothesis based on the current state of knowledge. It is the a priori statement of hypothesis (that is, prior to experimentation) that enables learning to occur at the follow-up. We either confirm our current understanding or unequivocally find out that we have more to learn. Without the hypothesis, the scientific method quickly devolves into a guess-and-check or trial-and-error approach, with commensurate declines in learning.

With the planning function complete, the plan can be executed. Simply, the Do step must be accomplished for any change to occur. Although this may seem obvious, many organizations seem content to spend all of their time in meetings and never actually get around to doing anything about the problems identified. Toyota recognizes that the Do step is essential. It is the experiment to test the hypothesis.

The Check step, then, is where the individuals involved validate their current level of understanding, a vital step of the learning process. How do we know that what we think we learned and understood in the Plan step is actually accurate? Toyota does this by following Shewhart’s and Deming’s advice and measuring the actual results. If actual results correspond to those predicted, the problem- solving team confirms that what they thought they understood is probably accurate. In other words, the knowledge has been validated. If actual results do not match the predicted outcome, more investigation is needed to find out why. The team will also want to make sure the implementation plan took place as planned (if not, of course, then why not).

Finally, the Act step identifies any loose ends or modifications to be made based on the learning from the Check cycle. It is also the step where the change becomes institutionalized for improved system performance and learning is shared with appropriate parts of the organization. It would seem obvious that we would not want to institutionalize a change until we verify that the organization will actually perform better as a result; however, many organizations routinely institute system changes without a clear idea of the effects the change will have.

Thus we see that while many managers have heard of PDCA, most do not understand that it involves more than just getting problems resolved in a timely fashion. PDCA is a high-level methodology to raise both individual and organizational consciousness about what is known and what is not known in order to resolve the problems currently faced and to prevent future recurrence. At the same time, PDCA aims to improve long-term system performance, not just take care of a localized problem. As the preceding discussion demonstrates, to use the cycle effectively requires a certain level of discipline.

A short story might help illustrate the point. While one of the authors was a young trainee in Japan, he was asked to solve a quality problem on a precision grinding machine. Initially, he proposed changing several parameters all at once to “fix” the problem. Instead, his supervisor had him sketch the machine in considerable detail and make a Pareto chart of all the different types of defects on the machine. This exercise isolated one particular type of defect as the main cause for concern. The supervisor then had the trainee list all the potential causes of the defect one by one using a fishbone diagram.

Because a positive cause-and-effect relationship could not be established by deduction alone, the trainee was required to come up with a list of corrective actions to attempt for each potential cause. He tested them one by one. Finally, after several days, he had a breakthrough when the coolant lubricating the part during the grinding cycle was analyzed and subsequently changed out. Bacteria had contaminated the coolant. After replacing the coolant, defects dropped from 2.3% to under 0.2%.

The trainee proudly reported the good news to his supervisor and the manager of production. Expecting words of praise, he was somewhat chagrined when they ...