Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles The Toyota Way: Learn to Improve Continuously

The Toyota Way: Learn to Improve Continuously

Key Takeaways

  • The Toyota Way’s fourteen principles can help any organization, manufacturing or service, work toward excellence.
  • At the center of the fourteen principles are people thinking scientifically to creatively dream and test their ideas while striving for a clear future state.
  • Strategy, nested goals, and disciplined execution along with creative problem solving for the many unanticipated obstacles, are at the heart of a sustainable enterprise.
  • In the end, a sustainable enterprise must become a learning organization, constantly studying and adapting to a complex environment to better add value to customers and society.  
  • Digital tools can amplify the power of thinking people, but will not replace the dexterity or creativity of people who bring to bear all their senses in continuous improvement.

The book The Toyota Way, 2nd Edition by Jeffrey Liker provides a view of the Toyota Production System with fourteen management principles for continuous improvement and developing people. The book, including the 4P model (Philosophy, Processes, People, Problem solving) and principles, has been updated to reflect new insights in systems thinking.

InfoQ readers can download an excerpt from the book The Toyota Way, 2nd Edition.

InfoQ interviewed Jeffrey Liker about The Toyota Way.

InfoQ: What made you decide to create the 2nd edition of The Toyota Way?

Jeffrey Liker: In the seventeen years since the original Toyota Way was completed in 2003, I have learned a lot. I have visited or consulted with many more companies going lean across industries, and between manufacturing and services. I have worked closely with my former student Mike Rother and learned from his Toyota Kata about the centrality of scientific thinking and thinking in terms of systems. In general, I wanted the book to present a less deterministic and more fluid view of the Toyota Production System and a company’s transformation and emphasize even more the importance of developing people. And the auto industry and Toyota have changed dramatically, a "once in a century" transformation according to Akio Toyoda.  

InfoQ: What's new in the 2nd edition? What has changed?

Liker: I still organized the book around my 4P model and fourteen management principles,  but the model and some of the principles have been revised. As I mentioned, the changes start with a stronger emphasis on systems thinking.

For example, principle 1 focuses on long-term systems thinking which is recognizing that the world is complex and the future unpredictable, and there are unintended consequences of actions. A related concept is scientific thinking, which is about questioning assumptions and subjecting them to tests before drawing conclusions. I emphasize that transformation is not mechanistic, plugging and playing, but rather organic, evolving.  

There are more examples from outside Toyota in both manufacturing and service. New principles focus on the role of customer-facing strategy and the way Toyota cascades strategy through the entire global company through a process known as hoshin kanri. In general I portray the Toyota Way principles as leading to an adaptive, innovative organizational culture.

InfoQ: For whom is this book intended?

Liker: Of course all the lean geeks, and now Kata geeks are excited. And lean-six sigma-agile-continuous improvement folks throughout the world will find it interesting. The original book also played another role; consultants and internal lean folks gave it to senior management and executives, as well as to front-line supervisors to help them get a vision for what excellence can look like. They found the book readable, even engaging. Internal book clubs became common. So I would like to reach beyond the traditional lean folks. I am recommending those who appreciated the original version to read the new one as it is about 90 percent enhanced and enriched.

InfoQ: What does the updated 4P model look like?

Liker: The original model was a triangle with Philosophy as the base, Processes that deliver value to customers next, and then People are highly developed to continually improve through Problem-solving. There is a logic to this ordering, but the more I learned the more I realized this is a system and all the parts are interconnected.  So I now show it as puzzle pieces that are all connected together.  

A new central puzzle piece is "scientific thinking," because without it people tend to stay above the surface with theoretical models of what they would like to see happening and only a superficial sense of how different the reality is from these mental models. Imagine an obese person who wants to lose weight, but believes they are only mildly overweight. They won’t be motivated to improve much.

Scientific thinking starts with having a clear vision of what you want to achieve translated into clear and measurable challenges. We then confront the gap between what we desire to do and the current condition.  

Toyota always emphasized going to the gemba, where the thing you are trying to improve is created or improved. By grounding our improvements in reality, we have an understanding of the current problems, strengths and weaknesses.  

When we move forward it is not with rigid plans or presumed solutions, but rather with creative ideas that we want to investigate. We find simple and inexpensive ways to test our ideas through experimentation and learning.

InfoQ: What benefits can scientific thinking bring?

Liker: From a mechanistic perspective, we tend to think we know the solutions and the task is to implement them as we envision. People get in the way and we try to figure out how to "overcome resistance to change." Success is when people tolerate the changes and the changes seem to look like we hoped. The typical result is that what we implement does not work as we expected, and can get in the way of work. People doing the work try to work around them. Over time, people will tend to revert to what they were doing before and the changes are not sustained. The neat and organized workplace we set up degrades to a mess and there is no control over inventory.

The scientific approach to improvement looks less like a mechanistic process of implementing known solutions, and more like a process of experimenting and learning from each experiment. We engage the people doing the work and coach local leaders through leading the change process. The goal is that they understand the changes, have a personal stake, and then are coached to continue to improve as new problems surface. The result is sustainable change through continuous improvement, and we can better meet challenging goals. The result is that our product and processes better meet customer needs, at lower cost, and with timely delivery. And in the process, safety and morale improve.

InfoQ: Why is scientific thinking so hard? How can we overcome the obstacles?

Liker: What makes it hard is our brains, which were hardwired hundreds of thousands of years ago for fast reactions to our environment. See the animal, kill the animal, eat the animal ... or be eaten. Those who thought quickly and acted immediately were more likely to stay alive and pass on their genes.  

The 21st century requires more slow-thinking people who study, analyze, think and experiment. We are simply not good at it and it feels uncomfortable to most. Developing the capabilities for scientific thinking requires deliberate practice which is also slow and onerous. Our fast thinking, primitive brain rebels and pain cells are activated to discourage us from wasting all that energy on slow thinking.  

It takes deliberate practice to create new neural circuits to make deliberate, scientific thinking feel natural. That is the basis of what Mike Rother calls the improvement Kata pattern which is practiced through various Kata--small, simple practice routines as in the martial arts.

InfoQ: How does one-piece flow work, and what benefits can it bring?

Liker: From a mechanistic perspective if you look at your current way of producing value you will see a lot of interruption, rework, and wasted activity, and that is bad. So you design a one-piece flow process that is based on pure value-added without waste- eliminate inventory, eliminate time buffers, do everything right the first time, and make sure all your people and equipment are in top form. Then you implement it and enjoy nirvana. But it does not work. The system that looks great on paper will fail, and you will put back inventory and time buffers.  

For Toyota, one-piece flow is not a thing to make work but an organic process of struggling to move in the direction of the vision of pure value added to the customer. It is a progressive process of idea-test-learn over and over, with some successes and some failures to learn from, and new sources of waste popping up all the time. The struggle is continuous improvement. The vision provides a true north to strive toward.  

The benefits are high value to the customer who will reward you with market share and profits to keep it all going into the future, and productively support your employees and business partners while benefiting local communities and society in general.

InfoQ: How can we design in quality in software development?

Liker: A customer-software company in Ann Arbor, Menlo Innovations, has gotten global recognition for its high-performance culture and processes and does a great job of that. They learned from the original concepts of Extreme Programming. This has been written about in Richard Sheriden’s book, Joy, Inc.

The starting point is a role they created, called the technical anthropologist, who lives with the users to understand their needs, work habits and how they use software. This is converted to a vision document through mind mapping. This leads to the development with the customer of key personas -- fictional with stories and photos -- and prioritizing the primary persona to design for. The overall vision for the software is developed as a set of features to be programmed and each feature is described on a story card with costs estimated and laid out as a scope of work. The cards are played one by one as programmers convert each card to code. Quality checks start with the practice of paired programming where there are two eyes on every step of the coding. Then unit tests are run to test each unit of code and compiled over night to test compatibility with the rest of the code. Then quality advocates check the results for each story card looking at it from the user’s perspective. And each week at the same time, the customer tries out the software without instruction, which can generate approval or rework that creates new story cards to be posted and worked on in the next week.  

Defects in coding, or poor design decisions, simply do not escape through all these checks, and by the last week of the project only the last week’s work needs to be checked and they often do not hear from customers for some time, as they on their own start using the code without problems.

InfoQ: How can visual control support people doing decision-making and problem-solving?

Liker: We remember a tiny percent of what we hear and about 90% of what we see and do. Seeing is one of the most important senses for doing work. Visual control means we have identified standard practices and the value-added worker can in real-time see gaps between the standard and actual conditions. Think of a traffic light: instant recognition. As we are doing work we usually shut out a lot of the outside world and only notice visuals that are clear, simple, and stand out, like traffic lights when we are busy driving.  

Menlo for example makes the story cards physical and posts them on a wall by day of the week and assignment to pairs of programmers. A colored thread shows what today is and colored sticky dots show the status of a card from starting to work on it, which is orange, to complete and checked by quality, which is green. At a glance, anyone can see if the cards above the thread are green or late.  

Menlo tried to computerize this and mostly returned to the physical, visual version. Toyota continues to use paper visuals as well, including in visual meeting rooms, called obeya, that on the walls show you gaps between planned and actual on timing, cost, and quality. The focus is on the red items first, and then the yellow, and adjustments are made in real-time. Toyota is doing more and more digitally, but at a slow deliberate pace working to maintain the key visual recognition and social dynamics of the manual approach.

InfoQ: What's Toyota's view on leadership and culture?

Liker: First, their view is that they go hand-in-hand. Leaders are role models for the cultural values and beliefs, and develop future leaders who live the values. Toyota’s model of the Toyota Way has two pillars -- continuous improvement and respect for people. They also go hand-in-hand- you can’t have one without the other. The foundation starts with people striving to meet each challenge, through kaizen, rooted in going to see the gemba first hand. Respect includes challenging people and coaching them to grow as they work to improve the processes and themselves. Teamwork is the fifth core and includes individuals taking responsibility to deliver their part to the team with enthusiasm and energy. This is translated to reality through leaders at all levels who care more about the team than their own egos and self interests and are more coaches than dictators.

InfoQ: How can we apply the Toyota Kata pattern to do one-piece flow problem-solving?

Liker: In Toyota Kata, Mike Rother views one-piece flow as a desired outcome rather than as an independent variable to be manipulated. One-piece flow problem-solving means we address each problem one-by-one rather than trying to solve big batches of problems. The example of Menlo starting with a vision of what users need, and breaking that down to pieces with specifications and coding those small pieces, checking small units of code one-by-one, is an example of one-piece flow problem-solving.

The traditional batch and queue way leads to big batches of problems that are addressed through long weekends or months added to the program to do rework. Menlo realizes that one-piece problem solving requires a strong culture of teamwork and people who care about user experience, and works deliberately to hire compatible people and lead them in continually reinforcing that culture. And they are always challenging their processes.

InfoQ: What's your advice for an organization that wants to become a lean learning organization?

Liker: The overall improvement Kata model is a good one. 

  1. Start with a vision and break it into measurable challenges (6 months to 1 year out);
  2. Study with brutal honesty the current condition,
  3. Set short-term target conditions (1-3 weeks out), and
  4. Run in series rapid experiments and learn.

The Kata approach is to assign projects to individuals who lead teams through this process, and each individual learner has daily coaching sessions to develop a scientific way of thinking. It is about many small changes in the direction of a big desired change.

About the Book Author

Jeffrey Liker is professor emeritus of industrial and operations engineering, University of Michigan, and president of Liker Lean Advisors. He is the author of the international best-seller, The Toyota Way (now in second edition), and co-wrote The Toyota Product Development System and Designing the Future.  In 2012 he was inducted into the Association of Manufacturing Excellence Hall of Fame, and in 2016 inducted into the Shingo Academy.   



Rate this Article