Thinking About Buffers and Production Systems (Birth of Lean Chapter IV – "The Evolution of Buffers at Toyota" – Kaneyoshi Kusunoki )
This week I continue my indulgence in exploring the Birth of Lean with a look at another chapter that some readers may tend to overlook. Instead of quoting heavily from the book – please read for yourself! – I'll share some personal observations.
Kaneyoshi Kusunoki was a key player in the initial establishment of Toyota production operations in North American in the 1980s. He was the first president of Toyota's production facilities in both Kentucky and Ontario (a commonly overlooked fact). Kusunoki was non-resident president, so day-to-day operations were left to executive vice presidents Fujio Cho in Kentucky and Aki Iwabuchi in Ontario.
No one, however, who worked at either of those facilities or at NUMMI where he also made critical production system decisions, will ever forget Kusunoki's visits. More than the gemba visits of anyone else, a pending assessment from Mr. Kusunoki would set the Japanese managers and trainers on high alert – they knew his review was always toughest of all the Toyota HQ executives.
Mr. Kusunoki is now 95 years old. I was fortunate enough to get to know him fairly well over the years and would still pay visits to him in his office throughout the 1990s. Each visit followed the same curve. I would ask him a question or two and he would totally enthrall me with stories of the old days, when the company was ramping things up after the war. Then he would turn the tables. No matter what topic we might be discussing, he would point out how the original and best thinking on it usually came from America. Then he would slip in the question, "So how come you can never get it right?" I never had an answer that satisfied me, let alone him.
How to Think
Mr. Kusunoki joined Toyota in 1946, the first year of hiring following the war. He told me he was one of a recruiting class of six, count'em, university graduates to join Toyota that year. Toyota was still a small company back then. To join Toyota was not to join a firm that could promise lasting job security. Entering into Toyota’s employ was even a risk – no one knew if the company would succeed or not.
Kusunoki spent most of his career in production engineering, not actual operations management like Ohno. Kusunoki’s challenges centered around how to get things to work, how to pour castings that wouldn't break, to stamp steel that would hold shape. As Kusunoki tells it in Birth of Lean, he came under Ohno's direct tutelage – or as he put it to me, "man-to-man training" – in the 1970s, as both men advanced in their careers. Kusunoki tells us he was,
"…the first person to head both production engineering and production control at Toyota, so I was the first person to have the chance to establish integrated policy [for buffering and other issues]." (Birth page 146)
Kusunoki’s history of the evolution of buffers at Motomachi and Takaoka plants speaks to much more than just buffers. His description illustrates the deep concern that Toyota leaders put into thinking strategically about production. Buffer decisions weren’t viewed as simple matters of inventory calculations. They were strategic considerations based on a complex set of factors: how to operate a production system, how to think about downtime and its causes, how to view the psychology of the managers and engineers, and how to utilize buffers as a motivational tool.
Kusunoki tells about developing early English explanations of TPS and of Kanban in particular. As Toyota became famous in the 1970s, the need arose to explain TPS to outside audiences. The company was besieged with “Kanban tourists”. The Birth of Lean reproduces the very first written English explanation of TPS, authored by Mr. Kusunoki along with Y. Sugimori, F. Cho, and S. Uchikawa for the Fourth International Conference on Production Research in 1977. In there, you’ll find buried one of the most telling illustrations of the thinking of TPS, right inside the simple Kanban formula:
y = [D(Tw+Tp) (1+a)] / a
where D is the demand per time unit presented by the following process, Tw is the waiting time per Kanban, Tp is the processing time, a is the number of parts per container, and a is the policy variable [or discretion]. Of special note here is a. Ohno-san taught us not to settle for the number of Kanban necessary to maintain production at some rate but to work continuously to achieve a leaner flow by reducing the number of Kanban as much as possible. He always said that steering that effort is the responsibility of workplace leaders. Without the policy discretion expressed by the a element, our formula would be a mere calculation that any clerk could perform. We knew that Ohno-san would never approve the paper if it did not contain that element." (Birth pages 141-142)
The calculation itself is simple. The telling portion of the story and the calculation is Ohno's insistence on including an alpha number, which stands for "policy", an "extra" amount of buffer to be determined based on the judgment of the human doing the calculation. Actually, Mr. Kusunoki told me that Ohno-san in fact did NOT approve the first draft of the paper and calculation, insisting on inclusion of the alpha value.
Ever the psychologist as well as engineer, Ohno refused to allow operations managers to fall back on simple numerical calculations that a computer could perform. He wanted his production managers to use judgment, to develop instinct, to pay attention, to never hide behind rules but to exercise their own discretion. The purpose of Kanban is, after all, not just to set inventory levels or even to reduce inventory, but to do Kaizen. Set the levels, use your judgment to determine an appropriate alpha value, and get to work doing Kaizen to reduce that number!
The point of inventory isn't simply that it should be zero – yes, less is better so zero must be best – but that buffer is necessary for specific reasons. (Good friend and one of the earliest TPS researchers Doc Hall knew that, when to his chagrin, the publisher insisted on the title Zero inventory for his 1983 book!) And buffers in various places are there for different particular reasons, serving different strategic purposes. For example, Toyota thinks about the buffer between stamping and body welding, between body welding and paint, between paint and final assembly each very differently. That difference resulted in, among other things, a painted body buffer of perhaps 40 painted bodies when I worked at Takaoka, compared with what would have been hundreds in a conventional automotive assembly plant.
Just two weeks ago, I visited the same Takaoka Plant that Kusunoki describes as important to the development of "buffer" strategy and where I underwent my own production training 25 years ago. Amazing. Pulling into the plant grounds, everything looks exactly the same as the last time I was there, about ten years ago. But, inside, the final assembly area is unrecognizable. The production line has been shortened beyond belief, the work of the workers made easier with very little walking, everything quiet and clean – the overall impression was more like an electronics plant than automobile.
That visit drove home to me, yet again, the intrinsic nature of constant evolution that comprises the deepest core of TPS. Kusunoki shares some of his thoughts about that with comments such as these:
"A defining characteristic of the corporate culture at Toyota is that managers don’t scold you for taking initiative, for taking a chance and screwing up. Rather, they’ll scold you for not trying something new, for not taking a chance. Leaders aren’t there to judge. They’re there to encourage people. That’s what I’ve always tried to do. Trial and error is what it’s all about!
We had no money after the war. We tried everything, like divvying up employees’ wages in small chunks, but we finally got to the point [in 1950] where we couldn’t meet the payroll and had to let people go. But even during those tough times, we maintained the spirit of trial and error.
When cash is short, your best shot is to speed the flow of material through your system. You don’t need capital spending to move things faster and shorten lead times. You simply need to improve the way you do the work."(Birth page 150)
The Kusunoki chapter is close to my heart. And writing this column has reminded me that I must send Mr. Kusunoki a copy of Birth of Lean! Don't be fooled by the chapter title, "Evolution of Buffering…". I was happy to hear from my co-author of Learning To See Mike Rother that the Kusunoki chapter is his favorite, not because of the technical discussion of buffers per se, but because careful reading displays some of the deep underlying thinking that Toyota put into designing an integrated automobile production system. And, as we all know by now, if we can get the thinking, we can design rational production systems for any product or service.
Lean Enterprise Institute, Inc.
Lean Enterprise Institute Responds to The Wall Street Journal's Mischaracterization of Just-in-Time
A message from LEI to the Lean Community
How the A3 Process Developed to Help Build Better Managers, Part Two
In this second of two articles, Isao Yoshino and John Shook explore how A3 emerged as powerful practice at Toyota for developing better managers.
How the A3 Process Developed to Help Build Better Managers
One of the hallmarks of a successfully executed A3 process is that it is a collaborative activity--a learning process for everyone involved: for learner and teacher, senpai and kohai, sensei and deshi, say authors Isao Yoshino and John Shook. Here's the first of two articles tracing the development of A3 thinking at Toyota.