Lean

Can We Have Too Much Scientific Thinking?

Avatar photo By Jon Miller Updated on July 28th, 2019

The term scientific thinking has been used with increasing frequency in the Lean management community over the past decade. The idea is not new. There is little doubt as to its importance in continuous improvement, good management and good decision-making. The introduction of the so-called Toyota kata approach to improving and coaching the PDCA cycle has popularized the notion of scientific thinking and efforts to practice it. However sometimes I can’t wonder if the community, or parts of it, haven’t glommed onto this “new” idea in a similar ways as we have done with many other buzzwords in the past. Namely, adhering fondly to a favored method at the cost of understanding its limits. Specifically, I am encountering more cases where people are touting scientific thinking when what the situation calls for is engineering.

Scientific thinking is a good thing. Any time we use our minds to critically examine a situation, practice logical reasoning and approach our presumed knowledge and assumptions with an attitude of skepticism, we improve the quality of our thinking. This rigor gives us a better chance at making discoveries that lead to improvement. Scientific thinking aims to reduce uncertainty by disproving hypotheses. This relies on data gathered from observation and experimentation. By their nature, such experiments are low-yield propositions. Most scientific experiments fail. The aim of science is to create knew knowledge. It is to turn the unknown into the known. This is all good, but not always what is needed when we are working to build, maintain or improve a modern management operating systems.

Engineering on the other hand, is the application of scientific principles to design and build things that do what we want them to do. This includes machines, lifeforms, structures, processes, systems and so forth. Engineering deals largely in the known and proven. Engineering is not about creating new knowledge, it is about inventing novel solutions to a problem or making improvements to existing solutions. Engineering works in the realm of applying well-understood principles to new or extended situations. Compared to scientific experiments, engineering efforts are higher yield, and often higher-states with failures being more expensive. For example, through scientific experiments we can test new materials for bridge-building in a lab and survive many small failures. One failed experiment in engineering a working bridge can be too expensive to even contemplate.

Scientific thinking which requires reproducibility, peer review and rigor with both method and data. The test of engineering is “Do we know that works?” and “Did it work in this application?” If an engineering solution reduces effort, difficulty, danger or even the current impossibility of a task, we call this a success. Understanding the underlying scientific theory is far less important to us as long as the mechanisms work consistently. It is the job of the engineer to come equipped with the technical knowledge and tools in such a way as to be able to recombine these in novel ways fit to the situation. This is why we rely on engineers to design and build things. Engineers in turn rely on scientists to do the fundamental research to give them better tools, methods, materials, theories and so forth.

The premise of Toyota kata is that there is threshold of knowledge that we must experiment our way through via scientific thinking. Should we always accept that this is so? Even if these threshold or gaps of knowledge exist for us personally, must we face them as scientists? When is it appropriate to call in the engineers who already know how to build bridges, figurative or physical, to span these gaps? In truth it is not either-or, and the current popularity swing towards scientific thinking may simply be a reaction to the decades of engineering-driven Lean. But to this I would pose the question of whether we scientifically understand the reasons that previous approach didn’t always work, and whether the solution is always kata.

One of the goals of Lean management is to enrich the lives of people. This means developing people. Among other things, this means making people better thinkers. Scientific thinking is one important upgrade for most of us. But is the goal of Lean to develop scientists? What if this comes at the expense of building systems that are less effective, more expensive and take longer to complete? When I see Lean coaches who teach kata and watch their teams develop something better but still far from textbook 101, it makes me wonder why we are doing the equivalent of high school science experiments in an aerospace workplace. It seems that scientific experimentation should be the approach only when we arrive at the very edge of available knowledge, after we have studied and adapted the state-of-the art methods for ourselves. This requires engineering. What I question is the “kata first” approach that skips basic benchmarking, study of the tools, research on application in similar environments, and mistakes unstudied ignorance for a true “threshold of knowledge.”

Some argue that it is better to let reinvent the wheel so that they experience the joy of finding the solution, even it if is already known by many. This can build buy-in and a sense of ownership of that solution. It also teaches scientific thinking. One problem with is that people arrive at a half-way solution, a flawed version of an established Lean method, and defend it rather than continue to evolve it to and beyond the state of the art. If we are developing people through thinking and practice, why not as engineers? Start by setting down the immutable scientific principles, such as flow and pull in Lean terms, and encourage people to invent ways to apply these principles and Lean tools to their situation. Scientific thinking can still be learned in the process of finding new applications rather than new knowledge.

In the Toyota management language, there is much talk about doing things scientifically. However they do not talk about turning their people into scientists or even scientific thinkers. The aim of all of that PDCA is to deliver better products and services to the customer, to contribute to society and thereby to make profit that sustains the business. In their manufacturing operations, scientific thinking exists to serve what they call seisan gijutsu or production engineering. Perhaps this point was missed by Western researchers because the term gijutsu can, depending on context, mean skill, engineering, technique, skillfulness, technology or art. In the plants, well-polished gijutsu at all levels is what allows Toyota to build reliable cars on-time at a low cost. Their practice of kata aims to develop gijutsu in people, processes and products.

Can we have too much scientific thinking? Sometimes I wonder. But as long as we are clear about what state of the art we are attempting to advance through our scientific thinking, we can always use more.


  1. Mike Rother

    August 2, 2019 - 9:21 am
    Reply

    Nice writeup Jon! Yup, as we navigate reality it’s interesting to see where the balance between testing vs. no-need-to-test lies. One extreme takes us to analysis paralysis and the other to falling off a cliff of incorrect assumptions.

    It’s exciting that humans are able to tap into and build on so much existing knowledge (like handbooks of machining speeds & feeds), though often we still have to check existing knowledge as used in a particular application. Engineers often make prototypes or models before production, though increasingly that’s done in computer simulation.

    Don’t be too hard on coaches teaching Toyota Kata, for encouraging simple experiments that, as you wrote, “…develop something better but still far from textbook 101” solutions. Whenever one is trying to develop new skill and mindset you tend to start with simple application practice of fundamental patterns, rather than trying right away to create standout solutions to tough-nut problems. Those coaches are probably not trying to teach Lean solutions, but a way of developing solutions.

    With regard to immutable Lean principles and tools, as you call them, what we found at Toyota is not Lean tools only or scientific-thinking practice only, but an elegant combination.

    Interestingly, our natural tendency to blindly jump onto new Lean buzzword ideas (it happened with Value Stream Mapping too) is something that can perhaps be counteracted by practicing scientific thinking. There’s always a knowledge threshold. Whether or not you have to explore it with testing is a different question, and the answer is certainly not “always.” It is simply healthy (and even enjoyable) to be aware of knowledge thresholds. Maybe that’s the goal of practicing scientific thinking.

    • Avatar photo

      Jon Miller

      August 2, 2019 - 7:46 pm

      Hell Mike. Thanks for your comment. I agree that as long as we maintain a certain level of self-awareness and skepticism of even our seemingly most effective practices, thinking approaches, tools etc. we can avoid it becoming part of the Buzzword Bingo.

  2. Tracy Defoe

    August 2, 2019 - 6:25 pm
    Reply

    Thank you Jon, for putting your thoughts out there.

    From my reading, I think you might be under-valuing the coach-mentor aspect of the Toyota Kata (TK) pair of interlocked patterns. Along with all that scientific thinking, the Improvement Kata (IK) and the Coaching Kata (CK) operationalize a way to respect and develop people and build trusting relationships, every day, while together you learn and get things done.

    It’s people who are practicing that scientific thinking, offering up a “What I learned,” and venturing a Next Step. The CK is a kind of standard work for leaders to pay attention to developing people and at the same time coach them to get past obstacles to desired outcomes.
    We wouldn’t need a workbook and a routine to learn this if leaders were already listening for meanings and teaching the people they manage in the ways they need them to act, think and yes, feel. Heard, safe to learn, curious – these are positive productive feelings to have at work or anywhere.

    Working with people is at least as complex as working with processes and machines. The TK helps make human relationships visible and actionable to anyone who might default to see aprocess and a machine before they notice the people working together.

    Are some people going to far? I don’t know and I hope not. That would be an indication of their learning curve, right? I think it’s common for some managers to feel they know the solution to an obstacle before a new improver/learner has figured out the current condition. That’s a problem for the coach, though not for the learner. As people gain skill as a coach, they get better at keeping the learner in scope – and on track.

    Many organizations can’t seem to sustain lean improvements. The learning doesn’t flow or it doesn’t keep flowing . If TK is a remedy to all the efforts that dry up and aren’t sustained, and if it develops people and builds better relationships, then I am not sure we can have “too much.”

    • Avatar photo

      Jon Miller

      August 2, 2019 - 7:44 pm

      Hello Tracy. Thanks for your comment. I hope I am not undervaluing the teacher-learner relationship in TK. I suppose my point was that sometimes the situation calls for engineering the solution with very little emphasis on developing people. Sure, a leader can still practice TK questions and make sure their engineers are practicing scientific thinking in their approach. But when coaches overly rely on the TK process without a firm foundation in the tools, engineering aspects to guide them, it can become an exercise in “so what do YOU think is the next step” rather than giving direction. It is important for a leader to be able to practice a variety of leadership styles and skills, based on the situation.

  3. Jeffrey Liker

    August 3, 2019 - 7:40 am
    Reply

    Hi Jon, I think you make some good points, but there also seem to be some misunderstandings.

    First, Toyota has said that the basis of TPS is Scientific thinking from the earliest documents in Japanese, I believe back in the 1950s.

    Second, what Toyota means by that is PDCA–that is you plan something, you try it –usually on a small scale– you check to see what happened and then you analyze what you learned from it, and then you plan your next step. It is through many small steps that you create something big. That is how Sakichi Toyoda invented the first fully automated loom, it is the way Toyota developed the Global Body Development System, and it is the way work groups find better ways of doing their daily work taking seconds at a time out of processes. It is very different from the big bang approach of deploying “known” solutions on the shopfloor.

    Third, Mike has worked hard to make clear he has introduced “starter kata,” for beginners who are trying to practice something to replace something else. It is practicing a logical reasoning process to formulate ideas clearly and test them instead of jumping to conclusions and making assumptions that are often faulty. We are all guilty of it. When anyone works hard to rewire their brain so they are thinking more systematically and breaking down a problem into manageable steps and testing their assumptions I think it is a good thing.

    Fourthly, engineering is applied science and the kata is applied scientific thinking–it has nothing to do with peer review or testing a hypothesis that is new to the world although there are similarities of the way of thinking.

    Fifth, when we face a new challenge that is big and beyond what we know, the kata assumes it is generally best to break it down into smaller problems. Challenges I have seen at Toyota include reducing cost of a manufacturing plant by 1/3, reducing the time to develop a production die by half, and reducing warranty costs for North America by 60%. At the outset the leaders assigned did not know how they could possibly achieve these goals, but they got the right people on board and broke down the problem and approached it systematically through PDCA–and in each case succeeded. Along the way there was a lot of engineering science that they knew and put to good use and there were a lot of ideas the teams created themselves and tested. When you are trying to meet a challenge like this you take good ideas, and proven knowledge wherever you can get it and test it in this context, modifying it as needed. Sometimes engineers are taking the lead on aspects and sometimes it is non-engineers. I do not see this as a dichotomy.

    • Avatar photo

      Jon Miller

      August 3, 2019 - 8:37 am

      Thanks for the clarifications.

      First, I agree that scientific thinking is the basis of TPS. My only statement in that regard was that they do not talk about turning people into scientists or scientific thinkers. Science is well revered in Japan and thanks to organizations like JUSE it is a a widely accepted norm in industry.

      Second, it is true that Toyota PDCA-ed their way to TPS. They had comparatively little in the way of prior knowledge to build on. I am not advocating for Big Bang. Nor am I standing with those who seem to think each company needs to rediscover their own way. Take the beaten path.

      Third point is well taken. The motivation for this article is witnessing kata morphing in some cases from “starter kata” to “kata is the tool through which we will practice all…” improvement activity, Lean implementation, or what have you. Potentially the worst of both worlds.

      Fourth point I will have to think about. I don’t know how kata can be scientific if we remove observation, experimentation, hypothesis testing, etc. from it. Maybe “science-like” is a better term.

      Fifth point, agreed. It is not a dichotomy. Kata and best practice solution adoption should go hand-in-hand. More kata practitioners should hear that message.

  4. Jeffrey Liker

    August 3, 2019 - 9:38 am
    Reply

    Two more points and then I will stop.

    With regard to the fourth point, what I said is that the IK is about practical, applied scientific thinking. So science created in a lab can certainly be inputs, but the process is to work on your specific challenges in your specific context and it is very much about observation, experimentation, hypothesis testing, etc.

    On point five, it is in most cases a bad idea to say that you do not have to think scientifically about how a particular practice applies in your situation. Most of the “best practices” of lean are not hard science such as the Bernoulli equation, and even that needs to be creatively applied to the design of a fluid system. Mostly the best practices are general concepts like 5S and pull and one-piece flow and standardized work and the application very much needs to be specific to specific problems in a specific workplace. That does not mean reinventing the wheel but blindly implementing best practices is antithetical to every aspect of the Toyota Way. I think more lean practitioners need to hear that message. Fourthly, I agree that you do not hear a lot in Toyota about turning all people into scientific thinkers. You do hear constantly about problem solving and about PDCA which is deeply embedded in the culture. I believe what Mike has promoted as scientific thinking is very similar, if not the same, as what Toyota calls problem solving. The reason for his choice of terms is that in so many companies problem solving has turned into fix a specific problem with a known solution, rather than find your way to a target based on exploration and learning. So he choose to use the more generic term of scientific thinking. But with a good understanding of problem solving in the Toyota Way of thinking that phrase could replace scientific thinking in my opinion. Toyota wants every one to be deeply trained in the mindset of problem solving and they are constantly coaching people against jumping to conclusions, assuming they know without evidence, and blindly implementing “best practices.” On the other hand there is no doubt there are kata practitioners who are viewing the IK and CK as fixed solutions to be implemented rather than exercises to practice.

    • Avatar photo

      Jon Miller

      August 3, 2019 - 8:28 pm

      Good points and much appreciated.

  5. Mike Rother

    August 3, 2019 - 10:08 am
    Reply

    Below are two interesting comments by Neil deGrasse Tyson: (from: http://www.youtube.com/watch?v=vqEmkp2zL6s)

    • “The goal here is not to make everybody a scientist. That’s not the goal. What a boring world that would be. You want artists, you want musicians, you want novelists, poets; you want the rest of us. What matters is whether they’re scientifically literate and maintain that literacy and that curiosity throughout their lives, no matter what becomes their profession.”

    • “People think science literacy is being able to recite facts, and it’s not that. It’s a part of it, but it’s not the main part. The main part is how do you look at the world. What does the world look like through your lens?”

    This certainly reminds me of what we found Toyota to be doing on the people side. I think as educators or consultants we should be able to distinguish between subtle goals like that and what initial practice (of any skill) may look like. At the beginning it’s often more structured (“Starter Kata”) and beginners may tend to interpret things in a less nuanced, more “buzzword bingo” way. That’s normal and particularly true outside of Toyota; in organizations whose managers (i.e., the teachers or coaches on the ground) are not yet seasoned in a scientific-thinking way.

    Just as a manager’s job is to accompany and coach their people through an arc of learning and increasingly sophisticated skill development, that’s our job too. And along the way we’re screwing up and learning and adjusting just the same as everyone else. I like that.

    • Avatar photo

      Jon Miller

      August 3, 2019 - 8:29 pm

      Good thinking, good products.

  6. Jon Wiederecht

    August 5, 2019 - 12:01 pm
    Reply

    What a tremendous dialogue by some of the pre-eminents lean thinkers. It forces folks like myself to think more deeperly. Clearly this is not a black and white situation – and it rarely is. Unfortunately many times we don’t have the appreciation/recognition that it’s not black and white and we do the copy/paste routine. It’s only when we get comfortable with this notion that we allow ourselves to explore more deeply; that building a lean management system is not like studying for a math test. I now have the task to read and think about this, then so back and reread. In essence, do a PDCA on my understanding. Many thanks again.

  7. Bob Emiliani

    August 6, 2019 - 6:19 am
    Reply

    It is likely that the earliest Toyota documents (in Japanese) referring to “scientific thinking” where informed by Scientific Management and the field of industrial engineering that emerged from it. I do not see sufficient discussion here about Toyota’s IE methods (which informed their kaizen methods), and which were the basis for creating TPS. While the words “scientific thinking” may be used, it has long looked to me more like “engineering thinking” given the emphasis on trial-and-error in the development and evolution of TPS. In his writings, Taiichi Ohno, talked mostly about “trail-and-error,” though sometimes “experiments” as well — the latter suggesting a more structured approach to problem-solving. My experiences in kaizen suggest that improvement is driven forward most rapidly by trial-and-error, a method that is more closely aligned with “engineering thinking.” Finally, the generally poor results of companies engaged with Lean over the last 30 years, and the atomization into constituent tools, suggests that the speed with which experimentation must take place (to keep up with changing times) is poorly understood among practitioners.

    • Avatar photo

      Jon Miller

      August 6, 2019 - 10:04 am

      Very good points. Thanks for the contribution Bob.

  8. Bob Emiliani

    August 6, 2019 - 6:47 am
    Reply

    I should add that the creation of Scientific Management was led by mechanical engineers, and its advancement was made in its first four decades within a professional engineering organization, the American Society of Mechanical Engineers (ASME). Therefore, the roots of progressive management practice are in engineering, not science (as in physical and natural sciences). And, like TPS, Scientific Management was also well known for making rapid improvements in organizations.

  9. Bella Englebach

    August 14, 2019 - 7:25 am
    Reply

    As someone trained as a scientist, and who has worked extensively with engineers, I often hear people say that science and engineering are fundamentally different. This does a disservice to both science and engineering. Yes, we can generalize and say that science is a search for knowledge and engineering is the application of that knowledge. But the first thing a scientist will do in exploring a hypothesis is to head to the the literature. What is already known, and have those results been confirmed? The experiments planned are to advance that knowledge. A design or product engineer does nothing different. A material may have application for a new purpose. Experiments are them conducted to determine if the material is appropriate. At Toyota, the knowledge base of previous experiments is carefully maintained, which is one of the ways that Toyota is able to outperform other manufacturers in product and process development. Reuse of knowledge is key. Whether one is a scientist or an engineer, the process for conducting an experiment is the same, and can be condensed as PDCA. As far as kata is concerned, there is no reason why exploring literature, going and seeing how others do things, or planning to advance known knowledge cannot and should not be part of the “P.” The answer to “what is your next experiment?” could very well be “go and learn who does this well,” or “read the literature on this topic.”

    Mike Rother is also correct in asserting that as someone learns the skill of kata they may run PDCA cycles that provide less than optimal results, and that is not a bad thing. Learning a skill (like kata) is like learning to ride a bike. You are not ready for the Tour de France after your first few attempts. You know nothing about the history of cycling, how to change gears, or maintain your bike. But you have learned to balance. The means are necessary before one can consider the end.

    • Avatar photo

      Jon Miller

      August 14, 2019 - 7:48 am

      Thanks Bella for the reminder that the place scientists start is with a review of the literature.

Have something to say?

Leave your comment and let's talk!

Start your Lean & Six Sigma training today.