One Point Lesson: Kamishibai

Prof. Jeffrey Liker has uploaded an excellent slideshow from 2005 titled The Toyota Way: A Sociotechnical Learning Organization in Action. The image above is from this presentation in which Liker touches briefly on the kamishibai board and its use.

What is a kamishibai?

In Toyota’s system of jidoka or building quality in to each process, each person checks their own work. But this is not enough to guarantee 100% quality. It is a myth to think that “lean means no inspection” when in fact there are checks upon checks within a build in quality system. Kamishibai is an example of randomized as well as scheduled audits of those process checks and standards.

Among the visual management tools such as hour by hour boards, andon lights, gemba walks and shift-to-shift communication meetings which promote a workplace-focused lean management culture, the kamishibai system is lesser known. In fact the only complicated thing about it is the name, although from the point of view of the discipline required to actually use it, we could say that it is not a tool for lean beginners. The kamishibai board is particularly useful when there is a will and desire for managers to practice genchi genbutsu (go see what’s really happening) but they are unsure how to structure day or even what to do when they are on the shop floor. The kamishibai system formalizes, prioritizes and schedules the checks to be made on the gemba. It is a simple and flexible visual tool to ensure the required checks are being completed.

Why do we need a kamishibai?

The main aim is not to catch people doing something wrong. Before using the kamishibai an organization needs a healthy “why?” culture and not a “who?” culture. For a manager, the proper use of a kamishibai is to train your eyes to see problems (deviations from the standard), identify improvements while they are still small, and teach others to see and solve these problems. In the kamshibai system the process is just as important as the result. In other words, faithfully completing the audits is as important as the result of the audit itself. The purpose is not to find faults, although problems should certainly be made visible; the purpose is to get in the habit of checking each day.

How is the kamishibai used?

Team leaders and above also have the responsibility to check the work of others working within their team or span of control. The work itself and the checking must be based on standard work. The checks performed by the team leader are specifically to audit the line and the performance of the team members working there. The key areas checked include safety rules and the proper use of personal protective equipment, adherence to standardized work, maintenance of accurate documentation, general workplace organization, as well as whether or not other routine activities such as TPM (lubrication, cleaning and checking machines) is being done.

The team leader picks out a card at random from the kamishibai board. This randomness is important to prevent the checks from becoming predictable. If the checks followed a predictable pattern in theory it would be possible for the team members to defeat the system if they so chose, by following the standard during the check but not at other times. The person pulling the card reads the instructions for the corresponding daily, weekly or monthly check and goes to that process to perform the check. After the check, the card is turned over and returned. If there are abnormalities, these are noted on a problem board. A two-color system can be used to indicate that a problem has been found.

While the team leader audit is focused on their immediate work area or zone and the daily, weekly and monthly check cards are pulled on a random basis, the kamishibai process for the group leaders, area managers and above differs slightly. They have a wider span of audit (multiple lines or zones for group leaders, sections or department for area managers, etc.) and the timing of the cards and checks are based on a schedule.

This schedule is built into the leader standard work which identifies when the supervisors and managers spend time directly on the shop floor during the day. The checks performed by the group leader tend to focus focus more on systems such as kanban, andon response, or hourly performance tracking. There may also be specific checks on critical to quality processes.

How do we get started with kamishibai?

A basic condition for starting to use kamishibai is that people hold each other accountable for following standards. Without this, don’t bother with kamishibai. This type of standard work for leaders is based on a longer cycle than an operator’s standard work but the goal is the same: insure that work is performed in the best known way for a safe, productive and high quality outcome. Although the cycle of checking may be longer, the response to nonconforming conditions should be swift and immediate. Needless to say, these countermeasures (action plans) should be made visual at the scene of the problem and the responsible persons must follow up frequently until the issue is resolved.

A site could have as many kamishibai boards as it has team leaders, group leaders and area managers (per shift). Within Toyota factories which use the kamishibai system there may be as many as 100 such boards. One of the keys to success is to pilot this system on a limited basis across several zones or a section, rather than the entire site at once, in order not to create an immediate and unreasonable pull on the managers to respond to and solve problems. The pull should be persistent and strong but not unreasonable. The condition of the kamishibai board will tell you not only how robust your processes are, but how well lean culture has become part of management behavior.


  1. Daniel Markovitz

    July 7, 2009 - 8:51 am

    Nice explanation of the use and power of the kamishibai board. But in keeping with the visual management system you’re starting to implement for your own work (as you described in a couple of earlier posts), you could even use a board like this for all the administrative/knowledge work that white-collar workers are doing. Imagine having a board that audits the quality of meetings, for example. Or the amount of time and effort spent on valueless email. I think it would go a long way towards helping people and companies spot that kind of non-tangible waste.

  2. Jon Miller

    July 7, 2009 - 9:18 am

    Hi Dan,
    Thanks for your comment. I agree that in principle the visual management of the checking of standards applies just as well to the office as to the factory floor. However I think job #1 and the biggest challenge for lean in the office is simply setting such standards which are the subject of kamishibai audits. For repetitive and transactional work this is easy to do but where clear standards are lacking, kamishibai is not recommended, so I would caution against taking this tool too quickly into the office.

  3. Tim McMahon

    July 7, 2009 - 12:57 pm

    Jon this is very similiar to what is done in automotive industry with something called Layered audits. The concepts is checks are done by every level of the organzation from operators to GM/presidents. The closer you are to manufacutring process the more frequent the checks where as the higher your position the less frequent but more strategic in scope they might be. I like the idea of using this visual board to support that process. I can see a real application for this in my organization. This would be good for behavior based safety audits as well like the STOP audits for example.

  4. Yildiz Biner

    July 7, 2009 - 4:22 pm

    Call it what you want, this is still inspection. How does it add value from a customer’s perspective? Would a customer really pay for the time and motion expended? Consider that the people performing the audits are virtually certain to be payed more than the person doing the work that is being audited. That’s waste. All kamishibai does is add cost. No, actually, it adds one more thing: mistrust in workers. How can people be expected to take pride and joy in their work with such an audit system? In a system needing audits, what makes you think those doing the audits can be trusted to conduct the audits faithfully? This system even encourages checking on those checking! This isn’t work! It’s all waste! [See your article “There is No Such Thing as Wasteful Work.”] Deming would have railed and roared at such nonsense as kamishibai. A board should be used if it visually provides needed information that can be pulled by those working the process, when it is needed by them. A kamishibai board takes up space, and adds no value. It also generates the waste of overproduction — namely by overproducing the time spent on checking the process and the documentation accompanying it.

  5. Jon Miller

    July 7, 2009 - 4:35 pm

    Hi Yildiz,
    You raise some interesting questions. From the end customer’s point of view very little of the work done in an organization adds value, including inspection. In some cases there are non-customer stakeholder such as the Food and Drug Administration or the Federal Aviation Administration which require manufacturers to inspect and log quality at key processes. We as individuals and companies also need to do accounting and prepare taxes even though this adds no value. Even though preparing quality or tax documentation adds no value, the alternative of going to jail or having an operation shut down means that we will do this non value adding work.
    In the same way, there are are many (perhaps all) acts of management, sales and service which by your definition are waste. Only those actions that result in a good product arriving in the customer’s hand add value, this is true. But we also have the internal customer. In some cases these management wastes can be eliminated, while in others they would result in even greater wastes if skipped. Quality assurance is in the latter category. There is a basic assumption in lean that zero defects is attainable with capable processes and checks upon checks to make sure those processes are performing as intended. The kamishibai audit is not final inspection, but rather making sure that the process is working as intended, that the built-in self-checks are being performed properly.
    One very important point to add to the article above is this: when a kamishibai system is installed, there is an expectation that these checks (targets of the audit) are being 100% performed, and that the standards are followed. If a kamishibai audit results in the discovery of many non-conformities, this is a sign that the process is unstable or that one of the 4M elements (man, material, machine, method) is too variable in the workplace. Such a workplace is not ready for kamishibai, you are absolutely correct. Kamishibai is not a tool for finding kaizens, it is a tool for maintaining a stable, reliable, high performance system. It is not the first tool to implement on a lean journey, it is one of the last.
    Inspection has internal value in that it prevents mistakes from passing downstream. We must check our work to protect the customer. I don’t think Deming would have argued with that.
    All systems need audits. It is not a question of trust.
    The kamishibai system works because problems are discovered while they are still small, the problems are corrected and people learn to prevent these problems from happening again.
    In a lean system there are checks upon checks. This seems incredibly redundant and wasteful. It is counter-intuitive. Your reaction is completely normal, and it is a typical reason many people resist much of what lean has to offer.

  6. Jack Parsons

    July 8, 2009 - 6:46 pm

    Yes, the “layered audit” is an adaptation of kamishibai, but there is a danger of adopting tools to pass an OEM audit without adopting the spirit and the intensity that should be behind this approach. I have seen many layered audit programs that when audited show that the LA is not catching non standard work.

  7. Yildiz Biner

    July 8, 2009 - 6:51 pm

    Jon, thanks for the response.
    After much reflection on your reply, I must say that, sadly, I remain firmly opposed to using the kamishibai system, even with your added clarifications. Let me try to explain why, by addressing your argument. You rightly point out the need for meeting the requirements of non-customer stakeholders such as the FDA through necessary but non-value added work. I also agree in essence with your comments on management’s actions re internal customers. Managers are often expected to perform many functions not immediately directly related to producing value for the end customer, but which better enable the worker “on the floor” to provide such value. Time spent by management in designing and improving systems, developing employees, etc. are examples.
    You go on to say: “In some cases these management wastes can be eliminated, while in others they would result in even greater wastes if skipped.” Ironically, this is the basis for one of my key arguments against kamishibai. I don’t believe kamishibai results in greater wastes if skipped. I think the opposite is true. I think using it leads to greater waste than skipping it, because of the extent to which it robs the worker of joy in work and pride in workmanship, and in the way it subtly but insidiously provokes mistrust. Just look at the reason you give for choosing a card at random: “If the checks followed a predictable pattern in theory it would be possible for the team members to defeat the system if they so chose, by following the standard during the check but not at other times.” Implicit in this is the theory (not in theory, but the theory) that workers can’t be trusted.
    It was Deming’s belief that if employees understood and could see (e.g. via a flowchart) where they fit into the system, how what they do matters, who depends upon them, that was the key to doing quality work. His belief was that people want to do good work, they just don’t always understand what is the aim of the system they work in, how they contribute to that aim, why they are doing what they are doing, and the like. Kamishibai, or any system that implies mistrust, defeats intrinsic motivation (joy in work). Ask people what problems they are having, and act upon it. Don’t audit them. Management must improve the system, with feedback from those who work in it. You say: “There is a basic assumption in lean that zero defects is attainable with capable processes and checks upon checks to make sure those processes are performing as intended.” First of all, this definitely runs counter to Deming’s stated beliefs; just read “Out of the Crisis” and his several comments on the fallacy of zero defects.
    Aside from that, I don’t believe that is a basic assumption in lean anyway; certainly not the lean I leaned and teach. I think the relevant basic assumption is that perfection is an ideal to work towards, without expecting its achievement. Reducing variation should be the primary aim. One can reduce defects and still increase costs (just widen the specs). And checks upon checks may be the way some practice lean, but I think they do so wrongly. Checks upon checks is pure waste. And also, for what it’s worth, antithetical as well to the teachings of Deming. He specifically addresses this in “Out of the Crisis,” explaining that it leads to a system where “no one has a job” because each checker thinks the other checkers do their job well, and he/she can depend on them, making their own “checking” perfunctory. He calls this divided responsibility, and argues (with evidence) that it creates more defects, accidents, etc., not less. You say: “The kamishibai audit is not final inspection, but rather making sure that the process is working as intended, that the built-in self-checks are being performed properly.” This can be done without the boards. The boards, as I tried to argue in my original post, are overproduction and take up space. Furthermore, randomizing, or choosing cards, leaves to chance what will be looked at. This is no substitute for knowledge. It denigrates the process, turning it into a semi-lottery.
    A side issue: you mention that when an audit turns up many non-conformities “that one of the 4M elements (man, material, machine, method) is too variable in the workplace.” No. It could also be the case that the environment is too variable. I remember a story that Shingo tells about an individual who recognized that a certain defect was produced every time the local railroad train roared by causing vibrations, a cause that had been missed by everyone else. I might add, one of the purposes of 5S is to reduce variation in the work environment, to make it a less variable element, and help standardize that environment. It could also be that information is too variable an element. Does it flow, without burdening? Can it be pulled consistently when needed? Is it uniform from shift to shift? Etc. Copying 4M from Toyota (conforming in checking non-conformity) is not yokoten. We should recognize there are other salient categories as well. “Inspection has internal value in that it prevents mistakes from passing downstream. We must check our work to protect the customer. I don’t think Deming would have argued with that.” As I’ve indicated, he would have argued (and did argue) who was responsible for the checking, and by what method it should be done. “All systems need audits.” No! “It is not a question of trust.” Yes it is, ultimately—with some exceptions. But it is the exception, not the rule. “
    The kamishibai system works because problems are discovered while they are still small, the problems are corrected and people learn to prevent these problems from happening again.” No, it fails because in a statistically stable system, this amounts to tampering, increasing variation (this is what’s counterintuitive, as opposed to what you assert to be counterintuitive). Again, read Deming, or any article on what is known as the funnel experiment. “In a lean system there are checks upon checks.” No. In a system with checks upon checks there are checks upon checks. They are not required to do lean. They are a method (a wasteful one) of attaining an outcome (e.g. furthering standardization). But those outcomes can be achieved without checks upon checks, and more effectively than by such brute methods.
    “This seems incredibly redundant and wasteful.” It not only seems so, but is so. “It is counter-intuitive.” No it isn’t. What is counterintuitive, is that taking action on a statistically stable process can make it worse (and is waste), and that doing less inspection and auditing improves quality. “Your reaction is completely normal, and it is a typical reason many people resist much of what lean has to offer.” ActualIy, I think your reaction to my post is completely normal, and it is a typical reason many lean adherents resist improving what they have learned as the way to do lean. I believe you defended what you already believe, your mental models as it were, without sufficiently critically examining an expression of disagreement offered by me. Perhaps I am wrong, but it is my interpretation, especially given the short interval between my post and your response. I am a huge proponent of lean; have been for many years. I have taught it without reliance on checks upon checks. Doesn’t make me right; but it grounds my beliefs.
    Respectfully, Yildiz.

  8. Pete Hinzy

    July 9, 2009 - 10:06 am

    I struggle with Yildiz’s concern with the use of audits within a lean environment. Many of the lean concepts and tools such as 5S, TPM, kanban, etc. require ongoing audits to ensure that they are functioning correctly and sustained. The purpose of the audits is not to identify “who” is not doing “what” but “why” the process is not working as planned. It is another way to understand the plan and variances to the plan so that corrective actions can be taken. In addition, the audits are signifying to the owners of the standard work that the process and following the process is important. It is not a matter of trust but of priority.

  9. Yildiz Biner

    July 9, 2009 - 4:40 pm

    Pete: I am not arguing for eliminating all “audits.” Not at all. Actually, I am not even sure we have an operational definition of “audits,” hence my use of quotation marks. What I am arguing against are certain types of “audits,” while also urging us to be very careful with the use of audits (whatever that term may mean) in general, by ensuring that(as Jon indirectly suggests) we use audits only when we judge that the potential loss from not doing so is greater than the waste inherent in the audit itself. In general, I have no problem with self-audits, such as those used to monitor process performance. Audits, however, conducted on a group by someone not in that group require more caution, though again I am not calling for their annihilation either.
    We need to distinguish between audits based on how and why they are conducted. Any management audit that uses processes or techniques which imply mistrust of subordinates, explicitly or tacitly (e.g. kamishibai’s use of randomized cards) is, I believe, a bad way to audit. Any audit used in lieu of actually working on improving a process or product/service, is usually a waste of limited critical time. Any audit that uses more than the absolute minimum expenditure of resources (time, people, money, equipment, energy, etc.) necessary to accomplish its purpose, is a bad form of audit. As a corollary, any audit that uses “checks on checks” (e.g. kamishibai) which divides responsibility, is a bad type of audit. [Note: This does not include all forms of multiple checks. See “Out of the Crisis” for the concept of “joint responsibility.”]
    As another corollary, audit processes that use physical artifacts which take up space permanently or long-term (as with kamishibai boards) need to be looked at very carefully, to see whether such use is essential to accomplishing the audit’s purpose. Audit processes that unduly burden those doing the auditing or those being audited should not be used unless the consequences of not doing so are clearly negative. On the other side of the argument, in processes that affect health, safety, national security, and the like, audits are normally necessary, since in these critical cases the consequences of process failure are too great to be concerned about waste inherent in the audit. It is also important to use the PDCA Cycle (or as Deming preferred to call it, the PDSA Cycle) when conducting an audit, so that one can potentially gain new knowledge(or affirm current knowledge) from it.
    With any audit, I would ask some questions to ensure that using the audit process is the right thing. E.g. “What is the purpose of this audit?” “Is there another way to accomplish this purpose?” “Is this purpose still a valid one?” “Can we do this with less waste?” “How does this audit help us?” “Who can be harmed by it?” “What is likely to happen if we don’t do this audit?”

  10. Sunil Pantoji

    July 13, 2009 - 11:51 pm

    Dear Yildiz
    From the customer point of view Kamishibai audits help.
    1. Customer cannot come for audit in all the cases and some one within the organization has to represent the customer.
    2. This also helps if ( higher paid ) executives visit the actual place of work every day to understand the difficulties faced by the working people and it reduces the danger of ” Standard Job sheet” becoming a theoretical document to satisfy the fancy of the so called higher paid executives. These audits do provide the framework for further improvement.
    3. Some kind of negative motivation is also required and balance is reqd. ( I differ with Deming on this point.)
    4. Your objection for audit by someone outside the group doesn’t stand as these are audits by the supervisory staff and they are very well part of the team and have a vital role to play.
    Sunil Pantoji

  11. Sharma

    July 14, 2009 - 5:24 am

    Dear Jon,
    I have my sincere sympathies with Yildiz.(I am afraid that he will take even my words as criticism and won’t be able to digest, but I really don’t want to hurt him, but anyways I won’t be surprised if he is rude to me.) I have the following points :
    1) First and foremost, some meditation will help him a lot.
    2) He should implement A3 in writing his responses(in this article his words exceed the article itself.)
    3) He should start his own blog.
    4) He says that he teaches Lean, but when I compare you and him in the way you put across your points, I am sorry to say that you(Jon) are polite to say the least and he is at the blame-game to say the least. He is aggressive instead of being assertive.
    5)”The Toyota Way” by Liker says that TPS is not about cherry-picking your favourite tools which it offers, but applying all the tools at all levels at all times.
    6) He assumes that Kamishibai auditors will abuse the “WHOs” for non conformity. But it can be a way to review his(auditor’s) own responsibilities.
    7)He stresses again and again on “checks on checks are wrong, checks on checks are wrong” but he fails to understand that we all agree with him. TPS has to be in letter and spirit.
    8)If we go by his definitions even customers should not complain if they find any defect in our products, as this can hurt the manufacturers feelings. And the manufacturer should not check or ask for the “CHEQUES” or payments from the customers as this can hurt their feelings.
    9) His response should have been like this “Hey Jon, with my past experience, and for the benefit of the readers I would like to add that auditors should not score points, as this will defeat the audit purpose.” Period!
    10) Lastly, I salute your humility and the maturity with which you have answered him. Most of the people would have either not showed his comments our would not have bothered to answer him.
    Phew! This reminds me of my A3 comment above and now I should conclude.

  12. Jon Miller

    July 14, 2009 - 6:03 am

    Hi Sharma
    I want to stress that all views are welcome here, especially views opposing those we might hold precious. The lean philosophy should never be dogmatic and always tempered with skepticism.
    I admire Yildiz for speaking out and sticking to his point, on this forum where his views may be in the minority.
    Far more important than settling right or wrong, we should work to bring understanding, and the collision of opposing viewpoints is a great way to do that.

  13. Joe Molesky

    July 15, 2009 - 5:20 am

    I would like to take a brief moment to respond to comments posted by Yildiz. I noticed that Yilidiz implied that auditing in this format creates a culture of distrust between management and employees. If performed properly that is not the case. As I performed audits in my past life as a supervisor I found that when I audit a “process” more often than not the process is flawed and the individual is not. Individuals may not be following SW to the letter and often there is a very good reason. When discrepancies are noticed the person auditing must discuss the finding immediately with the employee and decide what the remedy should be. This is why Kamishibai is only meant for a company that has a deeply ingrained CI culture. A culture where employees and management are working toward a common goal of bettering the company. In an effort to keep this entry short I would like to sum up my thoughts in one sentence. You must establish a distinct CI culture before you implement Kamishibai

  14. Steve Burkhalter

    July 22, 2009 - 12:22 pm

    Hi Jon,
    As a former Group Leader within the Toyota culture Kaimishibai audits are essential to ensure the system (TPS) is functioning per Standard Work. At Toyota we manage by walking around & the audit cards are represented only as a tool to gauge all processes & functions within the production system (office environment as well) & to help us (the management team) to understand current state. Kamishibai is also used at the Team Leader level but at more detail. In the machining environment kaimishibai establishes vital specification checks (samplings) to ensure stability & variation of the machining application. All members are trained in the Kamishibai System & help to establish the program therefore trust in our members is not an issue.
    Steve Burkhalter

  15. Bob MacPherson

    July 28, 2009 - 5:28 am

    Working in a multi-cultural organization for the past three years has given me a new opportunity to learn about the challenges of sustaining a lean system. I learned lean from work experience at Ford, GM and Toyota suppliers and in every case, there were questions about “trust”. The real issue was not so much about trusting (which implies a potentially value system) but about sustaining the culture needed to support a robust lean system. The most successful companies I worked with are those that use some form of Kaimishibai. Those companies also exhibited the highest levels of communication and trust between the workforce itself and their leadership. The trust was demonstrated by the acceptance that this was just another tool to help them improve. This particular workforce was more mature in their lean approach since it was ingrained into their day to day work. Lean was not something new to do, lean was what they did. And since it was so important to their survival, the lean rhythm was balanced around the use of this tool to maintain the balance and flow of the workplace.
    People are individuals with strengths and weaknesses. The machines and processes they use are subject to some of the same changes that they face. We readily accept measurement of process variation in machines yet feel we need to somehow defend people against such measurements? Is that more of an emotional response or a scientific response? (I mean, after all, with the exception of one of my oxygen generators on board my fourth submarine, machines don’t get hurt feelings, do they?)
    In the best lean systems I have worked with, people are part of the system and see the value in ensuring we have not developed unplanned variance in any area of the system. As others have said, trust has different values depending on the maturity of the people involved. In a true lean system, we find trust is about the process, not the processor.

  16. Luiz Martins

    December 21, 2010 - 4:59 am

    Jon Miller,
    Thanks and congratulations for the explanation, you were too didactic.

  17. Eric Kahn

    August 12, 2011 - 6:19 am

    Am I the only person who noticed that Yildiz is cranking out his diatribes while he is likely at work? Perhaps if they had an audit system at his office or plant they’d find him to be off-task most of the day arguing some finer point rather than making improvements. Just sayin.