Walking the Gemba

Duke left a great question as a comment to “A Morning Market.” He asked:

What’s the key differences between the [morning] market meeting and a gemba walk? Maybe I should ask what is the purpose of a gemba walk?

Good question.

I’ll start with one of my (other) favorite quotes from an old friend, Dave:

“It isn’t about shaking hands and kissing babies.”

That is to say, a lot of leaders think that heading out to the work area to “show the flag,” and do a meet-and-greet with the team members is “walking the gemba.” While they may be at the gemba, and they may actually be walking around, this isn’t it.

Walking the gemba is part of “Check” in Plan-Do-Check-Act. It is the process of carefully observing to see where things are not as they should be. Sometimes there is less walking involved, and more just standing and watching.

This, of course, begs the question: Watching for what? This, ultimately, is what makes it different than even just walking around looking.

Try this exercise. Go to some place where any activity is taking place. It really doesn’t matter what. If you aren’t in a factory, just go watch someone doing data entry or computer work. Watch the receptionist process walk-in customers and phone calls.

Note: Whatever you choose, please talk to people and tell them what you are doing, otherwise this is going to seem creepy to them if they aren’t used to it.

Get a feel for what is happening. More importantly, try to envision what is supposed to be happening. If this process were going perfectly, how would it look? Try hard to visualize in your mind a smooth, totally value-adding work flow.

Now, as you watch, become totally conscious of what is really happening. Why is this process other than how you visualized it? What disrupts the work? Where could mistakes be made? What keeps those mistakes from being made? Is it just vigilance? Or is there some mechanism to either prevent the mistake or either cue the person on the correct way, or alert them on the incorrect?

Is there any backtracking, rework, looping around? Are things where they are actually needed? Do people have to look around for things?

How do they know what they should be doing? What is their source of information? Do they have to hunt it down, or worse, guess at what should be done? Or is the “right thing” and the “right way” chrystal clear to even the casual observer (that would be you).

Is there a pace to the work? Even if it isn’t traditional takt-time work (especially if it isn’t), there is some kind of deadline. How does the person know whether things are on time or not? When do they learn they are behind?

If the person encounters some kind of problem, something unexpected, something needed but not there, what happens? Is there a support system to get this person back on-process? Or is he or she left to their own devices to just figure it out?

As you watch, keep your standard very high. If something is not obviously as it should be, then question whether it is. “Control” has the burden of proof here, not chaos.

When you see these things, focus in on one of them. Ask yourself “Why?” does this condition exist? Where does the problem originate? Go there. Study some more. How is it that this process fails to support its customer? Do they have a clear understanding of what is expected? (Probably not, most don’t, especially administrative processes.) How do they know they delivered (or didn’t) what their customer required? Did they think their output was defect free (according to their standards), but was really causing problems for their customer?

Often it helps to look at one person doing a single task, find a single issue, and follow the trap line upstream until you arrive at the origin of the problem. Often, again, you will have a fixable root cause right in front of you.

Now the fun begins. As a leader, it is not your job to fix the problem. That is not what you are looking for.

When you walk the gemba, you are assessing how well your organization is tuned to seeing these issues, clearing them, finding their causes, and solving them. If you see these opportunities, then your job now is to teach.

Grab whoever it is whose responsibility bounds both the effect and the origin of the problem. Guide him through the same things. If this issue was unaddressed, there are a couple of problems this leader must address.

First, there is someone downstream likely coping with an issue that he should be raising. (Or worse, there is no process for raising the issue, or worse still, no process to respond if he does.) If that isn’t happening, help this leader understand that this is the shortfall here, not the actual issue. It is now his turn to teach, in turn, either the Team Member, or whatever leader is between him and the Team Member.

Second, upstream, there is a process which either does not understand what “defect free” is, or does not have (or does not use) a positive way to verify it before sending it along. Same issue. This is a leader training process until the education reaches the level that should be doing the checks and the verifications.

The best way to learn how to do this is to walk your flow with someone who has the skill. It is simply a skill, and it can be taught. But learning also requires some humility, and the student must bring that to the “classroom.”

So, to recap in two short statements.

  • The gemba walk is a “Check” of Plan-Do-Check-Act.
  • You are checking the health of your leadership systems by looking at how they engage their people and processes.

Walking the gemba is a process of developing your people.

26 Replies to “Walking the Gemba”

  1. Very good article.!!! Perfect.

    One problem we have here where I work is, if the General Manager (He is also the owner) goes out to walk the Gemba, the Gemba changes. Mainly because the workers will work better and faster while he is there. Also when he sees a problem he wants to fix it instead of involving the workers in the fix. However he usually gives the problem to me to fix.

    I am the Lean Manager. I spend a lot of time in the Gemba. Fortunately the workers are comfortable with me there and I can see what is really going on.

    Thank you for your insight on Gemba. I learned some things…..

    1. Jim –
      So when you get a problem to fix, make it your job to teach others to fix it rather than fixing it for them. That is closer to your job as the “Lean Manager” – someone who should be improving the skills and knowledge of the organization by developing them in others.

  2. First, thanks!

    Second, the current state of our gemba walks is that they occur on two levels. The first is a daily manager/supervisor/manufacturing engineer walk that is basically what you describe, but not a rigid in problem solving. The ‘data dump’ to the manager still exists in this meeting and the problems are limited to what that group can solve. Then there is a staff level gemba walk with all the managers(including purchasing, engineering, materials, scheduling, quality). This walk is once a week and is more of a data dump to the plant manager. Certainly the ‘where are we, what are we going to ship’ question leads to most of the problem solving. The idea of ‘watch and observe’ is replaced with ‘how to we solve the current line stopages’.

    I would say that the gemba walk does not address the root cause of some of our big system problems, just puts a band aid on what is going on today. How do you shift focus to the ‘elephant in the living room’? How do you replace the ‘data dump’ to the various managers or wein them from this source of information?


    1. Duke –
      Big system problems are usually the cumulative effect of little niggling ones that tend to get dismissed. The net result is a culture of “normalized deviance” where problems that happen every day become part of the background hiss.

      If you look at “how do we solve the current line stoppages?” as a core question, though, it isn’t a bad one. Focusing on that, alone, would get you a long, long way. The issue is how it is addressed, not whether it is the right problem to work on.

      Consider this thought experiment. When your line was set up, it was set up to run, not to stop. For better or worse, your best minds applied their accumulated experience and knowledge to set it up, design the processes, and in general, make it all work.

      Then you ran it.

      And it stopped.

      So, yeah, clear the stoppage. But solving the problem means to ask some questions. Your initial assumptions called out conditions for success. Were those conditions present? If so, then what have you discovered that also needs to be in place? You have learned something new. Or, if those conditions were not present, then why not? What needs to go into place to make sure they remain, or to at least warn you earlier?

      Lack of parts? Your procurement / inbound logistics system was set up the same way. I would assume that it was set up to get the right parts to the right place at the right time. It didn’t? OK – was there something nobody thought of? Great – you have learned. Or was there something that just didn’t happen? Also great – now we know it was really important, but we need to make sure we know it does happen.

      If there are things out of your control that are tanking the plans, then also great. Now you know the limits of your plans, but you also know that your initial assumptions about what would happen need to be looked at. Either make the system more robust, or eliminate the source of the disruption.

      At the working level, the Team Member does not have what he needs, how quickly does the organization respond? If it takes 10 seconds to get him what he needs, that is a 10 second line stop. If it takes 45 minutes, or he has to spend another 45 minutes working around the problem, then that is a 45 minute line stop. Line stops are valuable sources of information if the organization can respond quickly to them.

      When you start looking at how to tackle the problems that are stopping the line, this is the kind of thought structure that gets you there.
      – Assume you set up the best line you could, applying everything you knew, and that you really thought it would work. (After all, you didn’t set it up knowing it would fail, did you?)
      – When it doesn’t work, you have a source of valuable information.

      Bottom line – the today’s line stoppages are going to happen tomorrow, for the same reasons, unless there is a little more digging. Put a few of these on the morning market board, and carve out some time to run a few of them to ground. Every one that gets tackled will be one that does not come back. That will start to shift the culture. It isn’t going to happen all at once but the effect is cumulative.

  3. Leaders Standard Work without visual management tools results in a marginally effective social event at the Gemba. Only when used together, visual management tools and Leaders Standard Work set a foundation for a sustainable, self-correcting, continuous improvement environment that is characteristic of true Lean.
    recently we had implemented this at our site, In case you wish to have more info pls call at 0915620379

  4. The original questions was comparing the marketing meeting to gemba. In new product development (a key marketing concern), we do not have a gemba to walk around. In these cases, the gemba shifts from our workplace to the customer’s workplace or life (depending on whether your product is BTB for BTC). Then, the gemba visit can bring out the latent customer needs that you may not know even existed. This use of gemba requires special analysis since we have no authority in the customer’s gemba to demand explanation. In Blitz QFD®, a unique tool set has been developed in recent years to aid this activity. There are several case studies on my http://www.mazur.net website that explain how the process was tailored for different companies.

    1. Glenn – I am very happy to have you as a reader and commenter!

      Actually the original comments were about the “Morning Market” which is a process for managing shop floor problem solving, but I can see the source of confusion since the context is only implied. I made a small edit to the post to clarify.

  5. I was doing a google search on Gemba and ran across this thread. I read it with great interest.

    I am a Black Belt but not a Lean practitioner. What this seems to describe seems a lot like Deming’s insistence in acquiring “profound knowledge” about a process to really understand it. He did warn against “managing by walking around” because it tends to only look at the surface and make employees even more wary because of their boss’ attentions (the opposite of Deming’s point about: “drive out fear”). Anyway, this has been enlightening and I wanted to let you know that as a passerby.

  6. First, thanks for all information shared on this GEMBA WALK talks. Second,GEMBA WALKS works, at least in here there should be a compromise from hihg level management to take part of this process. As we know, Managers all the time are involved in their Jobs and GEMBA goes to second level, but GW must be 100 % supported by all the Managers ad will be part of the Quality System, giving Us proud of the solutions that all the team is giving to solve any single “opportunity” caught in the GW. In a few words,GW is very important to gave to operators (The ones that touch our material and transformed at final product) confidence in our resolutions to solve their opportunities that at the end will be part of all of US. GW means Continues Improvement in a high level.

  7. What are your thoughts on machine operators presenting at the department KPI boards instead of the supervisor during the GW?

  8. Kevin –
    A structured presentation in front of a KPI board may well be appropriate for a manager “show and tell” but in my definition, it isn’t part of what I would call a “gemba walk” at all.

    As described, a gemba walk should not require people to alter their normal work patterns. In fact having them do so defeats the whole purpose. The key point is to compare what should be happening vs. what you observe so that you can check if the processes and systems are working correctly.

    Checking KPIs doesn’t, in my mind, require anyone to present anything. They are there to see.

    Now – if it may be that your machine operators are engaged in improvement activities that are intended to impact the KPIs.

    If that is the case, then it is entirely appropriate to have the present:
    – The target condition.
    – The starting condition.
    – What problems and barriers are preventing the target condition.
    – What they are working on.
    – What results they anticipate.
    – What results they achieved.
    – Explain any differences, and what they learned about the process.

    This would be part of the people development process, and would be an appropriate check to see how well the Team Leader is teaching it – by checking how well his students (the operators) have learned it.

  9. Mark,
    We have a relatively large plant and have four departmental KPI boards. We start our walk at a spacific time and only allow a certain time frame for each board. During each stop, the supervisor presents to the leadership team only the areas in which they did not achieve their goals or any highlights for the last 24 hours. At this point they explain why goals were not achieved and if further resources are necessary to solve the problem. After the walk, we have a wrap-up meeting to discuss any misses and if any further resources need to be allocated.

    1. Kevin –
      That sounds like a potentially great process.
      You may be doing this already – so pardon me if that is the case – but the next step would be to tighten down the structure of the problem reporting and problem solving so that the supervisors (and by extension the machine operators) are systematically driven to gain deeper and deeper understanding of the process. If you can’t tell by reading my recent posts, I really like Mike Rother’s coverage of this process in Toyota Kata.

      The more the organization practices the process of engaging with each other in this way, the better everyone will become at doing it + the better the operation will become.

  10. Hi,
    I am a lean manager at a pharmasutical industry. I would like to organise Gemba Walks in a company that there is almost next to no experiene in Lean or appriciation in Lean. Top Management pushes for Lean but to a new comer like me it looks like nothing has happened so far, except the people to dislike Lean.

    My experience is relatively small (almost 5 years in Process/Lean management) and new at this company (almost a month). My idea of conducting Gemba walks as a 1/2 hour walk at the production and packaging shop floor, with the heads of dept and tech depts, is to try and make them think more in a Lean perspective in general as also assist me to understand better the situations.

    What I would like to know is…what is the best way to conduct the Gemba Walks in order to make the most out of them, and make people think that Lean actualy works and can make their life easier if they try. Carrently they just watch it in separation of the rest of the production something like extra work to please the managers and not as a tool to solve problems.

    Many thanks in advance…

    1. Hi Asotosios –
      I think I would start by doing what is outlined in the post you commented upon. The last couple of paragraphs give you a good starting point, and the 18 comments that precede yours offer some additional insight.

      Your question is actually two questions, and they have separate answers.
      “The best way to conduct a gemba walk to make the most of them” is outlined in the original post – though I can go into more detail, it is best to try it first and see what you can learn by doing what is described.

      The second question – “…and to make people think that lean actually works and can make their life easier if they try” is more problematic. You simply can’t “make” someone think or believe anything. They have to do that for themselves.

      What you can do is to teach them to visualize what an ideal work flow would look like, then ask them to observe an actual one and describe to you what they see that is preventing the worker from doing the job the best way. IF a person is willing to learn and is teachable, you can spend time with them and teach them to do this by observing yourself, and asking them questions to direct their eyes until they can see the same things.

      Ultimately, it is a BIG help if a few people can gain this experience with someone who has done this for a long time, a “sensei” if you will – a coach in common English. That can reduce the learning time considerably.

      Like any skill, acquiring this one takes practice in an environment where the right kinds of encouragement and advice are offered. It takes some struggling and correction. It takes some drive and motivation, and those things must come from within to be truly effective.

  11. I’m investigating Gemba Walks to create a training module for leaders at my company. We have spent many years training and practicing lean concepts. The question lately is how do we become teachers and coaches? I believe the Gemba Walk is exactly how this is done and the purpose of the concept. I have read many articles on the subject and we currently perform many Sensei walk activities. I learned a concept from a company in the mid-west (AutoLiv) that mentored us through the beginning of our lean journey. called GUS Follow-up…

    Grasp The situation
    Understand the problem
    Set expectations

    Gemba Walks are all about this, the reason I do a Gemba walk is to understand what the current condition is with my own eyes so that I honor the real situation.

    While I’m there I have to understand the problem… But that is not good enough I have to admit to myself that there is waste and I must teach others that waste does exsist so that they can see.

    Now I must set expectations and this is were it gets tricky. I have to be careful not to dictate the solution, but to try and coach the leader or employee to see the need for something to improve the situation. At the beginning this could be a visual that tracks the problem so that I know how offten this happens or what the number one issue is. In later stages it could be the formation of a team to begin the PDCA workshop or DAMIC project. (I should not try to solve the issue on the floor)

    Then I must follow-up constantly. This is the hardest part in my opinion. If you are treating the Gemba Walk as a one time event you will fail and make lean a program of the month. Don’t start if you are not prepared to continue. Setting expectations with out follow-up is worse than setting no expectation at all… Food for thought…

    Thanks for listening.
    Douglas Miller

  12. Interesting article. The Gemba walk believe will be strongly related to the method called Genchi Genbutsu. Gemba means place and Genchi Genbutsu is going to Gemba to check on the genbutsu (object). Another interesting method is Management by Walking About which is to some degree similar to Genchi Genbutsu.

  13. I realize this is an old string, but I was taken back by the post Kevin made concerning KPI boards. There should never, ever, be a “presentation” on a Gemba Walk. In an effort to drive true continuous improvement in a very large production facility, this has been my greatest frustration. So much pressure is put on Operators to utilize precious time in completing their hour by hour boards when this information is readily available through a sophisticated online system. Then, the Operator has to abandon their responsibilities to “visit” with the Management Team. Mark is spot on that there should be no staged presentation. No daily walk should ever be routine. There should be less talk, more observation, and considered more check of the PDCA process.

    1. I think there is a huge difference between a staged presentation (which accomplishes little) vs. an active coaching session (which should be conducted with the responsible leader).

      To that end, there *is* an improvement story board of some kind, though it wouldn’t be a traditional “KPI board.” Rather, it would be outlining the target condition, current condition, obstacles in play, and summarizing the last round of PDCA / daily improvements / experiments that have been done by the area leader.

      See this more recent post: https://theleanthinker.com/2013/03/18/policy-deployment-and-the-coaching-chain/ for a working example of how this goes.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.