How to Run a Gemba Walk: Step-by-Step Plan and Agenda
Disclosure: M365.FM host Mirko Peters works at iFAKT GmbH. This page describes the company factually; paid partnerships and links are labelled as advertising.
By Mirko Peters ·
How to do a gemba walk in eight steps: a clear plan, a 45-minute agenda, what to observe and ask at each process step, and the mistakes that waste the walk.
To do a gemba walk, pick one product flow, tell the team what you will do, walk the flow from start to finish, ask questions about what you see, write down facts and agree one action afterwards. This article is for lean managers who need a clear plan they can follow and teach to others. It gives you eight steps, a 45-minute agenda, a list of what to observe at each step and a worked example. Gemba is Japanese for "actual place", the place where work happens and value is created. The gemba walk meaning is simple: leaders go there to observe. The guide is written for leaders and managers who want results, not theory.
Key points
- Run a gemba walk in eight steps: set a purpose, pick one flow, brief the team, walk it, ask questions, note facts, debrief and follow up.
- A 45-minute agenda works well for a first walk: 5 minutes to align, 30 minutes on the floor and 10 minutes to agree actions.
- Follow one product family from start to finish and look at each step through the eyes of the customer and of the people doing the work.
- Ask what first, then why, then what if. Asking why too early skips the work of understanding the situation.
- Measure what you see, such as queue size or waiting time, so the debrief rests on facts and one root cause, not on opinions.
How to do a gemba walk in eight steps
The Lean Enterprise Institute describes the habit behind a walk as "go see, ask why, show respect". Lean writers also use the plain form "go and see". The eight steps below turn that phrase into a plan. This is the gemba walk process in short form.
- Set the purpose. Write one sentence: "I want to understand why orders for product family X ship late." A walk without a purpose becomes a tour.
- Choose the flow. Pick one product family, one line or one customer-facing process. Jim Womack, who wrote Gemba Walks (see his introduction to the book), follows a single product family or process from start to finish and looks at each station through the eyes of the customer and of the people creating the value.
- Brief the people. Tell the team leaders the day before. Say what you will look at and that you will not judge individuals.
- Walk the flow. Start at the first step and follow the work in the order the product moves. Stay at each station long enough to see several cycles.
- Ask questions. Ask what is happening, then why, then what would happen if something changed. Listen more than you talk.
- Note facts. Write down times, counts, queue sizes and what people tell you. Do not write conclusions yet.
- Debrief. Within 10 minutes of finishing, sit down with the leader of the area. Share what you saw and pick one thing to change.
- Follow up. Name an owner and a date for the corrective action. Go back to the factory floor and check whether it worked. That visit is your next walk.
These gemba walk steps repeat. The second walk starts where the first one ended.
The 8 steps rest on direct observation, which is the main source of facts in lean manufacturing. Repeated over weeks, they support continuous improvement: each walk finds opportunities for improvement, and the follow-up turns the best one into a change. You do not need special tools. You need a clock, a notebook and the time to walk the production floor itself, not the office. Regular gemba walks give you a deeper understanding of the actual work. A separate article in this series offers a gemba walk template for your notes.
A 45-minute gemba walk agenda
Use this agenda as a starting point. Change the times to fit your site.
- Minutes 0 to 5, align. Meet at the board or at the first station. Restate the purpose and the flow. Agree who asks and who writes.
- Minutes 5 to 35, walk. Spend about 30 minutes on the floor. If the flow has four steps, that is a little under 8 minutes per station. Keep to the clock, or you will spend the whole walk at the first one.
- Minutes 35 to 45, debrief. Sit down. Each person names one fact they saw. Agree on one action, one owner and one date.
The 45-minute total is simple arithmetic: 5 + 30 + 10 = 45. Once the team knows the routine, you can cut the agenda to 20 or 30 minutes and still keep the same three blocks.
An effective gemba walk is short, focused and ends with an action. Do not try to solve every problem you see. The Lean Enterprise Institute warns against jumping to conclusions and against issuing prescriptions from a distance. Diagnose and decide close to the work, ideally with the people doing it. Its article How to go to the gemba describes this in more detail.
What to observe and ask on the shop floor
Womack's Lean Enterprise Institute article on checklists suggests five questions for a first, big-picture walk. They work well as your core list: What is the purpose of this process? Who is responsible for it? Is it achieving its purpose and, if not, how is it failing? What causes the failures? How are the people who touch the process involved in improving it? You can read his full advice in Gemba walk checklist.
At each station, look at three things:
- The work itself. What does the operator do, in what order, and how long does one cycle take?
- The flow around it. Where does material wait? Where does the operator wait? What arrives late? Waiting, searching and extra walking are typical inefficiency, and removing them raises productivity without extra people.
- The people. Do they have what they need? Who do they ask when something goes wrong?
Womack also asks how more value can be created with less muda (waste), mura (unevenness or variation) and muri (overburden). Use these three words as a lens to identify waste, and note an example of each.
Keep your questions open. Start with "what", for example "What happens when this part is missing?" Then ask "why". Later, ask "what if". The Lean Enterprise Institute notes that asking why too early skips the work of understanding the situation. You may also ask whether corrective actions from earlier walks are visible at the station. Ask the team leaders what they think is the root cause of a problem, not only what the symptom is.
Good questions improve communication between management and frontline staff. They show that you want to learn the real work and to understand their challenges, rather than relying on reports or data alone. That builds mutual respect, which is a core lean principle.
Example: reading the queue in front of one station
This is an illustrative example with made-up numbers. During a walk on a machining flow, the lean manager reaches step three, a drilling cell. She counts what is waiting in front of it and what the cell finishes.
- Queue: 12 pallets of parts, each holding 18 parts. That is 12 × 18 = 216 parts.
- Output of the cell: the operator reports about 54 parts per hour. She checks this against 10 minutes of observation and counts 9 parts, which fits.
- Waiting time: 216 parts ÷ 54 parts per hour = 4 hours.
Every part that enters the queue waits about four hours before it is drilled. The drilling itself takes about a minute per part. The previous step releases parts in large batches, so the queue builds up in bursts.
She asks the operator why the pallets are so large, and learns that the previous cell changes its tool setup rarely and therefore runs long batches. That is a root cause worth a small kaizen, a practical improvement made close to the work. In the debrief, the team agrees to test smaller batches for one week and to count the queue again on the next walk.
One number, a count and a division, turned a general feeling ("this cell is always flooded") into a fact the team could act on.
Common mistakes
- Walking without a purpose or a flow. You collect impressions and no facts.
- Skipping the briefing. People who are surprised feel inspected and show you a polished version of the work.
- Talking more than listening. If you explain the problem to operators, you learn nothing from them.
- Solving on the spot. A quick order from a manager who has watched for five minutes often misses the cause. Save decisions for the debrief.
- Taking a large group. Five or more visitors at one station disturb the work and the answers.
- Forgetting the follow-up. An action without an owner and a date disappears. Review it on the next walk.
How software can help
Notes from a walk often end up on paper and get lost. A simple action list with owners and dates is enough, but plant software can also help to show people what the plan for the day is. iFAKT's Polaris Operator is a module of the Polaris production software. It brings production plans to the shop floor as time-based tasks and feeds progress back to planning in real time. On a walk, such a view gives you a shared picture of what should be happening at each step, so you can compare it with what you see.
Next step
Choose one product flow today, write your one-sentence purpose and book 45 minutes for the first of your eight steps. After the walk, use the same agenda again next week, because how to do a gemba walk well is something you learn by repeating it.