Gemba Walk Questions for Production Managers, by Process Step

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 · 2026-11-11

Gemba walk questions for production managers, by process step: material supply, machining, assembly, test and shipping, plus how to spot the seven wastes.

Good gemba walk questions are open, specific to the process step and aimed at the work, not at the person. This article is for production managers who want questions they can use on the line today. It starts with four questions that fit every step, then gives questions for five typical steps in a discrete manufacturing flow and a way to spot the seven wastes. Gemba is Japanese for "actual place", the place where the work happens.

Key points

• Ask the same four things at every process step: what comes in, what work happens, what goes out and what gets in the way. Then add questions specific to the step.
• Sort questions by process step, such as material supply, machining, assembly, test and shipping, so you can follow the flow in the order the product moves.
• Ask what first, then why, then what if. Asking why too early skips the work of understanding the situation.
• Use the seven wastes as a lens for observation: overproduction, waiting, conveyance, processing, inventory, motion and correction.
• Count what you see, such as minutes spent searching or waiting, so the answer to each question is a fact the team can act on.

Gemba walk questions that work at every process step

Four questions work almost anywhere. Ask them at each step, in the order the product moves:

1. What comes in? Which material, information or tool does this step need to start work?
2. What work happens? What does the operator do, and how long does one cycle take?
3. What goes out? What does the next step expect, and does it get that?
4. What gets in the way? What makes the job harder than it should be?

The Lean Enterprise Institute suggests an order for your wording. In its article How to go to the gemba, it recommends asking what, then why, then what if, then why not, and warns that asking why too early skips the work of understanding the situation. So first learn what happens. Then ask why. Later, test ideas with "what if".

A value stream mapping exercise uses the same logic on paper: it follows one product family through its process steps and records what comes in, what is done and what waits. Your walk is the same thing on foot.

Gemba walk questions by process step

The questions below are practical examples assembled by the author from common lean practice. They are not a published standard. Choose three to five questions to ask per step, and adapt the words to your plant. They suit the shop floor of a discrete manufacturer, and you can invite the shift supervisor to join and review the answers with you.

If time is short, do not walk every step. Start at the step with the longest queue in front of it, because work piles up where the process is slowest. That step is often the bottleneck, the resource that limits how much work can pass through the system in a given period. Ask your questions there first, and then walk one step upstream and one downstream to see what feeds it and what it feeds. This gives you a short walk with a clear reason for each stop.

Material supply and receiving

• How does material reach this station, and who decides when?
• What happens when a part runs out?
• How far does material travel before it is used?
• Which items are searched for most often?

Machining or fabrication

• How long does one cycle take, and how long does the machine actually run in an hour?
• What stops the machine most often?
• How long does a changeover take, and what does the operator do during it?
• How many parts wait in front of this machine, and why that many?

Assembly

• Does the operator have everything within reach, and does the station meet its 5S standard (a method to keep a workplace sorted, tidy and clean)?
• Where do operators walk, search or wait during a cycle?
• What do operators do when a part does not fit?
• How is the standard work shown, and does the operator follow it?

Inspection and test

• What does this step find, and what does it send back?
• How long does a finished unit wait for inspection?
• Who decides what happens to a part that fails?
• Which defect appears most often, and where does it start?

Packing and shipping

• How do orders reach this station, and in what order?
• What do packers wait for?
• What makes an order late at this point?
• How does the team find out that a customer is waiting?

Use only a few of these on one walk. A walk with 20 questions at every step leaves no time to watch.

How to ask so that people answer

The wording matters, but so does how you stand and listen. Four habits help production managers, and all of them favour open-ended questions over yes-or-no questions:

• Stand beside the operator, not across. You both look at the work, which feels like a shared task and not like an interview.
• Ask people to show you, not tell you. "Show me how you load this part" gives you more than "How do you load this part?" You see the work as it is.
• Follow one order or one part. A real part gives the walk a thread, and it gives every question a context.
• Keep notes short and write them after the step. Writing while the operator talks breaks eye contact. Step away, then note counts, times and quotes without names.

Leave when you have your answer. Thank the operator, and say what you will do with it. If you promise a follow-up, put it in the action list the same day.

How to spot waste at each step

Waste observation is easier with a list in your head. Taiichi Ohno of Toyota identified seven wastes typically found in mass production. The Lean Enterprise Institute lists them in its lexicon entry on the seven wastes:

• Overproduction: making more than the next process or customer needs.
• Waiting: operators idle while machines run, equipment fails or parts arrive late.
• Conveyance: moving parts or products unnecessarily.
• Processing: doing unneeded or incorrect work.
• Inventory: holding more stock than a controlled pull system needs.
• Motion: unnecessary body movements, such as searching for parts, tools or documents.
• Correction: inspecting, reworking and scrapping defective output.

Match them to the steps. This keeps a gemba walk focused. At material supply, look for conveyance and inventory. At machining, look for waiting and overproduction. At assembly, motion. At test, correction. The Japanese word for waste is muda. Jim Womack also asks about mura, unevenness or variation, and muri, overburden, as described in his introduction to Gemba Walks. Use the three words as a lens, but always count what you see before you name it.

Example: counting waste at one assembly station

This is an illustrative example with made-up numbers. A production manager watches an assembly station for 10 minutes and notes what the operator does with a stopwatch.

• Assembling: 6 minutes.
• Searching for a fastener or a tool: 2.5 minutes.
• Waiting for a part from the previous step: 1.5 minutes.

The three add up to 6 + 2.5 + 1.5 = 10 minutes. She calculates the share of each:

• Assembling: 6 ÷ 10 = 60 percent of the time.
• Searching (motion): 2.5 ÷ 10 = 25 percent.
• Waiting: 1.5 ÷ 10 = 15 percent.

So 40 percent of the observed time is not assembly. She then asks three questions: "Where do you usually find this fastener?", "What happens when the part from the previous step is late?" and "What would you change first?" The operator names a missing parts tray and a batch release from the previous cell. The team agrees to test a tray at the station for one week and to measure again.

Ten minutes is a short sample, so she does not treat 40 percent as a fact about the whole shift. She repeats the observation on two more days before drawing conclusions.

Common mistakes

• Asking all the questions at every step. The operator stops working and starts answering.
• Asking about people. "Why are you slow?" ends the conversation. "What slows this step?" starts it.
• Accepting the first answer. The first reason is often a symptom. Ask why once or twice more to reach the root cause.
• Naming waste without counting it. "There is a lot of motion" is an opinion. "Two and a half minutes of searching in ten" is a fact.
• Skipping the step that feels fine. Problems often start at a step that looks calm, such as material supply.
• Not reporting back. If answers vanish, operators stop giving them.

How software can help

Notes from walks show waste at one station. To see how steps connect, a value stream view helps. iFAKT's Polaris VSM is a module of the Polaris production software for data-based value stream mapping. It models material and information flows collaboratively and makes waste and bottlenecks visible. A model like this can point to the steps where your questions are most useful. The answers still come from the people on the line.

Next step

Pick one product flow, write three questions for each process step and take them on a walk this week. Count what you see, bring the facts back to the team and use them to prepare better gemba walk questions for the next round.