Complete firstProcess Mapping
A rough process map or value-stream sketch of the area exists, so the walk visits target stations instead of moving randomly.
Without: Without a map, observation drifts to notable points and misses critical handovers.
RunsheetOperations
Gemba Walk
ComplexityLow
Time30-120 min
Participants2-6
FormatWorkshop
MaturityCanonical
01
Prerequisite
What needs to be finished first
Complete firstStandard work documentationnot in catalog
A current standard or target description of the work at the place of value creation exists so deviations can be identified.
Without: Without a standard, each observation is anecdotal and the gap between should and current state stays invisible.
02
Preparation
What needs to be ready before start
Materials
Notebook or tablet with prepared observation grid (time, station, observation, question, idea); stopwatch or phone timer; camera for sketches or photos after approval; safety equipment on site.
People / roles
Walker (leader or coach who observes and asks questions); host per station (operator or team lead explains); optional scribe who writes notes; no consultants or external guests without process relevance.
Pre-read
Purpose of the walk in one sentence; station list with time windows; standard cycle times and current metrics; planned questions per station; safety, hygiene, or access constraints.
Time needed
45-90 min walk, plus 30 min debrief
Setup
Share walk plan with stations and time windows in advance. Print the observation grid. Announce that the walk is about observing, not grading. Plan safety briefing at entry.
03
Core question
The one question this method answers
Which gaps between standard work and real work become visible on site but do not appear in reports or dashboards?
04
Flow
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Briefing | 10 min | Walker, host, and optional scribe meet at start point. Repeat purpose. Confirm observation grid and station sequence. Clarify rules: no on-site judgments, no quick fixes by walker. | If the walker cannot explain the purpose in one sentence, the walk is not ready. Do not skip briefing, otherwise the team may perceive the walk as an inspection. |
2Phase 2: Walk through stations | 45-60 min | Observe 5-10 min at each station. Record notes in order: observation, question, idea. Ask operators to explain their work without judgment. Walk active transitions between stations. | Transitions are where the learning is, not the station itself. If the walker gets stuck at one station, he loses sight of system behavior. Use a timer to keep station time. |
3Phase 3: Question loop | 10-15 min | At the end, clarify open questions with operators on site. Briefly validate observations with directly involved people before the debrief hardens the picture. | Operator correction is valuable, not embarrassing. If a walker defends an observation instead of validating it, they have left walk mode. |
4Phase 4: Debrief | 30 min | Walker and optional scribe sort notes into three columns: confirmed standard deviations, hypotheses needing more data, ideas for coaching or A3. Name one owner and next step per point. | Quick fixes are prohibited on station, allowed in debrief. Hypotheses need a second data source before they become measures. |
05
Artifact
What comes out at the end
Form
Gemba Walk note as wiki entry or PDF with date, walker, station list, observations per station in three columns (observation, question, idea), debrief sorting, and follow-up list with owner and deadline.
Versioning / ownership
One entry per walk with date and area in title. Follow-up walks are linked, not replaced. Follow-ups get a status (open, in progress, done) and are explicitly referenced in the next walk.
Tool alternatives
- Paper observation grid with photo scan to wiki
- Confluence page in operations space
- Notion database per walk with linked follow-ups
- OneNote or Obsidian notebook in team vault
gemba-walk-working-template.md
Compact working template for Gemba Walk with context, input, output artifacts, and next step.
Gemba Walk Working Template
Goal
Go to where the work actually happens and observe the real process.
Context
When and for what do we use this method?
Input
Which data, observations, decisions, or materials are available?
Execution
Short notes along the runsheet.
Output artifacts
- Observation Notes:
- Improvement Ideas:
Assumptions and open questions
- ...
Decision / Next step
Owner, date, and success signal.
06
Example output
Concrete filled scenario, fictional example
gemba-walk-beispiel.md
Concrete filled scenario, fictional example
Gemba Walk Line 3 (12.05.2026, 09:00-10:15)
Walker: @stefan (Plant Lead). Host: @maria (shift lead).
Stations: Receiving, assembly prep, Station 4 fastening, final inspection.
Fastening Station 4 (09:35-09:50)
- Observation: Torque wrench hangs on floor instead of designated holder. Operator picks it up 11 times in 12 minutes.
- Observation: After reset, display shows 4.2 Nm, standard is 4.8 Nm. Operator did not know maintenance swap occurred on 17.04.
- Question: Who informs shifts 2 and 3 after a tool swap?
- Idea: Initiate an A3 for maintenance calibration.
Debrief sorting
- Standard deviation: Calibration at 4.2 Nm instead of 4.8 Nm is confirmed in the display.
- Hypothesis: 73% scrap increase after 17.04 correlates with tool swap (run Pareto review, owner: @maria, by 14.05.).
- Idea: Adjust tool holder layout at station 4 (owner: @stefan, by 19.05.).
07
Pitfalls
Recognize symptoms and steer against them
Trap
Walk feels like inspection
Symptom
Operators are defensive, answer like in audit training, and show only the showcase station.
What to do
Explain upfront that no evaluation takes place. Walker stays in observation and questioning mode and gives no instructions. When trust grows, walk the same station without prior announcement next time.
Trap
Walker performs quick fixes
Symptom
Walker intervenes directly, moves tools to the right place, corrects notices.
What to do
Record the observation and move fixes into debrief. On-site intervention removes team learning and creates dependency on leader.
Trap
Trying to see too much
Symptom
Walker visits 8 stations in 30 min, notes stay shallow, no handovers observed.
What to do
Maximum 4-5 stations per walk. Two walks are better than trying to do everything in one.
Trap
No follow-up
Symptom
Notes stay in notebook after walk, no wiki entry, no deadlines, no A3.
What to do
Fix the debrief immediately after the walk, not later. Add at least one follow-up with owner and 24-hour deadline in the tool.
Trap
Walker without domain context
Symptom
Questions feel uncertain, operators explain basic terms, little learning happens.
What to do
Before first walk, do a context briefing with shift lead. Walker reads standards and latest metrics in advance. Coach brings domain terms.
Trap
Observation without standard
Symptom
Walker documents impressions but cannot separate should and is.
What to do
Do not walk without standard documentation. If no standard exists, first action is documenting standard work instead of conducting a walk.
08
Stop criteria
Done signals checkable in under a minute
No host on site is available, operators do not explain process from their own perspective.
Area is in exception mode (disturbance, audit, or redesign), and observation does not show normal process.
Safety or access rules prohibit access to critical stations.
Walker has no clear purpose and the walk becomes a mandatory exercise with no learning need.
No standard for work exists, so deviations are not identifiable.
Walker is also escalation receiver and cannot stay in observation, so the walk turns into intervention.
Finished the runsheet?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.