When Is a Gemba Walk Worth Doing? Choose the Right Focus for Stable and Unstable Processes

A gemba walk is worth doing when it has a defined learning purpose tied to an important process. Gemba means going to the real place where work happens and value is created. If the process is unstable, walk to expose variation and workarounds before auditing compliance. If it is stable, walk to observe standard work, coach thinking, and refine the process. Results vary by process conditions and context, but a walk without that distinction is not a leadership practice. It is calendar activity.

The purpose should fit the leader’s role and the condition of the work. A first-line leader may need to understand why a handoff repeatedly fails. A manager may need to see whether an agreed action changed the work. Neither purpose is the same as management by walking around. Effective walks clarify purpose, build process understanding, and engage the people closest to the work.

A defined purpose also helps protect limited leadership attention. It gives people a clear reason for the visit and helps the leader distinguish useful learning from a routine presence that produces no clearer next step.

The contrarian point is simple: do not start every walk by checking whether people followed the standard. If there is no dependable path through the work, there is little useful compliance to audit. First make the variation visible. Then strengthen the current best way.

Why reports alone cannot tell you what the process needs

Reports are useful, but they cannot reliably show the conditions that produced the result. As information moves up a hierarchy, details are often filtered. Leaders who are several levels away may receive a problem statement without seeing the missing input, delayed part, unclear approval, or cross-functional handoff that created it. Direct observation adds context before corrective action is chosen.

That matters because a completed task can disguise a broken process. Someone waits for information, calls a personal contact, uses an unofficial spreadsheet, or gets a verbal approval from a person who happens to be available. The work gets done, so the dashboard may stay green. But the process depended on an alternate path.

What most people get wrong is treating that workaround as proof of resilience. It may be evidence of commitment, but it is also variation. Over time, the alternate path becomes normal enough that no one names it as a problem. Leadership attention then goes toward the visible symptom, while the repeated condition that created it remains untouched.

Then again, the exact opposite happens when a leader arrives certain that the report is wrong. That posture turns observation into a hunt for confirmation and damages trust just as quickly. The point of going to gemba is to improve the quality of the question, not to arrive with a solution.

Use open, nonjudgmental questions such as: What types of problems occur during this work? How do you know when a problem exists? Can you show an example of what you are describing? What has been learned from actions already taken?

These questions help people explain the work without being pushed toward a preferred answer. They also show whether the walk is a learning conversation or a disguised inspection. For a closer look at that distinction, see Most Gemba Walks Fail for a Quiet Reason: They Become Inspections, Not Learning Loops.

If the process is unstable, walk to make variation visible

An unstable process repeatedly relies on alternate paths, waiting, missing inputs, unclear approvals, or inconsistent execution. The exact form differs by setting. A production team may be missing components. A service team may be waiting for information. A knowledge-work team may not know who can make a decision. In each case, people either wait or create a workaround to finish the job.

That is the walk’s focus: see the actual sequence, ask where the normal path breaks down, and invite people to show the recurring workaround. Workarounds often remain hidden precisely because people are busy getting the work completed. Yet each one introduces variation, and accumulated variation makes process performance less dependable.

Do not turn this into a broad hunt for every defect. Leaders cannot stabilize every troubled process at once. Choose the process that matters most in the current situation, especially where it affects a critical customer outcome, a major operational priority, or a repeated cross-functional failure. In healthcare, for example, the source uses medication administration, record updates, and patient rounds to illustrate selecting critical processes rather than attempting to fix every variation at once.

During the walk, follow one real piece of work through the process. Ask where people must wait, improvise, or seek an alternate approval. Note which workaround repeats and who needs visibility of it. Notice whether the same condition appears at a handoff, decision point, or information source. Then identify the bounded next action rather than declaring a root cause on the spot.

A walk will not always reveal the root cause. It can still do something valuable: make the variation visible to the people who can help address it and raise awareness that stabilization is needed. The cost of skipping this step is false confidence. Leaders may audit a paper standard while the real process runs through side channels.

This is why stabilization comes before compliance auditing when work has no dependable baseline. Once a current best method is visible and repeatable enough to observe, the walk can shift from exposing variation to improving standard work.

If the process is stable, walk to strengthen standard work and follow-through

A stable process gives the team an observable current best way to work. That makes a standard-work observation meaningful. Standard work is not a rigid script written away from the work. It should be built with the people doing the work, focus on the few actions that matter when performed consistently, and remain open to improvement.

Treat the standard as an experiment, not a permanent rulebook. The current method is simply the best way the team knows at this moment. A stable-process walk can therefore focus on one purpose: observing standard work, reviewing progress on existing actions, or seeking improvement ideas against a target condition. These are distinct purposes, even when they occur in the same area.

Ask what problems still occur, how the team recognizes them, what evidence can be shown, and what actions have taught them so far. Those questions support coaching because they ask people to think through the work rather than defend themselves.

Record the specific learning or issue. If action is appropriate, clarify the next step and owner. Then revisit whether the walk fulfilled its chosen purpose. A useful gemba walk is a focused learning loop, not a leadership ritual. The sharper question is not, “When is the next walk?” It is, “What does this priority process need leaders to learn next?”