Here the street widens, because this station is not about a place. It is about a practice — the specific thing Womack did on arriving at a factory, reconstructed concretely enough that you could do it tomorrow. It has three moves.
Walk backwards. Start where the value leaves — the shipping dock, the customer’s hands — and trace the stream upstream, against the direction of flow. Walking backwards is not a ritual quirk. Walked forwards, a process explains itself to you the way its designers intended, each step introducing the next like a well-rehearsed tour. Walked backwards, every step must answer a colder question: what did the step after you actually need, and is that what you delivered? The self-justifying narrative falls apart, and what remains is what is actually happening.
Ask the three questions. You have been carrying them since the prologue; here is where each one goes to work.
Purpose comes first, and it is asked of the whole stream: what problem for the customer is this process solving? Ask it on a real floor and, more often than seems possible, nobody can answer — not because the people are careless but because the purpose evaporated somewhere between the business plan and the work instructions, leaving the process running on habit.
Process is asked of every single step, five ways. Is the step valuable — would the customer pay for it? Is it capable — does it produce a good result every time, or does quality depend on heroics? Is it available — ready when the work arrives, or a queue? Is it adequate — sized to the demand, neither starving the flow nor drowning it? Is it flexible — able to change over when the mix changes? Five small questions, and a step that fails any of them is a place where the stream silts up.
People is asked last and matters most: is every person touching this stream engaged in making it better, or merely enduring it as they found it? A process can score perfectly on paper while the people inside it have quietly stopped believing anything they notice will ever be acted on. Womack’s blunt formulation was that a good walk evaluates whether the work is designed so that the people doing it can succeed — and whether anyone has asked them lately.
Show respect. The third move is the etiquette that makes the other two possible, and it is easiest to define by its violations.
- A walk is not an audit. An audit arrives with a checklist and looks for deviations from standard; the standard is assumed correct. A walk assumes nothing and looks for understanding — including the understanding that the standard itself may be the problem.
- A walk is not a tour. Tours are staged. The floor is swept, the metrics boards are updated, the awkward corner is behind a curtain, and everyone knows the visitors will be gone by lunch. The gemba on tour day is not the gemba.
- A walk is not an inspection of people. The walker’s founding assumption is that waste is produced by the system — the layout, the batch sizes, the incentives — and not by the worker standing in it. The operator walking four steps to the pallet is not slacking; the pallet is in the wrong place.
Watch what the questions do to the walker’s attention. An auditor looks at compliance. A tourist looks at what is pointed at. A walker cross-examines the stream itself, step by step, backwards, with the people who live in it — and writes down what was actually seen, not what was supposed to be there.
One more thing belongs in the method, and Womack modeled it himself. The later essays in Gemba Walks include his own corrections — places where he says plainly that his earlier books had emphasized the wrong thing, underweighted purpose and the role of management, and been too optimistic about how easily the practice would take root outside Toyota. The field’s founding walker published his own errata and kept walking. This site tries to hold itself to the same standard: its later stations make claims that no lean canon backs up, and the badge each of those stations wears says so plainly — argued, not sourced.
Here the street does something no street in 1950 ever did. Look down: the yellow line is still under our feet, but the concrete is beginning to change. The next station is four decades long, and by the end of it the floor we have been standing on will not exist.