Lean, before the tools
Lean is not a toolbox: it is a way of looking at work. The tools come afterwards, and only when a problem calls for them. This page sets out what to have in mind before opening a topic. Ten minutes, free and without an account.
What it is
Lean is a way of organising work with a single aim: to give the customer what they expect, when they expect it, using as few resources as possible — and to get a little better every day.
It is not a piece of software, nor a headcount-reduction plan, nor a certificate to obtain. It is a discipline: go and see what actually happens, measure, change one thing, check, start again.
Where it comes from
Toyota, in the 1950s. The Japanese market was small and capital was scarce: mass production on Ford's scale was out of reach. Taiichi Ohno and Shigeo Shingo built a system that turned that constraint into an advantage — small batches, fast changeovers, low stock, defects dealt with at source.
The word "Lean" is not Japanese: it was coined at MIT in the late 1980s to describe that system from the outside. It is worth knowing, because much of what is called Lean today is a Western, partial reading of the Toyota Production System.
Value is judged from the customer's side
Everything starts with one question: what would the customer agree to pay for? An operation that moves the product towards what they expect creates value. Inspecting, moving, storing, waiting, reworking: it does not. Those activities may be necessary today, but they are still not value.
On a production line the ratio is often brutal: a few minutes of value added for several days of lead time. That is not a judgement on how people work — it is a judgement on the organisation they have been given.
The seven wastes
Ohno named seven forms of waste — muda. Naming them is what makes them visible: what has no name is never raised in a meeting.
- Overproduction — making earlier, or more, than needed. The worst of them, because it hides all the others.
- Waiting — for a part, a piece of information, a decision, a machine.
- Transport — moving a product does not transform it.
- Over-processing — doing more than the customer asked for.
- Inventory — capital tied up, and defects that will only surface late.
- Motion — searching, bending, turning, walking back.
- Defects — redoing what was already done.
An eighth is often added: unused skills — the idea held by the person doing the job, which nobody listened to.
The two that get forgotten: mura and muri
Muda made a career; mura and muri much less so. Yet they are often the causes.
Mura, unevenness: a sawtooth workload forces you to size for the peak, therefore to hold stock, therefore to overproduce.
Muri, overburden: a machine or a person pushed beyond what is reasonable breaks down, gets hurt, or produces defects.
Hunting waste without levelling the load is emptying a leaking bucket.
The two pillars
Just-in-time: produce what is asked for, when it is asked for, in the quantity asked for. That is what makes stock unnecessary — and, by the same token, what makes problems visible, since there is no longer any stock to hide them.
Jidoka: give both the machine and the operator the right and the duty to stop as soon as something is wrong. Stopping is expensive at that moment; letting it through costs far more later.
Respect for people
This is half of the system, and the half most often forgotten. Improvement is made by the people doing the job, not to them. Whoever does the work knows the problem better than anyone: their role is not to apply a solution handed down from elsewhere, but to find one and try it.
A deployment that ends in job cuts destroys, in a few weeks, the trust without which nobody reports anything again. It is the surest way to stop continuous improvement.
So why tools at all?
Because a problem that keeps coming back deserves a proven answer rather than improvisation. Each Lean tool answers one precise question:
- I do not know where the time goes → value stream mapping, spaghetti diagram.
- I lose time at every changeover → SMED.
- I cannot find anything at my workstation → 5S.
- Too many causes, not enough time → Pareto.
- The same defect keeps returning → the 5 whys, poka-yoke.
- My output is all over the place → kanban, levelling.
The order matters: the problem first, the tool second. Picking a tool because it is fashionable and then looking for somewhere to apply it is the most common way to fail.
What makes a deployment fail
- 5S reduced to a cleaning campaign, without the fifth S — the one that makes the rest hold.
- The visual board nobody looks at after a month, because it is not connected to any decision.
- Indicators chosen because they are easy to measure rather than because they describe the problem.
- The spectacular one-week workshop with nothing to hold the gain: no standard, no follow-up.
- Lean presented as a cost-cutting programme. The word is then burnt for ten years.
Where to start
Go to the floor and watch one workstation for an hour, changing nothing. Write down what stops the operator. Count. By the end of the hour you will have a list of problems more accurate than any audit — and it is that list, not a catalogue of tools, that should decide what comes next.
Only then will the matrix tell you which tool answers what you saw.