Few management approaches have been adopted as widely, or as unsuccessfully, as lean.
The boards nobody updates
The pattern is recognisable in almost any plant that has run a programme.
Boards on the wall with last quarter's figures. Shadow boards with outlines for tools that are elsewhere. Standard work documents in a folder nobody has opened. A kaizen event that produced actions with no owner. Floor markings that no longer match the layout.
Everything visible about lean was installed. Nothing invisible about it was built.
Three reasons this happens, and none of them is that the workforce resisted.
The tools were treated as the thing. The visual controls and standard work are artefacts of a practice. Installing the artefacts without the practice produces decoration.
It was done to people rather than with them. An external team maps the process, redesigns it and hands it over. The people doing the work were not asked, do not own it, and revert as soon as attention moves.
There was no time allowed. Improvement requires time that is not producing, and a plant running at full utilisation with no slack cannot improve, only comply.
What the direction covers
The scope: cutting waste, process improvement, statistical control, Kaizen and operational excellence.
Four areas.
Flow and waste. Seeing the process as material and information moving, and identifying what does not add value.
Stability and standard work. Making the process repeatable before trying to improve it, which is the step most often skipped.
Problem solving. Structured methods for finding cause rather than treating symptom.
Statistical methods. Process capability, control charts and designed experiments, which is where Six Sigma lives.
Tools versus practice
The distinction that determines whether a programme survives.
The tools are the visible things: value stream maps, 5S, kanban, standard work, control charts, A3 reports.
The practice is harder to photograph. People at the work noticing problems and being expected to solve them. Supervisors who respond to a raised problem rather than absorbing it. Small changes made frequently rather than large ones occasionally. Going to see rather than discussing in a room.
Four requirements for the practice to exist.
Time. Not much, but some, protected and regular.
Permission. To change something without a lengthy approval, within defined bounds.
Response. When someone raises a problem and nothing happens, they stop raising problems. This single behaviour kills more improvement programmes than any other.
Skill. Structured problem solving is learnable and not innate, and expecting people to do it without teaching them is a common and unfair failure.
Notice that three of the four are supervisor behaviours. The capability that makes lean work sits with front-line supervision, and organisations train belts instead.
Where this sits in the domain
Lean manufacturing and Six Sigma is the fourth of ten directions in Astra Trainer's advanced manufacturing domain, connecting to quality engineering and reliability for the statistical layer, to manufacturing engineering for flow and layout, and to supply chain and production operations for the planning half.
Because the practice depends on supervisor behaviour rather than tools, partners frequently scope it with the leadership world in the consumer catalogue. Lessons are five minutes and work on a phone, which suits people whose improvement time is measured in minutes rather than days. You can see the ten directions here.
Where Six Sigma is genuinely the right tool
Worth stating precisely, because Six Sigma is both over-applied and unfairly dismissed.
It is the right approach for a specific class of problem: a variation problem where the cause is genuinely unknown and the data to find it can be gathered.
Four conditions that make it appropriate.
The output varies and you do not know why. If you know why, fix it. The structured method is for when the cause is not apparent.
Multiple factors may interact. Designed experiments genuinely outperform changing one thing at a time, and this is where the statistical machinery earns its place.
The problem is worth the effort. A full project takes months. Applying it to something a technician could fix in an afternoon is a waste of both.
Measurement is trustworthy. If the measurement system is the source of variation, every conclusion is wrong. Measurement system analysis exists for this and is skipped constantly.
Where it is the wrong tool: obvious problems, flow problems, problems where the data does not exist, and organisational problems dressed as process problems. Applying the method there produces long projects with modest results, which is where the reputation comes from.
The belt system, honestly
The certification structure deserves direct treatment because it is widely misread.
Belt certification signals training completed. Depending on the provider, it may or may not require a completed project with verified savings. The variation between providers is large and the certification itself does not indicate which.
There is no single governing body. Unlike engineering registration or welding certification, Six Sigma certification is issued by many organisations to many standards, which means a belt on a CV tells you less than it appears to.
Capability comes from projects, not courses. Someone who has completed several real projects with a mentor is capable. Someone who passed an exam may not be.
Organisations over-certify. Training large numbers of green belts who never run a project produces cost and no capability, and it is a common way to spend a training budget with nothing to show.
The useful version is fewer people trained more deeply, each with real projects and a mentor, plus basic problem-solving capability spread widely among supervisors and operators. That is the opposite of how most programmes are structured.
The roles, named
Continuous improvement engineers and managers.
Black belts. Full-time improvement specialists, where the role is genuinely resourced.
Industrial engineers. Flow, layout and capacity, overlapping with this direction substantially.
Production supervisors. Where the practice lives or dies.
Quality engineers. Sharing the statistical toolkit.
Operations managers. Who decide whether improvement time is protected.
Value stream managers, where the organisation is structured that way.
Who can be trained into it
Production supervisors. The highest-return group, because their response to a raised problem determines whether anyone raises another. Structured problem solving plus the behaviours above changes more than a belt programme does.
Operators. Know the problems in detail and are almost never taught a method for solving them. Basic problem-solving capability spread widely is more valuable than deep capability concentrated in a few.
Quality engineers. Already hold the statistics, needing the flow and practice layers.
Maintenance staff. Structured problem solving applies directly to recurring faults, and connects to the maintenance direction.
Planners and schedulers. Flow thinking is their subject from a different angle.
Office and administrative staff. Process improvement applies outside production, and transactional processes are frequently where the largest untouched waste sits.
Where improvement meets regulation. Changes to processes in regulated manufacturing, including pharmaceutical, medical device, food and aerospace production, are subject to change control and may require validation or regulatory notification. An improvement implemented without following change control is a compliance breach regardless of its merit. Training builds improvement capability and awareness of where change control applies. It does not authorise process changes in any regulated environment.
What to take from this
Lean fails when the artefacts are installed without the practice, and the practice needs time, permission, response and skill.
Three of those four are supervisor behaviours, which means front-line supervision is where the capability belongs and where it is least often built.
When a raised problem produces no response, people stop raising problems. That single behaviour ends more programmes than resistance does.
Six Sigma is right for unknown-cause variation problems with trustworthy measurement and enough value to justify months. Elsewhere it is the wrong tool.
And belt certification signals training rather than capability, standards vary widely, and over-certifying green belts who never run a project spends budget without building anything.
Why do lean programmes fail?
Because the visible tools are installed without the practice underneath. The practice requires protected time, permission to change things, supervisors who respond to raised problems, and taught problem-solving skill.
What kills improvement fastest?
A raised problem that produces no response. People stop raising problems, and the programme becomes compliance rather than improvement.
When is Six Sigma the right tool?
For a variation problem whose cause is genuinely unknown, where multiple factors may interact, where the value justifies months of work, and where the measurement system is trustworthy.
Is belt certification meaningful?
It signals training completed. There is no single governing body, standards vary widely between providers, and capability comes from completed projects with a mentor rather than from an exam.
Who should be trained?
Production supervisors first, because their behaviour determines whether the practice exists, then operators with basic problem-solving method spread widely rather than deep capability concentrated in a few.