Respect for people: the forgotten pillar of Lean
At Toyota, respecting somebody means asking them to think. Not to obey. And that's where most of the deployments I've seen fall over.
Lean stands on two pillars: continuous improvement and respect for people.
Enter your email and you will get an alert when something new is published. Unsubscribe anytime.
The first is easy to copy. Boards, events, indicators, photos for the intranet. The second almost never gets copied. And that's why so many deployments end up as scenery.
Let's be clear about what "respect" means here, because it isn't what it looks like.
It isn't kindness. It's demand.
In the Toyota tradition, respecting somebody is not treating them well and sparing them trouble. It's taking what they know seriously: handing them the problem and expecting them to bring the solution.
And its opposite isn't harshness. It's paternalism: deciding for the person because you've assumed they won't know how.
Think of the supervisor who solves every problem his team has. He believes he's helping. He's doing the exact opposite of respecting them.
| What people usually understand | What it means in lean |
|---|---|
| Good treatment, good atmosphere | Trusting their professional judgement |
| Sparing them problems | Giving them the problem and the means |
| Listening to complaints | Their proposals changing things |
| General training | Training to decide at their station |
How you know if you have it
You won't measure this with a climate survey. It shows in three very concrete things.
Whether the improvements you implement were proposed by somebody at the station, or always come from outside. Whether the first question when a problem appears is "who?" or "why did the process allow it?". And whether somebody can stop the line on seeing a defect without asking anyone's permission.
That last one — the andon, the authority to stop — is the most honest test I know. An organization that doesn't grant it is saying out loud that it doesn't trust its people, however many posters it hangs in the corridor.
The bridge to 2026
This is where industrial artificial intelligence wins or loses its acceptance. And I'm telling you this having watched it fall both ways.
An AI system in your plant can be set up in two ways.
As surveillance: it measures the operator, counts his times, flags his mistakes. Or as a tool: it warns him before the problem reaches the end of the line, takes the paperwork off him and leaves him to decide the doubtful cases.
And watch out, because technically they can be the same system. The difference is in who sees the data, what it's used for, and whether the operator can contradict the machine and have that count for something.
In the deployments where the operator could correct the AI and his correction went back into the model, the system was adopted on its own. Nobody had to push. Where it was installed to control him, it ended up quietly ignored — which is how a project dies in a factory, without a sound.
That's why I talk about strengthening rather than replacing. And it isn't sentimentality on my part: it's the only version that survives a year on the floor.