Plenty of lines have defined setpoints. Quality still varies. Scrap still climbs on certain products. Experienced operators still correct by feel.
The problem is rarely that target values are missing. The problem is that centerlining isn't run as a living leadership system.
A setpoint isn't a process window
A target value says where the process should ideally run. A process window says which combinations of material, temperature, speed, pressure, humidity, tool condition or environment are stable.
In variable processes, a single value often isn't enough. Teams have to understand when a deviation is harmless, when it becomes critical, and which response is appropriate.
Experience compensates for poor visibility
In many plants, experienced shift teams know when they need to nudge the process slightly. Those adjustments work, but they're hard to scale. New people learn them slowly, and the company can barely check whether the corrections really stabilise anything.
This is where Jidokai can help make patterns visible: which process states preceded good quality? Which combinations often lead to drift? Which interventions stabilise, and which only shift the problem elsewhere?
Centerlining needs a response logic
The key question isn't only: where should the process run?
It's: what does the team do when it leaves the range?
A good response logic clarifies:
- which deviation is only observed
- which deviation triggers a check
- who's allowed to decide
- when QA or engineering gets pulled in
- when a standard has to be adjusted
Lean plus data
Lean brings standard work, a leadership rhythm and problem-solving. AI brings earlier signals and better pattern recognition. Together you get centerlining that doesn't end in an Excel sheet but lives in operations.
The takeaway
Setpoints matter. But stable production only emerges once process windows, the team's response and learning come together.