Stellantis has been telling anyone who will listen that its quality war rooms and cost cuts are paying off. I do not doubt the first half. I have run those rooms – the real ones, with a customer escalation burning a hole in the week and a plant manager who could not leave until the disposition was signed – and they work. Containment that used to drag for ten days compresses to twenty-four hours once the decision-maker is physically in the room. That is exactly the problem. Anything that works that visibly never gets switched off, because the organisation remembers the rescue and forgets the arson. The fire gets fought brilliantly. The fuel sits untouched.

Two decades in automotive and aerospace, and one observation about escalation structures never changes: they are all downstream. Every one of them exists to catch what a process has already produced. Fine as a stopgap. Fatal as an operating model.

What a war room actually buys you

Done properly, an escalation room buys three things at a speed nothing else in the plant can match. Forced priority – when the room convenes, everything else on the agenda dies. A single owner – one name against containment, one against root cause, no committee diffusion. And a decision-maker present, which is the entire point. Most containment drags not because engineers are slow but because dispositions sit in inboxes. In the room, the signature happens while the problem is still warm.

The anatomy is boringly consistent wherever I have built one: a 24-hour containment clock, visible boards on a customer-facing cadence, twice-daily stand-ups while the customer is live, and the unspoken rule that nobody defends their function. Air freight and 100% sorting will happily burn €150,000 in a week before anyone touches root cause, so compressing that window is real money. I have walked customers back from the edge in that setting. De-escalation is a craft, and I respect the people who are good at it. It is not a strategy.

Where it curdles

The decay starts quietly. The room survives the crisis that created it – it worked, after all, and somebody senior liked the energy. Six months later the metrics have shifted: meetings held, actions closed, average containment time. All of them measure the fire brigade. None of them measure the arsonist. Meanwhile first-pass yield on the line has not moved, and nobody notices, because the room's dashboard has become the dashboard.

Then the cost cuts arrive, and they always arrive in the same order. Supplier development gets trimmed. PFMEA reviews slip a quarter. Maintenance windows are deferred, training hours get "optimised". Prevention has no photographs – nobody films the recall that did not happen – while a war room delivers daily theatre of visible action. So prevention loses the budget fight and the room keeps its screens. Leading quality initiatives across a 2,000+ workforce, multi-site, I watched the pattern repeat too often to call it coincidence: every site that kept a permanent war room also kept its failure modes. The room had become the process.

A standing war room is not a muscle – it is a scar that learned to book meetings.

How to make the room redundant

The exit exists, and it is not a better room. It is response moved onto the line itself.

  • QRQC at the source. Solve where the defect occurs, in hours, with the operator who found it standing next to the engineer. Give the line stop authority that is real – an operator who hesitates before pulling the cord has already been taught that stopping costs more than passing the defect on.
  • A3 to remove the fuel. Every recurrence that reaches the war room is a problem statement somebody declined to finish. The A3 exists so the same defect never earns a second meeting.
  • Q-Wall discipline. At Witte Automotive the Q-Wall was our visible room, and it performed – QRQC plus A3 plus that wall delivered a substantial failure-cost reduction. But the wall's real job was to push every response downward until the wall itself ran thin. A Q-Wall that stays busy is not a success. It is a confession.

None of this is exotic. It is the oldest arithmetic in the trade: a defect stopped at the station costs cents, and the same defect in a customer sorting bin costs air freight.

At SNOP I built the greenfield quality function for a 900+ employee plant and made a deliberate choice on day one: the QRQC cadence lived on the line, not in a conference room. We ran a full quarter with zero critical customer escalations, and the 70% defect-cost reduction that followed came from fixing the fuel – design, process, supplier – not from fighting fires faster. Faster firefighting has a ceiling. Fuel removal compounds.

Key takeaways

  • A war room compresses containment brilliantly and changes nothing about why containment was needed. Treat it as scaffolding, not architecture.
  • When the room's metrics – meetings held, actions closed – replace the process metrics – first-pass yield, escape rate – the room has started lying to you politely.
  • Cost cuts hit prevention first because prevention photographs badly. Defend PFMEA reviews, supplier development and training hours by name, with money attached.
  • The only healthy trajectory for an escalation room is a shrinking caseload. A steady-state agenda means your process is manufacturing the calendar.

The best escalation room I ever ran scheduled its own dismantling. We set the exit condition on day one – recurrence at the source below threshold for three consecutive months – and when we hit it, the room ended and the cadence moved to the line, where it belonged. That is the only version of success I trust. A war room with a standing calendar invite and a birthday is not a quality system; it is a subscription to your own defects, paid monthly in firefighting hours. Stop admiring the brigade. Go count the fireplaces.