Fifty war rooms is not a strategy. It is a diagnosis. Stellantis has said publicly that it is running around fifty of them at once to pull quality back under control, and that number admits more than any transformation slide ever will. Here is an uncomfortable observation from two decades in automotive and aerospace quality: a plant that needs one war room has a problem, and a company that needs fifty has a prevention layer that failed quietly, years ago, while everyone was busy hitting the plan. I have run these rooms. I have closed them. The closing is the hard half. The war rooms are not the scandal; the scandal is how early the alarms were silenced.

What a war room is actually for

A proper war room is a scalpel, and a scalpel is designed to be put down. At SNOP, where I built the quality function of a 900-plus-employee greenfield plant, the rapid-response discipline was QRQC, and it ran to a simple, unforgiving rhythm. Eight o'clock, at the defect, not in a conference room – the part in hand, the operator who found it standing next to the engineer who has to fix it. One owner, named on day one, in front of witnesses. Containment inside 24 hours: segregated stock, protected shipment, the customer informed before they discover it themselves. And an exit condition written when the room opens, not negotiated when people get tired.

The exit condition separates a crisis cell from a committee. Ours was blunt. Three consecutive weeks with zero escapes on the specific defect. A verified countermeasure folded back into the control plan and the layered audit. Then the meeting dies. That willingness to kill our own meeting is part of what took the plant to a 70% reduction in defect costs and a quarter with zero critical customer escalations. The rooms closed. That was the entire point of them.

Triage is not treatment

Containment is what you do while you decide; corrective action is the decision. A war room exists to hold the space between the two, measured in days, not quarters, because in that gap sorting, rework and expedited freight will happily burn a six-figure invoice in a single bad month. And the moment a war room starts "discovering" new failure modes week after week, it has stopped being a rapid-response instrument. It has become a live re-derivation of a PFMEA nobody maintained, performed one escaped defect at a time, at the highest possible price per lesson.

Containment without correction is a subscription, not a cure.

I saw the healthy version at Witte Automotive, where the fight on failure costs ran on QRQC, A3 and a Q-Wall – a physical wall where every open quality issue lived until it died. The wall worked because items came off it, visibly, with dates. The day a wall becomes a display of busyness rather than a countdown, you have built a monument. And failure modes that "surprise" a war room were almost always known; they sat in a launch-era PFMEA nobody revisited when volumes, suppliers or variants changed. The honest response to a room full of surprises is not more meetings. It is a re-baselined PFMEA and a rebuilt control plan.

When fifty open at once

Now do the arithmetic on fifty. Each room eats a quality engineer, a production lead, a technician and slices of their managers – call it four to six people. Multiply. Two to three hundred of your best problem-solvers, pulled out of the very work that prevents problems: updating control plans, running layered process audits, challenging capability at the PFMEA table. A war room consumes exactly the people who should be making the next one unnecessary. Worse, visible activity is seductive, and leadership starts to mistake it for control. A dashboard with fifty active rooms looks like grip. That is a fever chart.

The KPI inversion is the tell. Measure issues churned per room, and rooms will churn issues forever – the metric pays for its own perpetuation. The metric that matters is rooms closed per quarter, with zero recurrence verified afterwards. In de-escalation work I treat the step-down as designed, never drifted: contain fast, over-communicate for two hard weeks, then deliberately move the customer from daily contact to weekly to the standard management routine, with explicit agreement at each step. The crisis cell closes in front of the customer. You cannot leave it running as wallpaper – wallpaper is what a meeting becomes when nobody has the authority to delete it.

Key takeaways

  • Open every war room with its closure criterion already written: a named owner, containment inside 24 hours, a verified countermeasure, and the date the meeting dies on the wall from day one.
  • Measure rooms closed per quarter with zero recurrence – not issues processed per room. Throughput inside a war room is a cost, not an achievement.
  • If the room keeps "discovering" failure modes, stop firefighting and re-baseline the PFMEA and control plan; you are re-deriving a document that already exists.
  • Step the escalation cadence down deliberately – daily to weekly to the standard routine – with the customer's agreement at each step. A crisis cell that fades by drift becomes permanent furniture.

The best war room I ever ran made itself unnecessary in six weeks. The countermeasure held, the layered audits picked up the surveillance, the meeting was deleted and nobody mourned it. That is the whole craft. Build the cell temporarily, design its death on day one, and pour every lesson back into the boring instruments – PFMEA, control plan, layered audit – so the next failure dies at a table where it costs an hour instead of a quarter. An hour at the PFMEA table is the cheapest quality intervention in existence; fifty parallel rooms are what you pay for skipping it. And if your war room has a five-year lease, it is not a war room. It is a department with worse coffee.