Here is an uncomfortable observation from two decades in automotive and aerospace quality: a greenfield site is the only factory on earth with no inherited failures, which means every escape it will ever ship is self-inflicted. Hanwha Aerospace has broken ground on a $2.2 billion munitions campus in Arkansas, and the renders are what you would expect – robots, dashboards, ribbon-cutting optimism. What no press release tells you is that the decision determining whether that campus produces has already been taken. Or quietly deferred. It is not in the machine orders. It is whether quality architecture gets engineered now, while the plant exists only as drawings, or gets scheduled for year two at full price.
The cheapest engineering you will ever buy
Legacy plants can at least blame history. Stellantis is reportedly running fifty war rooms to chase its quality problems, and I have sympathy, having run such rooms myself – but the greenfield question is why a new site should ever need one. The weeks before start of production are the only period in a plant's life when a change costs €0. Pour the machine foundations and the same decision becomes a €500k change order plus a schedule slip, at which point everyone develops a sudden interest in risk acceptance.
I learned this building the quality function for SNOP's greenfield plant: an empty building, 900+ people to hire, everything to decide exactly once. I sat in layout reviews while the plant was still CAD. That seat was worth more than any inspection budget I ever controlled. Containment space next to the line, not in a remote room nobody walks to. Escape paths for suspect material that never shared a corridor with good parts. Error-proofing designed into the cell, because a poka-yoke fixture specified at concept is hardware, while the same intention after installation is a work instruction with a laminated cover. Suppliers were qualified in sequence with their purchase orders – PPAP milestones written into the PO, VDA 6.3 process audits before first delivery, not after first rejection. Suzuki is now asking its suppliers to reserve a day a week for maintenance; a greenfield writes that discipline into the contract before anyone needs it. And a greenfield is every supplier's newest customer. Nobody knows where your weak points are. Including you.
A legacy plant inherits its failures; a greenfield plant elects them.
The hiring order is the quality strategy
Most organisations hire quality at ramp, and they hire police: inspectors to catch escapes, a manager to sign whatever the customer demands, an organisation built to react. The cheaper sequence is the reverse. Quality engineers and first-line leaders in place before volume, so their first QRQC sessions and A3s run on our own startup scrap – where a bad decision costs a bin of parts, not a customer escape. We rehearsed escalation before we rehearsed anything else: who stops the line, who calls whom, what gets communicated in the first hour. By the time volume arrived, problem-solving was muscle memory. The numbers people ask me about – a 70% reduction in defect costs, 98% customer satisfaction, a quarter with zero critical customer escalations – were decided by that hiring order, not by the year they appeared on a chart.
What cannot be retrofitted when the product is ordnance
A munitions campus raises the stakes past commerce. In automotive, an escape costs money and meetings. In defence production it can cost lives, and the traceability that answers for it cannot be reconstructed after the fact. Lot traceability from part one. First-article discipline on every process change, not just the ones the customer watches. My AS9100 and EN 9100 work taught me that the audit certificate is the receipt, not the system – it documents what you built; it builds nothing. The system is lot integrity, first-article inspection and process control, executed daily by people hired early enough to practise it. VDA 6.3 supplies the companion habit: audit the process before it ships, not the paperwork afterwards.
A few million decides whether $2.2 billion produces
Nobody models the retrofit economics. In my consulting work I have watched the counterfactual play out repeatedly: containment space value-engineered out of a layout and rebuilt at six figures; error-proofing downgraded to instructions, then reinstated after a scare; supplier audits conducted after the first truck of bad parts instead of before the purchase order. What a greenfield gets for free, a running plant buys back at retail. That is the most expensive habit in manufacturing.
Key takeaways
- Layout is a quality decision. Containment space, escape paths and in-cell error-proofing cost €0 in CAD and about €500k once foundations are poured. Fight for them before the machines are ordered.
- Hire architects before police. Quality engineers and first-line leaders in place before volume, rehearsing QRQC and escalation on internal scrap instead of discovering both on a customer escape.
- Sequence suppliers with the purchase orders. PPAP milestones in the PO and process audits before first delivery – a greenfield has no supplier history to lean on.
- In safety-critical work, traceability starts at part one. Lot genealogy and first-article discipline cannot be backfilled; the AS9100 certificate is the receipt for the system, not the system.
To plant managers standing up new sites: the machines will arrive whether you are ready or not, and the ramp will consume every hour you surrender to it. Spend the paper weeks on quality architecture – it is the only engineering you will ever get at €0. And to recruiters learning to read greenfield CVs: certifications are common. The rarest line is the one saying a candidate built a quality function from an empty building and can tell you exactly what existed before the first machine was ordered. Ask that question. The answer tells you whether the $2.2 billion produces – or merely occupies ground in Arkansas.