Capacity is procured. Competence is grown.

Embraer outlines an expansion of its Chihuahua operations at a nearshoring summit. Hical Technologies opens 60,000 sq ft of aerospace floor in Bengaluru. Nelcast's board approves a ₹75 crore machining investment in Gudur. Mississippi stands up a composites training hub and collects a state designation with it. One week, four announcements, and every number in them – square footage, rupees, headcount, enrolment – is a purchase order. I have commissioned a quality organisation inside an empty building, so I can tell you what none of these releases will ever announce: the one thing a new plant actually runs short of cannot be procured. Square footage photographs well. Judgement does not, which may explain its absence from the announcements.

What the greenfield actually runs short of

I built the QA/QC department of a greenfield plant from an empty shell. The site grew past 900 people, hired into a building that had been a field the first time I walked it. Inspectors were the easy part. We wrote the control plan, built the training blocks, ran capability studies on every gauge, and the gauges and CMM programmes arrived on trucks with lead times we could plot on a Gantt chart. I had a functioning inspection operation inside a quarter.

The constraint appeared in week three of launch, when three deviations landed inside one shift: a burr sitting at the top of print tolerance on a stamped flange, torque drifting toward the low limit on a weld nut, an oil film on a batch that should have arrived dry. My inspectors escalated all three at equal volume, because to them the three were equal. To the customer they were three different animals – one cosmetic, one a functional trend, one a traceability exposure. The gap between those two readings was not knowledge. They had the knowledge; the certificates said so. The gap was calibration – knowing which of three simultaneous signals earns a 02:00 phone call and which waits for the morning report. You can hire knowledge in a hiring season. Calibration you can only grow, in the person standing in front of the problem.

Judgement compounds by launch cycle, not by course

My own certificates – IATF 16949, AS9100, VDA 6.3, Six Sigma – sit in a drawer, and I have taught enough of that material to say plainly what a course delivers: vocabulary. What a deviation is called, which clause governs it, what a capability index claims. Necessary. Nowhere near sufficient.

Calibration arrives by launch cycle, in instalments. The first launch gives the vocabulary a face – you learn what a deviation looks like at 06:10 on a Monday. The second gives it a price: the €40,000 containment month, the customer line-stop your trend quietly became. The third gives it foresight, which is knowing which early whisper turns into a recall letter and which dies at the andon. Five years, three launch cycles, and you have an engineer you can leave alone with a ship/no-ship call. The industry keeps trying to compress this with classroom hours, and classroom hours are not the currency. Look back at those four announcements: each will be tracked against schedule, none against learning. Capacity gets a Gantt chart. Maturity gets nothing – until the day it is missing.

Certificates teach vocabulary. Launch cycles teach you which word to say at two in the morning.

Treat competence as a lead time, because it is one

Once you accept that competence has a lead time, run it like one. The plants that grow engineers fastest do not buy more training hours. They run more judgement repetitions under load.

QRQC is the cheapest competence machinery most plants own, and most of them waste it. We ran it at the line every morning and rotated the chair. A chair gets ninety seconds, three simultaneous facts and a decision about what matters – a judgement gym with zero capital cost. A3 in the same spirit: at WITTE Automotive, the A3 and Q-Wall work delivered substantial failure-cost reduction, and half the value was never the defect work itself. It was engineers learning to argue a chain of causation out loud until the argument carried weight.

Rotate 8D authorship. When the same two veterans write every report, you have a knowledge hoard; when authorship rotates with a senior coach behind it, you have a distribution system. Stage escalation drills once a quarter – three simultaneous deviations, one red-stamp scenario, timed triage. Then put the output on the plant scorecard, next to OEE and PPM:

• Time to first solo: weeks from hire to a ship/no-ship decision that survives audit;
• Time to first chair: weeks to a chaired QRQC;
• 8Ds closed as author, not as contributor;
• Escalation calls overturned on review – the honest one, because it measures calibration directly.

Across a 2,000-plus multi-site quality organisation I could tell within one conversation which plants had grown their engineers and which had merely staffed them. The staffed plants escalated everything and resolved little. The grown ones escalated precisely and de-escalated cleanly. Every serious de-escalation I have run – including a quarter closed with zero critical customer escalations – ended the same way: not with added headcount, but with one person whose judgement the customer had learned to trust, walking the floor and separating the fatal from the cosmetic.

The growth is real and welcome – nearshoring in Mexico, a deepening Indian supply base, composites capability moving to where the work is. The Mississippi hub is the closest of the four announcements to the true constraint, and even it counts its ambition in enrolment figures. So celebrate the floor space. Then open the business plan and give judgement its own line: a capital item with a decade of lead time and no expediting option. You can pour a floor in a season. The judgement that stands on it takes ten years to cure.