The slide had two flags on it. Vietnam on top, Cambodia underneath, a dotted line joining them, and the word resilient sitting in the footer. The steering committee nodded the way people nod at a finished thing. I asked to see the quality file behind supplier number two. One signature deep. A PPAP approved off a desk audit, a capability study run at a third of contracted volume, a VDA 6.3 that had been "in planning" for two consecutive quarters. That was the whole insurance policy. This week's crop of Vietnam+1 briefings, selling Cambodia as cheap diversification, took me straight back to that room.

Here is an uncomfortable observation. Every "+1" country is a greenfield quality system wearing a purchase order: full PPAP, VDA 6.3, capability at rate – usually on a compressed timeline, because the resilience deadline came from procurement, and it bears no relation to how long a quality system takes to stand up. A backup supplier that has never run at volume is not resilience. It is an untested hypothesis, and switch day is the live experiment.

a second source is not a second capability

I built a greenfield QA/QC department for a 900-plus-employee plant from an empty office and a forklift aisle. The business plan carried the building, the tooling, the headcount. What it did not carry – what no diversification deck I have seen since carries – is the man-months of the quality system itself: control plans for every part family, MSA on every gauge, calibration loops, inspector training until reaction plans lived in people's hands rather than in a binder, escalation paths wired back to the customer. The better part of two years passed before that system could hold rate without me standing in the middle of it. That number, expressed in man-months, is the true price of any "+1", and it appears nowhere in the sourcing spreadsheet.

Consulting inside large sourcing organisations through FOREAST – ArcelorMittal among them – showed me how the spreadsheet sees it: tonnage, tooling cost, freight delta, tariff exposure. All real. All countable. Qualification hours do not fit the format, so they default to zero, and the map gets its second flag.

Meanwhile the ground under the new flag is still being poured. Cambodia's audit infrastructure is young enough that factory heat stress is a live compliance topic, not history. Aerospace is pouring concrete in the same window – new plants rising in Texas and Bengaluru – and each building will rediscover the same arithmetic: capacity is concrete and capital, capability is man-months.

the backup that never runs at rate

Sample lots are theatre. Controlled material, the best operator on days, an engineer hovering over the gauge. Rate is different physics: tool wear at cycle counts the pilot run never reached, second-shift operators meeting the control plan for the first time that morning, feedstock from a sub-supplier nobody re-audited, poka-yoke that has never been provoked at takt. Cpk measured in a sandbox is not Cpk at contracted volume.

At Witte Automotive I lived through what compressed qualification produces. A supplier approved on a Friday pilot lot; the first volume order a month later was, functionally, a different part – different shift, tooling worn into its second personality, a coating thickness from a sub-supplier nobody had thought to re-verify. We caught it at our Q-Wall, ran QRQC on the line the same day, closed the A3 weeks later. The failure costs were contained; the lesson was not. The supplier had passed a stamp and failed a system.

And when switch day comes – a port closure, a border tax, a second wave – the backup that has idled for two years meets rate for the first time on your production schedule. Leading quality across a 2,000-plus workforce, multi-site, taught me the pattern plainly: the quarter we logged zero critical customer escalations was not luck. It was the quarter in which second sources had been exercised, at volume, on purpose, before anyone needed them.

You buy resilience in qualification hours, not in countries.

qualification is a system, not a stamp

PPAP is not a document. It is evidence that a system exists behind the part: a PFMEA connected to a living control plan, MSA that proves your gauges can see what you claim to measure, capability at rate, reaction plans a night shift can execute without phoning anyone, traceability that survives a 3 a.m. containment decision. An on-site VDA 6.3 – scored, with open points closed before the first production release – tells you whether any of that holds when the customer is not watching. A signature on a desk audit teaches you nothing except who owns a pen.

Compress the timeline and you do not skip the cost – you convert it. The €40,000 saved on a three-week qualification returns as €200,000 of sorting, containment, premium freight and a line-down week. I have watched that exchange rate more than once. It never improves. Procurement's deadline was real. Physics did not attend the meeting.

The resilience talk this week says collaboration and visibility matter. It is right. But visibility that has never watched the backup run at rate is a photograph of resilience, not the thing itself.

Key takeaways

  • Exercise the backup at contracted volume on a fixed rhythm – a quarterly run-at-rate lot is the cheapest insurance in sourcing. Price it against one week of containment before you call it expensive.
  • Refuse desk-audit PPAPs for critical parts: require an on-site VDA 6.3, scored, with open points closed before the purchase order goes live.
  • Carry the man-months in the business case – control plans, MSA, inspector training, escalation wiring – not just tooling, freight and tariff deltas.
  • Split volume deliberately, roughly 80/20, so the second line stays warm and drift surfaces while switching is still a choice rather than a rescue.

The steering committee's two-country map was not wrong to exist. It was wrong to be celebrated. Diversification is a decision to spend qualification hours – the only currency resilience accepts – and until those hours are booked, run and repeated, supplier number two is decoration on a slide. Run the experiment yourself, on your schedule, at your volume. Leave it to switch day, and switch day will run it for you, at the worst possible moment, with your customer already on the escalation call.