A minister cuts a ribbon, somebody says the word sovereignty, and the photographer gets the shot in front of the new line. Six weeks later the first container off that line fails incoming inspection on three characteristics, and two engineers spend a fortnight explaining to a customer why a part that was fine for a decade is now scrap. Nobody lied. Nobody cheated. The capability was never in the CAD file. It never is.

The trade press has been running two stories side by side: reshoring triumphs on one page, offshoring pressure on the next. Both treat location as the decision. Having run quality through launches and transfers across automotive and aerospace for two decades, I can tell you location is the easy half. Moving production home is not a logistics event. It is a requalification event, and it resets a decade of offshore process maturity to zero.

What travels, and what never does

The visible layer moves in weeks. Drawings, 3D models, CNC programs, tooling, gauges, the PPAP folder – a terabyte of PDFs and a few crates on a truck. What stays behind is everything that made the process capable:

  • dwell times nudged by two tenths of a second at a time, over years, by operators who never wrote any of it down;
  • the fixture someone reworked with a hand file and never redrew;
  • the press hand who hears the process go wrong before any gauge sees it;
  • a heat-treater two villages away whose certificates nobody has double-checked since 2014;
  • two hundred small fixes that exist in no document yet made the difference between scrap and shipment.

The PPAP folder records the process as it was qualified, not as it actually ran. The delta between the two is maturity. Maturity does not fit in a container.

Capability is the one part number that never appears on a packing list.

The requalification bill nobody budgets

A transfer is a launch wearing a logistics costume. New people, new presses, different humidity, different sub-suppliers – the failure modes changed with the postcode even though the drawing did not. So run it as phase-0 APQP or do not run it at all. New PFMEA, because the site changed the physics. Fresh MSA, because the gauges were recalibrated after transit. New capability studies, because last year's Cpk belongs to a line that no longer exists. Run-at-rate at contracted volume on the receiving line before the first shipment leaves.

Then put honest numbers beside the business case – the one that carried freight, cranes and duty, and nothing else. For a mid-complexity machined or moulded component, I budget 10–15% of annual purchase volume on top of freight and capex: PPAP requalification, scrap through the learning curve, duplicate gauging, containment, third-party audits. On €4m of annual spend that is €400k–€600k the spreadsheet never carried. The crane company's invoice is itemised. The capability reset arrives as a footnote.

How to move a process without breaking it

Start joint engineering months before the last shipment, not after it. Put the receiving site's engineers at the outgoing supplier's machines and run the final batches together, logging every deviation into the control plan while the people who hold the knowledge are still on payroll. Once that supplier's line stops, the knowledge starts leaving with them.

Then go and see. Sit at the machine. Film the operators' hands, pull the real cycle data off the controller, compare it with the standard – the actual dwell times will surface within an afternoon. Interview the old hands before they retire. That is how tribal knowledge becomes a controlled condition: a number, a tolerance, a check, written where the next site can hold it.

Audit the receiving line with VDA 6.3 before the first part ships. A process audit exists precisely to expose the gap between what the control plan claims and what the new line does. My rule from transfers and launches: anything under 90% overall is a no-go for first shipment, and a weak support-process element is a warning, not a rounding error.

I learned the size of that invisible layer at SNOP, building a QA/QC department from bare concrete for a 900-plus-person greenfield plant. There was no tribal knowledge to lose, because none existed. Every capability figure we claimed had to be built from the first stroke of the press, and the 70% defect-cost reduction we delivered came from rebuilding problem by problem – never from importing anything. At Witte Automotive the failure-cost reduction we drove with QRQC and A3 obeyed the same law from the opposite direction: the process earned its capability back through structured problem-solving at the line, day after day, until it grew its own scar tissue. Capability is rebuilt. It is never shipped.

Key takeaways

  • Treat every transfer as a launch: phase-0 APQP, a new PFMEA, fresh MSA and capability studies, run-at-rate at contracted volume before first shipment.
  • Buy the invisible layer early – joint engineering and go-and-see at the outgoing site, months before the last shipment, turns tribal knowledge into controlled conditions.
  • Audit the receiving line with VDA 6.3 before cut-over; the gap between control plan and reality is where capability leaks out.
  • Budget requalification at 10–15% of annual part spend above freight and capex, or pay twice that later in containment, expedited freight and customer escalations.

Reshore for resilience – the supply-risk argument is real, and the sovereignty one is not silly. But fund the move as a quality programme with a production line attached, not the other way round. The firms and countries that grasp this will keep the repatriated work. The ones that treat it as a moving job meet the true price at the customer's dock, denominated in sorting, containment and lost trust. The part comes home in a container. The process stays behind until somebody rebuilds it – deliberately, measurably, with a budget line to prove it.