This week's manufacturing press ran two stories that belong on the same page. One argued that India's factories must move from making to inventing. The other reported a chemicals plant collecting two platinum awards at the CII National Six Sigma competition. Here is an uncomfortable observation: most executives read the first headline and picture an R&D building – glass walls, patents, a logo over the entrance. In two decades of automotive and aerospace plants, the highest-return invention I ever witnessed happened at a PFMEA table and a Q-Wall. The two stories are the same story. The industry refuses to file process capability under invention, so it funds the building and starves the floor.

The invention that never gets a patent

The 70% defect-cost reduction we achieved at SNOP – a greenfield plant of 900-plus people, where I built the QA/QC department from an empty room – never appeared in any patent database. It was invention nonetheless. We reinvented how the plant detected failure and how it responded to it.

The mechanism was unglamorous. QRQC first: fifteen minutes, twice a day, at the line, with the part, the operator and the data – failure analysed within 24 hours where it occurred, not in a meeting room three days later. Then A3: one problem, one owner, one page, root cause verified by turning the failure off and on again rather than by consensus. Then the control-plan rebuild – reaction plans rewritten so the process shut itself down on first-signal drift instead of waiting for an escape to reach the customer. What did it cost? Hours, not capital. An hour a day per team leader. One A3 per engineer per month. A standing PFMEA review cadence. Nobody cuts a ribbon on a rebuilt control plan, which is exactly why boards underfund it.

Why making-only plants quietly bleed

Contract assembly looks like a business. Mostly, it is a lease. Build to someone else's print and you inherit their process assumptions: their tolerance stack-ups, their tooling history, their risk appetite, a PFMEA written for their factories, not yours. You own the costs and the escapes. They keep the know-how. Every defect you pay for is a lesson delivered to their engineers, not yours.

The consulting work through FOREAST proved the point twice, once with ArcelorMittal and once with Meopta. The savings never came from launching a product. They came from re-engineering a process – a designed experiment that located the real process window, an inspection sequence rebuilt around failure modes instead of tradition – and the payback ran to multiples of the fee. At Witte Automotive the Q-Wall worked for the same reason: it priced failure in front of everyone every morning, and a wall that shows what scrap costs you this week invents faster than a quarterly review ever will.

The same week's ransomware headlines make the point from another angle. Manufacturing is now the most-targeted sector – and I say that as a certified ethical hacker who reads the threat reports, not as a quality man guessing. A plant that inherited its process assumptions has also inherited failure modes it never mapped. Resilience, like margin, is a process-invention outcome.

Making to print earns the order. Reinventing the making earns the margin.

Fund the floor like a lab

If invention lives on the floor, then lead it the way you would lead an R&D function. Budget and protected time, for a start – improvement hours ring-fenced on the schedule with the same authority as maintenance windows, because anything that yields to production pressure was never a policy. A dojo next: a corner with a duplicate fixture where operators and engineers run designed experiments without touching takt. And QRQC as the daily invention engine – every containment is a hypothesis, every A3 a research paper written in scrap rate.

The greenfield build at SNOP remains my proof that you can design that culture from day one instead of importing it from a certificate. We hired quality engineers for curiosity. We trained them to interrogate process capability. We measured them on the movement of defect cost per unit, not on audits passed. An IATF 16949 certificate tells a customer you can document. Capability tells them you can learn. Plants that confuse the two stay clever on paper and slow in fact.

Key takeaways

  • Treat PFMEA, QRQC, A3 and DOE as R&D – budget them in hours, protect the hours, review them at the same table as the product roadmap.
  • If you assemble to someone else's print, process invention is your only route to owning know-how instead of renting it: variation studies, DOE on your own tooling, reaction plans rewritten for your realities.
  • Measure quality and process engineering by capability movement – defect cost per unit, Cpk, escape rate – not by audits passed or trophies collected.
  • Hire for invention: a CV line that says "made this process capable", with numbers attached, is worth more than one that says "launched this product".

A nation – or a plant – that treats making as execution is renting its future, and the rent compounds. The ones that fund process invention own their margins, because capability cannot be outsourced to the customer who wrote the drawing. So a note for recruiters, and for anyone assembling a manufacturing CV: the most valuable line on it is not a product someone launched. It is a process someone made capable. The contract rewards making. The margin rewards invention – and invention, done properly, clocks in at the PFMEA table at 06:30.