The spec didn't change. The material did.

The aluminium industry has been lobbying hard for higher import duties on low-grade scrap. Their argument is straightforward — inferior feedstock degrades the integrity of the final product, and tariffs would force the market to clean up its act. They are protecting their product. Fair enough. I am more interested in what is already sitting in your incoming queue right now, with a certificate that says compliant.

Building the greenfield QA/QC function at SNOP — 900+ employees, IATF 16949, starting from bare concrete — I learned something no textbook covers. The most dangerous defect is not the one your control plan was designed to catch. It is the one that arrives wearing a valid certificate, produced by a process nobody told you had changed.

When your supplier's scrap pool changes, your PPAP is a historical document

Your PPAP was validated against a specific material, from a specific supplier, using a specific process. It documents a moment in time. When that supplier quietly shifts their scrap sources — domestic clean scrap swapped for cheaper imported low-grade, maybe their own supplier base consolidated — the chemistry has not necessarily moved outside the spec window. But the distribution within that window has. The mean shifted. The variance widened.

The Cpk values you submitted with confidence were calculated on a population that no longer exists. Your PFMEA identified failure modes based on material properties describing the old lot, not the one currently on your dock. And your control plan — the document telling operators what to check and how often — was written for a world where incoming material behaved the way it did during your initial capability study.

This is where quality and sourcing talk past each other. A spec is a range, not a point. Material anywhere inside that range is technically compliant. But material near the upper tolerance limit behaves differently in your forming process than material near the lower limit. If your supplier's scrap quality drifts toward the edges — more silicon variability in an aluminium alloy, more tramp elements from mixed scrap streams — your dies wear differently. Cycle times shift. Scrap rate climbs, and nobody can pinpoint why because the material is in spec.

A material cert verifies the chemistry of a sample. It says nothing about the consistency of the lot.

A material certificate proves that a sample, taken at some point, from some part of a production lot, met a specification. Not a statistical guarantee about the entire lot. Not proof that the material at the bottom of the pallet matches the top. And not assurance that the next lot from the same supplier will behave identically in your process.

A certificate is a photograph of a moment. Your process lives in the video.

At SNOP, 70% of our defect-cost reduction came not from adding inspectors — we did not add a single one — but from tightening supplier material discipline. Lot-level traceability, non-negotiable. We correlated incoming material data against in-process defect rates, and when we found correlation, we went back to the supplier with hard data, not complaints. A supplier will argue with an opinion. They cannot argue with a scatter plot that shows their material drifting in lockstep with your scrap rate.

At Airbus, working under AS9100, the stakes scale differently but the failure mode is the same beast. A supplier changing scrap sources without notification is an undocumented 4M change — Material, one of the four pillars. In aerospace, that is a reportable event, and it should be. The downstream consequence of a latent material defect is not a cosmetic blemish. It is a structural failure at altitude.

Catching material-driven variation before it becomes a field failure

The most effective tool is not a more elaborate inspection regime. It is QRQC — Quick Response Quality Control — and the reason it works is the discipline it imposes on time. Response within hours, not a report within weeks. When you see statistical drift on a critical characteristic, the question is not whether to investigate. The question is what changed in the material stream, and when.

That speed matters because material drift is patient. It does not announce itself with a sudden spike in nonconformances. It shows up as slow erosion — your Cpk slides from 1.67 to 1.33 over a quarter, and by the time someone flags it, you have shipped product built on assumptions that quietly stopped holding three months ago.

What actually works in practice: incoming verification that tests for the characteristics your process is sensitive to, not just the ones the spec enumerates. Lot traceability granular enough to connect field returns to specific material batches. And a written, enforced, audited protocol that treats any new scrap source from a supplier as an engineering change requiring re-validation of the affected PFMEA and capability studies.

Most organisations stumble on that last point. They treat sourcing decisions as commercial events. They are also engineering events. Every time a supplier changes their input material to save cost, they have changed your process. The fact that they did not tell you does not make it less of a change. It makes it an undocumented one.

Sourcing saved €0.03 per kilo by approving the cheaper scrap stream. Quality inherited a process change nobody filed. The savings show up in one department's quarterly review. The warranty claim shows up in another department's inbox, months later, with a customer's name attached. That structural disconnect — between the function that changes the input and the function that owns the output — is where the real cost lives. Fix the handshake between sourcing and quality, or keep paying for the gap between them. Every kilo, every lot, every quarter.