A supplier quote lands 35% below the incumbent. Procurement signs the purchase order in three weeks because the unit price looks irresistible and the ISO 9001 certificate is attached. By month two, your incoming inspection line is rejecting one in eight parts. No containment in the control plan. No PFMEA shared during qualification. No Cpk data anywhere in the package. BESS developers currently sourcing battery storage from Chinese manufacturers at aggressive price points are learning this in real time. The factory audit passed. The cost crisis is deferred.

The certificate cleared customs. It didn't clear your line.

An ISO 9001 certificate confirms that a documented quality management system exists somewhere on the premises. Someone filed the paperwork. That is what it confirms. Whether the laser weld parameter on the cell tab attachment was ever validated — the certificate is silent. Whether the electrode coating thickness runs at Cpk 1.67 or 0.85, whether the supplier's own SPC charts show a process drifting toward the tolerance limit with every die maintenance cycle — the certificate does not know and cannot tell you.

I have walked into this failure mode too many times. At WITTE Automotive, the QRQC and A3 escalations that consumed our engineering hours were rarely cases where a part was inspected incorrectly. The root cause was a process parameter that had never been validated: a clamp force nobody had verified, a heat-treat cycle that drifted with ambient temperature, a torque value pulled from a drawing but never confirmed against capability data on the actual production line. The gauge was fine. The process was broken. Nobody had checked.

Supplier qualification without process capability verification is not cost reduction. It is cost relocation, and your quality system pays the difference every month until someone has the courage to re-source.

What inherited process variation actually costs

Real numbers. At SNOP, where I built the quality function for a 900-employee greenfield plant from zero, we cut defect costs by 70%. Not by adding inspection — inspection is the most expensive possible response to poor capability — but by driving supplier process capability at source. When you stack inspection layers on an unvalidated process, you pay for the parts, you pay for the inspectors, you pay for the sort, and you still pay for the escapes that get through. Triple cost stacking with no upstream fix.

The cost stack when you inherit unverified variation:

  • Scrap at incoming: 12.5% rejection on parts you already paid to ship across an ocean, with no recovery mechanism in the commercial terms.
  • Rework labour: hours consumed fixing what the supplier's process should have controlled, pulled from your direct labour budget.
  • Line-down events: the assembly cell stops because the incoming buffer is contaminated and the next conforming part is two inspection cycles away.
  • Customer escapes: the defects your sampling plan missed, reaching the end customer, generating warranty claims and field returns — in aerospace, potential airworthiness concerns.

At Airbus, I measure this through Routing Verification KPIs. The 97% reduction in internal lead time came directly from eliminating the inspection-and-sort loops that existed for one reason: supplier processes were never validated to capability before serial delivery. Every one of those loops was a hidden tax on a purchase order that looked cheap.

VDA 6.3 tells you about the process. ISO 9001 tells you about the filing cabinet.

The procurement articles I have read this week treat factory audits as though a two-day production walk delivers a serial production guarantee. It does not. A factory tour confirms the building exists, the machines are painted the same colour, and the canteen is clean. Whether the process can hold tolerance across a 40,000-unit run — that is a different question entirely.

What answers it is a VDA 6.3 process audit. I have run these across multi-site supplier bases in automotive and aerospace. They work because they interrogate the process, not the paperwork. P6 questions do not ask whether a procedure exists. They ask whether the process parameter is defined, validated, monitored, and demonstrably capable. Whether the operator at the workstation has an instruction that matches what the PFMEA says. Whether the deviation rate correlates with a specific shift, a specific tool cavity, a specific material lot. They expose the gap between what the certificate promises and what the process delivers.

If I am qualifying a battery cell supplier today, the ISO 9001 certificate gets them through the door. The VDA 6.3 scorecard, the initial process studies, the Cpk data on critical-to-quality characteristics — that gets the purchase order signed. Everything else is hope dressed up as procurement strategy.

The total-cost model procurement should have run

Start with unit price — attractive. Add incoming inspection cost for 100% screening, because no capability data exists to justify sampling. Add sorting and rework labour. Add scrap at the demonstrated rejection rate. Add line-down risk weighted by probability and multiplied by the cost of lost production. Add warranty reserve for escapes. Add engineering hours for QRQC cycles, 8D investigations, and containment actions every time the supplier's process drifts. That total almost always exceeds the incumbent's unit price by 8–15%. I have seen it exceed the incumbent price by 30%.

The procurement savings were never real. They were borrowed from the quality budget, the operations budget, and the customer relationship, with interest compounded monthly.

Key takeaways

  • Demand Cpk data on critical characteristics as a non-negotiable gate in supplier qualification — no data, no PO.
  • Run a VDA 6.3 P6 process audit before serial commitment; an ISO 9001 certificate alone is not evidence of process capability.
  • Model total cost of poor quality — inspection, scrap, rework, line-down, escape risk — into the sourcing decision, not just ex-works unit price.
  • Drive capability improvement at the supplier source rather than adding inspection layers internally; every inspection loop is a cost admission that the process was never validated.

The purchase order is the easy part. The 18 months of QRQC cycles, containment actions, engineering investigations, and customer explanations that follow — that is where the actual price of that cheap quote gets paid. I have closed enough supplier crises across automotive and aerospace to recognise the pattern in the first incoming inspection report. Supplier qualification without process capability verification does not reduce cost. It relocates cost from procurement's quarterly savings target into your scrap rate, your warranty reserve, and eventually your customer's trust. The BESS story is not new. It is the same invoice, presented to a different industry.