The ledger is one-sided. The supply chain isn't.

Here is an uncomfortable observation. At one plant I ran quality for, procurement re-sourced a stamped mounting bracket to a cheaper supplier and booked €118,000 in annual savings. Nine months later I sat in the monthly quality review doing the arithmetic in reverse: air freight on four emergency shipments, a sorting cell of twelve agency workers for eleven weeks, a re-PPAP after the supplier quietly moved the part to a sister tool. Containment had consumed just over €300,000 — more than two years of the saving, spent in nine months. Nobody lied. Every figure was real on the paper it was printed on. The cheapest supplier on the purchase order is rarely the cheapest in the P&L, and the gap lives in ledgers nobody reads.

Where the money actually goes

Deloitte's research on the hidden costs of poor supplier performance put a number on what every plant manager already knows: the damage almost never appears as "supplier quality" in anyone's chart of accounts. Across two decades of automotive and aerospace plants, I watched it land in the same five places.

Premium freight comes first. The charter, the expedited truck, the weekend delivery — logistics pays it, logistics gets blamed for it, and the supplier who caused it never sees the invoice. Then line-side sorting: agency workers at a bench behind the line, three shifts, sometimes for months, booked as indirect labour and buried in plant overhead. Containment adds third-party sorters, quarantine warehouses and blocked stock that ties up working capital while finance wonders why the inventory turns dropped.

The nastiest line is re-PPAP after silent change. A tool relocated, a resin substituted, a heat-treat cycle "optimised" without notification — the part still measures in spec until the day it doesn't, and then you requalify everything under fire. Underneath it all sits engineering firefighting: your best design and quality engineers on supplier calls instead of their own work. The most expensive people in the building doing the least value-adding work in it.

Each line sits in a different budget with a different owner. The logistics manager owns freight but not sorting. The controller owns indirect labour but cannot attribute it. Engineering absorbs the firefighting as overhead and calls it Tuesday. Nobody's bonus depends on adding it up, so nobody adds it up.

Why the scorecard can't see it

PPM counts defective parts per million. OTD counts delivery adherence. Both are part-level metrics aimed at a system-level problem. A supplier can run 12 PPM and still cost you six figures a year — every dimension in spec, the process behind them out of control. I watched this at scale while leading quality initiatives across a 2,000-plus workforce on multiple sites: the same supplier green on every dashboard, red in three cost centres, and no report in the company able to connect the two. The costs land in the wrong ledgers, so the total signal dies before it reaches anyone who can act on it.

The purchase order buys a part. The P&L buys a supplier.

That is the mechanism Deloitte is pointing at, and it is not a measurement problem in the narrow sense — we measure plenty. It is an ownership problem. Until someone owns the sum, the sum does not exist.

Audit the process, not the part

The trade press fills weekly with ranked lists of certified suppliers — IATF-stamped machining shops, audit-compliant factories — as though the certificate were the product. A certificate proves a supplier could pass an audit on the day the auditor visited. It says nothing about the Tuesday night shift that runs out of coolant.

A VDA 6.3 process audit tells you considerably more. You walk the line, not the conference room. You look at change management, reaction to a failed control plan, escalation rules, whether the person who finds a defect can stop the line. Silent tool moves and quiet substitutions reveal themselves there — in the discipline of the process, not the wall of frames in reception.

I have sat on both sides of this table. As the customer — at WITTE, then at SNOP — I learned that punishment does not work but joint problem-solving does. At WITTE, QRQC, A3 and a Q-Wall drove our failure costs down hard. At SNOP we cut defect costs by 70% and ran a full quarter without a single critical customer escalation, and none of it came from threats; it came from joint A3s with named owners on both sides and an escalation clock the supplier helped set. Later, through my consulting agency FOREAST, I worked the same problem from the outside with clients such as ArcelorMittal and Meopta, and learned what a customer's demand actually costs when it lands on a supplier's desk — which demands change behaviour, which merely generate paperwork. The two sets barely overlap.

Sometimes the answer is simply to pay for capability. A supplier quoting €0.38 more per part, with a capable process and real change discipline, is usually the cheap one. Procurement just needs a ledger that can prove it.

Do the unglamorous thing first. Ninety days, one consolidated cost-of-poor-quality number per supplier, built from invoices and time bookings you already have. Then force one sourcing decision onto that total. Then give the quality director the pen on the number procurement quotes. The ledger has been one-sided because we built it that way. The supply chain was never as cheap as the purchase order claimed — the difference was simply scattered across five budgets where nobody was paid to add it up. Count it once, honestly, and the sourcing argument takes care of itself.