The shortage is temporary. The interim approval is permanent.

The Financial Times reports that the world's largest carmakers are scrambling to avert a motor oil crisis. The headline reads as a logistics story – tankers, allocations, a supply gap with a number attached. Here is an uncomfortable observation from two decades running quality in automotive and aerospace plants: shortages rarely hurt plants. Panic substitutions do. I have stood in a plant manager's office at 06:40 while he asked whether the alternative oil was "the same stuff". The answer is never yes. It is "show me the data". Within seventy-two hours of any shortage, procurement's inbox fills with equivalent alternatives, each one an uncontrolled change wearing a purchase order as a disguise. Factory-fill oil, stamping lubricants, drawing compounds, corrosion preventives: engineered products, all of them. "Compatible" is a supplier claim until your PFMEA says otherwise.

what a compatible oil actually changes

Motor oil is the visible case, but the mechanism generalises to every chemical that touches the line. Friction chemistry lives in the boundary layer: change the additive package and you change the friction coefficient, and friction is a process input, not a laboratory curiosity. In forming, a new drawing compound changes draw-in, press tonnage and burr height. Burr height on a cyclically loaded bracket changes crack initiation. Crack initiation changes field returns. The physics does not read the purchase order.

The same chain runs through tool wear, corrosion behaviour and cleanliness. A "drop-in" rust preventive can move salt-spray performance from 480 hours to 90 while the datasheet summary stays identical, because film weight is not a line suppliers volunteer. A cleaner with a different residue shifts e-coat adhesion, weld porosity, bond strength. Then the arithmetic. Regrinding a progressive die costs €15,000–20,000 plus a fortnight of press capacity. An engine warranty claim clears €4,000 a unit. A paint-adhesion campaign runs €80–120 per vehicle across a production year. The substitution, meanwhile, saved €0.02 a unit. No supplier datasheet in history has ever said "marginally worse".

interim approvals never had a sunset

Under commercial pressure the deviation system floods. I have watched an open-deviation log grow from eleven items to more than a hundred in six weeks, every one carrying a scaled-down validation signed at midnight by an engineer who needed the line running at six. IATF 16949 saw this coming: clause 8.5.6 asks for review, verification and – where the risk warrants it – validation, with records retained. It cannot ask for courage. When supply normalises, nobody re-runs the full PPAP. The interim material is working fine, engineering hours have moved to the next fire, and the interim approval quietly becomes the standard.

A deviation is signed at midnight and remembered by the bill of materials for a decade.

Nobody behaves badly in that story. That is what makes it dangerous. Every local decision is defensible; the accumulated result is a permanent product change nobody ever approved as a whole.

a substitution gate that runs at crisis speed

Build the substitution gate before the first crisis. You never get the budget afterwards. I learned that building a QA/QC department from the ground up for a 900-plus-employee greenfield plant, and the scaled PPAP-with-sunset process we built paid for itself in year two. A supposedly chemically identical forming oil, imported under shortage pressure, passed every datasheet comparison we could construct. Three weeks in, our daily QRQC board flagged tool-wear drift: the die-maintenance interval on a deep-draw station had fallen from 40,000 to 26,000 strokes, and burr height was walking toward the limit. An A3 traced the cause to film weight in the new oil inside a week. The material was pulled. Not one suspect part reached a customer, and we closed the quarter with zero critical customer escalations – not luck, but a gate that had made the substitution visible on day one.

That gate had four working parts:

• A 48-hour technical screen, driven by the PFMEA – friction coupons, tool-wear trials, salt-spray panels, gravimetric residue – with pass/fail limits owned by quality and stamped on the purchase order before it is raised.
• Scaled validation, not full PPAP – bench and wet tests aimed at the failure modes this chemistry touches.
• Hard sunset dates – typically 90 days, automatic reversion to the original specification, extension only through a full PPAP re-run with a named owner who cannot delegate it.
• Batch-level traceability – every substituted batch tied to date codes, serial ranges and shipments, so when field data whispers you know which shipments to listen to.

I have spent enough of my career on customer phones, de-escalating the fallout from ungated substitutions, to know how this ends. The ones you don't gate are the ones you negotiate eighteen months later – windowless room, 8D on the table, their line down and your signature on the deviation.

The tank refills. The BOM remembers. Supply will normalise, the spreadsheets will close, and every unexamined substitution will stay behind – in the product, in the tooling, in the warranty reserve. Quality's job in a crisis is not to slow substitution down; the line will run one way or another. The job is to make each substitution visible, bounded and reversible: visible in a register nobody can route around, bounded by tests aimed at the real physics, reversible because the traceability exists to pull it back cleanly. Do that, and the shortage stays what it always was – a logistics story. Skip it, and it becomes your warranty story, arriving on an eighteen-month lag.