Congratulations arrived last week: TIGHITCO, an aerospace structures and systems supplier, has been certified at the "Excellent" level through Caterpillar's Supplier Excellence Program. Somewhere in Indiana or South Carolina a plaque is going up on a wall. I have stood on the receiving end of those plaques and on the buying side of them, and here is an uncomfortable observation from twenty years of both: the suppliers with the best certificates are often the slowest to raise a hand when something has already gone wrong. Not because they are dishonest — because the certificate measured the wrong thing. It sampled their system on a good day. It never measured their clock on a bad one.

What an excellence audit actually samples

I have sat on the auditor's side of the table, running process audits in the VDA 6.3 tradition. Two days, sometimes three. A sampling plan through project management, supplier management, process control, corrective action. You interview people, you walk the line, you trace a part number from the control plan back to the PFMEA and forward to the rework log. Done honestly, it is one of the best snapshots our industry has invented. Done at "Excellent" level, it certifies that the management system was mature and functioning — on Tuesday.

What it cannot do is watch the plant at 02:40 when a press drifts out of tolerance mid-shift and the operator is not sure whether the last ninety parts were good. An audit samples a system in still frames. A plant lives in continuous motion. The gap between the two is where supplier risk actually lives, and no rating scheme on the market prices it in. Leading quality across a 2,000+ multi-site workforce taught me this: within ten minutes of a site visit I could tell which plant would call me early and which one would "handle it" until the customer's dock caught it. The certificates on both walls looked identical.

A certificate proves the system survived the audit. It says nothing about how fast the system confesses.

Reaction latency, the unmeasured variable

Call it what it is: reaction latency, the elapsed time from detection to containment to honest escalation. Three clocks inside one number. Detection — how long between the defect existing and someone knowing. Containment — how long between knowing and physically stopping the flow. Disclosure — how long between knowing and telling the customer the truth, before the customer's receiving dock does it for you.

The third clock is the one that ruins relationships, and the one no rating captures. I have run de-escalation work for years, including a quarter with zero critical customer escalations, and I will tell you plainly: that record was not earned by certificates. It was earned by QRQC reaction discipline — same-hour containment, same-shift root-cause huddles at the line, and a standing rule that the customer hears about a suspect batch from us, not from their inspector. The quiet, unrated supplier who calls you at 16:20 about a suspect lot already on the truck is worth three decorated ones who send you a beautifully formatted 8D ten days after your line stopped.

At SNOP, building a greenfield QA operation for a 900-person plant from the ground up, we cut defect costs by roughly 70%. The tempting narrative is better paperwork. The real mechanism was shorter response time — detection to containment compressed from days to hours, and disclosure compressed from "after the customer finds it" to "before". The cost of a defect is mostly a function of how far it travels before somebody stops lying to themselves about it.

Buy the clock, not the plaque

So what should a buyer actually contract for? Not the certificate — ask for that and you get it for free. Contract for the supplier's clock.

  • Contractual containment times. 24 hours from detection to verified containment, written into the purchase agreement with defined notification triggers. Make late disclosure a commercial event, not a cultural preference.
  • Escalation drills. Once a year, plant a defect scenario and watch the clock. I would trade a full VDA 6.3 audit for one honest drill any day — the drill tells you what the audit cannot.
  • Reward early self-reporting. Suppliers who disclose get lighter consequences than suppliers who get caught. Most scorecards punish both identically, which is how you teach silence.
  • Demand QRQC evidence over audit trophies. Reaction-time trend data, containment logs, three recent self-disclosures. If they have none, they are either perfect or quiet — and you know which is more likely.

None of this is anti-audit. Process audits remain the best structural x-ray we have, and I will keep running them. But an "Excellent" rating in the TIGHITCO mould is a door-opener, not a risk transfer. The buyer who treats it as risk cover has simply moved the risk onto their own receiving dock, with a delay attached.

Key takeaways

  • Excellence certifications sample a management system in snapshots; supplier risk lives in continuous time.
  • What decides whether a defect becomes a crisis is reaction latency — detection to containment to honest self-disclosure.
  • Contract the clock: 24-hour containment clauses, annual escalation drills, lighter consequences for self-reporting than for getting caught.
  • Ask suppliers for QRQC reaction-time records and recent voluntary disclosures before you ask for the next certificate.

Plaques open doors. Clocks keep plants alive. TIGHITCO may well deserve every letter of that "Excellent" rating — but the buyer downstream of them is not protected by it, and should stop pretending otherwise. Audit the seconds, not just the system, and you will find out who your suppliers actually are before the truck does it for you.