Every site acceptance test ends with a punch list. The difference between a two-week closure and a six-month standoff is decided by how the list is written on the last day of the SAT.
No site acceptance test ends with zero findings, and none should be expected to. The SAT exists to surface the gap between the factory-tested system and site reality — the sensor that needs repositioning, the sequence that needs a timing change under real load, the alarm text the operators want reworded. The punch list is the SAT working as designed. What matters is whether the list closes.
Lists fail to close for predictable reasons, and most are baked in on the last afternoon of the test. Items written as complaints rather than actions. No severity classification, so a cosmetic label and a safety interlock queue in the same lane. No owner per item. No agreed retest method, so even a fixed item cannot be declared fixed.
Our punch list format is deliberately rigid. Every item carries a category — A blocks production, B blocks final acceptance, C is cosmetic or documentation — plus an owner by name, a due date, and the verification method that will close it. Category A items block handover outright; B items permit conditional production with a defined closure window; C items ride the final invoice, not the start of production.
The classification argument is worth having on site, loudly, on the last day. It is the cheapest argument of the whole project. Every item left ambiguous becomes a slow email thread between people who were not in the room, and slow email threads are where acceptance certificates go to die.
Closure needs the same discipline as capture. Each fix is verified by the method written on the item — a retest, a measurement, a signed document — not by the sentence "this has been resolved". We schedule a single closure review rather than trickling confirmations, because a defined event forces both sides to prepare and finishes the list in one pass.
For plant engineers on the receiving side: refuse punch lists without categories and owners, and refuse closure claims without the agreed verification. Not because suppliers act in bad faith, but because ambiguity punishes both sides equally — and the list format on day one determines whether anyone is still talking about it in month six.
— GANI Lifecycle engineering team