The eleven sheets everyone was building to and nobody knew were dead
Every drawing on a modern job carries a little block of text called the status code. Understanding what these codes mean and what behaviours they should trigger is critical. A sheet shared for review and a sheet issued for construction are asking to be treated in two quite different ways and in the flow of a busy job the behaviour around them ends up much the same, because the drawing is there, it's the newest one anyone has and the work is waiting. One of those codes is an instruction. The other is a conversation and you can pour a lot of money into a conversation before anyone notices it's not ready to be acted upon.
The job I'm going to describe loosely was a design and build shed, steel frame with a big envelope package, the kind of project every contractor in the country has experience of. The drawings lived where drawings always live, which is to say in two places at once. The current set sat on the client's document platform, doing everything properly, revisions landing, transmittal emails pinging away. And the working set lived in the site folder, printed and pinned and saved to somebody's desktop, updated the way site folders have always been updated, by whoever happened to be least busy at the time.
A sharp engineer caught the first one by luck. A dimension felt wrong to him, he chased it back and the drawing he'd been given turned out to have been superseded for three weeks. Fair play to him, that's a good catch and on most jobs the story stops right there, the engineer gets a pat on the back and everyone carries on assuming that was the only one.
On this job somebody did the boring version instead and went through every sheet in the working set, title block by title block, against the register. Not the suspicious ones. All of them. Eleven drawings in that folder were at review status, shared for comment, never issued for construction and were being built from anyway. And there was a second finding tucked in behind that one which I'd argue matters just as much: another nine sheets where the status couldn't be made out at all. It's tempting to file those nine under fine. They weren't fine, they were unchecked. Nothing wrong and nothing checked are two very different facts that most filing systems will happily blur into one.
There's a money mechanism under all of this, because there always is. Work built to a superseded drawing ends one of two ways. If the current revision was properly issued to you and your folder simply never caught up, the rework is yours to pay for. If it was never properly issued, you may well be owed for the change instead. Either way the answer lives entirely in dates and records: what was issued, when, through what route and what your working set actually held on the day the work went in. Without the boring check, nobody even knows the question is there to be asked.
The discipline that comes out of this is gloriously dull. Once a week, a check of the working set against the current issue, title block by title block, with a separate list for the ones you couldn't tell, because a sheet you couldn't check belongs on a list, not in the pile marked fine. It's exactly the sort of job that gets done brilliantly for three weeks and then never again as busyness takes over.
How many of the sheets your crews are working to right now are at a status nobody has looked at? If the honest answer is that you don't know, there is either live risk to be managed or value to be extracted.
Details changed to protect the parties; the pattern is real.