Hlídat každý prošlý termín na jednom místě

Open the deadline register and switch it to overdue. It gathers the dates each module already tracks, correspondence response deadlines, corrective actions off non-conformance reports, open punch items, and groups them by module with the days late on each, so you read one list instead of opening three.

Proč: A project rarely slips in one obvious place. It slips a few days here and a week there, in registers nobody reads together, and by the time it is visible in the programme the reasons are two months old. One list turns a vague feeling that things are drifting into a count you can do something about this morning.

The overdue sweep runs across the same register and pushes a reminder to whoever owns each late item. Once an item has sat past the grace window it escalates to the project managers, once and only once, and the register shows you what it moved.

Proč: Very little slips because somebody decided to ignore it. It slips because nobody told them it was theirs. Escalating on a grace window rather than on day one is what keeps the escalation worth reading, because a manager copied into everything reads none of it.

Take the item sitting at the top of the overdue list and raise a task against it with a named owner and a date you actually believe. Work it until the underlying record closes, at which point the entry drops out of the register on its own.

Proč: A reminder tells somebody a date has passed. It does not tell them what to do next, and the ones that have gone badly overdue are usually the ones where that is not obvious. The register only shortens when the source record closes, so the task has to point at the real work rather than at the reminder about it.