Debrief within the week or not at all
Memory of specifics decays fast and the team disperses.
Almost everything worth learning from an event is known only to the people who were there, and it is specific: this queue formed here at this time, this supplier arrived late, this radio channel was unusable. Within a fortnight most of it is gone and the team has moved on.
A debrief in the week following, with the people who actually ran things rather than only the leads, captures it. It has to include the things that went wrong, which requires the meeting to be safe enough that people will say them.
The output should be specific enough to act on next time: not that communication could improve, but that the two radio channels overlapped and one should be assigned differently.
It should be written and stored somewhere the next event's team will find it, which is a genuine obstacle for organisations running events irregularly. A single document per event, in a known place, is a low bar that many organisations do not clear.
The other reason to debrief with the people who ran things is that the leads have a distorted view. Anybody who spent the day in a control position saw the event through radio traffic and reports, and the person on the gate at four o'clock saw something the control room never heard about.