Write down what you checked and found nothing
A number over budget got a proper investigation and turned out to have no fault behind it. Recording the dead end took ten minutes and saves every future look at that row from starting from scratch.
One of our own pages sits slightly over the speed target we hold ourselves to. It has done for weeks. A few days ago it got a full investigation: where the time goes, what the page weighs, whether the server or the network or the device is responsible.
The answer was that nothing is wrong. The page is small, the server answers quickly, the time is going somewhere ordinary. It is over the target and there is no defect behind it.
That is an unsatisfying result, and the temptation is to leave it unrecorded — nothing happened, so there is nothing to write down. We wrote it down anyway, with the figures, and it is the most useful ten minutes of that whole investigation.
Because the alert does not go away
The row is still on the board. It will be on the board tomorrow. Without a note, the next person to look at it does the same investigation, reaches the same dead end, and spends the same hour. Do that enough times and the outcome is not a fixed page — it is a team that has learned to scroll past that row, which means they also scroll past it on the day it changes.
Every venue has these. The walk-in that reads two degrees warm and has since it was installed. The card terminal that prints a duplicate receipt on the first sale of the day. The supplier whose invoices are always a week behind their delivery notes. Somebody looked into each of those once. If what they found is not written somewhere the next person will look, it was not an investigation, it was a private experience.
What a useful note contains
Not "checked, fine". That is indistinguishable from nobody having checked. Four things:
The date, because a finding about a system has a shelf life.
What you actually measured, in numbers. "The fridge reads 4.1 against a set point of 2, with the door shut and a full load" is a finding. "The fridge is fine" is a mood.
What would change your mind — the reading that would mean this has stopped being a dead end and become a fault. Ours is a change in the number of samples behind the row. A walk-in's might be five degrees, or a compressor that runs continuously.
Who to ask if it turns out to matter after all.
The instinct that a negative result is not worth recording is strong and it is wrong. Half of operating anything is knowing which of the things nagging at you have already been looked at.