Three days are not data

The month began this morning across most of the region, depending on where the sighting was accepted, and by nine the dashboard has been opened more times than on any other day of the year.

Almost nothing in it is worth reading.

The morning is dead by design, so the day opens on a floor. The night that matters has not happened yet, and when it does, much of it lands after midnight and gets cut in half by a reporting day drawn for a market that sleeps at eleven. The year-on-year line points at an ordinary February Tuesday, because last year the month began on the first of March. And the countries on one sheet may not have begun on the same morning, so a regional total is adding a first day to a last ordinary one.

By noon somebody senior has asked what is being done about it.

That question is the whole risk of the month, and it is not a stupid one. It comes from a person who cannot see the shape of thirty nights yet, looking at the only number in front of them, in a company where saying nothing is never defensible.

So something gets done. Budget comes off a daypart on one dead morning. Creative gets swapped before a single table has seen it. Whatever the first week taught was a study of a reporting artefact.

The decision on day one is not what to change. It is who is allowed to say nothing yet, and whether that was agreed in writing before today.

The company that has given one named person the authority to declare the first three days not data has a better month than the one that reacts to them, and the whole cost is a sentence in a plan.