The segment nobody can build

The most valuable audience for the month expected to begin around the eighteenth is the customer who bought during the last one, and in most businesses I have worked inside, nobody can produce that list.

Not for want of data. The purchase is in there, with a timestamp, in a system that can tell you what somebody ordered at nine in the evening two years ago.

The problem is the question. Every audience builder in every email and customer platform I have used offers the same shapes. Last thirty days. Last ninety. This month against the same month last year. The holiday period. All of them are windows cut into a fixed calendar. There is no field for a month that begins on a different date every year, and no way to ask for last Ramadan, because last Ramadan was in March, a few years back it was in April, and one of those years ran across two calendar months.

So it gets done by hand. Somebody looks up the dates for the last few years, writes out a range for each, joins them, and holds the answer in a static list that is out of date the moment a new order lands. Most years it does not get done, because it sits with the person who is also building the campaign, in the fortnight when everything is due at once.

What runs instead is the default. The customer who ordered on most of the last month's thirty nights gets picked up by the acquisition campaign as a cold prospect, and the welcome flow greets them as a stranger.

These tools were built where the peak has a fixed date on a printed calendar. The region's most predictable customer is invisible to them, every year, for exactly that reason.