The account learns who stayed

For years my advice to a smaller advertiser going into the summer was the same sentence. Keep something on. Do not go dark, keep the account warm, and September will not have to start from nothing.

I still think going dark is expensive. I had the reason half wrong, and the half I had wrong is the part that costs money.

An account left running is not resting. It is learning.

The schools are expected to close at the end of this month and the cities empty behind them. By July the households with the most to spend are in London or Cairo or on a coast somewhere, and the account carries on with the settings it was given in April, buying whoever is still here to answer it.

Three months is a long lesson. Every day the system is being told, correctly, that the person who responds in July is the customer. It has no way of knowing that person is the exception this month rather than the rule. It optimises towards them, because that is the only honest thing it can do with what it has been shown.

So September arrives. The budget goes back up, the creative gets refreshed, and the account is aimed with real confidence at somebody who is not coming back. The audiences it trusts, the placements it favours, the offer it has learned to lead with: all of it was earned in an empty city, against the people who could not leave it.

Warm was never the question. Warm only describes whether the account is switched on.

The question is who it has spent the summer learning, and an account left to decide that on its own will decide it with whoever was left behind.