The pause is not free

Every June the same email arrives. The account is pausing until September.

It is a reasonable email. The customer has left, the number is falling, and spending into an empty city looks like burning money. The pause looks free, because the cost of not spending is, by definition, nothing.

From the platform side you get to see the second half of that decision. It arrives in September, and it has a price.

An account that has been dark for ten weeks does not restart where it stopped. The learning that priced its audiences has gone stale. The creative that worked in June is a picture of a season that has ended. So the account restarts cold, and in the one week when every other account in the region is restarting too, with the back-to-school money, the National Day money and the fourth-quarter money all landing at once. The auction is at its most crowded and the account at its least informed. It pays the highest price of the year to relearn what it already knew in June.

I have watched this from three seats, and the accounts that came back strongest in September had one thing in common. They never fully left.

Not at full budget. A retention layer for whoever is still ordering. A small always-on line that keeps the learning current. An email programme that costs attention and little else. Enough signal that the platform still knows who the customer is on the day the customer comes home.

The pause saves the July budget and spends it again in September, at a worse rate, with interest.

So the summer question was never whether to spend. It is how little it takes to keep an account warm, and the answer is usually a fraction of the budget. It is never none of it.