Paying for the same start

An advertiser turns a campaign on, and by the eighth evening has changed something in it every day.

The budget nudged up on Sunday. An ad paused on Monday because there were no sales in it. The audience widened by three years on Tuesday. A new headline on Wednesday, because the old one looked tired. Every one of those decisions is defensible on its own, and most of them were taken by somebody being careful with their own money.

From the platform side I have watched what that sequence actually costs.

A campaign spends its first days working out who responds, and it does that at the worst price it will ever pay, because it knows nothing yet. A material edit sends it back to the start of that process. The evidence is not carried forward. The account does not resume. It begins.

So that campaign is not eight days old. It is eight first days in a row, and a first day is the most expensive day a campaign ever has.

The advertiser then reads the cost per result and concludes the platform does not work at this size. What they bought was the opening day, eight times, at the opening price.

In this region the person doing this is usually the owner. It is their money, in an account they can open on a phone between two other jobs, and leaving it alone for four days does not feel like discipline. It feels like negligence.

That is the real price of an edit. Not the change, which is usually small and often sensible, but the days of evidence it deletes on the way through.

Doing nothing was the cheapest thing available in that account all week, and it was the one thing nobody was willing to buy.