Marketing practice, written small
Practice 06
Paid distribution buys one thing: the arrival of a reader who would not otherwise have arrived, at a moment chosen by the buyer rather than by the reader. That is a genuine and substantial thing to be able to buy, and it is frequently asked to do work it cannot do.
It cannot buy a reason to stay, an argument that holds, or a category that the reader was not already in the market for. Where those are missing, additional spend increases the number of people who discover that they are missing.
The value of paid arrival is largely in its timing. A reader who arrives at a moment when the relevant question is live behaves differently from one who arrives when it is not, and most of the variance attributed to creative is variance in how well the placement selects for that moment.
A paid arrival lands on a unit, and that unit either continues the specific promise that produced the click or restarts the argument from the beginning. Restarting is the ordinary case and it wastes the most expensive part of the transaction. The repair is to read the promise and the destination as one unit, in that order, and check that the second sentence follows from the first.
Distribution costs rise for reasons entirely outside any single buyer's control, and it is tempting to read every increase as a failure of execution. Sometimes it is; often the placement is simply being bid for by more people. Distinguishing the two requires knowing whether the material's behaviour changed, which requires having read it at unit scale before the costs moved.
If a sequence does not work for readers who arrived on their own, paid arrivals will not make it work; they will make it fail faster and at a known price. Testing the sequence on the traffic that already exists is cheaper, slower and almost always the correct order of operations.