The ceiling is a census

An account climbs for five months, and then it stops.

Same offer, same creative cadence, same people running it. The cost of a customer starts drifting upward. Nothing was broken and nothing has been fixed.

I used to read that flat line as a fault, so I would go looking for it. The bidding. The audience settings. The landing page. Creative that had gone stale. A pixel firing twice. There is usually something wrong. In an advertising account there is always something wrong, which is what makes the search so satisfying and so misleading.

What changed my mind was having a lot of small accounts in view at the same time. The same shape, over and over, across categories, run by capable people. A climb, a shoulder, then a line that no amount of rebuilding moved.

The flat line was not a fault. It was a count.

In any city there is a finite number of people who want that particular thing, at that particular price, and who can be reached for less than they are worth. The number is smaller than the dashboard implies, because the dashboard reports a potential audience for a whole country while the business serves one delivery radius, in one language, at one price point. An account that has plateaued has usually met its number and started paying to speak to the same people a second time.

That is harder to say in a review than a fault is. A fault has a fix. A count does not.

There are two ways past a count. A different group of people, which means a new proposition, new creative and a stretch of months where the numbers look worse before they look better. Or a different thing to sell to the people who already said yes.

Both are slow, both cost real money, and neither is what gets asked for. What gets asked for is more budget, which buys the same people again at a higher price, and then the market gets blamed.