Sample size of the advice

Every piece of advice about an advertising account arrives with a sample size attached, and nobody says what it is.

The platform has seen thousands. Thousands of budgets, thousands of creative refreshes, thousands of campaigns that stalled at the point yours has. When the rep says accounts like yours usually do better with a broader audience, that is not a guess. It is the pattern. What the platform has never seen is your business: the margin, the warehouse, the founder's tolerance for a bad week, the fact that the product sells to mothers in Jeddah and to nobody else. The advice is true on average and blind to you.

The agency has seen a dozen. A dozen clients in your category over a few years, close enough to know what a bad month feels like, far enough to have watched a few different answers work. The advice has texture, and it is also whatever worked for the last client, because a dozen is enough to form a habit and not enough to test one.

The in-house team has seen one. This account, every day, for years. They know why the number dipped that spring and what the founder said about it. They know everything about the account and nothing about whether it is normal. The advice is precise and cannot tell a problem from a pattern.

I have sat in all three seats and given all three kinds of advice, with the same confidence each time. The confidence was never the problem. The problem was that nobody in the room asked where the advice had been learned.

That is the question. Not whether it is right, but where it was seen. A thousand accounts, a dozen, or this one. Each seat is right about a different thing and wrong in a predictable direction, and the direction is the sample size.