On a roster the constraint stops being how well you run an account and becomes which accounts you look at — attention goes to whoever emailed, not to whoever is losing money, and the accounts that quietly break are the ones nobody is complaining about.
One account is a craft problem. You know the offer, you recognise the numbers, and you notice when something is off because you were looking at it yesterday. Everything an agency learns about running Meta ads is learned in that mode.
Forty accounts is not that skill applied forty times. It is a different problem wearing the same clothes, and the thing that runs out is not expertise. It is the number of accounts a person can hold in their head in a week.
The Monday tab-through
The default method is to open them one at a time and look. Forty accounts at five minutes each is three and a half hours, and five minutes is not enough to see anything subtle — it is enough to check nothing is on fire.
So in practice nobody does forty. They do the six they were already thinking about, plus whoever emailed on Friday. The other thirty-four are marked as checked because nothing about them came up.
An account with no complaints is not an account with no problems. It is an account with no one watching.
Attention follows noise, not cost
This is the structural failure and it has nothing to do with competence. Attention on a roster is allocated by who asks for it. The client who calls weekly gets weekly attention whether or not their account needs it. The client who trusts you and never calls gets looked at when something goes badly enough wrong that they finally do.
That ordering is close to backwards. The quiet accounts are quiet because the client is not watching either, which is exactly the condition under which a campaign can spend for three weeks against a form that stopped submitting.
What actually gets missed
It is rarely dramatic. Nobody misses an account that stopped delivering entirely — that produces a phone call. What gets missed is the slow kind:
- Spend continuing on a campaign that has not produced a result in a week, on an account nobody opened.
- Cost per lead drifting up a little each week, where no single week looks like anything.
- Budget quietly concentrating on one placement because that is where the auction was cheapest.
- An ad set delivering well outside the area it was targeted at, which looks like healthy volume.
- Frequency climbing past the point where the same people have stopped responding.
Every one of those is obvious when you are looking at that account. None of them is obvious when you are looking at thirty-nine others.
Why another dashboard does not fix it
The usual answer is a reporting layer that puts every account on one screen. It helps with the tab-through and not with the actual problem, because a dashboard still requires somebody to look at it and to know what normal looks like for each of forty different businesses.
That last part is the hard bit. A number that is alarming for a gym is a good week for a property developer. A grid of forty accounts with one colour scale across it is a grid that is wrong about most of them — which is the same mistake an industry benchmark makes, drawn as a heat map.
What actually changes the shape of the week
The only thing that helps is inverting the default. Instead of looking at every account to find the few that need you, have something look at every account on the same schedule and hand you the few — judged against each client’s own history rather than one shared line, so the shortlist is short enough to be real.
Four accounts out of forty needing a person today is a morning. Forty accounts needing a check is a week you will not finish, and the accounts you drop will not be the least important ones. They will be the quietest.