Scale, wait or kill turns a table of ad performance into a decision. This guide is for whoever manages budget at the ad level. It covers how the verdict is reached, why spend matters as much as cost per acquisition, and the most expensive mistake people make with it.
How the verdict works
Every ad is plotted on two axes: how much it has spent, and what it costs to acquire a customer.
Cost per acquisition alone is not enough to decide anything, because an ad that has spent very little has a cost per acquisition built on almost no evidence. An ad with one conversion from twenty dollars looks extraordinary and means nothing. Spend is what converts a ratio into a result you can trust.
So the verdict combines both. Ads with enough spend to be credible and a cost per acquisition below your target are scale candidates. Ads with enough spend and a cost above target are kill candidates. Ads without enough spend to judge are wait, regardless of how good or bad they currently look.
Wait is a real answer
Most people ignore the wait verdict, and it is the one doing the most work.
The instinct with a new ad showing a poor cost per acquisition is to switch it off quickly to limit the damage. The instinct with a new ad showing a good one is to raise its budget immediately. Both act on evidence too thin to support them, and both are how accounts end up churning through creative without learning anything.
An ad needs enough conversions for the number to be stable. Below that, you are reading variance. Leaving it alone until it has spent enough to be judged is usually the highest value thing you can do.
Why your target matters more than the tool
The verdict is only as good as the cost per acquisition target you set, and that target should come from your lifetime value curve rather than from your first order margin.
If a customer is worth substantially more by month six than their first order, you can afford to acquire them above first order break even, provided you can fund the gap. Setting the target at first order margin will kill ads that are profitable over a customer lifetime.
Setting it at twelve month value without the working capital to bridge the gap is the opposite error and is how profitable businesses run out of cash.
Scaling without breaking it
An ad that earns a scale verdict will not hold that performance at any budget. Raising spend widens the audience, and the additional people reached are by definition less likely to convert than the ones the algorithm found first.
Raise gradually and watch what happens to cost per acquisition rather than to total conversions. Total conversions will rise even while efficiency collapses, which is why a scaled campaign can look successful and lose money at the same time.
Watch frequency too. Rising frequency with falling performance means the audience is saturated and no budget change will fix it. The answer is new creative, not a different bid.
What to do with it
Work the kill list first, because that is money currently being wasted. Then the scale list, gradually. Leave the wait list alone. Then check the creative library for what the winners have in common, because the durable gain is producing more of what works rather than reallocating between what already exists.