The request that arrives with almost every stalled account is some version of "we need better audiences". The client has usually spent weeks in the targeting settings, stacking interests, testing lookalike percentages, building exclusion rules of real complexity.
The audience is almost never the problem. Meta and Google both decided years ago that they are better at finding the right person than you are, and on that narrow point they were right. What you still control is what the right person is shown.
Which means that once the structure is consolidated and the tracking is honest, creative is the only lever with any travel left in it. And most accounts are not short of creative. They are short of things to say.
Angles and executions are not the same thing
This distinction is the whole framework, and getting it wrong is why most creative testing produces no learning.
An angle is a reason to buy or an objection to overcome. "It arrives before the weekend." "It does not need replacing in two years." "You will stop having the same argument with your supplier."
An execution is a particular way of expressing one: a founder talking to camera, a split-screen comparison, a static with a bold claim, a customer filming their unboxing.
Most accounts test executions and believe they are testing angles. Six videos go out, all variations on "high quality, great price, free shipping", and when they fatigue the conclusion is that video is not working any more. It is not that the format died. The account only ever had one thing to say and ran out of ways to say it.
Build the angle inventory first
Before any production, write out every reason somebody buys from you and every reason they hesitate. Not what you wish they cared about — what they actually say.
Three places to get this honestly:
Sales calls. The objection that comes up every time is an angle. So is the sentence that makes people relax.
Reviews, yours and your competitors'. Read the three-star ones. People explain exactly what nearly stopped them, which is an objection you can answer in an ad before it forms.
Support tickets. What people misunderstand before buying is an angle, because the ad can remove that misunderstanding earlier and cheaper than your support team can.
A mature account should have fifteen to thirty angles written down. Most have two, both invented in a meeting.
How much to test, and how to judge it
A practical cadence for an account spending around $10,000 a month is three to five new concepts a week. Fewer and the account has nothing new to learn from. Far more and you cannot judge any of them properly — and judging is the point.
Run new creative in its own testing campaign so it has a fair trial without disturbing what is already working. Give each concept enough budget to reach a decision rather than spreading it so thin that everything dies inconclusive.
Then judge on the metric that matches the job. A hook is judged on whether people keep watching. A whole ad is judged on cost per purchase or cost per qualified lead. Click-through rate on its own has sold a great many bad ads to people who were measuring the wrong thing.
Most importantly: tag every ad with its angle and report by angle as well as by ad. Without that tag the data cannot answer the only question that matters when something fades.
Reading fatigue correctly
When an ad's performance drops, there are two possible explanations and they lead to opposite actions.
If the execution is tired, the message still works and the audience has simply seen this version too often. Rebuild the same angle in a different format and performance usually returns.
If the angle is tired, making another version is wasted money. You need a different reason to buy.
You can only tell them apart if results are grouped by angle. When every ad for an angle declines together, the angle is spent. When one ad declines while others carrying the same message hold up, it was just that execution.
Rising frequency alongside a falling conversion rate is the usual signature of execution fatigue in a finite audience. A broad audience that has stopped responding to one argument but still responds to another is angle fatigue.
When something wins
Two mistakes get made with winners, equally often.
The first is duplicating the winning ad set five times to "scale it". That splits the conversion data and makes every copy worse than the original was. Scale by raising budget in steps and letting it settle, not by cloning.
The second is treating the winner as finished. A winning angle is a direction, not a destination. Rebuild it in three other formats, aimed at slightly different people, and you will usually find the angle had more in it than the first execution captured.
The mistakes I see most
Judging on day two. Early numbers swing violently and mean very little. Decide in advance how long a test runs and how many conversions it needs, then honour that.
Changing six things at once. If a new ad has a different hook, a different format, a different offer and a different landing page, a win teaches you nothing you can repeat.
Letting the brand team own this alone. Beautiful work that says nothing specific loses to plain work that says something true. The people reading sales calls should be in the creative brief.
Testing creative on broken tracking. If reported purchases do not reconcile with your store, you are ranking creative on fiction. Three platforms will give you three different numbers, and that gets resolved before any of this is worth doing.
The short version
Targeting is largely automated and your competitors have the same automation. What you still own is the argument. Write down every reason people buy and every reason they hesitate, build ads for each, tag everything by angle, and judge the angle separately from the execution.
Accounts that do this keep finding room long after accounts that are still testing audiences have concluded the market is saturated.
This only works on top of a structure that lets data concentrate — the account architecture that makes scaling possible comes first. And if the account will not spend what you give it, delivery is the thing to fix before creative.
Creative strategy is part of our Meta Ads work. The ad account audit includes a read of your creative performance by angle, which is usually the first time anyone has looked at it that way.