The targeting controls that agencies used to compete on are mostly gone. Detailed interest stacks, lookalike ladders, elaborate audience exclusions: platforms have spent several years removing or automating them, and the accounts that cling to the old structures generally underperform the simple ones.
What is left as a lever is the creative. Not as a slogan, but literally: on paid social, what you show is now the main way you decide who sees it.
Why creative became the targeting
Delivery systems on Meta, TikTok and Snapchat all work roughly the same way now. You give them a broad pool and a conversion objective, and they find the people most likely to convert. The single strongest input they use to decide who those people are is who engages with the specific asset you uploaded.
A video that opens on a price appeals to a different person than one that opens on a problem, and the algorithm will deliver them to different pools without being told to. That is targeting. It just happens through the creative rather than through a checkbox.
Which means the practical question is no longer "who should we target" but "how many genuinely different things can we put in front of the system this month".
Velocity is not volume
The common mistake is to interpret this as needing more assets, and to produce fifteen versions of the same idea with different colour grades. That is volume. It teaches the algorithm nothing, because every variant appeals to the same person.
Velocity is the rate at which you test genuinely different concepts. It helps to separate the two explicitly:
- Concept. The angle. What problem it opens on, who it speaks to, what it promises. Changing the concept changes who responds.
- Variant. The execution of a concept. Different hook line, different first frame, different edit length. Changing the variant changes how efficiently that group responds.
You need both, and in that order. Concepts find pockets of demand. Variants squeeze them. Testing variants of a concept that never worked is the most common way to waste a production budget.
A schedule the account cannot outrun
Most accounts do not fail at creative because the ideas are bad. They fail because production is slower than fatigue, so there is a fortnight every month where the account is running on assets everyone already knows are tired.
The fix is boring: put creative on a fixed cadence and treat it as a supply chain rather than a project.
- Weekly brief. Written from last week's data: which hooks held attention, which search terms and comments revealed an objection, which concepts are decaying.
- Batch production. Six or more assets per drop, covering at least two distinct concepts. Batches are cheaper per asset and remove the temptation to skip a week.
- Ship on the same day every week. Predictability is what stops the gap appearing.
- Retire deliberately. Decide in advance what gets turned off when the new batch lands.
Knowing when something is dead
Fatigue is usually diagnosed too late, because people watch the campaign average rather than the asset. Signals worth watching per asset:
- Hook retention falling while the rest of the video holds. The audience has seen the opening before.
- Click-through decaying against a stable conversion rate. Fewer people are interested, but the ones who click still buy.
- Frequency climbing without a matching lift in results.
- Cost per result drifting up while a fresh asset in the same campaign holds steady. That comparison is the clean one, because it controls for auction conditions.
Track fatigue per hook, not per campaign. A campaign average hides one exhausted asset that is quietly absorbing most of the budget.
The short version
Clever targeting is mostly a memory. The accounts that win now are the ones that reliably put something new and genuinely different in front of the algorithm every week, and turn off what is finished without arguing about it.
Want this applied to your own accounts? Book an audit; the first read costs nothing.