The short answer
On Meta today the ad itself tells the system who it is for, so the opener and the awareness level work as your targeting. But literal filters underperform: on our board, ads that open with "If you..." reach the top tier at 3.2%, against about 10% overall. Winners target with a specific situation, a named role, or the level of knowledge they assume.
What the phrase means
"Creative is the new targeting" became a slogan after Meta moved delivery toward broad audiences and machine learning. The mechanics behind it are real. Meta's retrieval system (Andromeda) shortlists ads for each person, and its ranking models use a representation of the creative itself as an input, according to Meta's GEM engineering post. The system learns who responds to each ad and goes looking for more people like them.
So the audience you reach is decided less by the interests you tick and more by who stops on the ad. The opener filters. The awareness level filters. The person on screen filters.
The obvious conclusion is "call out your audience in the first line." Our data says that is only half right.
What 2,443 winners do with their first line
We classified the opening line of every ad in our October 9, 2026 snapshot (2,443 live ads, $50k+ in estimated spend each). Top tier means the top 10% by Ad Radar's winner score (live 21+ days), so about 10% is the baseline.
| How the opener selects its audience | Ads | Share reaching top tier | Median estimated spend |
|---|---|---|---|
| Conditional filter ("If you...", "When your...") | 126 | 3.2% | $235K |
| Names an age or gender ("women over 60", "guys") | 133 | 11.3% | $232K |
| Speaks about "your dog / your cat" | 57 | 10.5% | $343K |
| Names a role ("dog owners", "nurses", "cat parents") | 19 | 10.5% | $248K |
| Direct yes/no question ("Do you...", "Are you...") | 17 | 17.6% | $296K |
| Board average | 2,443 | 9.6% | $223K |
The conditional filter is the outlier. 126 winners open with "If you..." or "When your...", and only four reach the top tier. The result holds outside health too: 3.2% in health ads, 3.1% everywhere else, where the baseline is 11%.
The groups that name a role, a pet, an age or a gender sit around the baseline. The small group of direct questions does better, but 17 ads is too few to lean on.
Why would the conditional underperform? We can only guess. An "If you have X" line tells everyone without X to leave, and it often describes a condition rather than a person you can picture. A line that names a group or paints a scene lets the right people recognize themselves while staying interesting to the people next to them. That gives delivery more engagement to learn from.
Awareness level is targeting too
The second filter is how much the ad assumes the viewer knows. An unaware ad opens with a story or a surprising fact; a product-aware ad opens with the offer. They reach different people even with the same audience settings.
825
winners written for unaware readers (34%)
496
for solution-aware readers (20%)
11.5%
top-tier share for solution-aware ads
8.6%
top-tier share for unaware ads
Unaware ads cast the widest net and are the largest group on the board. Solution-aware ads speak to people already comparing options and reach the top tier more often. Running both is how one product reaches two audiences without a second ad set. Our awareness levels guide covers each stage.
The person on screen targets too
The third filter is who appears in the ad. A viewer decides in a second whether the person talking is "someone like me," and delivery learns from that decision like any other. Among ads with a tagged production format:
| Production format | Ads | Median estimated spend | Share reaching top tier |
|---|---|---|---|
| UGC selfie | 78 | $592K | 12.8% |
| Footage + voiceover | 177 | $345K | 14.1% |
| Founder or expert | 145 | $296K | 12.4% |
| Product demo | 97 | $301K | 8.2% |
UGC selfie ads carry the highest median estimated spend of any production format on the board, about $592K, more than double the board median. A real-looking person filming themselves is a strong audience signal: their age, setting and way of talking tell viewers (and the system) who the product is for before a word of the script lands.
In practice, swapping the person on camera is one of the cheapest ways to reach a new audience with a proven script. A 60-year-old and a 30-year-old reading the same lines are two different ads to the people watching.
Four winners that target with the creative
Most cat parents give their cat water in a bowl or a plastic fountain.
- Est. spend
- $1.3M
- Days live
- 235
- Format
- Video, 120s
Names the audience as an observation, not a filter. Every cat owner sees themselves in the first line, and nobody is told to leave.
Anytime you see guys looking good in plain blank T-shirts, they're not wearing ones like this with loose sleeves, a loose collar, and super boxy and baggy right here.
- Est. spend
- $1.4M
- Days live
- 108
- Format
- Video, 26s
- Variants
- 5
Targets men who care how a tee fits, through a scene they recognize. Top tier, with five variants.
The three biggest mistakes dog owners make.
- Est. spend
- $823K
- Days live
- 358
- Format
- Video, 53s
A role plus a promise of a list. Almost a year live on one version.
If you have a cat older than seven, you'll want to hear this.
- Est. spend
- $1.9M
- Days live
- 308
- Format
- Video, 130s
The conditional opener can work: this one has run 308 days. It names a group of people (owners of older cats), where many "If you" ads on the board name a symptom instead.
Turning this into a targeting plan
- Write one concept per audience you want. If you sell to new dog owners and owners of senior dogs, that is two openers, two scenes, possibly two people on camera. Don't build two ad sets with two interest lists and one ad.
- Name people, not conditions. "Dog owners," "nurses," "guys who wear plain tees" give people an identity to recognize. "If you suffer from X" gives them a test to fail.
- Show the situation. A scene the viewer lives through (bread going stale, a cat ignoring its water bowl) does the filtering without a call-out at all.
- Use awareness as your second axis. For each audience, run one unaware and one solution-aware version. That doubles reach from the same idea.
- Keep the campaign broad. Let the creative do the sorting. Our guide to broad targeting covers when that works and when it doesn't, and problem-aware ads shows openers for people who already feel the problem.
To find the openers competitors use to reach a specific audience, search Ad Radar for the role or situation ("nurses", "older cat", "senior dog"). Search covers headlines, body copy and video transcripts, so you see every winner that speaks to that group, across brands.
Figures marked as estimated spend come from Ad Radar's model of engagement on public Meta Ad Library ads. They are estimates, labeled as such, and are best used to rank ads against each other.