How DTC Brands Run Meta Ads: Patterns from 749 Brands

The board in one table

This is an aggregate view of how direct-to-consumer brands run Meta ads that work. The data is the Ad Radar board on October 9, 2026: 2,443 active Meta ads from 749 brands, each past $50k in estimated spend and live at least 30 days, each tagged on ten dimensions. "Top tier" means the best 10% by winner score (days live, variants, engagement growth), so a group that reaches it 14% of the time is beating the board.

The pattern in that table: the most common choice is rarely the one that outperforms. Brands crowd into the default (root cause, product page, medium proof), and the edges sit in less crowded spots. The rest of this page breaks each row down, with real ads.

How many winners a brand needs

Winning ads are concentrated in a few advertisers, but most brands on the board have only one.

Two readings. First, a single winning ad is enough to sustain a brand on Meta: more than half the brands here run one. Second, brands with five or more winners are better at producing ads that keep growing, likely because they test more and scale what works with variants. The scaling side is covered in variant count as a scaling signal.

High-volume brands also spread their ads across Pages. On the board, 62% of ads point to a website that also gets ads from at least one other Page: creators, makeup artists, review Pages, celebrity partners. One beauty brand on the board receives winning ads from its own Page and from at least two makeup creators.

Format and production

Video is 81% of the board, but category moves that number. Travel winners are 95% video and pet and food winners 91%, while health sits at 75% because image ads with long copy still work there.

UGC selfie ads carry the highest median estimated spend on the board, and footage with voiceover reaches the top tier most often. AI-generated ads make it to $50k, but they rarely keep growing and are replaced sooner (97 days median). We look at why in AI UGC ads.

Awareness: who the ads talk to

Eugene Schwartz's five awareness levels describe what the viewer already knows. DTC brands on Meta lean hard toward the cold end.

Broad targeting pushes brands toward unaware openers, because the algorithm finds the buyer and the ad has to interest someone who has never heard of the problem. Problem-aware ads run longest and carry the highest median estimate. Solution-aware ads reach the top tier most often.

Video length shrinks as awareness rises. The median unaware video runs 71 seconds, problem-aware about 72, solution-aware 61, product-aware 55 and most-aware 43. Cold viewers need the story; warm ones need the offer. The full breakdown is in awareness levels for ads.

Angles: the argument behind the ad

The angle is the reason the ad gives for buying. Root cause ("the real reason you have this problem") dominates, but it is crowded and reaches the top tier less often than average.

Cautionary tales (what happens if you don't act) are rare and reach the top tier most often. Insider angles carry the highest median estimate in the table. Story and systematic breakdown sit in the middle on volume and above average on results, which makes them the safest second angle to test after root cause. See root cause ads and story and case study ads for examples of each.

Proof: how much evidence the ad shows

We tag proof on three levels: unproven (claims only), medium (reviews, demos, testimonials) and strong (data, clinical or third-party evidence, measured before/after).

Only 4% of ads show strong proof, and they carry the highest median estimate and the highest top-tier share of any proof level. The difference between unproven and medium is small. The jump comes from real, specific evidence. Whatever level you pick, the claims are yours to substantiate. More in social proof ads.

Landing pages: where the click goes

(Not every ad has a landing tag, so shares don't add up to 100%.)

Most DTC brands send Meta traffic straight to a product page, even from unaware ads: 49% of tagged unaware ads land on one. The less common funnels do better. Quiz funnels carry the highest median estimate and run the longest, and listicles reach the top tier 14% of the time. The advertorial sample is small (19 ads), so treat its number as a hint. Our guides to quiz funnels and listicle landing pages cover how to build them.

Product form: what is being sold

Supplements are a quarter of the board and the most crowded category, with the lowest top-tier share in this table. Food and drink brands are fewer but carry the highest median estimate and the best top-tier share. Subscription meals and fresh pet food lead that group, often through creator Pages.

What to take from 749 brands

1. Start where everyone is, then move one step off. Root cause plus product page is the default because it works. Your second test should change one thing: a cautionary tale angle, a story, or a listicle instead of the product page.
2. Invest in proof before production. Strong proof beats every production format on top-tier share. A measured result filmed on a phone beats a polished ad with no evidence.
3. Match length to awareness. Long for cold traffic, short for warm. If your unaware ad is 20 seconds, it probably skips the part that makes a stranger care.
4. Use more than one Page. Creators and partners run a large share of winning ads. Build that into your plan, not as an afterthought.
5. One winner is enough to start. Most brands on the board have exactly one. Find it, then scale it with variants and new hooks.

For a single-brand version of this analysis, the competitor ad research process shows how to tag a competitor's ads weekly. And if you want to run these cuts on your own niche, Ad Radar lets you filter the board by every dimension in this article.

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.