The short answer
Long-form video ads are a normal part of winning Meta accounts: 421 of the 1,897 winning videos on our board run between 2 and 5 minutes, and 105 more run past 5. They target cold audiences, teach before they sell, and work best as expert, story or demo videos. As listicles, they almost never reach the top tier.
How many long ads are winning right now
The usual advice says Meta video ads should be short. The ads that brands keep paying for say otherwise. On the Ad Radar board (October 9, 2026), we have 1,897 live winning video ads with a known length, every one with $50k+ in estimated spend.
421
winners between 2 and 5 minutes
105
winners longer than 5 minutes
28%
of all winning videos run past 2 minutes
$227K
median estimated spend of 2-to-5-minute ads
| Length | Ads | Top-tier share | Median est. spend | Median days live |
|---|---|---|---|---|
| Under 60s | 888 | 9.0% | $245K | 117 |
| 2 to 5 min | 421 | 8.3% | $227K | 122 |
| Over 5 min | 105 | 5.7% | $191K | 99 |
The 2-to-5-minute band performs almost the same as short ads: a similar top-tier share, a similar median spend, slightly longer runs. Past 5 minutes, results soften but don't collapse. Six of those 105 very long ads sit in the top tier.
For the full distribution of lengths, see the best video ad length for Facebook. This piece looks at the long end: who runs it, how the ads are built, and what separates the long ads that win from the ones that just run.
Who runs long ads
Health brands, mostly. 270 of the 526 long ads (51%) are in health, followed by pets (86), beauty (40) and apparel (34). Food and fashion barely appear. If the buyer can understand the product by looking at it, the ad doesn't need three minutes.
Brands that commit to the format. Long ads cluster in a few advertisers. One digestive health brand alone has 54 long ads on the board. Others have 10 to 17 each. When a brand finds that long form works for its product, it builds a library of them, testing new openers and new angles on a proven structure.
Why does Grandma's tummy look pregnant?
- Est. spend
- $941K
- Days live
- 378
- Format
- Video, 492s
- Variants
- 2
Over eight minutes long and live for more than a year, from the brand with the most long ads on the board. The question is everyday and slightly funny, which buys the first few seconds of an eight-minute watch.
Long ads talk to people who aren't ready to buy
| Long ads (2 min+) | Short ads (under 60s) | |
|---|---|---|
| Unaware audience | 36% | 31% |
| Problem-aware | 26% | 18% |
| Solution-aware | 18% | 22% |
| Most aware | 2% | 12% |
The difference is at the edges. Long ads almost never target people who already know the product (2%), and they over-index on people who know their problem but not the solution. Those viewers need the cause explained and the alternatives dismissed before the product makes sense. That takes time.
How long ads open
| Hook pattern | Long ads | Short ads |
|---|---|---|
| Problem/Benefit Opener | 35% | 27% |
| Question Opener | 21% | 14% |
| First-Person Story | 16% | 14% |
| Specific Number Opener | 11% | 22% |
| Offer-First Opener | 6% | 11% |
Long ads open on a problem or a question far more than short ones. They rarely open on an offer: the offer comes at the end, after the case is made. Short ads lean on numbers and offers because they have no time to build up to them.
If you have a cat older than seven, you'll want to hear this.
- Est. spend
- $1.9M
- Days live
- 308
- Format
- Video, 130s
A ten-month run on a callout. The first line picks the audience (cat owners with an older cat) and promises information, which is the contract a long ad has to make in its first sentence.
How long ads are built
Our transcripts cover the start of each ad, and the tags cover the whole thing. Together they show a consistent structure:
- Hook: a problem, a question or a personal moment (these three patterns cover roughly 70% of long-ad openers).
- Cause: why the problem happens. Root cause is the most common angle (27%), and expert-based is a close second (21%, against 6% of short ads).
- Failed fixes: what the viewer has already tried and why it didn't work. This is the "wrong question" and "insider" material that short ads have no room for.
- The product as the answer: usually framed as a discovery or a better mechanism.
- Proof: long ads carry strong proof (studies, numbers, demonstrations) at 8%, more than twice the rate of short ads (3%).
- Offer: at the end, and often on a dedicated landing page.
That last point shows in the landing data. 30% of long ads send traffic to a dedicated landing page, against 16% of short ads. A viewer who has watched four minutes is warm, and brands give that warmth a page built to close.
Would you spend $100 on this book?
- Est. spend
- $1.2M
- Days live
- 1999
- Format
- Video, 166s
Live for more than five years. The opening question frames the price before the pitch begins, and the next two and a half minutes justify it.
Which long formats win
The video type matters more in long form than anywhere else.
| Video type (2 min+) | Ads | Top-tier share | Median est. spend |
|---|---|---|---|
| Expert authority | 60 | 11.7% | $346K |
| Story narration | 74 | 10.8% | $309K |
| Demo of the mechanism | 59 | 10.2% | $209K |
| Interview clip | 12 | 8.3% | $417K |
| Voiceover listicle | 48 | 2.1% | $274K |
Expert, story and demo videos all beat the board's 8.8% video rate when they run long. A voiceover listicle that runs long reaches the top tier once in 48 tries. A list has no tension: once the viewer has seen two of the five points, they know what the rest will look like. A story or an expert explanation builds toward something. We cover the list format and how to fix it in voiceover listicle ads.
By production format, founder or expert (90 ads) and footage + voiceover (86) make up a third of all long ads. AI-generated ads are next with 44, most of them long stories, though AI reaches the top tier far less often than human formats. See AI UGC ads.
Are your chin hairs multiplying?
- Est. spend
- $1.4M
- Days live
- 103
- Format
- Video, 265s
- Variants
- 2
Four and a half minutes on a problem most people wouldn't say out loud. A long ad gives room to normalize an embarrassing problem before selling the fix.
When to keep it short instead
Long form is a tool, not a default. Skip it when:
- The product explains itself. Food, apparel and accessories rarely need two minutes. Median lengths in those niches sit under a minute.
- The audience already knows you. Most-aware buyers need an offer, not an education. Only 2% of long ads target them.
- You only have a list. If the script is "five reasons," keep it under a minute or turn it into a story.
- You have no proof. The extra time is where proof goes. Without it, the long ad is just a slow short ad.
A long-form video ad script outline
Use this as a starting point for a 2-to-4-minute ad:
| Time | Beat | What to say |
|---|---|---|
| 0:00 to 0:05 | Hook | The problem as a question or a personal moment |
| 0:05 to 0:30 | Agitate | What the problem costs the viewer day to day |
| 0:30 to 1:15 | Cause | Why it happens. An expert voice or a story helps here |
| 1:15 to 1:45 | Failed fixes | What they've tried and why it can't work |
| 1:45 to 2:30 | Product | The answer, shown working |
| 2:30 to 3:15 | Proof | A number, a test, a before and after, a customer |
| 3:15 to end | Offer | Price, guarantee, where to go next |
Then cut a 60-second version from the same footage and run both. If the long version holds viewers past the first minute, it usually earns its length. For more on writing the script itself, see AI ad script writing, and for long primary text to pair with long video, see long-form ad copy. The Ad Radar board lets you filter winners by length and read their transcripts if you want real long-form scripts to study.
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.