Swipe file

Facebook video ad examples: four winners, taken apart.

Meta publishes no spend and no results, so there is exactly one public signal that an ad worked: how long the advertiser kept paying for it. Nobody funds a losing ad for sixteen months. These four ran between 172 and 516 days.

Below, each one is broken into the three decisions underneath it, with the reason the structure holds and what you would need to shoot to rebuild it.

A note on what this is built from. Every headline, body and CTA quoted here is the advertiser's own text, pulled from the Meta Ad Library. The videos themselves are not described shot by shot, because Meta's video links expire within hours and these were captured in June. Each ad links to its permanent Ad Library page, so you can watch it yourself and judge whether the reading holds.

How to read one

A video ad is three decisions, not one.

An ad reads as a single object, so it gets copied as a single object, and that is why most copies fail. It is three independent parts, and each one can be replaced without touching the other two. That is what makes a winner worth studying instead of worth admiring.

The text is also the brief for the footage. A headline that names an enemy needs a shot of that enemy; a body that promises no streaking needs a shot of fabric staying clean. Read the copy and you know what the camera had to do.

1. Hook

Who it stops

Rarely about the product. It names a person or a problem, and everyone else scrolls past on purpose.

2. Body

Why it is believed

Mechanism, demonstration, or the objection answered before it is raised. This is where the click is earned rather than requested.

3. Ask

Where it sends you

The button and the page behind it have to agree. When they disagree, the ad works and the funnel does not.

The examples

Four ads, longest first.

Most collections of Facebook video ad examples are screenshots of ads someone liked. These four were chosen by runtime alone, so the selection contains no taste and no opinion: they lasted, and that is the whole criterion.

Example 01

Bro Glo

516 days Self-tanner, men Video
Headline

Be A Man With A Natural Looking Tan

The whole ad is aimed by this one line. Self-tanner is a category men are quietly embarrassed to shop in, so the headline hands them permission before it hands them a product. It does not sell tanning; it removes the reason not to buy tanning.

Body

Tired of the pasty white look? Don't want to lay in the sun for hours? Use Bro Glo Sunless Tanner.

Two questions, and both are enemies rather than benefits. The pale look is the thing you already dislike; the afternoon in the sun is the alternative you already rejected. The product arrives as the only door left, in one sentence, with no feature list.

Ask

Shop now → the product page, with Free Shipping over $75 in the link description

Straight to the product, not to an article. The free shipping threshold does the work of an upsell before the click: $75 is more than one bottle, so the basket size is being set in the ad.

The part worth stealing

Bro Glo ran three different bodies under the same headline, all inside the same 516 days. One opens on the enemy, one answers the objection (Does that mean you have to live with a pale body? Not a chance.), one leads with ease of use (This easy to apply foam is the answer to pasty skin). Two of the three carry the same concrete proof: won't streak or stain clothes and sheets.

That is a hook held fixed while the body is tested against it. It is the cheapest real test there is, because the expensive decision, the one that determines who stops scrolling, has already been made and is not being disturbed.

What you would shoot

Almost nothing. A pale arm, the foam going on, and a white shirt or a bed sheet staying clean, because that last one is the claim doing the persuading. Three clips, then the same three assembled against one headline three different ways.

2025-01-16 → 2026-06-16 thebroglo.com Watch in Ad Library →
Example 02

Orivelle

248 days Nail care Video
Headline

Say Goodbye to Toenail Fungus

Deliberately plain. In a category where the sufferer is already searching, the headline's only job is to confirm the topic. Cleverness here would cost recognition, and recognition is the entire hook.

Body

Tired of hiding your nails? Say hello to open-toe confidence in just 2 weeks. The Orivelle Nail Pen doesn't just cover up the problem, it goes deep to fight the fungus at its source. Targets the nail root. Easy 10-second daily application. Noticeable results in days. Backed by science and a 30-day guarantee.

The opening is social, not medical: hiding your nails, open-toe confidence. The shame is the buying trigger, and it gets named before the mechanism does. Then a competitor is attacked without being named, since doesn't just cover up is aimed at every polish and lacquer the reader has already wasted money on. Everything after that is objection removal in order: how long, how hard, how soon, what if it fails.

Ask

Learn more → tryorivelle.com/articles/nailfungus-aj02

Note the destination: an article, not a product page. The soft button and the editorial page agree with each other. This buyer is not ready to check out from a video, and the funnel is built to admit it.

What you would shoot

The 10-second application, once, in close-up. The whole ad rests on that number being visibly true. Everything else is text on screen over it.

2025-09-01 → 2026-05-07 tryorivelle.com Watch in Ad Library →
Example 03

Peta Jane Beauty

193 days Self-tanner Video
Headline

Natural Looking Tan Year Round

The promise stated flat, with no audience named and no enemy. It is the same product as example 01 and the opposite decision about who to stop.

Body

Getting a flawless, sunless tan has never been easier. Try our award-winning self-tanner and love the skin you're in.

Two sentences, one claim, one piece of borrowed authority. There is no mechanism and no objection handled, which puts the entire burden of persuasion on the footage. Whatever this video shows, it is doing the work the text declines to do.

Ask

Shop now → the homepage

The homepage, not a product page. The visitor arrives having been told a category, not a product, and has to choose one on landing.

Read it against example 01

Same product, same market, same season. Bro Glo names a buyer and an enemy and ran 516 days. Peta Jane states a benefit to everyone and ran 193. Both are winners by any normal standard, and the gap is still more than double.

This is not proof, since two ads with different budgets and different products cannot prove a rule. It is worth noticing anyway: the narrower ad lasted longer, which is the opposite of what the instinct to address everybody predicts.

What you would shoot

With copy this light, the footage is the entire ad. Application and result, in good light, with nothing between them. There is no argument to cut away to.

2025-12-04 → 2026-06-15 petajanebeauty.com Watch in Ad Library →
Example 04

Emuaid

172+ days Nail care Video
Headline

One Ointment, Fast Relief. Goodbye Nail Fungus.

Format named in the first two words. Against a market full of pens, lacquers and pills, one ointment is positioning disguised as a description: it says simple, and it says familiar.

Body

EMUAIDMAX penetrates deep where other treatments can't reach, targeting nail fungus at its root. The formula contains no steroids, yet delivers powerful relief by reaching the source of stubborn nail fungus.

The differentiator is a negative: no steroids. That is aimed at a reader who has already been to a doctor and did not like what was offered. It is a later-stage buyer than example 02, and the copy knows it.

Ask

Shop now → emuaid.com/pages/nail-fungus-cndf01

A dedicated condition page rather than an article or a catalogue. Ad, button and landing all name the same condition, so nothing has to be re-explained after the click.

What you would shoot

Nothing new. This ad is a claim and a category, and the footage exists to keep eyes on screen while the claim is read. Product in hand, application, done.

2025-12-26 → still running at capture emuaid.com Watch in Ad Library →

The pattern

What four winners have in common.

Four ads is a small sample and a narrow one, two self-tanners and two nail treatments. Three things still show up clearly enough to be worth carrying into your own work.

1. The strongest claim in the category is already taken

Orivelle and Emuaid are different companies selling different formats, and both built their body copy on the same sentence: it penetrates deep and reaches the root, where the things you have already tried do not. Between them that claim ran for over 400 days.

Which means it works, and it means it is spent. Entering that market with the same line puts you third in a queue behind two advertisers with more history. The opening is not a better version of the claim; it is the claim they both skipped.

2. Shame outsells improvement

Tired of hiding your nails. Tired of the pasty white look. The two longest-running ads here both open on something the reader dislikes about themselves, not on something the product does. The version that promises a better outcome to everybody, example 03, ran less than half as long.

3. The button and the landing page were chosen together

Shop now goes to a product page. Learn more goes to an article. In all four, the softness of the ask matches the page behind it, and both match how much convincing the buyer still needs. The mismatch that kills funnels, a hard ask sending a cold reader to a checkout, does not appear once.

Questions

Working from examples.

How many video ads should I be testing?

More than the number you can edit, which is the actual constraint for most operators. The useful shape is the one Bro Glo used above: keep the hook that already works, change one thing underneath it, and let the runtime tell you whether the change was an improvement. That is cheap in decisions and expensive only in edits, which is exactly the part worth automating.

What makes a Facebook video ad keep running for months?

Nothing visible in the first frame. An ad survives when its cost per result stays under what the advertiser can pay, and the four here suggest that comes from a specific person being addressed rather than from production value. The most expensive-looking ad in this set is not the one that ran longest.

Where do I find more Facebook video ad examples myself?

The Meta Ad Library is public and free. Search a competitor, filter to video, and sort your attention by the launch date shown on each ad: the old ones are the ones that worked. Everything on this page came from there, and the only thing added was the reading.

Can I copy one of these structures?

The structure, yes. That is what a swipe file is for, and none of these four invented anything: a named enemy, a demonstration, a matched ask. What does not transfer is the footage, because their clips show their product. The skeleton is reusable, the body has to be yours.

You already own the footage for your next one.

Look at what these four actually needed on camera. A pale arm. A sheet staying clean. Ten seconds of an application. Nothing that required a crew, and nothing that required a synthetic actor pretending to be a customer.

raw2ads takes clips you already have and composes them into that shape, then into the variations that let you hold one hook fixed and test the rest, the way example 01 did for sixteen months.

Start free 7-day trial · no card · no editor