How to make a dropshipping ad
How to get from a saved product to three to five ads worth testing, with what the ads we track have in common next to each step.
The short answer
An ad that sells shows one person one problem being solved by your product, and it shows it in the first seconds. Decide that angle before you pick up a camera. Film it on a phone with the sample you ordered, in 9:16 for Reels and Stories. Keep the caption short, use the Shop Now button and make three to five versions that differ in one thing, so the test tells you what worked. Then read every line for the claims Meta rejects before you post.
Most of this costs time rather than money. The expensive mistake is paying for clicks on an ad that never gave anyone a reason to stop scrolling.
Start with the angle, not the camera
An angle is one sentence: who it is for, what bothers them and what changes once they have the product. “Pet hair on the sofa every evening, gone in one pass” is an angle. “Amazing lint roller, 50% off” is not. It names a product and a price and leaves the viewer to work out why they should care.
Write three angles for the same product before you film anything. They usually come from three places.
- The problem. The moment the buyer already hates. Show it first and then the fix.
- The result. What things look like afterwards. It works best when the result is visible, like a clean car seat or a tidy drawer.
- The doubt. The reason someone would not buy, answered on screen. “Does it really fit any car?” followed by three different cars.
Read the reviews on the supplier’s listing and on competing stores for the words buyers use. The complaint in a one-star review is often your best angle, because it says what people hoped the product would do. The product description guide shows how to build copy from that kind of evidence instead of inventing it.
The first seconds decide whether anyone sees the rest
People scroll past an ad before they know what it sells. So the opening has one job: show the problem or the product working, straight away. No logo, no “hey guys” and no slow pan across a box.
- Open on movement. A hand using the product beats the product sitting on a table.
- Put the angle on screen as text in the first frame. Plenty of people watch with the sound off, so the words have to work without it.
- Keep it close. On a phone screen a product filmed from across the room is a blur.
- Get to the result quickly. If the payoff needs twenty seconds of setup the ad is too slow.
What to film, with a phone and one sample
You do not need a studio. You need the product in your hands, which is why the Launch Builder has you order one in its Source step. Filming it also shows you whether it works and how long it really took to arrive, two things a supplier’s listing will not tell you honestly.
A shot list that covers most products:
- The problem in a real setting: the messy drawer, the tangled cable, the stiff neck at a desk.
- The product in hand, close enough to see what it is.
- The product doing its one job from start to finish, without a cut that hides anything.
- A close-up of the detail that matters: the clip, the fabric, the seal.
- The result, held long enough to register.
- An unboxing if you like. It reads like a real customer rather than an advert.
Film each shot a few times in good daylight. One afternoon of footage can be cut into several ads with different openings, which is how you get three to five ads out of one sample.
The supplier’s own video is tempting because it is free and polished. The trouble is that every store selling the product can download the same clip, so your ad looks like theirs. Use it while your sample ships if you have to and replace it once you can film your own. Image ads work in that gap too, and the Launch Builder suggests starting with them while the sample is on its way.
Video or image
Of the 33,886 Facebook ads we read for our dropshipping statistics, 63.9% were videos. Image ads had been running longer, a median of 34 days against 18 for video. Part of that gap is age: video ads are newer on average because more stores now test video first. So read it as how the two formats are used and not as proof that one sells more.
What it does say is that an image ad is not a beginner’s compromise. Plenty of stores keep one running for months. A clean photo of the product in use with the angle as one line of text is cheap to make and worth testing next to your videos.
The caption: shorter than you think
We compared the 6,796 ads that ran 90 days or more with the 5,844 that stopped in their first week. The long runners had a median caption of 261 characters against 466. They also mentioned a discount more often (9.8% against 3.4%) and used words like “limited” or “today only” more often (7.4% against 2.8%).
That shows what long-running ads have in common, not what makes an ad last. A store that keeps an ad running for months has usually already found what works for its product. But it points one way. 261 characters is about three short lines, and a caption that works usually goes like this:
- Line one: the angle in the buyer’s words. It is the only line many people read.
- Line two: what the product is, and the one detail that answers the main doubt.
- Line three: the offer and the delivery time, stated honestly.
A discount or a deadline is fine when it is real. A countdown that starts again for every visitor is not, and nor is a “was” price the product never sold at. The FTC’s 2022 staff report on dark patterns gives countdown timers that reset as an example of design that tricks buyers.
The button
52.1% of the ads we read used Shop Now and 19% used Learn More. For a store selling a product Shop Now is the plain choice, because it says what happens when you tap. Learn More suits a product that needs explaining before anyone buys, and it is worth a test if your ad gets clicks that the product page then loses. These shares tell you what stores pick, not which button converts, so let your own test decide once you have the clicks.
Sizes, and where Meta covers your ad
Meta’s Ads Guide recommends 9:16 for Reels and Stories. For the feed, current spec guides point to 4:5 with square 1:1 still accepted. Film vertically and you can crop a 4:5 version from the same footage. Meta changes these specs from time to time, so check its Ads Guide before you export.
In Reels and Stories the caption, the button and your page name sit on top of the bottom and the edges of the video. Keep your text and the product near the middle of the frame so nothing important ends up underneath them.
Make three to five ads and change one thing
A test only teaches you something when the ads differ in one way. Keep the same footage and caption and change the opening: one starts on the problem, one on the result and one on the doubt. When one of them wins you know it was the hook. Change the hook, the music, the caption and the price at once and a winner tells you nothing you can repeat.
How long to run each one and when to stop is the next part of the work. The traffic guide covers the $20 a day test, and how long to run a TikTok ad gives stop rules in money spent rather than days.
Claims that get ads rejected
A new ad account gets little patience, and one rejected ad can hold up the whole test. Read every line of your ad and your product page for these before you post.
- Saying something about the viewer. Meta’s advertising standards do not allow ad text that asserts or implies a personal attribute, such as someone’s health, weight or money troubles. “Tired of your back pain?” talks about the viewer’s body. “Built for long days at a desk” talks about the product.
- Health and body claims. “Cures” and “treats” draw rejections fast. Meta’s health and wellness rules do not allow side-by-side before-and-after pictures for weight loss products, and ads must not suggest someone should feel bad about their body. Say what the product does, not what it does to a body.
- Reviews you do not have. Since 21 October 2024 the FTC’s rule on consumer reviews and testimonials bans fake reviews, AI-written ones included, and lets the FTC seek civil penalties. Put a star rating or a quote in an ad only when a real customer gave it.
- Delivery you cannot keep. “Fast shipping” on a parcel that takes two weeks from China is a promise you break on every order. The first orders guide explains why US law cares about that.
In SpotPeaks, Check before you post reads your ad, headline, script and product page for these and offers a rewrite for each one.
Where SpotPeaks fits, honestly
Ad Creative in SpotPeaks writes angles, captions, headlines, audiences, a phone-shot script and ad images around the product you picked, and every product page shows the ads other stores already run for it. It does not film the product for you, and nobody can tell you in advance which hook will win. That is what the test is for.
FAQ
What makes a good dropshipping ad?
One clear angle shown in the first seconds: the problem the product solves or the result it gives, filmed with the real product. Then a short caption, a plain Shop Now button and three to five versions that change only the opening, so the test tells you which hook works.
Should a dropshipping ad be a video or an image?
Test both. Of the 33,886 Facebook ads we read, 63.9% were videos, but image ads had been running longer: a median of 34 days against 18. Video ads are newer on average, so part of that gap is age. Treat an image ad as a cheap extra test, not a fallback.
Can I use the supplier's video in my ad?
Only to tide you over while your sample ships. Every store selling the product can use the same clip, so your ad looks like theirs. Film your own once the sample arrives, which also shows you whether the product actually works.
What size should a Facebook or Instagram ad be?
9:16 for Reels and Stories, which Meta's Ads Guide recommends, and 4:5 for the feed with 1:1 still accepted. Film vertically and crop the feed version from the same footage. Check Meta's Ads Guide before you export, since the specs change.
How long should an ad caption be?
Shorter than most beginners write. Ads that ran 90 days or more had a median caption of 261 characters, against 466 for ads that stopped in their first week. That is what long runners share rather than proof of cause, but three short lines is a sound place to start.
See the ads that already run
Every SpotPeaks product page shows the ads stores run for it, and Ad Creative writes angles, captions and a phone-shot script around your product. Research is $29.99/month. Launch is $49.99/month with a 7-day free trial.