YouTube Shorts ads are vertical 9:16 video at 1080x1920, accepted up to 60 seconds, and most direct response offers do their best work between 15 and 30 seconds. The spec that matters more than any of those numbers is the one nobody lists: sound is on far more often on YouTube than in a social feed, so the voiceover can carry argument that a TikTok cut has to burn into the frame. That single difference changes the script more than any resolution setting will.
What are the YouTube Shorts ad specs?
Below is the shape of the requirements plus what we actually export, which is not always the same thing. Google moves placement requirements without much announcement, so confirm the current numbers in Google Ads before you commit a large batch.
| Spec | Shorts requirement | What we ship |
|---|---|---|
| Aspect ratio | 9:16 vertical | 9:16 only, never a padded 16:9 |
| Resolution | 1080x1920 recommended | 1080x1920, 1080p on every self-serve plan |
| Duration | Up to 60 seconds | 15 to 30 seconds for direct response |
| File type | MP4 with H.264 is the safe default | MP4, H.264 |
| Audio | Optional but expected | Always present, voiceover plus a bed |
| Captions | Not enforced by the platform | Burned in, inside the safe area |
| Frame rate | 24 to 60fps | Conformed to one rate before export |
| Headline and CTA | Rendered by YouTube over your video | Kept clear of the lower third |
Two rows cause most of the rejected or wasted uploads. The first is aspect ratio: a 16:9 master uploaded to a Shorts placement gets letterboxed into a small strip with black above and below, and that reads as a repurposed TV spot within half a second. The second is audio. Shorts will run a silent file, but it will lose to anything with a voice on it, for reasons covered further down.
Why 15 to 30 seconds is the working range
The 60 second ceiling is a technical limit, not a recommendation. Two things go wrong at the long end and something different goes wrong at the short end.
Under about 10 seconds you can state a claim and ask for a click, but you cannot make the claim believable. There is no room for the evidence beat, and what you see in reporting is acceptable click-through with weak downstream conversion. Past roughly 35 seconds the problem inverts: you have said enough, and every additional second is another exit point. Retention on Shorts is unforgiving because the next video is one thumb movement away.
Where inside 15 to 30 seconds depends on how much explaining the offer needs:
- 15 to 20 seconds: an obvious visual product, a discount, a before and after.
- 20 to 25 seconds: a claim plus one handled objection. Most DTC sits here.
- 25 to 30 seconds: anything with a workflow to demonstrate, subscription pricing to frame, or a testimonial doing the persuading.
Treat length as a test axis rather than a house rule, the same way you would treat a hook. We cover the mechanics of that in the piece on YouTube ad lengths by placement.
What does sound-on actually change about the script?
This is the part that gets copied wrong from TikTok playbooks. On YouTube, viewers arrive already listening. They were watching something with audio, the Shorts shelf keeps that audio state, and headphones are common. In a Meta or TikTok feed you write assuming the first second is silent, which forces the claim onto the screen as text and forces the visual to do all the persuading.
With sound on by default, three things change:
- The build can be slower. You get roughly a second and a half of visual setup before the claim has to land, where a muted feed gives you none. A slightly slower open often reads as more confident and less like an ad.
- The voiceover can carry the specifics. Numbers, guarantees, pricing framing and the name of the thing can live in the read rather than competing with the product for screen space.
- Tone becomes a variable worth testing. Pace, accent and energy level move performance on Shorts in a way they simply cannot in a feed watched on mute.
None of that means audio is guaranteed. Plenty of Shorts viewing happens muted in public, so the ad still has to make sense with the sound off. The correct posture is that audio adds a layer, it does not replace one.
Captions are still required
Write captions as if they are the only channel. Burn them in rather than relying on auto-captions, which arrive late, break lines in the wrong places and are not styled to your ad. A muted viewer should be able to follow the claim, the evidence and the ask from text alone.
Placement is where captions fail on Shorts specifically. The bottom of the frame carries the channel handle, the ad description, the call to action button and the subscribe row, and the right edge carries the like, comment and share icons. Keep caption text in the middle third of the frame, well inside the safe zone, and keep it to a line and a half at a time so it can be read at a glance.
Shorts versus in-stream, in one table
People search for Shorts specs and then upload the same file to skippable in-stream, which is a different ad with a different constraint.
| Shorts | Skippable in-stream | |
|---|---|---|
| Frame | 9:16, 1080x1920 | 16:9, 1920x1080 |
| Length that works | 15 to 30 seconds | 15 to 30, longer if retention holds |
| Hard constraint | The scroll, at any moment | The skip button at 5 seconds |
| Sound expectation | On, often headphones | On |
| Median video CTR | Not covered in our benchmark | 0.42 percent |
| Median hook rate | Not covered in our benchmark | 22 percent |
| Weekly frequency | Treat 3 to 7 as the ceiling | 3 to 7 per week |
The YouTube figures in our 2026 fatigue benchmark are in-stream medians drawn from published platform and agency sources, not Shorts numbers, and we are not going to present them as if they were. What the report does say across placements applies here: most creative is effectively dead within three weeks, and brands shipping 15 to 50 variants a month see 3 to 5 times longer campaign lifespan than teams refreshing quarterly.
Producing to spec without a manual resize pass
Genyad is our product, so read this with that in mind. It takes footage you already own, transcribes and tags every clip, then builds each variation as a fresh script, shot selection, voiceover and caption set rather than a re-cut of one timeline. Every variation exports to 9:16, 4:5, 1:1 and 16:9 in one pass, at 1080p on all self-serve plans and 4K on Enterprise, with no watermark on any plan. Captions are placed inside the safe area by default, which is the step most often skipped when a 16:9 edit is resized by hand. The Shorts ad maker page has the detail.
What it does not do: it does not publish to YouTube, Meta or TikTok, so you still export files and build the campaign in Google Ads. There are no AI avatars or synthetic presenters, no static banner formats, no product URL import and no predicted performance score attached to a variation, because we do not believe that number can be produced honestly before an ad has spend behind it. Scripts are written natively in English, German, French, Spanish, Italian or Hindi rather than translated, which matters for voiceover on a sound-on placement.
Frequently asked questions
What size should a YouTube Shorts ad be?
Export vertical 9:16 at 1080x1920 as an MP4 with H.264. Do not upload a 16:9 file to a Shorts placement, because it gets letterboxed into a narrow strip with black bars and immediately reads as recycled TV. If you need the same concept for in-stream as well, export a separately framed 16:9 version at 1920x1080.
How long can a YouTube Shorts ad be?
The placement accepts up to 60 seconds, but 15 to 30 seconds is where most direct response offers perform. Below about 10 seconds there is no room for an evidence beat, so click-through can look fine while conversions do not follow. Past roughly 35 seconds every extra second is another chance for the viewer to scroll.
Do YouTube Shorts ads need captions?
Yes, in practice. Sound is on more often on YouTube than in a social feed, but plenty of viewing still happens muted, and auto-captions arrive late and break lines badly. Burn your own captions in, keep them in the middle third of the frame away from the interface overlays, and make sure the ad still argues its case with the audio off.
Can I use the same video for Shorts and in-stream ads?
You can use the same concept, not the same file. Shorts is framed vertically against a scroll, while in-stream is 16:9 and has to survive a skip button that appears at 5 seconds. Build one variation, then export both frames with subject framing and caption placement handled per ratio rather than cropping the centre out of a wide edit.