In-stream and Shorts are two different ads that happen to sit in the same account. In-stream is 16:9 at 1920x1080, interrupts something the viewer chose, and lives or dies at the skip button that appears after 5 seconds. Shorts is 9:16 at 1080x1920, sits inside a scroll the viewer controls, and has to keep earning attention every few seconds rather than surviving one decision. The common and expensive mistake is uploading one file to both placements.
What is the difference between in-stream and Shorts ads?
| Skippable in-stream | Shorts | |
|---|---|---|
| Frame | 16:9, 1920x1080 | 9:16, 1080x1920 |
| Viewer state | Interrupted, waiting for content | Browsing, choosing what to keep watching |
| The threat | Skip at 5 seconds | Scroll at any second |
| Useful length | 20 to 30 seconds for response | 15 to 30 seconds, rarely past 35 |
| Median video CTR | 0.42 percent | Not covered in our benchmark |
| Median hook rate | 22 percent | Not covered in our benchmark |
| Weekly frequency | 3 to 7 per week | Treat 3 to 7 as the ceiling |
| Interface over the frame | Minimal, a skip button and a CTA strip | Heavy: handle, description, action icons, subscribe row |
| Sound | Usually on | Usually on, but must work muted |
| What framing has to survive | Wide composition, small on mobile | Tight composition, overlays eating the edges |
The in-stream medians come from the platform figures in our 2026 fatigue benchmark, which is a synthesis of published platform and agency data rather than our own testing. The report does not break out Shorts separately, so the honest answer is that we do not have a Shorts median to give you, and we are not going to reuse the in-stream numbers as if they applied.
Skip mechanics versus scroll mechanics
The behavioural difference is worth stating precisely, because it decides where the effort goes in the edit.
In-stream gives the viewer one moment of power. Until second five they cannot leave, so what you are buying is a guaranteed five seconds followed by a vote. That makes in-stream front-loaded by nature. The claim, the audience disqualifier and the product all belong before the vote, and the middle of the ad is written for a smaller, more interested crowd who chose to stay. A 22 percent median hook rate tells you how ruthless that vote is.
Shorts gives the viewer continuous power. There is no protected window at all, and no single point where retention collapses. What you need instead is a reason to still be watching at second four, at second nine, at second fifteen. In practice that means a visual or verbal beat every two to three seconds: a new shot, a number appearing, an objection named, a demonstration landing. An ad that says everything important in the first three seconds and then coasts will lose viewers steadily rather than in one drop.
There is a second-order difference in intent. Somebody watching a 20 minute review video is in research mode and can be sold to at greater length. Somebody working through the Shorts shelf is in entertainment mode, which is not worse traffic, just traffic that needs a shorter path from claim to click.
The pillarbox failure mode
This is the single most common technical mistake in a YouTube account running both placements, and it happens in both directions.
Upload a 16:9 master to a Shorts placement and it sits as a narrow horizontal strip in the middle of a vertical frame, with black bands above and below. Your subject occupies maybe a third of the available area, captions become unreadably small, and the ad announces within half a second that it was made for somewhere else.
Run a 9:16 file in a 16:9 in-stream slot and you get the reverse: a tall pane with black pillars down each side, which on a TV screen looks like a mistake and on desktop looks like a repost.
The lazy fix is worse than the problem. Centre-cropping a 16:9 edit into 9:16 keeps the middle 56 percent of the width and throws the rest away, so a two-person interview loses one person, a product held at the edge of frame loses the product, and any lower-third graphic gets cut in half. The subject was composed for a wide frame and the vertical crop does not know that.
Doing it properly means reframing shot by shot, with the subject repositioned per cut, and keeping text clear of the safe zone where the Shorts interface sits: the bottom band with the handle, description and call to action button, and the right edge with the action icons. Caption text belongs in the middle third of the vertical frame, not in the broadcast lower-third position that 16:9 habits produce.
What actually needs to change between the two versions
- Framing per shot, not one crop applied to the timeline.
- Caption position and size, because the readable area is different.
- The opening beat: a disqualifier plus claim for in-stream, motion plus claim for Shorts.
- Pacing, since Shorts wants a beat every few seconds and in-stream wants front-loading.
- Length at the margins, with Shorts capped tighter in practice.
Everything else, meaning the argument, the offer and the proof, can stay identical. That is what makes running both placements cheap in effort, if the export handles ratios properly.
Which placement should you start with?
If you are selling something visual at an impulse price, start with Shorts. Production is cheaper, the feedback loop is faster, and vertical footage you shoot for it also works on Reels and TikTok.
If you are selling something that needs explanation, start with in-stream. The 16:9 frame holds a screen recording or a demo without cropping, viewers are further into research mode, and the longer tolerated runtime lets a demo actually demonstrate.
If you are already running Meta or TikTok successfully, do not assume the winning file transfers. Sound-on is more common on YouTube, so an ad built to work muted often lands as under-scripted: the voiceover is thin because the original design put everything on screen. Rewriting the read is usually a bigger lift than reframing the picture.
Whichever you start with, both placements fatigue. Our benchmark puts most creative as effectively dead within three weeks, with week three sitting 45 to 70 percent below the launch baseline, and finds that brands shipping 15 to 50 variants a month see 3 to 5 times longer campaign lifespan than quarterly refreshers. Running two placements from one concept library is one of the few ways to raise variant count without inventing new claims every week.
Exporting both from one variation
Genyad is our product, so weigh this accordingly. It takes footage you already own, transcribes and tags every clip, and builds each variation as a fresh script, shot selection, voiceover and caption set rather than a re-cut of one timeline. Each variation exports to 9:16, 4:5, 1:1 and 16:9 with framing and caption placement handled per ratio, so the in-stream and Shorts versions of one argument come out of the same job rather than out of two edit sessions. Exports are 1080p on self-serve plans and 4K on Enterprise, with no watermark on any plan, and re-exporting costs no credits. There is more detail on the Shorts ad maker page.
Where we fall short: Genyad does not publish to Google Ads, so you still upload the files and build the campaign yourself. It has no AI avatars or synthetic presenters, no static banner formats, no product URL import and no predicted performance score, which means it will not tell you whether your in-stream or your Shorts cut is going to win. That answer comes from spend.
Frequently asked questions
Can I run the same video as an in-stream and a Shorts ad?
You can run the same concept, but not the same file. A 16:9 in-stream master placed in a Shorts slot gets black bands above and below and the subject shrinks to a third of the frame, while a vertical file in a 16:9 slot ends up pillarboxed between black bars. Reframe per shot and reposition captions rather than cropping the centre out of a wide edit.
Which YouTube placement performs better for direct response?
Both work, and they suit different products. Shorts tends to win for visually obvious, impulse-priced items because the path from claim to click is short, while skippable in-stream tends to win where the product needs demonstration and the 16:9 frame can hold a screen recording without cropping. Test both from the same argument before deciding.
How is a Shorts ad script different from an in-stream script?
In-stream is front-loaded because the viewer cannot leave until second five and then votes with the skip button, so the disqualifier, the claim and the product all belong in those first five seconds. Shorts has no protected window, so it needs a fresh beat every two to three seconds all the way through. The argument can be identical, the sequencing cannot.
Does sound-on change what I should write for YouTube?
Yes. Sound is on far more often on YouTube than in a muted social feed, so the voiceover can carry numbers, guarantees and pricing detail that a Meta or TikTok cut has to burn into the frame. Keep captions anyway, because a share of Shorts viewing is still muted, and the ad has to make its case either way.