Facebook Reels and Facebook feed need two different cuts of the same footage, not one file cropped twice. Reels is 9:16 at 1080x1920 with the interface drawn over your frame, so it wants fast shots, an offer inside the first five or six seconds and every message burnt into the middle of the frame. Feed is 4:5 at 1080x1350 with Meta's headline and CTA rendered below the video, so it tolerates longer holds, an offer around second eight, and it punishes burnt-in headlines that duplicate the ad's own. The cost of ignoring that is not aesthetic: a recrop loses either the composition or the pacing, usually both.
What actually changes between the two placements
| Facebook Reels | Facebook feed | |
|---|---|---|
| Ratio and size | 9:16, 1080x1920 | 4:5, 1080x1350 |
| Interface | Over the video, top, bottom and right rail | Below the video, in its own block |
| Working margins | Top ~14%, bottom ~20% | Bottom ~12%, to survive trims |
| Burnt-in headline | Required, nothing else carries it | Usually wrong, it stacks with Meta's headline |
| Shot length we cut to | 1 to 2 seconds | 2 to 4 seconds |
| Offer lands by | Second 5 or 6 | Second 8 to 10 |
| Total length we ship | 8 to 20 seconds | 10 to 25 seconds |
| Audience posture | Swiping, expects to leave | Scrolling, expects to browse |
| Median video CTR | About 1.78 percent | About 1.62 percent |
| Frequency ceiling | About 2.5 weekly on prospecting | About 2.5 weekly on prospecting |
The margin figures are working margins rather than published constants. Meta moves interface heights, and Facebook Reels puts its CTA at a different height from Instagram Reels, so build with room to spare rather than to the pixel. Our safe zone reference has the detail.
What breaks when you recrop one into the other
Recropping is the default because it is free, and it fails in four predictable ways.
Feed to Reels loses resolution and gains nothing. A 1080x1350 master cropped to 9:16 keeps 1080 of width and needs 1920 of height, so you either scale up and go soft or you pad and ship black bars. Both read as a repurposed ad, and the bars count as part of your creative.
Reels to feed loses the subject. A 1080x1920 frame cropped to 4:5 takes 570 pixels out of the vertical, and because your Reels composition put the face in the lower middle to stay clear of the top strip, the trim usually removes the top of the head or the product in the hands. It looks fine in the editor at full height and wrong in the preview.
Text drifts into the wrong band. Captions positioned at 78 percent of a 9:16 frame sit safely above the Reels CTA. Crop to 4:5 and they land near the bottom edge, where Meta's own headline block starts and where any further trim takes from. You now have two headlines, in different type, saying similar things.
Pacing goes with the wrong placement. A feed cut placed in Reels feels slow because the viewer arrived by swipe and will leave the same way. A Reels cut placed in feed feels frantic and, worse, the offer has already gone past before a browsing viewer has settled into the ad.
The tell that a team is recropping: the same creative shows a decent hook rate in one placement and a collapsed one in the other, while CTR among viewers is similar. That is a composition problem, not an audience problem.
The two cuts, second by second
Same footage, same claim, two builds. This is roughly how we lay out a product demo for each.
| Time | Reels cut, 9:16 | Feed cut, 4:5 |
|---|---|---|
| 0 to 1s | Motion in frame one, close on the problem | A held establishing shot, product visible |
| 1 to 3s | Two cuts, caption states the claim | One cut, caption states the claim |
| 3 to 6s | Product in use, offer on screen by second 5 | Product in use, held longer, detail visible |
| 6 to 10s | Two more angles, a second proof point | Offer arrives, spoken and captioned |
| 10 to 15s | Result shot, loop back to frame one | Second proof point, testimonial or detail |
| 15 to 25s | Not used | Result shot, then a full-frame end card |
| End | Hard cut to frame one, no fade | End card, no burnt-in CTA, Meta's button does that |
Three things in that table do real work. The Reels cut has roughly twice the cuts in half the time, because a static frame in a swipe feed reads as a dead ad. The offer moves by four seconds, because in feed a viewer who has already stopped scrolling will wait for it and in Reels they will not. And the ending differs completely: Reels ends on a hard cut so the loop counts as another view, feed ends on an end card because there is no loop to earn.
Which placement should get the budget?
Run both, in the same campaign, and read them separately. Meta will happily give you a blended placement result that tells you nothing, and the CTR medians are close enough (1.78 percent on Reels, 1.62 percent on feed) that the difference in your account will come from the creative, not the placement.
Our view after doing this a lot: Reels is where a new hook proves itself fastest, because the hook is doing almost all of the work and a bad one dies visibly. Feed is where a considered proposition converts, because the format allows the extra 10 seconds a demo needs. If you can only afford one cut, make the Reels one and accept a capped feed result: a 9:16 asset at least renders without bars in feed, while a 4:5 asset in Reels arrives padded.
Fatigue arrives at a similar pace on both. Our 2026 fatigue benchmark, a synthesis of published platform and agency figures, puts CTR decline at 15 to 20 percent over the first two weeks, week three at 45 to 70 percent below the launch baseline, and most creative effectively dead within three weeks, with weekly frequency around 2.5 on Meta prospecting marking where decline begins. Two cuts of one concept do not last twice as long. They cover two placements properly for the same three weeks.
Exporting both without editing twice
This is the practical objection to everything above: nobody has time to build two cuts of every concept. That is the problem Genyad exists to solve, so treat this as a disclosed product note.
It works from a library of footage you already own. Upload once, it transcribes and tags every clip, then each variation is generated as its own script, shot selection, voiceover and caption set rather than a re-cut of one timeline. Because the shot selection happens per ratio, the 9:16 version can use two more cuts and a tighter framing than the 4:5 version of the same script, and captions are placed inside the margins for the ratio being written. Our multi-format export writes 9:16, 4:5, 1:1 and 16:9 from one variation, and the Reels ad maker page shows the vertical side of it. Editing and re-exporting cost nothing, so pulling a shorter Reels version out of a working feed cut is free.
What it does not do: no AI avatars, no static banner formats, no product-URL import, no predicted performance scores, and no publishing to Meta, so you upload both files in Ads Manager yourself. If all you want is one master and an automatic crop, a manual editor like CapCut does that for nothing, and for some accounts that is the right trade.
Frequently asked questions
Can I use the same video for Facebook Reels and Facebook feed?
You can upload it to both, but one of the two will be compromised. A 9:16 file in feed wastes height and a 4:5 file in Reels arrives padded or upscaled, and in either case the text positions that were safe in one placement are wrong in the other. If you have to pick one file, a 9:16 master does less damage in feed than the reverse.
What is the main creative difference between Reels and feed ads?
Where the message lives and how fast the edit moves. In Reels the interface covers the top, bottom and right of your frame and nothing else carries your headline, so it has to be burnt into the middle. In feed Meta renders its own headline and CTA below the video, so burnt-in copy duplicates it, and the pacing can run two to four seconds per shot instead of one to two.
When should the offer appear in each placement?
By second five or six in Reels, and by second eight to ten in feed. Reels viewers arrive by swipe and leave the same way, so anything after the first third is shown to a shrinking group. Feed viewers have already stopped scrolling, which buys a few more seconds before the offer has to land.
Do Reels and feed ads fatigue at different rates?
Not meaningfully in our experience. Our benchmark report puts CTR decline at 15 to 20 percent in the first two weeks on both, with week three landing 45 to 70 percent below launch, and both placements showing decline from a weekly frequency of about 2.5 on prospecting. Two placement cuts of one concept cover more surface, they do not extend its life.