You're usually looking up specs for Facebook carousel ads at the worst possible moment. The campaign is built, the ad set is ready, the designer exported three square cards and two vertical ones, Ads Manager is lagging, and the preview doesn't match what you expected. That's when small spec mistakes turn into wasted time, ugly crops, and avoidable QA loops.
So this is the reference I'd want open beside Ads Manager. Not a generic walkthrough. A practical sheet for performance teams who need the numbers fast, then need the actual trade-offs Meta's docs tend to bury.
Table of Contents
- Quick Reference Carousel Ad Spec Sheet
- Image Card Technical Requirements
- Video Card Technical Requirements
- Text Copy Links and CTA Specifications
- Placement-Specific Adjustments and Compatibility
- How to Optimize Carousel Ad Workflows at Scale
- Avoiding Common Carousel Ad Errors and Rejections
- Performance Best Practices Beyond Technical Specs
Quick Reference Carousel Ad Spec Sheet
A carousel can look approved in the design file and still break once it hits mixed placements in Ads Manager. The usual failure point is not the headline or CTA. It is format discipline. One card gets exported differently, one video comes in heavier than expected, or one placement sneaks in and starts cropping the first asset in ways the preview did not make obvious.
Use this as the working spec sheet before launch, not just a compliance checklist.
Facebook Carousel Ad Specs At-a-Glance
| Specification | What to set | What it means in practice |
|---|---|---|
| Cards per ad | 2 to 10 cards | Two to five cards is the safer operating range for prospecting. You can run more, but longer carousels usually need a clear sequence or product set to justify the extra swipes. |
| Minimum resolution | 1080 × 1080 pixels or higher | Treat this as the floor, not the target. Small exports often pass upload checks and then lose clarity after Meta compression. |
| Recommended aspect ratio for feed | 1:1 or 4:5 | Pick one ratio at ad build and keep it fixed. Square is easier to scale across placements. 4:5 can win more feed space, but it creates more QA work. |
| Max file size | Images up to 30 MB. Videos up to 4 GB | Hitting the limit is rarely the goal. Heavier files slow approvals, create versioning mistakes, and make handoff harder across creative and media teams. |
| Placement support | Feed, Stories, Reels, Messenger, Marketplace | Support does not mean equal presentation. The ad may serve broadly, but crops, text visibility, and card behavior still vary by placement. |
| Ratio consistency rule | All cards must use the same aspect ratio | This is the setting teams break most often. One off-ratio card can turn a clean build into a crop-check marathon. |
| Video length | Video cards support a wide range of lengths, with Meta format details documented in its Ads Guide | Shorter cuts are easier to hold attention inside a swipe unit. If a card needs sound, setup, or explanation to work, it usually underperforms in carousel. See Meta's Ads Guide for carousel formats. |
The practical rule is simple. Build the whole carousel around the first card's format decision.
If card one is square, everything stays square. If card one is 4:5, the full unit follows 4:5. That sounds obvious, but it is the spec that saves the most QA time because it affects image exports, video edits, preview checks, and placement eligibility in one move.
A second workflow rule helps just as much. Decide early whether the ad is being built for broad placement coverage or for a narrower feed-first setup inside Ads Manager. Teams lose time when creative is designed for one environment and trafficking later tries to force it into another.
Image Card Technical Requirements
Image cards cause more QA churn than video cards because the failures look minor in build and expensive after launch. The usual pattern is simple. Creative exports pass a desktop check, the ad gets built in Ads Manager, then one placement preview reveals a crop that cuts the headline, product, or price lockup.
The base setup is straightforward. Carousel image ads run with 2 to 10 cards, and teams should export at 1080 x 1080 pixels or higher as a working floor. In practice, the spec is only half the job. The actual work is choosing a format your trafficking workflow can keep consistent across every card.
For teams managing broad placement coverage, 1:1 is still the default I trust most. It is easier to standardize across design templates, easier to review in bulk, and less likely to create avoidable crop checks once the ad starts serving outside the main feed.
Use 4:5 for feed-first builds when the extra vertical space helps the product or offer. The trade-off is operational. You get stronger feed presence, but you also increase the chance that previews need manual review card by card, especially if different designers or editors touched the asset set.
A practical image spec sheet usually looks like this:
- Card count: 2 to 10 cards
- Export floor: 1080 x 1080 pixels or larger
- Primary ratio: choose 1:1 or 4:5 before creative production starts
- File size: keep each image within Meta's upload limit
- Safe composition: keep text, logos, and product edges away from the crop boundary
The format decision matters because Ads Manager treats the first card like the template for the rest of the unit. If card one is square, later cards built taller will get forced into that shape somewhere in delivery or preview. That is where clean-looking exports turn into last-minute revision requests.
The failure mode is easy to recognize inside production:
- The first card is exported in 1:1.
- A later card arrives in 4:5 because it looked better in the designer's artboard.
- Another card has copy too close to the top edge.
- The feed preview looks acceptable.
- A different placement trims the later cards and the ad loses readability.
This is why ratio checks belong upstream, not during trafficking. Build a simple export rule into the handoff process. One campaign, one aspect ratio, one safe-area standard. If the creative concept needs mixed orientations, split it into separate ads and preserve the framing instead of forcing one carousel to absorb conflicting assets.
A few image habits save time repeatedly:
- Keep framing consistent. Product scale, logo position, and text blocks should feel aligned from card to card.
- Design for the crop zone. Leave breathing room around headlines, badges, and packshots.
- Use card one to sell the swipe. Lead with the strongest product, offer, or visual proof point.
- Review in actual placements. Ads Manager previews catch problems that folder thumbnails hide.
The teams that ship clean carousels do not rely on the platform to smooth over loose creative files. They lock the ratio early, export to a repeatable template, and treat image QA like part of media buying, not just design cleanup.
Video Card Technical Requirements
A carousel video can be technically valid and still perform like a bad build. The usual failure is simple. The editor exports a nice standalone video, trafficking drops it into a carousel, and nobody checks whether it still works as a swipe card.
Meta gives video cards a wide tolerance on length and file size, but carousel behavior is much tighter in practice. In Ads Manager, video cards work best when each one delivers a single point fast: product use, before-and-after, feature proof, offer reveal, or a quick testimonial moment. If the clip needs setup, pacing, and payoff, it usually belongs in a single-video ad, not inside one card of a carousel.
A real trade-off shows up with duration. We have tested longer explainer cuts against short product demos in carousel. The longer version looked better in review. Swipe-through usually got worse because card one asked for too much attention before the user had a reason to keep moving. For carousel, short wins more often because it respects the format.
Build for card behavior, not standalone video behavior
The safest workflow is to edit carousel video as card-native creative from the start. That means:
- Front-load the point in the first seconds. Users decide quickly whether to keep watching or swipe.
- Assume muted autoplay. Captions, supers, and visual demonstration matter more than voiceover.
- Use a strong poster frame. In preview and low-attention environments, the thumbnail does part of the selling.
- Keep motion readable at small size. Fine product detail, UI text, and fast transitions often collapse inside the card.
One operational mistake causes repeat rework. Teams cut a polished vertical video for Reels, then try to reuse it in carousel without re-editing text size, crop safety, or opening frame. The file uploads. The card still underperforms because the composition was built for immersion, not for a swipe unit surrounded by other cards.
Ratio discipline matters more with video
Video exposes framing problems faster than static cards because movement pulls attention toward the crop. If the aspect ratio shifts across cards, the ad starts to feel unstable even before users notice the exact issue.
| Build choice | What usually happens |
|---|---|
| Square image first, vertical video later | The video gets forced into the square frame and key action sits too close to the edges |
| Vertical-first concept mixed with square cutdowns | Product scale and text treatment feel inconsistent card to card |
| Different safe areas across exports | Captions, logos, and focal points jump around as users swipe |
Set the ratio before editing starts. Then keep the same framing rules across every card, including subtitle position and product scale. Checking this only at upload is late. By that point, the fix is usually a re-export, not a setting change.
Ads Manager checks that catch problems early
Before publishing, review carousel video cards inside the actual ad preview, not just in the media library. Pay attention to:
- Thumbnail quality on card one
- Subtitle placement near the edges
- Whether the first frame makes sense without motion
- How the card looks beside static cards in the same unit
- Whether card-level destination logic still matches the video message
The last point gets missed a lot. A product demo card that clicks through to a generic collection page usually leaks intent. If the card is specific, the landing page should be specific too.
Video cards work best when each one acts like a compact proof point. Show the product, show the result, then let the next card carry the next job. That is how carousel video stays clear, easy to swipe, and worth scaling.
Text Copy Links and CTA Specifications
Most carousel build issues aren't visual. They're structural. The copy hierarchy is unclear, card headlines repeat the same line, and the destination URLs don't match the user's swipe intent.
The practical way to think about carousel text is by layer. There's the Primary Text above the unit, then the card-level text below each asset, then the CTA and destination behavior. Each layer needs a different job.
How to write for the actual ad surface
Use Primary Text to sell the swipe. Don't waste it summarising what's already obvious from the cards. In most performance setups, the opening line works best when it frames the offer, problem, or angle before the user interacts with card one.
Card Headlines should do one of two things:
- identify the specific product or angle on that card, or
- complete a sequence if the carousel is telling a story in order
Card Descriptions are less reliable as a communication layer because they aren't consistently prominent across placements. Use them only if the extra context offers a substantial benefit.
Write card headlines so they can survive partial visibility. If one card gets seen out of context, the line should still make sense.
Link logic matters more than most teams think
One of carousel's biggest practical advantages is card-level destination control. You can route each card to a different URL. That's useful for product collections, feature stacks, comparison journeys, or localized landing pages.
But it also creates easy QA failures. The common ones:
- Wrong destination mapped to the wrong card
- All cards pointing to the same page by accident
- UTM inconsistencies across cards
- Headline promise not matching the landing page
If you're running broad testing, that mismatch gets expensive fast because the click itself may look fine while the post-click experience collapses.
CTA selection inside Ads Manager
In Ads Manager, the CTA should match the landing page type and user intent. Keep it boring and accurate. Performance teams often get into trouble when they use a more aggressive CTA than the page can support.
A clean check before publish:
- Read the headline on each card.
- Click the destination assigned to that card.
- Confirm the CTA feels consistent with that page.
- Preview on mobile, not just desktop.
What works is alignment. What doesn't work is using carousel to fake product depth or narrative depth that the landing experience can't deliver.
Placement-Specific Adjustments and Compatibility
A carousel can pass spec checks, clear review, and still break once placements expand. The usual failure happens after someone leaves Advantage+ placements on, previews Feed only, and ships a square-first build into Stories or Reels. Then the ad starts serving with cropped branding, weak opening frames, or cards that make sense only in sequence.
Carousel can serve across Feed, Stories, Reels, Messenger, and Marketplace, but eligibility is not the useful question. The useful question is whether the creative still communicates after Meta adapts it for each surface.
What actually changes by placement
Feed is still the cleanest environment for carousel logic. People expect the horizontal swipe, card order is easier to follow, and product comparisons or step-by-step framing usually hold together. This is the placement where a true sequence has the best chance of being understood as a sequence.
Stories and Reels are less forgiving. Users move fast, the first frame has to do more work, and vertical-safe composition matters. A common miss is a 1:1 product shot that looks fine in Feed, but the top portion gets cropped in Stories and the logo disappears. Another is a headline baked into the upper edge of the image that ends up partially hidden once the placement UI sits on top of it.
Messenger and Marketplace tend to reward simpler builds. Product-led cards, clear pricing, and obvious visual hierarchy travel better there than narrative-heavy concepts. If card one is vague, interaction drops quickly because the user has less patience for figuring out the sequence.
Here is the practical version teams use in QA:
| Placement | What usually works | Common problem |
|---|---|---|
| Feed | Sequential storytelling, product comparisons, square or vertical assets with clear first-card framing | Later cards carry too much of the message |
| Stories | Vertical-safe creative, simple offer structure, fast brand recognition on card one | A 1:1 product shot gets adapted and the top area crops out the logo or headline |
| Reels | Motion-led cards, obvious subject focus, strong first-second hook | Static-first concepts feel transplanted from Feed |
| Messenger and Marketplace | Clean merchandising, price-forward cards, direct value prop | Card one is too soft to earn the swipe |
How to handle this inside Ads Manager
Use the placement preview menu for every version before publish. Check the ad at the ad level, not just the asset in isolation. In practice, I look at three things first: whether card one explains itself without the rest of the sequence, whether any text or branding sits in risky crop zones, and whether the adapted frame still looks intentional.
The workflow decision is usually simple.
- Ordered story across cards: use fewer placements or upload placement-specific assets.
- Each card can stand alone: broader placement coverage is usually safer.
- Heavy motion concept built for vertical viewing: verify Reels and Stories first, then decide whether Feed is still worth keeping.
- Catalog-style carousel: Feed, Marketplace, and Messenger often make more sense than forcing every surface.
Automatic placement expansion is useful, but only after the creative survives preview. If the concept depends on sequence, tiny text, or precise framing, broad compatibility on paper does not help much in delivery.
The carousel that performs in Feed often needs adjusted framing, different opening emphasis, or tighter card independence before it deserves Stories and Reels spend.
How to Optimize Carousel Ad Workflows at Scale
Building one carousel in Ads Manager is manageable. Building dozens is where the friction shows up. The problem isn't that the format is hard. The problem is that Meta's manual workflow makes repeatable production slow and error-prone.
Where Ads Manager slows teams down
At scale, the bottlenecks are predictable:
- Asset uploads become repetitive: each card gets loaded one by one
- Naming drifts: ad names, card structures, and testing labels stop matching
- Placement variants get messy: feed and vertical assets end up mixed in the same build queue
- QA takes too long: the reviewer has to inspect every card manually
This gets worse when one account manager is trafficking for multiple brands or markets. Carousel is powerful, but it asks for discipline. Ads Manager doesn't help much with that.
The most frustrating part is that small mistakes are easy to miss until after publish. One incorrect card order, one wrong URL, one asset with the wrong ratio, and the ad is technically live but operationally broken.
A cleaner production workflow
The teams that move fastest usually standardize around a simple build system:
- Assets are exported into ratio-specific folders before upload.
- Card order is locked in a naming convention.
- Copy is approved in a structured sheet, not improvised in the UI.
- The reviewer checks previews by placement before anything goes live.
That workflow matters more than any individual spec. It reduces human error, which is the primary challenge once you're launching volume.
A few workflow habits pay off quickly:
- Separate feed assets from vertical assets at the file level
- Name assets in card order
- Keep URL mapping in the same worksheet as headline copy
- Use consistent ad naming so breakdowns stay readable later
The goal isn't elegance. It's reducing the number of places where Ads Manager can inadvertently turn a clean creative plan into a cluttered trafficking task.
Avoiding Common Carousel Ad Errors and Rejections
Most carousel mistakes don't trigger dramatic warnings. They pass into review, get approved, and then underperform because the build itself is flawed. That's why a pre-launch checklist matters more than another list of specs.
Pre-flight checks that catch most problems
Run these checks before publish, every time:
- Card order check: confirm the sequence inside Ads Manager matches the strategy doc or creative sheet
- Destination check: open every URL from the ad build, not from a separate spreadsheet
- Crop check: review placement previews and make sure key visual elements survive
- Headline check: verify the headline on each card still makes sense beside that card's asset
- Policy check: inspect thumbnails, text overlays, and claims before sending the ad to review
A lot of “Meta issues” are really assembly issues. The platform gets blamed, but the ad wasn't QA'd tightly enough.
Here's the embedded walkthrough if you want a visual checklist in parallel with your own QA process.
The settings drift problem inside Meta
There's another class of error that has nothing to do with creative files. It comes from default settings and automated enhancements changing the ad after your team thought the build was locked.
The biggest one to watch is Advantage+ creative-related automation. If your goal is strict creative testing, any automatic enhancement that changes layout, presentation, or rendering logic can muddy the result. That's not always bad for performance, but it is bad for clean testing.
What to do in practice:
- Review enhancement toggles before publish
- Document whether the test requires fixed creative conditions
- Re-check duplicated ads, because duplicated builds can inherit settings you didn't intend
- Preview after duplication, not just after the original ad is created
A carousel can be spec-compliant and still be the wrong test object if Meta changes how it presents the creative.
The teams that avoid expensive mistakes don't just inspect assets. They inspect settings drift.
Performance Best Practices Beyond Technical Specs
Once the ad meets the technical requirements, the real question starts. Why would anyone keep swiping? Specs keep the ad eligible. Structure makes it perform.
Build for swipe behavior not for design approval
The strongest carousel creatives usually follow a simple logic:
- Card one hooks
- Middle cards prove
- Final card resolves
That can mean product benefits, objections, use cases, ingredients, social proof, offer framing, or a step-by-step demo. The exact structure varies by account. The principle doesn't.
A few patterns repeatedly hold up:
| Creative pattern | Why it works |
|---|---|
| Hook, proof, offer | Gives the swipe a reason to continue |
| Product range by problem | Helps broad audiences self-select |
| Before, process, after | Makes transformation easy to understand |
| Feature stack with final CTA card | Useful when each feature needs its own visual space |
What usually fails is flat sequencing. Ten cards with the same layout and same message don't feel like a journey. They feel like indecision.
When to keep card order fixed
Meta gives you the option to let the platform prioritize stronger cards. That can help when each card stands alone and the account is focused on response efficiency over narrative precision.
But if the carousel is telling a specific story, fixed order is usually the better choice. Otherwise your “proof before setup” or “offer before context” sequence can get scrambled.
Use ordered storytelling when the sequence is the idea. Use performance-led ordering when each card is independently persuasive.
A few final creative rules are worth keeping close:
- Don't bury the best visual on card four
- Don't assume users will reach the last card
- Make each card earn its place
- End with a card that can convert, not just decorate
Good carousel ads respect both realities at once. Some people will only see card one. Others will swipe deep if the structure rewards them.
If your team is spending more time assembling carousel ads than testing them, Rapid Ads is worth a look. It's built for media buyers who need faster bulk uploads, cleaner naming conventions, easier multi-account management, and tighter control over Meta settings that tend to drift inside Ads Manager.
