Amazon Product Video Requirements, Plus a Review Pass Before You Upload
What Amazon checks in a listing video, why so many get sent back, and the pre-upload review pass I now run in my agency for every ASIN in a catalog.
Amazon listing videos are where my video-editing agency made a review mistake I still think about, on a batch for an ecommerce brand that I signed off myself on a Friday evening. I had read the Amazon product video requirements a few months earlier, right, and I basically assumed my team knew them as well as I thought I did. One video had an end card calling the product a best seller, another had the price burned into the last frame, and both came back from Amazon rejected within days.
The catch here is that the text was burned into the export and the same end card template sat on every video in the batch, so there was no quick fix, and we reworked every end card, exported everything again and uploaded it all again. The brand lost several days of live listing time and we ate the re-edit ourselves, because at the end of the day it was my sign-off that let it through, and to be very honest the edit itself was good, so what actually cost us was a review pass that did not exist.
So this post does two things, right. It summarises the rule categories Amazon checks, and then it walks through the review pass I now run before any listing video goes near Seller Central, set up inside PlayPause so it holds across a whole catalog.
Amazon product video requirements at a glance
When people search for Amazon listing video guidelines they usually want one neat table of numbers, but the product video specs Amazon publishes have shifted more than once, and they can differ between the listing itself, brand content and ad placements. So I'm deliberately not printing resolution, length or file size figures that could be stale by the time you read this, and I'd rather give you the categories, right, because the categories have stayed very very stable and they are what your checklist should be built on.
The first category is technical, covering the file container and codec, the resolution, the aspect ratio, the frame rate of the export, the maximum file size and the running length. In my experience this fails least, because editors export from saved presets, and when a preset is wrong it is wrong on every file, so somebody notices by the second video. The one trap I keep seeing is a team reusing a preset built from YouTube's recommended upload encoding settings, when that page only describes what YouTube wants and Amazon's own spec page is the one that decides your listing.
The second category is content, and this is where the real trouble lives. Amazon wants the video to be about the product on that page, so the rules basically push against anything that sends the shopper somewhere else or makes a promise Amazon cannot verify. For instance, that means prices, discounts and limited-time offers, mentions of competitors, review quotes and star ratings, website addresses and social handles, and claims like best seller or number one, while health, safety and performance claims get extra scrutiny depending on the category, and so on.
The third category is the metadata around the file, meaning the video title, the thumbnail and which products the video gets attached to, and I'm pretty sure this is the one most teams forget exists, because it gets filled in at upload time by whoever is uploading and not by the editor who knows the video.
In practice, before a batch starts, one person opens Amazon's current help page in Seller Central and writes a single dated spec entry, with the container, codec, resolution, aspect ratio, frame rate, maximum file size, maximum length and thumbnail rules copied word for word, plus the page they came from. If the brand also runs video ads, that gets a separate entry, because ad specs live on a different page. The editor builds the export preset from that entry and nothing else, right, so the check later is simply whether the export matches the dated entry, which anyone can confirm in under a minute.
Specs for format, length and file size change, so check the current seller help page before each batch instead of trusting an old screenshot.
The most common reasons listing videos get rejected
When an Amazon video rejected notice lands, the first reaction on most teams is to blame the editor, and I did exactly that the first time, right, but when I went back through our own rejections the editing was almost never the problem.
Burned-in text is the biggest one in my agency, a price in a corner, a discount badge, a best seller line on the end card, or a website address carried over from the brand's Instagram cut of the same video. Next is anything that points off Amazon, like a QR code or a social handle, and then there is the background, for instance a competitor's box sitting on a shelf behind the product in a lifestyle shot, which nobody sees until they pause on that exact frame.
Then there is the spoken stuff, where the voiceover promises something the listing copy never claims, like durability or a health benefit, and the mismatch gets caught. Finally there is the attach mistake, where the video showing the blue variant ends up on the green listing, and it is the same family of problem I wrote about in catching wrong product shades before launch.
What ties all of these together, right, is where they sit, in the last few seconds, in a lower third that flashes past, in one sentence of voiceover or in a field on the upload form. Nobody watches the final seconds of the fourteenth video in a batch with full attention, trust me on any level.
The mistakes that get a listing video bounced almost never live in the part of the video people actually watch closely.
Building a pre-upload review checklist
After that Friday mistake the first thing I changed was to stop relying on anyone's memory of the rules, including mine, and to write them down where the reviewer is already looking. In PlayPause that place is Playbooks, which on the Agency plan and up give you a per-client document of creative style and direction. The part that matters here is that a checklist inside a playbook is not done until every item is actually ticked, and that is the backbone of how we do video proofing for listing content.
We keep two layers, right. A brand playbook holds the permanent rules, meaning the Amazon content categories above, the dated spec entry and anything the brand's legal team told us never to say. A folder playbook sits on each catalog folder with what changes per launch, like approved claims for this product line, the variant names and which ASIN each file belongs to. Because a playbook can be shared as a no-login page, the brand manager can read exactly what we check without an account.
The checklist stays short on purpose, because a long checklist gets ticked without anyone reading it, and it borrows a lot from the general creative review checklist before sending anything to a client.
- No price, discount or offer text anywhere in the video
- No website, QR code, handle or contact details on screen
- No competitor products or logos anywhere in frame
- Every spoken claim matches the approved listing copy
- The variant on screen matches the ASIN it will attach to
- Export settings match the dated spec entry
The title and thumbnail get their own line on the upload sheet instead, because they are checked by whoever uploads and not by the person watching the cut, and does that make sense, right, you want each check owned by whoever can actually see it.
Flagging claims and on-screen text frame by frame
A checklist tells you what to look for, but whether the fix happens depends on how the note reaches the editor, and that is where our old habit of WhatsApp messages saying remove the price near the end fell apart. The editor would remove one price, miss the second one three seconds later, and we would go another round.
Now every problem gets a frame-accurate comment on that frame, so the editor lands on the precise moment instead of hunting for it. When a lower third stays up for a few seconds I use a range comment across the whole stretch it is visible, and when the problem is small, like a competitor logo in the background, I draw on the frame and circle it, because a circle removes the guessing. Threaded replies keep the conversation on that one note, so the editor replies fixed and I check that exact spot in the next version.
For voiceover claims, AI transcription on the Agency plan and up gives me a clickable transcript, so I read the script against the approved listing copy, click any line that makes a promise and jump to that moment to leave a note. If the editor cuts in Premiere Pro, the PlayPause panel inside Premiere brings those comments into Premiere itself, and clicking one jumps the playhead to that frame. For teams on DaVinci Resolve or Final Cut Pro the flow is export, upload, review, then work through the notes in the editor.
That last step matters more than it looks, right, because the new export stacks on the same card as MV2, and side-by-side version compare, on Agency and up, plays MV1 and MV2 together so I can confirm the re-export changed only the end card.
Reviewing videos for a whole catalog at once
One hero video is easy, and the hard part of any Amazon seller video program is the catalog, one video per ASIN across dozens of products and variations, all due on the same launch date.
Say a brand hands us eight kettles in three colours each, which is 24 ASINs and 24 videos, right. We make one folder per kettle, and its folder playbook lists the three colour names with their ASINs. Then we push all 24 exports up in one go with bulk upload for big batches, which takes 100+ files at once on every plan, so each ASIN lands on its own card with the ASIN in the file name.
The brand manager gets one share link, opens it in a browser with no account and no install, and comments on the videos they own, while every comment lands in our Slack channel so nobody on my side keeps refreshing. Who-watched analytics show who opened the link and when, so if the brand says they approved something I can see whether they actually watched it.
Link expiry matters too, because brand approvals can drag, and on Creator links last 30 days and files expire, while Agency links last 90 days and Enterprise links never expire. I wrote the longer version in reviewing hundreds of product videos, and our brand-side setup is on the page for brand marketing teams.
Tracking status per product listing
The last piece is knowing where every video stands, because once you have 24 ASINs in flight someone asks whether the green kettle video went live and nobody knows, right. We used to run this on a spreadsheet next to the review tool, and it was always a day behind reality, you see what I mean here.
Now each card carries custom statuses, and ours are roughly Needs claim check, Client approved, Uploaded to Amazon, Rejected by Amazon and Live. When Amazon bounces a video, the card moves to Rejected by Amazon, the reason goes in as a comment on the exact frame that caused it, the editor exports MV2, and the card walks back through the same path, so every ASIN's history lives in one place. I covered the approval side in product video approvals across SKUs, and if you also make store videos for apps, the app store preview review post applies the same idea of catching errors before you submit.
a spreadsheet of ASINs, notes in email and nobody sure which export went live
one card per ASIN, custom statuses and every note pinned to the exact frame
Pricing is flat per workspace and not per member, so Agency at $19 a month covers 50 members, 250 GB, unlimited projects and Playbooks, and if you want to check that against the re-edits you currently eat, the video review software ROI breakdown walks through the maths.
Frequently asked questions
What are the current Amazon product video requirements for format and length?
I don't print exact figures, because Amazon updates them and they differ between listing videos, brand content and ads. Before every batch, open the current help page in Seller Central, copy the format, resolution, length and file size limits into your brand playbook and date the entry, right, so nobody on the team trusts a screenshot from last year.
Why was my Amazon video rejected when it looked fine?
In my agency the answer is almost always something small that nobody watched closely, like a price or best seller line on the end card, a website on screen, a competitor product in the background, or a voiceover claim the listing does not support. Scrub the first and last seconds, read the transcript against your listing copy, and check the variant on screen matches the ASIN.
Does PlayPause upload the finished video to Amazon for me?
No, and I'd rather be clear about that. PlayPause is the review step before upload, so your team and the brand comment on the frame, approve the final version, and then whoever manages the seller account exports that approved version and uploads it through Seller Central. The custom status Uploaded to Amazon is simply how we track that the human step actually happened.
Which PlayPause plan fits an Amazon catalog?
For most catalog work I'd pick Agency, because it adds Playbooks for the rule checklist, AI transcription for voiceover claims, side-by-side version compare and 90-day links. Creator at $9 a month works for a small seller with a handful of ASINs, but its links last 30 days, it caps you at 30 projects and it has no Playbooks.
Where to start
If your listing videos keep coming back, put the checklist above into a brand playbook and run your next batch through it inside PlayPause. Every plan comes with a 7-day free trial, and you can compare Creator, Agency and Enterprise on the PlayPause pricing page.
So yeah. That's my way of saying it.
Saumyajit co-founded PlayPause after years watching review and approval quietly eat creative teams' deadlines. He writes about the workflow side of video, feedback, versioning, and getting to a clean sign-off.
Related resources
Keep reading
Bring your team into one review space
Centralize feedback, lock approvals, and deliver faster, start free today.
Sign Up for Free