Every tool that sells Amazon video marketing quotes a conversion lift figure. The figures range from 9% to 80%, none of them cite a methodology, and most trace back to a vendor's own blog citing another vendor's blog.
Ignore all of them. Here is what is actually true about video on Amazon, and how to find out what it does for your listing.
Where video actually appears
Four distinct placements, often confused with each other.
1. The image carousel
A video thumbnail among the product images, near the top of the listing. The most valuable placement, because it sits where every shopper looks.
Available to Brand Registered sellers. This is the one worth doing first.
2. Related Video Shorts
A carousel further down the page, below the fold. Your video can appear here — and so can competitors' videos, including on your own listing.
That last part is worth sitting with. A weak listing with no video of its own is real estate competitors can occupy.
3. A+ Content modules
Brand-owned content below the main listing. Video modules are available in Premium A+ for some sellers.
Lower impact than the carousel, because fewer shoppers scroll that far, but useful for products that need explanation.
4. Sponsored Brands Video
Paid placement in search results. Different animal entirely — this is advertising, not listing content, and it is bought by the click.
What the video has to do differently on Amazon
The context is not social. A shopper on Amazon has already decided they want the category and is choosing between options. They are not being entertained into interest; they are resolving a doubt.
This changes everything about the video:
- No hook needed. The attention is already yours. Spending three seconds on a scroll-stopper wastes three seconds.
- Answer the objection. Whatever your negative reviews complain about, show that thing. If people say it is smaller than expected, show it next to a hand in the first two seconds.
- Assume no sound. Amazon video does not autoplay with audio. If your video needs a voiceover to make sense, it does not make sense.
- Short. Fifteen to thirty seconds. This is not a brand film.
- Show scale early. The single most common Amazon return reason across categories is "not the size I thought." A hand, a coin, a familiar object — early.
The best-performing Amazon listing videos look almost boring compared to social ads. That is correct.
The technical requirements
Amazon's published requirements move around, so check Seller Central before producing rather than trusting any article including this one. As of writing, broadly:
- Formats: MP4 or MOV
- Aspect ratio: 16:9 for most placements; vertical is increasingly accepted
- Resolution: 1080p recommended, 720p minimum
- Length: under a few minutes technically, but 15–30 seconds practically
- Audio: must be present and not distorted; must not be silent for the whole clip
- No prohibited content: price claims, "best seller" claims, links elsewhere, contact details, time-limited offers
That last bullet rejects more videos than any technical issue. Amazon will not approve a video that says "on sale now" or shows a URL.
If your file is too large to upload, the free browser compressor will get it under a target size without sending it anywhere.
How to actually test whether it works
This is the part the conversion-lift statistics skip, and it is the only part that tells you anything about your own listing.
The honest constraint: Amazon does not give you a clean A/B test for listing video. You cannot split traffic between video and no-video. So you are working with before-and-after, which is weaker and easier to fool yourself with.
A defensible approach:
- Pick a listing with stable traffic. Not a new product, not one with a campaign running.
- Record two weeks of baseline. Sessions, unit session percentage (Amazon's conversion rate), and return rate.
- Add the video. Change nothing else. Not the title, not the images, not the price. One variable.
- Wait four weeks. Two is not enough; weekly seasonality will fool you.
- Compare unit session percentage, not sales. Sales move with traffic. Conversion rate is closer to a measure of the listing.
- Check return rate too. A good video can lift conversion and lower returns simultaneously — that is the real win, and it is invisible if you only watch sales.
Then repeat on a second listing. One listing moving is a coincidence. Two is a signal.
What will mislead you
- Seasonality. Compare against the same period last year if you have the data.
- Competitor price moves. Your conversion can fall for reasons entirely outside your listing.
- Running the test during a promotion. Everything moves; nothing is attributable.
- Changing the main image at the same time. The most common contamination, because people improve everything at once.
Where video helps most
Not every product benefits equally. From what the mechanism predicts and what sellers consistently report:
Strong candidates:
- Anything where size is misjudged
- Products with assembly or setup
- Items where texture or material matters — fabric, finish, flexibility
- Multi-function products where the second function is not obvious from photos
- Categories with heavy review complaints about expectations
Weak candidates:
- Commodity items bought on price
- Products fully described by one photo
- Replenishment purchases where the shopper already knows the product
If your listing is in the second group, your effort is better spent on the main image and the first three bullets.
Producing them at catalogue scale
The real obstacle for most sellers is not whether video works. It is that it works per listing, and they have two hundred listings.
A shoot amortises badly across a long tail. This is the case where generated video is the practical answer — not because it is better than a shoot, but because the alternative for SKU #147 is no video at all.
Faysell takes an Amazon listing URL or a product photo, pulls the selling points out of the listing, and generates a vertical ad. Worth knowing honestly: Amazon product pages sometimes serve anti-bot responses to URL imports, in which case uploading the image directly works. And close-up hand interactions — exactly the shots that demonstrate scale best — remain the weakest area of generative video, so those are better filmed if the product depends on them.
The summary
Video on an Amazon listing is an objection-handling tool, not an attention tool. The shopper is already interested; your job is to resolve the doubt that is stopping them, usually about size or material.
Ignore the conversion-lift statistics, including the flattering ones. Run a four-week before-and-after on a stable listing, watch unit session percentage and return rate, and repeat it on a second listing before believing the first.