You uploaded a clean product shot as your Amazon main image. Amazon suppressed the listing anyway. Or worse, the listing went live for a day, then the image quietly reverted to the old one you were trying to replace. If you have been fighting this for a week, you are not doing anything obviously wrong. You are tripping rules that Amazon enforces by algorithm, and those rules are stricter and more literal than the help pages make them sound.
We have fixed this exact problem hundreds of times across seller catalogs. Here is what is actually happening and how to get a compliant Amazon main image without another studio session.
The white background is not "white enough"
The Amazon main image rule is not "a white background." It is pure white, RGB 255, 255, 255, on every channel. That is the number the automated scanner checks against, and it does not round up.
Here is where sellers get burned. Your background looks white on your monitor. It reads as 252 or 254 in the corners. To your eye those are identical. To Amazon's system a 254 is a colored pixel, and colored pixels around a product on the main image are a violation. The listing gets rejected, or if it slipped through earlier you get a listing suppression on a later sweep.
Two things cause the near-white problem more than anything else:
- Editor "white point" adjustments that stop short of 255. A lot of one-click background removers lift the background to a bright grey, not true white, because it looks softer and more natural. Softer fails the check.
- JPEG compression casting the edges. Save a white background as a heavily compressed JPEG and the compression introduces faint color blocks along high-contrast edges. Those blocks sit at 250-253 and the scanner flags them.
The fix is not "make it whiter" by dragging a slider. You need the background forced to a flat 255, 255, 255 with a clean cut around the product edge, then exported without the compression artifacts that reintroduce color.
The 85% rule is why small products get pulled
Amazon expects the product to fill at least 85% of the frame. This is the rule that quietly kills listings for small items: jewelry, phone cases, supplements, cables. Sellers center a small product in a big white square because it looks tidy, and the product ends up occupying maybe half the canvas.
That reads as "product too small" to the scanner and as a weak thumbnail to shoppers. On a phone screen your listing is a fingernail-sized square in a grid of competitors. If their product fills the square and yours floats in the middle with white margins on all sides, you lose the click before anyone reads a word.
Fill the frame. The longest side of the product should run close to the edges, with only a thin margin of white so nothing looks cropped. This single change often lifts click-through more than any bullet-point rewrite.
No text, no logos, no props, no exceptions
The main image has to be the product and nothing else. That means:
- No text or graphics overlaid on the image, including "best seller" badges, size callouts, or your brand name.
- No watermarks.
- No props, accessories, or backgrounds that are not part of what you ship. If the buyer does not receive the wine glass, the wine glass cannot be in the main shot.
- No mannequins for most categories, and no color swatches or borders.
Sellers routinely lose listings because a designer added a small logo in the corner for brand consistency, or because a lifestyle shot got promoted to the main slot by accident. Save the badges, the infographics, and the lifestyle scenes for your secondary images, where they belong and where they actually help you sell.
Why the old image keeps coming back
This is the part that makes people think they are losing their minds. You delete the bad image, upload the good one, save. A day later the old image is back. You upload again. Same thing.
A few real causes, from the seller forums and from catalogs we have untangled:
- Browser and CDN cache. Amazon serves images through a cache. After a change you are often looking at the old cached file, not the live one. A hard refresh with CTRL+F5 clears your local view. The public image can still take hours to propagate.
- Ghost images on old SKUs. If the same product exists under a retired SKU or a variation parent, Amazon can pull the image from the wrong record and overwrite what you just uploaded. Remnants of old SKUs clog the catalog and keep resurfacing.
- Multi-region records. Sellers listing in several marketplaces sometimes have one region feeding an image back into another. You fix it in the US, the UK record pushes the old file back.
The workaround that clears most stubborn cases: remove every image on the listing, click save, wait about 10 minutes, then add the new main image, save again, and give it a full 24 hours before you judge the result. If it still reverts, the problem is a duplicate record and you need Amazon's Catalog team to merge or purge it, not another upload.
Get a compliant Amazon main image in minutes, not another shoot
Most rejection loops are not a photography problem. They are a preparation problem. The shot is fine; the background is a near-white grey, the product floats at 60% frame fill, and a stray reflection reads as a prop. Rebooking a photographer to fix that is slow and expensive, and the reshoot often comes back with the same near-white backdrop.
This is the gap we built ListingRVA AI to close. You upload the photo you already have. It cuts the product cleanly, forces the background to a true 255, 255, 255, scales the product to fill the frame to Amazon's 85% expectation, and exports at a resolution and compression setting that will not reintroduce edge color. You get a main image built to pass the automated check, without booking a studio to reshoot a product you already photographed.
We walk through how sellers use it end to end in the ListingRVA AI case study, including before-and-after examples of listings that were suppressed and then reinstated.
A pre-upload checklist that actually prevents listing suppression
Before you upload an Amazon main image, run through this. It takes two minutes and saves the week-long loop.
- Sample the background corners. Use the color picker in any editor. Every corner should read 255, 255, 255. If any channel is below 255, it is not compliant yet.
- Check frame fill. The product's longest side should run close to the edges. If there is a fat white margin on every side, the product is too small.
- Strip everything that is not the product. No text, logos, watermarks, badges, or props. Look at the corners specifically; that is where stray marks hide.
- Confirm resolution. Aim for at least 1600 pixels on the longest side so zoom works. Small files also compress worse, which is what casts the edges.
- Export clean. Save as a high-quality JPEG or PNG so compression does not add color to the edges you just cleaned.
Stop guessing at the algorithm
The frustrating truth is that Amazon's image scanner does not tell you which rule you failed. It just hands you a listing suppression, and you are left toggling files and refreshing the page. Once you know your Amazon main image needs an exact 255 background, an 85% frame fill, and a product-only frame, the fixes stop being guesswork.
If you are drowning in a large catalog, or you keep hitting the revert loop across multiple marketplaces, that is usually a workflow problem worth automating rather than a per-image chore. We help sellers and studios build exactly that kind of pipeline through our AI tools development services, from single-store fixes to bulk processing across thousands of SKUs.
Frequently Asked Questions
How do I check if my background is a true 255, 255, 255 white?
Open the image in any editor with a color picker, or eyedropper tool, and click each of the four corners. Read the RGB values. If every channel shows 255 in all four corners, you are compliant. If any corner reads 254, 252, or a light grey, that is the near-white that fails Amazon's scan even though it looks white to your eye.
Amazon suppressed my listing but did not say why. How do I find the actual reason?
Amazon rarely names the specific rule. Work the checklist in order: sample the corners for a true 255 background, confirm the product fills close to 85% of the frame, then scan every edge and corner for stray text, logos, or props. One of those three trips almost every main image suppression. Fix all three before re-uploading rather than guessing at one.
Does the 85% frame fill rule apply to every product category?
The 85% guideline is Amazon's general expectation for the main image, and it matters most for small items like jewelry, cables, and supplements that sellers tend to float in a big white square. A few categories have their own quirks, but you are almost never penalized for filling the frame. You get pulled for leaving fat white margins on all sides.
I uploaded the correct image but it still shows the old one. How long should I wait?
Do a hard refresh with CTRL+F5 to clear your local browser view first. The public image runs through Amazon's cache and can take several hours to propagate. Give any upload a full 24 hours before you decide it failed. If the old file genuinely reverts after that window, you are looking at a duplicate record, not a cache delay.
Why does my old image keep coming back even after I delete and re-upload it?
Persistent reverts usually mean a duplicate or retired SKU, a variation parent, or another marketplace record is feeding the old file back in. Try removing every image, saving, waiting ten minutes, then adding the new main image and saving again. If it still reverts, open a case with Amazon's Catalog team to merge or purge the ghost record. Another upload will not fix a duplicate.
Can I fix a rejected Amazon main image without reshooting the product?
Usually, yes. Most rejections are a preparation issue, not a photography one. The shot is fine but the background is a near-white grey, the product sits at 60% fill, or a reflection reads as a prop. A tool like ListingRVA AI cuts the product, forces a true 255 background, scales to fill the frame, and exports without compression color, all from the photo you already have.
Will saving as PNG instead of JPEG stop the edge color problem?
PNG helps because it uses lossless compression, so it will not introduce the faint color blocks that heavy JPEG compression casts along high-contrast edges. A high-quality JPEG at a low compression setting also works. The failure comes from aggressive compression on a small file, so keep the resolution up at 1600 pixels or more on the longest side and avoid maxing out the compression slider.
If your listings keep getting suppressed and you want a second set of eyes on why, tell us what you are seeing and we will point you to the specific rule you are tripping.