Marketplace product image requirements, in one table

Every marketplace publishes its image requirements somewhere, and every marketplace publishes them differently. One states a minimum in pixels, the next a minimum for the longest side only; one says "pure white", another says "white or light"; one enforces the rules automatically at upload and another leaves them to a human reviewer who may or may not care this week.

If you sell the same catalogue across four channels, you are reconciling four documents in your head. This page reconciles them once, in one table, with the official source linked on every row and the date we last opened it.

Two things we will not do here. We will not restate a requirement we could not find on the marketplace's own seller documentation — the cell says not stated instead, because "we could not verify this" is more useful to you than a confident guess. And we will not quietly let this page rot: each row carries its own verification date, so you can see exactly how old a claim is before you act on it.

Which rules are enforced automatically, and which are not

This distinction matters more than the numbers, and no marketplace explains it clearly.

Enforced at upload, by software. Pixel dimensions, file format, file size, and on several marketplaces the pure-white background check. These fail immediately and visibly: the upload is rejected, or the listing sits in an error state. They are also the easy ones — they are pure measurement, which is why we built a free checker that measures them in your browser before you upload anything.

Enforced later, by a human or a sweep. Text and logos on the main image, watermarks, props that are not part of the product, misleading colour. These pass upload cleanly and surface weeks later as a suppressed listing, usually on your best-selling SKU, usually without a useful explanation. Nothing automated catches them reliably — including our checker, which says so rather than pretending.

Never enforced, but costly anyway. Recommended sizes. A marketplace that accepts 1000 px and recommends 2000 px is telling you where zoom stops working. Nobody will reject the listing; shoppers simply do not zoom, and a shopper who cannot inspect a sofa's fabric does not buy the sofa.

The rule that fails most furniture listings

The requirement that the product must fill roughly 85% of the frame is the one furniture sellers break most often, and it is the least intuitive of the set.

Marketplaces measure the bounding box — the smallest rectangle containing the product — as a share of the image area. For a compact, roughly square product this matches what your eye expects. For furniture it often does not:

• A tall, narrow bookcase in a square frame can fill the height completely and still occupy well under 85% of the area, because the sides are empty.
• An L-shaped sectional has a large bounding box but little ink inside it. It usually passes, and it looks sparse.
• A dining set photographed as a group is measured as one bounding box, which is why group shots pass the fill rule and then fail a different one, for showing multiple products on the main image.

The practical fix is the same in every case: crop tighter, keep a small consistent margin, and shoot the product square-on so the bounding box and the product actually agree.

“White background” usually means one exact colour

Where a marketplace specifies pure white it means RGB 255,255,255, and it checks. A studio sweep photographed without enough light lands somewhere around RGB 240–250 — which looks white on your monitor, looks white next to the product, and fails.

Three things push a background off pure white without anyone noticing:

• Underexposure. The sweep is lit for the product, not for the paper behind it.
• JPEG compression. Edges between a dark product and a white ground get smeared into grey pixels, which is also why a background can pass in the corners and fail near the product.
• Colour profile conversion. An image exported in Adobe RGB and interpreted as sRGB shifts, sometimes just enough.

The check is one you can do in seconds: open the image, sample a corner pixel, and read the value. Or drop it into the checker, which samples the whole border and tells you whether the backdrop is pure white, plain but not white, or a scene.

Main image and gallery images are governed differently

Almost every rule above applies to the main image — the one that appears in search results. Gallery images are far more permissive, and that is where furniture actually sells.

The pattern that works across channels: a compliant white-background packshot in slot one, then room scenes, a scale or dimension image, a detail shot of the material, and where the marketplace allows it, an image showing the product from the back or underneath. The packshot satisfies the marketplace. The rest satisfies the shopper, who is trying to work out whether a 2.1 m sofa fits a room they are standing in.

This is also the answer to a question we get often: no, you do not have to choose between a compliant catalogue and a persuasive one. You need both images, and the requirements only constrain one of them.

The change already on the calendar: 31 January 2027

One dated change is worth planning for now, because it affects every seller who advertises and it has already been announced.

From 31 January 2027, Google Merchant Center requires every product image to be at least 500 × 500 pixels, in every category. Today the floor is 100 × 100 for most products and 250 × 250 for apparel; the separate apparel threshold disappears entirely. Images below the new minimum stop being eligible for Shopping ads and free listings.

If you sell furniture this is unlikely to bite on your hero images, which are already well above it. Where it bites is the long tail: legacy SKUs, supplier-provided thumbnails, swatch and accessory images, and anything migrated from an older platform at whatever size it happened to be. Those are exactly the images nobody looks at until the disapprovals arrive.

Two other dated items sit alongside it, and both matter specifically if you use AI imaging. Amazon requires a photorealistic AI-generated person in an image to carry the contains-synthetic-performer XMP tag. Allegro requires AI-generated or AI-modified images to be declared from 2 August 2026. Neither is about image quality; both are disclosure rules, and both are new enough that most sellers have not read them.

What actually fails, measured on real supplier images

We ran a set of real, supplier-provided furniture packshots — the kind a reseller receives from a manufacturer and uploads unchanged — through the same measurement the checker uses. The results were not what we expected.

Every one of them passed on the things sellers worry about. Clean pure-white backdrops, 1200 × 1200 pixels, correct file type. And the product filled 14%, 28% and 34% of the frame respectively, against a requirement of 85% on Amazon and 85–95% on OTTO.

This is the failure mode that hides. Resolution problems are obvious — the image looks soft. A background that is not quite white gets flagged at upload. But a product photographed small in the frame looks completely fine, passes every check a person would think to do, and quietly fails the one rule nobody measures by eye.

It is also the easiest to fix, because it needs no reshoot: the original file is large enough that cropping in gives you both a tighter frame and a compliant one. The only thing you need is to know it is happening.

How to keep up when the requirements change

Image requirements change quietly. There is rarely an announcement; a help page is edited and listings uploaded the following week behave differently.

Three habits catch it early:

• Watch the rejection reasons, not the rules. A sudden cluster of rejections on images you have uploaded successfully for a year is the signal a spec moved.
• Re-check before a bulk upload, not after. Checking one representative image before pushing 300 costs a minute and saves a re-export.
• Keep a master at higher resolution than any marketplace asks for. When a minimum rises — and minimums only ever rise — you re-export instead of re-shooting.

That last point is the cheapest insurance in this whole document. Every minimum in the table above has gone up over the past decade and none has come down.

Related: Wayfair image requirements in detail · Shopify furniture product pages · AI staging vs 3D rendering