Amazon UK, eBay UK and Google Shopping compared for feed specs

What to take away

The specs side by side

The table below sets the main limits next to each other. Treat the figures as the position in October 2026 and confirm them in each platform's own help pages before you build a feed, because all three change their rules.

Google has announced a higher image floor of 500 x 500 pixels for every category, with warnings since April 2026 and enforcement from 31 January 2027. Sizing every image to at least 1,000 pixels on the longest side satisfies all three platforms and keeps Amazon zoom working.

Category rules compared across Amazon UK, eBay UK and Google Shopping UK

Category rules are the first thing that breaks a shared catalogue. Each platform asks a different question about the same product.

Amazon UK decides where your product sits in the browse tree and whether you may sell it at all. Some categories are gated, meaning you need approval, invoices or brand authorisation before the listing goes live. The product type also decides which attributes are mandatory, such as size, colour or material.

A wrong browse node can leave a listing live but ranking badly or suppressed. Amazon UK sellers treat category selection as a commercial decision, not data entry.

eBay UK works from a category ID plus item specifics such as brand, type, size and colour. Categories are deep and change, so an ID that was right last year may now be obsolete. A product that fits two categories needs a deliberate choice, because the wrong one can lose the catalogue match to an existing product page.

Google Shopping UK cares less about a browse tree and more about attributes. You choose a Google product category, and the specification also asks for brand, GTIN, condition and availability. Google uses these to match your offer to a product on its side.

The three systems also differ on who maps the taxonomy. Amazon UK and eBay UK give you a finite list to choose from. Google expects you to map your own taxonomy to its product category tree once and reuse it.

For example, a homeware seller may find that Amazon UK wants a specific browse node, eBay UK wants a category ID for the same item, and Google wants a product category plus material and colour attributes. One product, three answers.

Store each platform's category fields in your central catalogue as separate columns. Do not derive one from another, because they are different taxonomies.

Feed formats, field names and identifiers

Feed formats differ in file type, field naming and update method. Amazon has retired legacy XML and flat-file listings feeds in its Feeds API, so UK listings now go through JSON_LISTINGS_FEED for bulk work or the Listings Items API one listing at a time. The Product Type Definitions API supplies the schema for each product type.

eBay UK accepts CSV uploads through Seller Hub File Exchange, where the key columns are Action, Category, Title and ConditionID, plus a product identifier column such as Product:EAN and item-specific columns prefixed C:, for example C:Brand. Copy the header row from the current eBay UK template, because it carries site and currency settings.

Google Shopping UK takes a file upload, a scheduled fetch or the Merchant API, which replaces the retired Content API for Shopping. Its core attributes are id, title, description, link, image_link, price, availability, brand, gtin, mpn and condition, with item_group_id for variants.

Amazon JSON listings use attribute names such as item_name, brand, externally_assigned_product_identifier and recommended_browse_nodes, together with the SKU and product type. The exact required list depends on the product type, so validate against the schema before you submit.

The good news is that the underlying product data is the same. A title, brand, GTIN, price, stock level and image are needed everywhere. The work is in the mapping, not in the data itself.

Choosing the tool that does that mapping is a decision in its own right. The shopping feeds selection checklist covers what to ask before you buy, from field coverage to error reporting.

Identifiers deserve care. Google asks for a GTIN where the manufacturer assigned one, and for brand plus MPN otherwise. Set identifier_exists to false only for goods that genuinely have none, such as handmade or custom items, and never fill the field with placeholders like "N/A". eBay UK accepts "Does not apply" for items without an identifier.

Prices must match the landing page on Google, and on every channel the feed has to show the price the buyer pays. The HMRC VAT rules for feed prices explain what that means in practice.

Variants complicate one catalogue. Amazon UK handles parent and child relationships through variation themes, eBay UK uses multi-variation listings with their own structure, and Google groups variants with item_group_id. A shirt in five sizes and three colours is fifteen child products, and each platform expects a slightly different parent and child structure.

A worked example helps. Take a single SKU: a stainless steel water bottle.

That is the whole method: one product record, three mapping sets, one update cycle.

Fees: the shape of the cost on each channel

Fees do not belong in the feed itself, but they change which channels are worth feeding. The three platforms charge in different ways.

Amazon UK charges a referral fee, a percentage of the sales price that varies by category. Fulfilment by Amazon adds fulfilment and storage fees, and Professional selling plans carry a monthly subscription.

eBay UK charges a final value fee when an item sells, a percentage that varies by category, plus optional promoted listings costs. Google Shopping UK takes no commission, so free listings and paid ads differ by cost per click rather than fee per sale.

Check the current rates on each platform's fee page rather than copying figures from an article.

Use the same cost-of-goods figure per SKU on every channel and compare it with each channel's fee. Google Merchant Centre has a cost of goods sold attribute for margin reporting. The cheapest channel on fees is not always the most profitable once returns and ad spend are counted.

Error handling, issue pages and feed diagnostics

Every feed fails sometimes. The difference is how quickly you find out and what you do next.

Amazon UK Seller Central shows listing quality alerts, suppressed listings and category-specific errors, and a feed submission returns a processing report. A suppressed listing is invisible to buyers, so it is urgent. Common causes are missing attributes, wrong images and pricing errors.

eBay UK shows upload errors in Seller Hub and in the File Exchange reports. Some errors block the listing and others are warnings. eBay UK also flags catalogue match problems and item specific gaps.

Google Merchant Centre has a Diagnostics page and an issues list, grouped by severity. Some issues affect a whole feed, such as a missing image link pattern, and others single items, such as a price mismatch. Fixing the pattern first clears the most errors. The Merchant Centre disapprovals guide covers the common ones.

A practical routine for UK sellers:

Most feed errors come from three sources: missing attributes, data that differs between the feed and the landing page, and category changes made by the platform. The shopping feeds quality checklist covers the checks that keep error rates low. When a listing disappears, check the error report, then the last upload, then any category change on the platform.

Planning a single catalogue that satisfies all three platforms

One catalogue, three outputs. That is the goal, and it is achievable with the right structure.

Start with a central product record that holds the data true everywhere: SKU, GTIN, brand, title, description, price, stock level, weight, dimensions, images and cost of goods. Then add platform mapping tables: browse node and product type for Amazon UK, category ID and item specifics for eBay UK, and Google product category plus required attributes for Google.

Keep titles flexible. Because the limits run from 75 characters on Amazon UK to 150 on Google, store a base title that fits the strictest channel, Amazon UK, plus platform overrides, and generate the final title per channel.

Stock must be near real time. Overselling on one channel because another sold the last unit is a common and expensive error, so use a single stock pool and push updates frequently. How orders then flow back is covered in how eBay and Amazon order sync differs.

A simple build order for a UK retailer:

A guide to shopping feeds tool implementation explains how to phase that rollout so mapping errors surface before the feeds go live.

Finally, document your mapping decisions. When a platform changes a category or an attribute, you need to know which SKUs are affected and why the original choice was made.

Common questions