Shopify product pages: the completeness checklist, channel by channel

Why can a product page that is “complete” on Shopify be rejected elsewhere?

Because each channel plays a different role. Your Shopify store is a storefront you control: a page with no barcode and no weight displays there without complaint. Google Shopping, Amazon or a marketplace are search and comparison engines: they have to match your product with other offers, file it in a category, filter it by size or color. So they demand structured data, and they check it.

The stakes grow with the number of channels. According to Fevad, the French e-commerce federation, marketplaces accounted for 32% of the business volume of online product sales in France in 2025, and 65% of French online retailers sell abroad. A small business that sells on its own store, on Google Shopping and on a marketplace is in fact managing three definitions of the “right” product page. As long as they live in separate Excel tabs, someone ends up copying by hand, and making mistakes.

What do you really need to fill in on Shopify?

On Shopify, the technical minimum is low, the useful minimum much less so. Five blocks matter.

The first is the title and description, written for the customer and not for the stockroom: a title that says what the product is, a description that answers the questions people would ask in a shop (material, use, care, dimensions).

The second is the product category. Shopify now requires each product to be linked to a category of its standard taxonomy; otherwise, it is saved as “uncategorized”. This is no administrative detail: according to the Shopify help center, the category is used to calculate taxes with Shopify Tax, to distribute to channels that expect standardized types such as Google or Facebook, for store filters and even to analyze return reasons.

The third is category metafields. A “shirt” category, for example, unlocks size, neckline or color as standard attributes. Shopify lets you link them to variant options: renaming “black” to “graphite” then updates every variant concerned. This is the embryonic structure that external channels will build on later.

The fourth is the variants and identifiers block: SKU, barcode (EAN), weight, and the options (size, color…). Weight is used to calculate shipping costs; the barcode will be requested by almost every other channel.

The fifth is media. Shopify accepts images up to 5,000 × 5,000 pixels and 20 MB, and states that a square 2,048 × 2,048 pixel format generally gives the best display. Add alt text to each image: it serves accessibility and search engine optimization.

Which fields does Google Shopping require?

Google Merchant Center is the most demanding and the best documented channel. According to its product data specification, consulted in October 2026, every product must include an identifier (50 characters at most, ideally your SKU), a title of up to 150 characters, a description of up to 5,000 characters, a link to the product page, a main image, and an availability and a price that exactly match what your site displays.

Product identifiers follow a precise logic. Brand is required for almost all new products, and Google prohibits filler values such as “N/A” or “Generic”. The GTIN(the 13-digit EAN in France) is required as soon as the product has one; failing that, the manufacturer part number (MPN) takes over. Declaring that an identifier does not exist when it does exposes the page to rejection.

For apparel sold in France, Google adds four required attributes: color, gender, age group and size. Variants of the same model must share the same group identifier (item_group_id), also required in France. You can attach up to ten additional images per product.

A recent change deserves your attention: since April 14, 2026, Merchant Center has been flagging images that are too small, and from January 31, 2027, all product images will have to measure at least 500 × 500 pixels; some will be optimized automatically, the others risk rejection.

If you distribute to Google from Shopify, these are the rules that apply: a page that is perfectly presentable on the store can be rejected for a missing color or a missing GTIN.

What does Amazon expect from a product page?

Amazon thinks in identifiers first. GS1 France points out that the GTIN is required there in most categories and that it is the product identifier recognized by the platform. Without a valid GS1 code, you have to go through an exemption procedure, reserved for certain cases, which delays going live. So check your EAN column before anything else.

Next comes the content. Up-to-date seller guides, such as the one published by the agency Nanakia in March 2026, cite a standard limit of 200 characters for the title, shorter in some categories such as clothing, five bullet points, and a main image on a pure white background, where the product fills at least 85% of the frame, at 1,000 pixels minimum to enable zoom. These rules vary by category and change over time: the reference remains Seller Central help, to be checked for your category before preparing your files.

For a small business, the essential point lies elsewhere: Amazon expects a set of attributes specific to each product type. A lamp is not described like a sweater. This is exactly what a PIM lets you anticipate with the notion of product family: the list of attributes expected for each type is defined once, not rediscovered at the moment of rejection.

And for a fashion marketplace such as Zalando or La Redoute?

Fashion marketplaces add a specific layer: the precision of variants. Each size and each colorway must be a separate variant, with its own EAN code, its own size label and its own visuals. GS1 France states, for example, that the GTIN is a prerequisite for selling fashion and home decor products on La Redoute. Size charts, detailed composition, care instructions and a consistent set of photos per colorway are generally expected; the exact requirements are set out in each platform’s partner guide, provided when the seller account is opened.

The classic pitfall lies in labels. Your store displays “Midnight blue”; the marketplace expects a color chosen from its own closed list. Your size chart mixes “38”, “M” and “One size”. Without a mapping defined once and for all, each new collection relies on a conversion table maintained by hand.

Which fields does each channel require?

* Secondary source (seller guide, March 2026): to be confirmed in Seller Central help for your category.

Which pitfalls come up most often?

The first pitfall is forgotten weight. It blocks nothing on the store, until the day the calculated shipping costs are wrong or a channel asks for it. Fill it in for each variant, in a single unit.

The second is the mishandled EAN: missing, duplicated, entered at model level instead of variant level, or converted by Excel into scientific notation. One EAN per variant, stored as text, checked once: that is the entry ticket to almost every external channel.

The third concerns secondary images. A single photo is enough on the store, not elsewhere. Plan at least a front view, a detail view and, for fashion, an on-model photo per colorway; check the resolution before Google’s January 2027 deadline.

The fourth, finally, is variant attributes: a color described in free text (“marine”, “navy”, “navy blue”) or inconsistent sizes. A closed option list, then a mapping table per channel, solve the problem at the source.

How do you track completeness without a spreadsheet?

By declaring once and for all what each channel requires, then letting the tool do the measuring. That is the logic behind cocoPIM’s completeness score by channel: for each channel you define the required attributes (for example GTIN, brand, color, size, gender and age group for Google Shopping), and each product page shows the percentage reached and flags the required fields that are still empty. The same page can be at 100% for your store and at 70% for Google: that is normal, and it is precisely the useful information.

Let’s take a fictional example with figures. A leather goods brand has 400 products on Shopify and wants to open Google Shopping. The score shows that all its product pages are complete for the store, but that 130 have no EAN and 90 have no color entered as an option. Instead of discovering these gaps one Merchant Center rejection at a time, the team knows which pages to complete, and publishes the ready ones first.

Around the score, Excel or CSV import brings in your existing data, the media library centralizes visuals, the Shopify connector syncs the store, and the built-in AI, powered by Mistral, suggests descriptions and translations based on the attributes entered: always to be reviewed.

Frequently asked questions

What fields are required for Google Shopping?

For all products: an identifier, a title of up to 150 characters, a description of up to 5,000 characters, a link, an image, availability and a price. Brand is required for most new products, and the GTIN whenever one exists. For apparel sold in France, color, size, gender and age group are added.

Why is my Shopify product page rejected in Google Merchant Center?

Because Merchant Center applies its own specifications, which are stricter than your store’s. The most common causes are a missing or invalid GTIN, a missing brand, apparel attributes left empty, a price or availability that does not match the site, or a non-compliant image.

Do you need an EAN code to sell on Amazon?

In most categories, yes: according to GS1 France, the GTIN is the product identifier recognized by Amazon. Products that do not have one must go through an exemption request, which slows down going live. It is better to assign GS1 codes to each variant before opening the channel.

How do you know whether a product page is complete for each channel?

By defining, for each channel, the list of its required fields, then measuring the gap page by page. A spreadsheet can do this for a few dozen products; beyond that, a completeness score per channel, like the one in a PIM, immediately shows what is missing and where.

The trademarks mentioned (Shopify, Google, Amazon, Zalando, La Redoute, Cdiscount, Microsoft Excel, Mistral AI, GS1) belong to their respective owners. cocoPIM is published by T2ND SAAS and is not affiliated with any of them. Platform requirements change and vary by category: check them in their official documentation before publishing your product pages. The leather goods brand example with figures is fictional.