Product structured data (JSON-LD)
Product structured data (JSON-LD)
feed attribute: structured_data (on the product page, not a feed column)
Google reads the Product markup on your product page to build merchant listings and to check the page against your feed. We read the JSON-LD on your product pages and tell you what is there — or that we could not read it.
Google Merchant Center
Google's merchant listing documentation: Product markup makes a page eligible for merchant listing experiences (the shopping knowledge panel, Google Images, popular product results, product snippets). name, image and offers are its required properties; brand.name and the gtin properties are recommended ones. It also recommends putting the markup in the page's initial HTML.
Meta
Meta's catalogue is a feed rather than page markup: Meta matches items on the attributes you submit and the link field, and asks for no JSON-LD. These two checks are therefore Google-only — a Meta catalogue is not held to them.
What the requirement says
- Product structured data is JSON-LD on the product page: a script block with @type Product and its properties (name, image, offers, brand, gtin or mpn, sku).
- Google lists name, image and offers as the required properties of Product markup, and brand.name, gtin, mpn, sku and description as recommended ones.
- Google's guidance for merchants: "If you're a merchant optimizing for all types of shopping results, we recommend putting Product structured data in the initial HTML for best results."
- The markup describes the page only if a shopper can buy the product there: pages that link out to other sellers are not eligible.
“When you add Product markup to your page, it can be eligible for display in merchant listing experiences on Google Search, including the shopping knowledge panel, Google Images, popular product results, and product snippets.”
Why products lose visibility when it fails
- Without a Product node the page is not eligible for merchant listing experiences — the item can only appear in the ordinary search results, not the product treatments.
- Google also uses the page to check what your feed claims: price, availability and identity are cross-checked against the page, and the markup is what makes that comparison automatic and reliable.
- A brand in the markup is how the page is tied to the same product sold elsewhere, and it is the brand your feed submits — a page that publishes none is a weaker match.
- Unparsable JSON-LD is worse than none: Search Console reports it as unparsable structured data, and the whole block is ignored.
The checks we run on it
Straight from the audit engine, with the basis each one is judged on. “Commonly suppresses” means the failure is the kind that stops a product serving, so it is weighted more heavily in your score.
| Check | Basis | Requirement | What to do |
|---|---|---|---|
No Product structured data on the product page minorstructured_data_missing | recommended | No Product structured data (JSON-LD) on this item's own product page. Google reads it to build merchant listings — the shopping knowledge panel, Google Images, popular product results and product snippets — and recommends putting it in the page's initial HTML; without a Product node the page is not eligible for those. | Add a JSON-LD block to the product page template: "@type": "Product" with name, image and offers, plus the brand and the gtin or mpn. In Shopify that is a product snippet in the theme; in WooCommerce an SEO plugin (Yoast, Rank Math) or a product-schema plugin emits it. |
No brand in the structured data minorstructured_data_brand_missing | recommended | The Product structured data on this item's page publishes no brand. Google lists brand.name as a recommended property of Product markup: it is how the page is matched to the same product sold elsewhere, and it is the brand your feed submits. | Add "brand": {"@type": "Brand", "name": "Your brand"} to the Product node — the same name you submit in the feed's brand field. |
Want to know which of your own products fail this requirement?
Audit my structured data (on the product page, not a feed column) free →How to fix it in Shopify
- Add a snippet to your theme's product template that prints the Product JSON-LD from the product object (liquid: product.title, product.featured_image, product.price, product.vendor) inside a script type="application/ld+json" block.
- If you use an SEO app (many themes ship with one), it usually emits Product markup already — check a product page for the script block before adding your own, so you do not print two.
- Theme updates overwrite snippets: re-check the markup after a theme upgrade or a migration.
- FeedGuard builds the page address from the product handle (https://your-store/products/handle), so the markup has to be on that page.
How to fix it in WooCommerce
- WooCommerce core does not print Product JSON-LD: an SEO plugin (Yoast SEO, Rank Math) or a product-schema plugin adds it. Install one and enable the Product schema.
- With Yoast, check Search Appearance → Content types → Products: the schema must be Product, not WebPage.
- A page builder that renders the product in a template of its own can bypass the plugin's schema output — check the live page source.
Worth knowing
- We read a sample — the first 10 distinct product page links in the catalogue, the same way image reachability is sampled. The report states how many pages were read, and a product page outside the sample is reported as not checked, never as a failure.
- A page we could not read (robots.txt asks us not to, an HTTP error, a JavaScript-only page that answers without markup) is also reported as not checked, with the reason.
- This check needs the page itself, so a feed URL or an uploaded CSV never fails it: those audits skip it and say so. Audit your store domain to run it.
- Meta does not ask for JSON-LD, so a Meta catalogue is never charged for it — both rules here are labelled Google-only.
How FeedGuard checks it
- We fetch each item's own product page — the link we built for it, on the store we are auditing — and read every <script type="application/ld+json"> block on it, including blocks inside @graph, arrays and nested nodes.
- A node counts as product markup when its @type is Product or ProductGroup; the brand is read from brand.name, and the GTIN, mpn and sku on the node are recorded as evidence.
- We honour robots.txt before reading a page, and we read at most 10 distinct product pages per scan, the same way image reachability is sampled.
- A page we could not read (robots.txt, an HTTP error, an answer that was not HTML, JSON-LD that will not parse) is reported as not checked — never as a failure. A feed URL or an uploaded CSV skips these checks entirely, because a file carries no page markup.
- GTIN is not charged twice: the page's GTIN is read as evidence, and the barcode itself is judged by the existing GTIN checks (presence, length, checksum, prefix, duplicates) on the value you submit.
Every failure names the field and shows the value at fault, counted over up to 250 products per scan. A requirement that does not apply to an item is left out of both sides of the score rather than counted as a pass, and a requirement our source cannot see at all is excluded from the score and listed as “not checked from this source” with the reason. Where a value can be corrected without inventing anything, the row is fixed in the corrected CSV you can download; anything we would have to make up (a real barcode, a brand, a price, a category) is listed for you to supply instead.
What each kind of scan can see
A store scan reads the public catalogue the store itself publishes (Shopify's /products.json). A pasted feed URL or an uploaded CSV is the file the shopping platform reads, so every attribute can be judged against it. This is the difference that decides whether a check can run at all.
| Attribute | Store scan (Shopify) | Your feed URL or CSV upload |
|---|---|---|
| structured_data | read from the store read from each item's own product page (JSON-LD), through the link we built for it | absent from your file Structured data (JSON-LD) lives on the live product page, not in a file, so a feed or an uploaded CSV cannot carry it. The structured-data checks are skipped for this audit and are never counted as failures — audit your store domain to check them. |
Read the engine's own rules for this requirement on the results page: every failing row links back to this guide. All 49 checks are listed on what we check.
Run the free audit on your store
Paste your store domain, paste your product feed URL, or upload a product CSV. You get a scored report of every requirement above, the products failing each one, the exact fields at fault, and a corrected CSV to download. No account, no card.
Run the free audit on your storeA store scan reads the public catalogue and skips the attributes it cannot see (barcodes and platform-only fields among them); a feed URL or CSV gets the full audit.