Product Schema Example for Shopify Apparel

Apparel product pages on Shopify get cited by AI shopping engines based on how precisely the JSON-LD describes the variant landscape. A shopper asking “linen shirt that runs true to size in medium” on ChatGPT, Perplexity, or Gemini needs the AI to recognise that the specific variant exists, is in stock, has a price, and is described distinctly from the L and XL siblings. Generic Product schema with no variant detail forces the AI to guess; ProductGroup with hasVariant gives it the data to cite confidently.

This example ships two copy-ready JSON-LD blocks: a ProductGroup for the variant matrix, and a BreadcrumbList for category context. Both are calibrated to pass Google Rich Results Test cleanly and to mirror the visible page content (the rule AI engines actually enforce in practice).

When to use Product vs ProductGroup#

Shopify’s default Liquid templates emit Product regardless of variant count. Upgrading to ProductGroup for any apparel store with multiple sizes is a one-time theme edit (or a Shopify App that handles it automatically); the AI-shopping citation lift is worth the effort.

ProductGroup JSON-LD (apparel with variants)#

Drop this into a product page template, replacing the placeholder values. The example shows a 2-variant subset; in production you’ll have one hasVariant entry per real SKU.

{
  "@context": "https://schema.org",
  "@type": "ProductGroup",
  "name": "Linen Relaxed Shirt",
  "description": "A breathable linen shirt in three colors, sizes XS–XL. 100% European linen, pre-washed for softness.",
  "brand": {
    "@type": "Brand",
    "name": "Example Apparel"
  },
  "productGroupID": "linen-shirt-001",
  "variesBy": ["https://schema.org/size", "https://schema.org/color"],
  "hasVariant": [
    {
      "@type": "Product",
      "sku": "LINEN-SHIRT-WHITE-M",
      "name": "Linen Relaxed Shirt — White / M",
      "color": "White",
      "size": "M",
      "material": "Linen",
      "image": "https://example.com/products/linen-shirt-white.jpg",
      "offers": {
        "@type": "Offer",
        "url": "https://example.com/products/linen-shirt?variant=white-m",
        "priceCurrency": "USD",
        "price": "79.00",
        "availability": "https://schema.org/InStock",
        "itemCondition": "https://schema.org/NewCondition"
      }
    },
    {
      "@type": "Product",
      "sku": "LINEN-SHIRT-BLACK-S",
      "name": "Linen Relaxed Shirt — Black / S",
      "color": "Black",
      "size": "S",
      "material": "Linen",
      "image": "https://example.com/products/linen-shirt-black.jpg",
      "offers": {
        "@type": "Offer",
        "url": "https://example.com/products/linen-shirt?variant=black-s",
        "priceCurrency": "USD",
        "price": "79.00",
        "availability": "https://schema.org/InStock",
        "itemCondition": "https://schema.org/NewCondition"
      }
    }
  ]
}

BreadcrumbList JSON-LD#

Pair the ProductGroup with a BreadcrumbList so AI engines understand the category hierarchy your product lives in. The breadcrumb URLs must match the page’s visible breadcrumb exactly.

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Home",
      "item": "https://example.com/"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "Shirts",
      "item": "https://example.com/collections/shirts"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "Linen Relaxed Shirt",
      "item": "https://example.com/products/linen-shirt"
    }
  ]
}

Common errors (in order of how often they trip stores up)#

1. Marking up variants that aren’t visible or selectable on the page. Schema must mirror visible reality. If your product page only shows S / M / L but the JSON-LD declares XS through XXL, AI engines will eventually try to cite a variant that 404s and downweight the store.

2. Using AggregateRating when no review widget exists. Either show the reviews visibly (use a Shopify Reviews app) or drop the schema field. Don’t fabricate.

3. Inconsistent currency, price, or availability between schema and the visible product page. Buyer sees $79 but schema declares 89.00 — AI engines pick the schema number, cite it, and the buyer feels misled. Worse, Google flags the schema as untrustworthy.

4. Reusing the same sku across variants. Shopify auto-generates unique SKUs but merchants often override them. Duplicate SKUs in hasVariant cause AI engines to merge variants incorrectly.

5. Mixing ProductGroup and Product such that variant-specific names get lost. If you emit both Product and ProductGroup for the same item, make sure the Product instance is the parent (no variant detail) and ProductGroup.hasVariant contains the variants. Don’t emit Product for each variant outside the group.

6. Adding medical or unsupported claims to apparel descriptions. “Posture-correcting” / “improves circulation” / “anti-bacterial fabric prevents skin disease” — these are regulated claims in most markets and trigger AI engine downweighting even when legally fine. Stick to fit and material descriptions.

7. Pointing schema URLs at non-canonical variant URLs. If the page uses one canonical product URL (e.g. /products/linen-shirt) and variants are query-param selected (?variant=white-m), the hasVariant[].offers.url should still point at the variant-specific URL — that’s where buyers (and AI agents) land.

How to verify#

1. Paste the deployed product URL into

   Google Rich Results Test

   . The Product and BreadcrumbList items should both validate.
2. Run the validation checklist above against each field.
3. Open the page in an incognito window and check that the visible product H1, price, availability, and breadcrumb all match what the schema declares.
4. Re-run the Schema Generator if you change product structure (new color, dropped size) and re-deploy.

The JSON-LD is one signal in the AI-shopping stack. Pair it with the Fashion llms.txt template for the content-map side, and verify that GPTBot can reach the product page using the Robots Analyzer.