Schema

Shopify's own structured data covers the basics and stops. This fills the gap.

Why it matters

Assistants read JSON-LD before they parse the page. Google's AI Overviews will not treat a product as eligible without a return policy and shipping details. A thin Product block is the most common reason a store is invisible to AI shopping, and it is also the cheapest thing to fix.

What to emit

Start from what Shopify already outputs, then add what is missing. Never duplicate an existing block, and never contradict it.

Product, the properties that matter beyond Shopify's default: material, color, size, audience, hasMerchantReturnPolicy, shippingDetails, aggregateRating and review only if real, gtin13 only if the barcode field is populated, isCustomProduct or a madeToOrder note where the store offers it.

Organization, on the homepage, where the store has none. Name, url, logo, sameAs for real social profiles only, and contactPoint if the store publishes one.

Article, on blog posts. headline, datePublished, dateModified, author, image.

Rules

• Emit nothing that is not true on the page. Schema that contradicts visible content is a manual-action risk, not a clever shortcut.
• Never fabricate ratings or reviews. If there are none, omit aggregateRating entirely. An invented rating is the fastest way to lose rich results permanently.
• Check reference/schema.md for the deprecated types before emitting anything. Some once-standard types no longer produce rich results and only add weight.
• Validate the JSON parses, and that every URL resolves, before handing it over.

Output

A single <script type="application/ld+json"> block ready to paste into theme.liquid above {{ content_for_header }}, plus a short list of which properties you added and where each value came from. If a property was left out because the data does not exist, say which and what would be needed.