Product schema is the structured data that tells a machine what a product is. A bare Product with only a name and a price gets a store listed, while brand, identifiers such as GTIN, accurate availability and real ratings give an AI enough to compare the product with others when a shopper asks.
Comparison comes before any recommendation. A machine can only weigh a product on the facts it was given. Each missing field is a question the machine cannot answer about your product.
What does a listed product look like?
Google's product snippet guide requires a name and one of offers, review or aggregateRating. A Product with a name and an offer with a price meets that floor. It tells a machine the product exists and what it costs, and almost nothing else. Two products with the same name and similar prices look identical at that level.
What does a product that can be compared look like?
Google's merchant listing guide lists brand, gtin and mpn among the recommended properties of a Product, and availability and priceCurrency on the offer. Each one answers a question a shopper might ask an assistant.
| Field | Listed (bare minimum) | Recommended (complete) |
|---|---|---|
| name / price | Present | Present |
| gtin / mpn | Missing | Declared. Matchable to the real item |
| brand | Missing | Declared. Comparable within a brand |
| availability | Missing | InStock / OutOfStock. Transactable |
| aggregateRating | Missing | Rating + review count. Trusted |
- Brand tells a machine who makes the product.
- GTIN and MPN match the product to the same item elsewhere, as GTIN, MPN and SKU explains.
- Availability says whether a shopper can buy it today.
- Price and currency let a machine compare cost.
- aggregateRating and review count give a rating, but only when real reviews are shown on the page.
Most Shopify stores stop short of this. In our June 2026 study of 216 Shopify storefronts, 72% failed Entity Integrity, our check of product and brand structured data. Most of the failing stores had product schema that was thin or broken.
What should I leave out?
Leave out every field you cannot back up on the page. Schema that claims a rating the page does not show, or a price that is not real, creates a mismatch between the schema and the page, which is worse than the missing field. A product made in your own studio has no GTIN, so the field stays out.
How do I read my own Product schema on Shopify?
- Open one of your product pages, choose View page source, and search for ld+json to find each structured data block.
- Find the block whose @type is Product. If you find two, read the guide to duplicate Product schema on Shopify before you change anything.
- List every field in that block: name, brand, gtin, mpn, sku, offers with price, priceCurrency and availability, and aggregateRating.
- Sort each missing field into two groups. Fields your product has but the page does not declare are your work list. Fields that do not apply to your product stay out.
- Fill the work list from your Shopify admin. Brand often comes from the Vendor field on the product, the GTIN from the barcode field on each variant, and availability from your inventory.
- Paste the product address into Google's Rich Results Test and confirm one Product with the right values.
What does a cycling apparel shop add so its helmets can be recommended?
A cycling apparel shop on Shopify has 150 products, and its theme prints a Product block with a name, an image and an offer with a price, and nothing else. The shop sells jerseys from its own label and helmets from two outside brands, and the helmets carry UPC codes on their boxes.
The owner fills Vendor on every product, types the helmet UPCs into the barcode fields, and asks the developer to add brand, gtin and availability to the theme's Product block in a duplicate theme. The jerseys get brand and availability and go without a GTIN, which is the correct answer for a label that never bought barcodes. The review app shows ratings on 40 products, so aggregateRating goes on those 40 and nowhere else.
Where does product schema fit with the rest of the page?
The structured data repeats what the page says, so the page has to say it first. Product titles AI can match and product descriptions for AI cover the visible text. Collecting reviews AI trusts covers getting real reviews onto the page before they go into your schema. When your theme and apps print markup that you cannot untangle yourself, AI Visibility Setup: Foundation fixes duplicate and broken markup at the template level, in a copy of your theme.
Questions owners ask
What is the minimum product schema I need?
Google's product snippet guide requires a name and one of offers, review or aggregateRating. A Product with a name and a price meets that floor, but it gives a machine almost nothing to compare.
Which product schema fields help AI compare my products?
Brand, gtin and mpn identify the product, and price, priceCurrency and availability describe the offer. Google lists these among the recommended properties for merchant listing markup.
Should I add rating schema if I do not have many reviews?
Add rating schema only when real ratings are shown on the page. Invented or inflated rating data creates a mismatch between your schema and your visible content.
Does a service or custom product need a GTIN in its product schema?
A GTIN belongs in your schema only when the product carries a real barcode. If it has none, leave the field out and declare brand and SKU instead.
Sources
We opened the Google sources below on November 9, 2026.
- Product snippet structured data, Google Search Central.
- Merchant listing structured data, Google Search Central.
- The State of AI Visibility 2026, Visibility Mesh, 216 Shopify storefronts scored in June 2026.