On this page
Structured product data is product information organized into named fields and defined values that software can read.
It separates details such as brand, model, size, weight, and seller-specific price rather than relying only on prose. APIs, feeds, and webpage markup can all carry this information. In SEO, 'product structured data' often specifically means webpage markup; the broader data concept also applies to catalogues, feeds, and APIs. A structured record still needs accurate facts and timely updates.
Description vs structured fields
A fictional product description says: 'TrailCo's black waterproof jacket, size medium, is available from Retailer A for USD 129.' A structured record separates those details into fields.
| Field | Illustrative value |
|---|---|
| Brand | TrailCo |
| Color | Black |
| Size | Medium |
| Seller | Retailer A |
| Price | 129.00 |
| Currency | USD |
Product vs variant vs merchant offer
Consider a fictional travel headphone model sold in black and silver. The black version is one variant. Retailer A and Retailer B can each sell that variant with different prices, availability, and return terms. The silver version should not be merged into the same purchasable record simply because the model name matches.
Schema.org Product describes the item; its offers relationship can connect it to seller terms described with Schema.org Offer. Keeping these layers separate helps a system compare the same variant across sellers.
Structured product data in APIs, feeds, and webpages
The same jacket facts can appear in a product API, a feed, and webpage markup. Each representation serves a different reader or system.
| Representation | Primary use | What to check |
|---|---|---|
| Product data API | Programmatic retrieval or management of product records. | Catalogue scope, supported operations, and update times; see product data API. |
| Product feed | Bulk or incremental data ingestion by another platform. | The receiving platform's fields, formats, and refresh rules; see product feed. |
| Schema.org Product markup | Describe a product on a webpage using shared vocabulary. | Markup matches visible product information and the intended search feature; see product schema markup. |
What fields make product data useful?
Useful structured data identifies the exact item and expresses the attributes needed for the intended comparison. Preserve units, value types, and source information rather than treating every field as interchangeable text.
| Layer | Example fields | Keep separate |
|---|---|---|
| Product identity | Brand, model, GTIN (trade item identifier) or manufacturer part number where available | A retailer SKU is not automatically a global product identifier. |
| Variant | Size, color, capacity, variant ID, parent/group ID | One product family can contain different purchasable items. |
| Specifications | Material, dimensions, weight, compatibility | Keep the unit with the value: 240 g is more useful than 240 alone. |
| Seller offer | Seller, price, currency, availability, destination | Two sellers can offer the same variant on different terms. |
| Source and update time | Source, observed time, validation status where supported | The source and update time help explain conflicting prices or stock records. |
Why structured product data matters for AI shopping
Explicit attributes let shopping systems compare options against a buying brief. A travel assistant can compare sourced headphone weight and battery life without extracting each value from promotional prose.
Structure also makes gaps easier to detect. The assistant can show 'not supplied' for compatibility rather than infer it from a similar model. Product entity resolution identifies which records refer to the same item; normalization aligns units and fields; enrichment adds supported information.
How to improve structured product data quality
Start with fields that change a user's decision, then reconcile them across the product page, feed, and API. Accuracy matters more than filling every field.
Validate identity and variants
Keep stable IDs and check exact version, pack size, condition, and region. Do not invent GTINs or merge records solely because titles look similar.
Preserve evidence for enrichment
Add attributes from reliable product sources. Label unresolved or inferred information internally and review it before publishing it as a factual specification.
Resolve time-sensitive conflicts
Choose which source to use when prices or stock records disagree. Record the seller and update time so the app can distinguish a recent price from an older observation.
Structured product data vs an AON structured Offer
Structured product data helps an AI app understand the item. An AON offer adds a way to act on it, such as following a referral to the merchant.
For a jacket, the product data might describe size, material, and waterproofing. The offer connects that information with the seller and a purchase link, helping the app move from comparison to a recommendation the user can follow.
Read AON Partner Offer FetchCommon questions about Structured product data
Is structured product data the same as Schema.org markup?
No. Schema.org Product markup is one representation of structured product information on a webpage. Catalogues, product APIs, and feeds also contain structured product data under their own contracts.
What is the difference between product data and offer data?
Product data describes the item and its variants. Offer data describes a seller's commercial terms or an actionable commercial option. In AON, the Offer contract also defines supported action and attribution context.
How does product schema help search and shopping systems?
Product schema gives software a shared vocabulary for reading a product page, including its name, brand, and offers. It helps systems interpret the page; whether they display it also depends on their coverage, data requirements, and relevance rules.
Can AI fill missing attributes in structured product data, such as size or compatibility?
AI can assist data enrichment, but generated attributes need reliable evidence and review. A model's plausible answer is not proof of a specification, compatibility claim, or product identifier.
Sources
- Schema.org ProductRelated standard
- Schema.org OfferRelated standard
- AON Partner Offer FetchAON documentation