GRAFF

How to read it

Every node is a Schema.org entity with a stable @id; every edge is a Schema.org property linking two entities. Labels show the entity name and its primary type.

• Organization: the business. Content points at it via publisher (articles, videos) or brand (products).
• WebSite: the site container. Pages declare isPartOf → WebSite; it declares publisher → Organization.
• WebPage nodes: one per template page. Each points at its subject via mainEntity and at its breadcrumb trail.
• BreadcrumbList: the Home → template trail Google shows in results and agents use to locate a page in the site hierarchy.
• Missing edges are usually schema.org domain rules, not omissions. A Product cannot declare isPartOf because it is not a CreativeWork.
• Store policies (returns, shipping, loyalty, variants) are nested objects inside the Organization and Product entities rather than separate nodes — click a node to see them summarized on its card.

Google publishes rules for each rich result type in Search Central: a set of required properties the markup must carry, and recommended properties that strengthen the result. A page qualifies when its markup uses a matching Schema.org type and every required property is present. The cards below track exactly that: Qualifies now means the requirements are met, Needs values means the structure is complete but placeholder values remain, and Missing data means a required property needs real nested content GRAFF won't invent.

Rich result eligibility is the baseline; there are more fields to fill than just the required ones. If all things are equal, the most complete, truthful entity graph wins the competition. Aim past "qualifies" and fill as many fields as completely as possible.

Markup that once earned a rich result but no longer does. GRAFF does not generate or screen for these; they are listed so you know where they went.

Each template gets one self-contained snippet: a @graph holding the page's primary entity (Product, Article, Recipe…), a WebPage node, its BreadcrumbList, and the shared Organization and WebSite. One block serves every page built from that template. Fill the per-page tokens ({{PAGE_URL}}, {{PAGE_TITLE}}) from the page itself and the site-wide tokens ({{ORGANIZATION_LOGO_URL}}) once.

• @id fragments: entities are named by URL plus fragment: site-wide ones on the origin (/#organization, /#website), page-scoped ones on the page URL (#webpage, #breadcrumb). Keep them exactly as generated: crawlers merge nodes with the same @id across pages, and this joins per-page markup into one site-wide graph.
• Duplicate Organization and WebSite blocks: every snippet repeats them so each page stands alone. Search engines de-duplicate them by @id.
• Nested objects: Offer, PostalAddress, Person, and ListItem shapes match what Google's documentation specifies for each feature.
• Breadcrumb hub levels: middle crumbs come from the example path (/gear/tent yields Home > Gear > page). Confirm each hub URL exists as a real page, or remove that level.
• References: values like {"@id": "…/#organization"} are pointers to another node in the graph, not text. Leave them unchanged.