A Structured Data Checklist That Actually Passes Validation
By Priya Chandran · Published · Updated
Most structured-data problems we find in audits aren’t syntax errors — they’re a schema type that doesn’t match what the page actually is, or required properties left empty. Fixing both takes five concrete steps, not a rewrite.
What is JSON-LD, exactly?
JSON-LD is a way of embedding machine-readable facts about a page — its type, author, price, publish date — directly in the HTML as a script tag, using the shared vocabulary defined at schema.org. Search engines and AI crawlers read it to understand what a page is without having to guess from the visible text.
Step 1: pick the type that matches the page, not the ideal
A pricing page is a WebPage, not a Product — the product is what you sell, the pricing page is a page about it. Choosing the closest matching real type avoids a validator warning for properties that don’t apply to your content.
Step 2: fill every required property for that type
Each schema.org type defines a small set of properties search engines expect to see: headline and datePublished for an Article, name and offers for a Product. A type with none of its required properties is barely better than no schema at all.
Step 3: use a real Person or Organization for authorship
An author field written as a plain string ("author": "Priya") is a common shortcut that discards an authorship signal search engines and AI systems increasingly weight. Structuring it as {"@type": "Person", "name": "..."} instead links the content to a real, identifiable author.
Step 4: don't duplicate the same type twice
Two Organization nodes on one page with different names or logos reads as inconsistent data, not extra information. If a page needs both a publisher reference and a broader company profile, reuse the same node with an @id rather than declaring it twice.
Step 5: validate before you ship
Run every new schema block through a real validator, not just a linter:
- Schema.org validator for pure schema-shape correctness.
- Google’s structured data policies for eligibility rules on top of valid markup.
Tip: Fixture Scanner checks JSON-LD syntax, required properties, type relevance to the page, and duplicate nodes automatically — see the product overview.
Key takeaways
- Match the schema type to what the page really is.
- Fill every property that type requires, not just the easy ones.
- Structure authorship as a real Person entity, not a string.
- Reuse shared entities with @id instead of duplicating a type.
Frequently asked questions
Why doesn't my JSON-LD show up as a rich result?
Valid JSON-LD is necessary but not sufficient — the search engine also decides whether your page qualifies for a rich result based on content quality and eligibility rules that vary by type. Passing validation only confirms the markup is well-formed, not that it will be displayed.
Can a page have more than one schema type?
Yes, and it usually should. A blog post commonly carries both a BlogPosting node and a BreadcrumbList node; a product page often pairs a Product node with an Organization node for the seller. What it should not have is two competing nodes of the same type describing the same entity differently.
Related reading: see our responsive images guide for another common source of avoidable audit failures.