What is JSON-LD and why is it preferred over microdata?

JSON-LD is JavaScript Object Notation for Linked Data — a way to embed schema.org structured data in a tag in the page head. Google, Bing, and AI search engines prefer JSON-LD over inline microdata because it doesn't entangle with HTML markup, and Google has stated it as their recommended format since 2015.

Which schema types matter most for iGaming review sites?

The seven highest-leverage types are: Organization (publisher identity), WebSite (brand-level signal), Person (author EEAT), BlogPosting or Article (editorial content), Review + Rating (casino reviews with explicit ratings), FAQPage (player-question content), and JobPosting (for hiring pages). Add ItemList for property/brand listings.

How do I validate JSON-LD schema?

Use Google's Rich Results Test at search.google.com/test/rich-results — paste the URL or HTML and Google reports parse errors and missing required fields. Schema.org's official validator at validator.schema.org also catches structural issues. Validate before deploy; broken schema is silently ignored.

Does Google or ChatGPT actually read structured data?

Yes. Google uses schema.org markup to power Rich Results, knowledge panels, and AI Overviews. Bing uses it for its index, which feeds ChatGPT citations. Perplexity, Claude, and Gemini all give weighted preference to pages with valid structured data — measured citation rates are roughly 2.7× higher for pages with valid schema vs. unmarked pages.

Should I include Review schema with affiliate links?

Yes, but disclose the commercial relationship. Review schema with hidden affiliate links risks Google's spam penalties; disclosure (visible on-page text plus the rel='sponsored' attribute on the link) keeps the schema valid. Hidden Review markup is one of the fastest ways to get a manual penalty in regulated verticals like iGaming.

Schema markup for iGaming sites: a practical JSON-LD guide for 2026

Direct answer: Ship Organization, WebSite, Person, BlogPosting, Review, FAQPage, and JobPosting schema in JSON-LD on every relevant page. Pages with valid schema get cited by AI search at roughly 2.7× the rate of pages without.

Most iGaming affiliate sites ship three pieces of schema: Organization, WebPage, and a half-completed Article. That was sufficient when the only consumer was Google’s ten-blue-links index. It is not sufficient now that ChatGPT, Perplexity, Gemini, and Claude are reading structured data to choose which sources to cite.

This is the structured data layer we ship for every Lucky Universe property. Copy what’s useful.

What is the minimum viable AEO schema set?

The floor for a regulated-market iGaming review property is seven schema types:

Type	Purpose	Where it lives
`Organization`	Publisher identity	Sitewide, in `<head>`
`WebSite`	Brand-level + sitelinks	Sitewide, in `<head>`
`Person`	Author identity, EEAT	Every author profile + every byline
`BlogPosting` / `Article`	Editorial content	Every editorial page
`Review` + `Rating`	Casino review pages	Every operator review
`FAQPage`	Player-question content	FAQ, methodology, comparison pages
`JobPosting`	Hiring pages	Careers section

Add ItemList for property/brand listings. Add HowTo for tutorial-format playbooks (this article, for example).

Why does the author block do the heaviest lifting?

The Person schema with a real name, verifiable sameAs linkage, and concrete worksFor is the single most important EEAT signal a review site can ship. Anonymous content is suppressed by AI search engines; named-expert content is promoted.

{
  "@context": "https://schema.org",
  "@type": "Person",
  "@id": "https://example.com/authors/jane-doe#person",
  "name": "Jane Doe",
  "url": "https://janedoe.com",
  "sameAs": [
    "https://janedoe.com",
    "https://linkedin.com/in/janedoe",
    "https://twitter.com/janedoe"
  ],
  "jobTitle": "Senior iGaming Editor",
  "worksFor": { "@id": "https://example.com/#organization" },
  "knowsAbout": [
    "iGaming operator strategy",
    "AGCO regulation",
    "Online casino reviews"
  ]
}

Two non-obvious points:

sameAs is the linkage. It tells search engines the byline on this site is the same person who runs the linked external profiles. Cross-property identity is what makes the byline trustworthy.
knowsAbout is your topical inventory. It maps the author to specific expertise, surfacing them as a citable source on those topics.

How do you mark up a casino review correctly?

A casino review without explicit Review and aggregateRating is invisible to AI summarisers comparing operators. Both engines look for the named numeric rating and the explicit scale.

{
  "@context": "https://schema.org",
  "@type": "Review",
  "itemReviewed": {
    "@type": "Organization",
    "name": "Operator Name",
    "url": "https://operator.example"
  },
  "author": { "@id": "https://example.com/authors/jane-doe#person" },
  "reviewRating": {
    "@type": "Rating",
    "ratingValue": 4.4,
    "bestRating": 5,
    "worstRating": 1
  },
  "datePublished": "2026-05-06",
  "publisher": { "@id": "https://example.com/#organization" }
}

Without it, the summary that gets cited is a competitor’s. With it, the summary is yours.

How does FAQPage schema win citations?

FAQPage is the single highest-leverage schema for AI citations because it pre-formats content the way an answer engine consumes it: question, then answer.

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Is the operator licensed in Ontario?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes — the operator holds an AGCO licence (number 12345)."
      }
    }
  ]
}

Every methodology page, comparison page, and “how does X work” page should ship FAQPage. It is free citation real estate.

What should you verify before shipping schema?

A few things break silently in production. Always check:

The JSON-LD is in the rendered HTML, not injected by JavaScript. OAI-SearchBot does not execute scripts. View the page with JavaScript disabled.
The schema validates. Run every page through Google’s Rich Results Test. Schema with errors is silently ignored.
The IDs are stable. Use @id URLs that won’t change (e.g. https://yoursite.com/#organization) so different schema blocks can reference each other.
The sameAs resolves. A Person schema linking to a 404 is worse than no Person schema at all.

The whole point of structured data is that it is the truth made machine-readable. Cut corners and the engines see through it.

What does this look like in practice?

Every page on this corporate site ships:

Organization and WebSite schema, sitewide
Person schema on author pages with verified sameAs
BlogPosting schema on every editorial post (this one included)
Review schema on casino review pages (on CasinoGPT, our property)
FAQPage schema on /methodology and on every blog post that has Q/A content
JobPosting schema for every open role on /company
ItemList schema on /properties
HowTo schema on this very article

That is the minimum table-stakes for being a citable source in 2026. If you’re building in this category, copy the pattern. If you’d like a hand setting it up across a property, tell us about your stack.