Structured data & schema

IntermediateDuration ~40 min hands-onTools Google Rich Results Test, schema.org docs, A JSON-LD validator, Your CMS or site template

Why this lesson

Google reads your page as words. Structured data hands it a second, machine-clean copy of the facts — this is an Organization named X, here is its logo, this is an Article, here is its author and the date it was updated. When those facts are unambiguous, Google can dress your listing up with a rich result (star ratings, a logo, a breadcrumb trail, an article card) and, increasingly, AI answer engines can quote you with confidence. Get schema wrong — or hide it behind JavaScript — and you get nothing, or worse, a warning. This lesson gives you the short list of schema types that actually earn something in 2026, the current-as-of-this-year rules about what Google still rewards, and the one delivery mistake that quietly makes all of it invisible.

The essentials

Schema markup has more folklore around it than almost any SEO topic. Half the tutorials online recommend markup Google stopped rewarding years ago. So this lesson is built from the primary sources — schema.org, Google’s own rich-results documentation, and the GEO schema framework — not from a video. The goal is a mental model you can trust: what the vocabulary is, which types are worth your time, and how to ship them so both Google and the AI crawlers can read them.

The explainer

What structured data actually is

Structured data is a standardized block of facts about your page, written in a vocabulary both you and search engines agree on. That shared vocabulary is schema.org — a joint project of Google, Microsoft, Yahoo and Yandex that defines “types” (Organization, Article, Product, Person…) and the properties each type can carry (name, logo, author, datePublished…).

You can express schema in three formats, but only one matters in practice. JSON-LD (JavaScript Object Notation for Linked Data) is a self-contained <script type="application/ld+json"> block you drop into the page. Google explicitly recommends it over the two older inline formats (microdata and RDFa) because it sits apart from your visible HTML — you don’t have to weave attributes through every tag. Every example in this lesson is JSON-LD.

The types that matter

You do not need dozens of schema types. A small set carries almost all the value:

• Organization schema — who the site belongs to: name, logo, and (the important part) sameAs links to its other profiles. On a physical business, its cousin LocalBusiness schema adds address, hours, and geo-coordinates.
• Article schema (and the BlogPosting subtype) — for editorial pages: headline, author, datePublished and dateModified. Mark the author up as a Person, not a plain string, and you connect the content to a real, credentialed human.
• BreadcrumbList — encodes your Home › Category › Page trail so Google can show it in the listing instead of a raw URL.
• Product — price, availability and review data for e-commerce pages, which can earn the star-rating and price rich results.
• Person — a standalone entity for an author or founder, with jobTitle, knowsAbout, and sameAs.

What actually earns a rich result in 2026

Here is where most tutorials are dangerously out of date. Two big changes you must know:

• HowTo rich results were removed in September 2023. You can still write the markup, but it no longer produces any visual enhancement in search. Don’t spend time on it for that purpose.
• FAQPage rich results were restricted in August 2023 to well-known, authoritative government and health sites. For an ordinary business or blog, FAQPage markup no longer shows the expandable Q&A in results. Valid markup, but no rich result.

So the rich results still genuinely available to most sites are the Organization/logo panel, the Article card, breadcrumbs, and Product (price, availability, review snippets). Mark those up; skip the retired ones.

The mistake that makes all of it invisible

Schema only counts if the crawler can see it in the page it receives. JSON-LD injected by JavaScript — added client-side after load by a tag manager or a framework — creates two problems. Google may process it late or not at all (Google’s own 2025 guidance flags the delay), and AI crawlers don’t run JavaScript at all, so to them the schema simply isn’t there. The fix is one rule: server-render your schema. It must be in the raw HTML that comes back on the first request, not painted in afterward. View source; if you can’t find your <script type="application/ld+json"> there, neither can they.

The bridge to Level 3

One property does more for AI visibility than any other: sameAs. It lists the URLs where your entity lives elsewhere — Wikipedia, LinkedIn, Crunchbase, your verified social profiles. That’s how an answer engine confirms “the X in this Organization block is the same X on Wikipedia,” which is the foundation of being cited. We’ll build that out fully in Level 3; for now, know that a complete sameAs is the highest-leverage line in your schema.

Do it

1. Pick one real page you control — ideally your homepage or a blog post.
2. Write an Organization block: name, url, logo, and a sameAs array with at least three of your real profiles (LinkedIn, a social account, a directory listing).
3. If it’s an article, add an Article (or BlogPosting) block: headline, author as a Person with a name, plus datePublished and dateModified.
4. Place both inside a single <script type="application/ld+json"> and make sure it is in the server-rendered HTML — check View Source, not just DevTools.
5. Run the page through Google’s Rich Results Test and a JSON-LD validator. Fix every error and warning.
6. Confirm which rich result (if any) the page is now eligible for, and log it in the Level 2 workbook. The downloadable schema starter templates (Organization, LocalBusiness, Article/author, Product) give you fill-in-the-blank JSON-LD to start from.

Level 2 workbook — schema starter templates (Organization, LocalBusiness, Article/author, Product) + a validation checklistlevel-2-workbook.pdf116 KBOriginal course material — free to use

Terms introduced

Terms introduced in this lesson

• Structured data
• JSON-LD
• Schema.org
• Rich result
• Organization schema
• Article schema
• LocalBusiness schema
• BreadcrumbList

Each term links to the glossary, where it is defined by the way you just used it.

Check

Check yourself

1. A developer adds valid JSON-LD to a page, but it is injected by JavaScript after the page loads. Why is that a problem in 2026?

   Google ignores all JSON-LD; you must use microdata insteadIt risks delayed processing by Google and is invisible to AI crawlers, which do not execute JavaScript — schema should be server-renderedJavaScript-injected JSON-LD is invalid and will fail the Rich Results TestIt only works if you also add the schema in the page title

2. You mark up a step-by-step guide with HowTo schema hoping for a rich result. What is the current reality?

   HowTo requires a paid Google accountHowTo rich results are the easiest rich result to earnHowTo rich results were removed by Google in September 2023, so the markup no longer produces a visual enhancementHowTo only works on mobile

3. Which schema property is the single most impactful addition for AI visibility, and what does it do?

   author, because it hides your competitorspriceRange, because it shows your priceskeywords, because it tells Google what to rank you forsameAs, because it links your entity to its other authoritative profiles (Wikipedia, LinkedIn, social) and confirms who you are

Check answers

You can move on when you can… add valid, server-rendered Organization + Article JSON-LD to a page, pass Google’s Rich Results Test, and correctly say which rich results are still available in 2026 (and which — HowTo, ordinary FAQPage — are not).

Mark this lesson complete

Go deeper

• Google Search Central — “Structured data markup that Google supports”: the authoritative, continually-updated list of which types earn which rich results. Treat it as the source of truth over any blog post.
• schema.org: the full vocabulary reference for every type and property.
• Next: 2.6 · Content & E-E-A-T — the trust signals that decide whether your well-marked-up content is actually believed.