Structured data doesn’t change a word your visitors read, but it changes how search engines and AI systems interpret, categorize, and cite your pages. This guide covers the schema types that still earn a genuine benefit in 2026, correctly implemented and validated — including an honest, current answer on FAQ schema, which most guides still get wrong. It builds on the broader technical SEO picture, since structured data only helps once a page is genuinely crawlable and indexable in the first place.

Structured data is machine-readable code, most commonly written in JSON-LD, that describes a webpage’s content so search engines and AI systems can parse and cite it accurately.

JSON-LD schema

JSON-LD is the format Google explicitly recommends for structured data, ahead of the two older alternatives, Microdata and RDFa. It’s a standalone <script type="application/ld+json"> block, usually placed in the page head, that describes the page’s content separately from the visible HTML — meaning you can redesign a page’s layout entirely without touching its structured data, and vice versa.

A minimal, correct JSON-LD block needs three things: a @context pointing to https://schema.org, a @type naming the schema (Article, Product, FAQPage, and so on), and the required properties for that specific type. Google’s own documentation is the authority on which properties are required versus recommended for a given type — the Schema.org vocabulary is broader than what Google actually consumes for rich results, so when the two disagree, follow Google’s structured data documentation rather than the general Schema.org spec.

Two implementation habits prevent most real-world problems: don’t let more than one plugin or system output the same schema type on a page (duplicate Article blocks from an SEO plugin and a separate schema plugin is a common WordPress mistake), and always view the rendered page source in production before trusting a validation result — some caching layers serve stale markup that looks correct in your editor but not on the live page.

Article schema

Article schema (or its more specific variant, BlogPosting) tells search engines the structured facts around a piece of content: who wrote it, when, and who published it. Implemented completely, it’s one of the more reliable schema types for author attribution and publication-date display in search results.

The core required and strongly recommended properties: headline, author (as a nested Person object with a name), datePublished, dateModified, publisher (as a nested Organization object with name and logo), and image. Missing the author field specifically is a common, avoidable gap — an Article schema without it won’t reliably support author attribution even if every other property is filled in correctly.

Apply this to every blog post and article-style page on the site, and keep the dateModified value honest — updating it without a genuine content change undermines the same trust signal it’s meant to provide, similar to how dishonest lastmod dates undermine a sitemap.

Product schema

Product schema is what makes star ratings, pricing, and availability appear directly in search results for e-commerce pages — genuinely one of the higher-value schema types still in active use in 2026, since it drives a visible, clickable rich result rather than just background parsing.

Required properties center on name, image, and offers (a nested Offer object with price and priceCurrency). Star ratings specifically require an aggregateRating property with real review data — a Product schema missing this property simply won’t show ratings, and fabricating review counts or scores that don’t match the actual page content is a fast way to lose rich-result eligibility entirely, not just for that property but potentially for the whole page.

Only mark up products you’re genuinely selling on that specific page, with pricing and availability that match what a visitor actually sees — Google’s quality guidelines are explicit that structured data must describe content that’s visible on the page, not content that exists elsewhere or doesn’t exist at all.

FAQ schema

Here’s the part that changed: Google deprecated the FAQ rich result in Google Search on May 7, 2026. The expandable question-and-answer dropdown that FAQPage schema used to generate in search results no longer appears for any site — this follows an August 2023 restriction that had already narrowed FAQ rich results to well-known government and health sites, and the May 2026 change removed the feature entirely, for everyone. Google is also retiring the Search Console FAQ reporting and Rich Results Test support for this type in June 2026, with API support ending in August 2026.

What this doesn’t mean: FAQPage is still a valid Schema.org type, adding it doesn’t break anything or trigger any penalty, and Google has stated that unused or non-rich-result-eligible structured data doesn’t cause errors. If your FAQ section genuinely helps readers and matches real, visible content on the page, there’s no urgent reason to remove existing FAQPage markup — it may still support how AI systems and Google itself parse the section’s meaning, even without a visible SERP accordion attached to it.

What this does mean practically: stop implementing FAQ schema with the expectation of a rich-result payoff, since that payoff no longer exists for any site type. Treat a genuinely useful, clearly written FAQ section as valuable for readers and machine parsing on its own merits — not as a SERP-decoration tactic. If your only reason for an FAQ block was chasing the old accordion display, that reason is gone; keep the section only if the content itself earns its place on the page.

Rich results test

Google’s Rich Results Test is the standard validation tool: paste in a URL or raw code, and it reports exactly which rich result types the page qualifies for and flags any missing required properties. Run it on every page after a schema change, before pushing to production, not after.

Passing validation is necessary but not sufficient. Google’s own structured data guidelines are explicit that markup must describe content that’s genuinely visible on the page, must reflect the page’s real content rather than something staged elsewhere, and that violating these quality guidelines can prevent an otherwise syntactically correct rich result from displaying, or in more serious cases, get the markup flagged as spam. A schema block can pass the Rich Results Test’s syntax check and still fail to produce a rich result because the underlying content doesn’t meet Google’s quality bar.

Beyond the Rich Results Test, Search Console’s Enhancements reports monitor structured data errors across your whole indexed site after deployment, which catches problems the single-URL test won’t surface — a template-level bug that only shows up on certain page variants, for instance. Check both: the Rich Results Test for pre-deployment validation, Search Console Enhancements for ongoing, site-wide monitoring.

FAQ

Is FAQ schema still worth adding in 2026?

Only for what it does beyond rich results. Google retired the FAQ rich-result display in Search on May 7, 2026, so there’s no longer a SERP accordion payoff for any site. The markup remains valid and harmless, and a genuinely useful FAQ section still helps readers and may support how AI systems parse the page — but implement it because the content is worth having, not because it will earn a rich result.

What’s the difference between structured data, schema markup, and Schema.org?

Structured data is the general concept: organized, machine-readable information about a page’s content. Schema markup is the code you write to add that structured data to a page. Schema.org is the shared vocabulary — the specific set of types and properties, like Article or Product — that almost all structured data uses. In practice, “structured data” and “schema markup” are used interchangeably.

Does adding schema markup improve my Google rankings?

Not directly. Google has been consistent that structured data isn’t a ranking factor on its own. What it does is make a page eligible for certain rich results (which can improve click-through rate) and gives search engines and AI systems clearer signals for parsing and citing the page’s content — an indirect, supporting role rather than a direct rankings lever.

What’s the easiest way to validate my structured data?

Google’s Rich Results Test is the standard tool: it checks your JSON-LD for syntax errors, missing required properties, and which rich result types the page currently qualifies for. Run it on any URL or pasted code before deploying a schema change, and use Search Console’s Enhancements reports afterward to monitor for errors across the whole site.

Do I need special schema for AI Overviews or ChatGPT?

No. Google has stated there’s no special Schema.org markup required specifically for AI Overviews or AI Mode. A 2026 Ahrefs study tracking pages that added JSON-LD found no meaningful citation lift in AI Overviews, AI Mode, or ChatGPT from schema alone — clear, well-structured, genuinely useful content is what drives AI citation; correct schema supports that content without substituting for it.