Schema.org structured data is one of those SEO investments that has spent fifteen years quietly being underrated. It is not flashy. It does not show up in dashboards as a single moving line. It just consistently makes sites get found, get cited, and get represented correctly across every surface that reads the web.
In 2026, that quiet underratedness has become a real opportunity. Most sites still have no structured data. The ones that do — and do it well — are getting cited by AI assistants, displayed correctly in search rich results, and treated as primary sources at a rate disproportionate to their domain authority. It is one of the few SEO levers where doing the work still gets you ahead, because almost no one does the work properly.
This is the practical version: which Schema.org types actually earn their place, how to implement them honestly, and what to skip.
What schema actually is, briefly
Schema.org is a shared vocabulary for describing things on the web in a machine-readable way. The recommended format is JSON-LD — a small block of JSON inside a script tag in the page head. The vocabulary is huge (700+ types). The useful subset for a business website is small.
The reason it matters: search engines, AI assistants, voice assistants, social platforms, and increasingly browsers all read structured data directly. It lets you tell those systems "this page is a Service of type X provided by Organization Y, with these benefits, this price range, in this area." That removes guesswork. Removing guesswork makes you more likely to be picked up cleanly.
The seven types worth your time
1. Organization (or LocalBusiness). Goes on the homepage, ideally @id-referenced from every other page. Includes name, URL, logo, contact, sameAs links to your real social/identity profiles, areaServed. This is the foundation. AI engines use Organization markup to decide whether a site is a coherent entity or a marketing surface that might be anyone.
2. WebSite + SearchAction. Lets search engines surface a sitelinks search box for your domain. Small but visible. Quick to implement.
3. Service. One per service page. Describes the service, who provides it, the area served, the service type. AI engines read these directly when a user asks "who does X near me" or "best company for Y."
4. Article (or BlogPosting). One per post. Includes headline, description, author, publisher, datePublished, dateModified, image. Required for the article to show up properly in Discover and in many AI citation contexts.
5. FAQPage. Where you have real Q&A on a page. Google has reduced rich-result display for FAQPage in some categories, but AI engines still read it heavily — and FAQ-shaped content gets quoted more often.
6. BreadcrumbList. One per page that has a breadcrumb. Helps engines understand site hierarchy. Cheap to implement, high-yield.
7. Product (e-commerce sites only). For each product page. Includes name, image, description, brand, offers (price, availability). Essential for any commercial visibility.
That is the list. Seven types, applied consistently across the right page types, will put a site in the top 5% of structured-data coverage on the open web.
What to skip
HowTo schema. Restricted in Google rich results since 2023. Not worth the implementation cost.
QAPage. Originally meant for forum-style Q&A pages. Largely ignored now. FAQPage covers the practical use case.
Most of the "marketing-emphasizing" types (AggregateRating that you can't honestly populate, Review schema for testimonials that are not actually structured reviews, Offer markup on services that don't have a real price). Implementing these dishonestly gets sites flagged. Implementing them when there is no real data to populate them is busywork.
Schema-plugin-recommended types that have no business being on your pages. A blog post is not a Recipe. A service page is not a JobPosting. Be picky.
How to implement it properly
Hand-rolled JSON-LD in the theme, not a plugin. Plugins write schema that looks correct in isolation but conflicts with the theme's own schema, duplicates @ids, and changes between plugin versions in ways that quietly break things. Hand-rolled schema in the theme stays under your control, integrates with the rest of the page, and is auditable.
One @graph per page. Group your structured data into a single JSON-LD block with an @graph array. This lets you @id-reference your Organization across every page without duplicating it. AI engines like the structure too — it makes the entity model clearer.
Use real data. If you do not have aggregate ratings, do not put aggregate rating schema. If your service does not have a price, do not put a price. The penalty for inflated schema is more than the upside of having it.
Validate. Google's Rich Results Test and Schema.org's own validator catch most mistakes. Run pages through both whenever you make schema changes.
Test with real AI assistants. The honest measurement loop. Ask the assistants the questions your buyers ask. Watch whether your structured-data work shows up as cleaner citations over the months that follow. Getting cited in ChatGPT goes deeper on the citation-measurement workflow.
What we typically find on a structured-data audit
When we audit a site for structured data, the pattern is consistent.
The homepage has Organization schema, usually from a plugin, often with the wrong name format or with sameAs links that no longer point at active profiles.
Service pages have no Service schema. Sometimes they have Article schema by accident, because the SEO plugin defaults to it.
Blog posts have Article schema, usually with the wrong author entity (set to the WordPress user "admin" instead of an actual person or to the organization).
No BreadcrumbList anywhere, even though the site has visible breadcrumbs.
No llms.txt at the root.
Fixing this is a one-to-three-week project for most sites, depending on size and how custom the theme is. The result is usually visible in citation tracking within a quarter.
Where this fits in the larger picture
Structured data is one piece of a broader AI-visibility posture. The full picture also includes technical SEO basics, answer-first content, real authority signals, llms.txt, and Core Web Vitals — covered in what modern SEO actually looks like now.
But schema markup is the place we usually start. It is cheap relative to content work, fast to implement, and the cleanest single signal we can give to the engines that decide who to cite. If your site does nothing else for AI visibility this year, do this.
Our SEO & AI Visibility work routinely includes a structured-data audit and rebuild as the first month of engagement. Tell us about your site if you want a read on what is currently in place and what would change the most.
