Schema Markup in 2026: How to Earn Rich Results Without a Penalty

Schema markup attracts more bad advice than almost any other part of technical SEO. It gets sold as a ranking lever it is not, bolted onto pages it does not describe, and then never checked again. The result is a lot of sites carrying broken or mismatched structured data they cannot see, because it renders invisibly in the code, while believing it is helping them.

This guide owns structured data as its own subject: what it does, what to implement, how to validate it in minutes, and where its real limits are. For how structured data sits alongside the other technical foundations, our technical SEO foundations guide gives the wider picture. Here we go deep on the markup itself.

What structured data is and what it actually does

Structured data is a block of JSON-LD that labels your content for machines: this is an article, this is the author, this is a product, this is its price, this is the rating. It does not change what a human sees. It tells a search engine, in a format it can read without interpretation, what the page is made of.

What it earns is eligibility for rich results: the enhanced listings that show star ratings, prices, breadcrumbs, recipe cards and the like, which take up more space in the result and can lift click-through. That is the upside, and it is real.

What it does not do is raise rankings. Schema is not a ranking factor in itself, and Google has said so repeatedly. Adding Product schema to a page will not move it from position six to position three. And eligibility is not display: valid markup makes you eligible for a rich result, but Google still decides whether to show one, based on quality, intent and the layout of the result. Two honest sentences that cut through most of the bad advice: schema earns eligibility, not ranking, and eligibility is not a guarantee.

The schema types worth implementing, by site type

Resist the urge to mark up everything. The types that earn their place depend on what your site is, and the right answer is usually a short list, not the full vocabulary:

• Organization and Person, for entity clarity, so Google understands who is behind the site and the content.
• Article and author markup, for content and publisher sites, identifying the piece, its author and its publish date.
• Product, Offer and AggregateRating, for ecommerce, where price, availability and genuine ratings can show in the listing.
• BreadcrumbList, for navigation, which shows the page’s place in the site structure in the result.

One important update, because a lot of older advice has not caught up. FAQPage and HowTo no longer earn rich results in Google. HowTo was retired earlier, and as of 7 May 2026 Google removed FAQ rich results entirely, for every site type, including the government and health sites that had kept them after the 2023 restriction. Both are still valid Schema.org types and leaving the markup in place does no harm, but neither produces a visual rich result any more, so do not add either expecting a SERP enhancement. If a guide still sells FAQ schema as a way to win dropdowns and take more space in the listing, it is describing a feature that no longer exists.

Why broken schema is worse than none

This is the part that turns a harmless oversight into a real problem. Schema that declares something the page does not show, a rating with no visible reviews, a price that does not appear in the copy, an FAQ block that does not match visible questions and answers, breaks Google’s core accuracy rule: structured data has to reflect what is actually on the page, visible to users.

Get that wrong and one of two things happens. The mild outcome is that Google ignores the markup and you get nothing. The serious outcome is a manual action for structured-data misuse, which removes all rich-result eligibility for the affected URLs, even where the markup is technically valid. Google’s spam policies are explicit that markup must reflect the content a user sees, and a mismatch is a policy violation, not a clever tactic.

The rule is one line: if a human cannot see it on the page, do not claim it in your structured data. Markup is a description, not a wish.

The reason most sites ship some schema broken and never notice is that it lives in the code, invisible to the eye. A plugin updates, a template changes, a price goes stale, and the markup quietly stops matching the page. Nobody sees it because there is nothing to see, until a rich result vanishes or a manual action lands.

Validating, the step almost nobody takes

Validation is a five-minute job that almost nobody does, and it is the single highest-return habit in structured data. There are two tools, and they answer different questions:

• Google’s Rich Results Test checks whether your markup is valid and which Google rich results the page is eligible for. Start here.
• The Schema.org validator checks generic syntax against the vocabulary itself, without Google-specific feature rules. Use it when you want to confirm the markup is well-formed regardless of what Google surfaces.

Both have a limit worth knowing: they validate structure and eligibility, not accuracy. Neither tool checks whether your markup matches the visible page, which is exactly the mismatch that earns a manual action. So a green result is the starting line, not the finish: the test confirms the code is readable, your own eyes confirm the markup describes what the page actually shows. For live errors across the whole site, the Search Console enhancement reports surface problems Google found in the wild, which the one-page tools cannot.

The cadence is simple and worth fixing as a habit: validate on publish, and re-check after any template or plugin change, since those are exactly when markup silently drifts out of line with the page.

Schema and AI search, the honest version

This is where the bad advice is thickest, so here is the plain version. Google has stated there is no special schema required to appear in AI Overviews or AI Mode, and no AI-specific markup. Pages get cited in AI answers with no FAQ schema, no special structured data, nothing. Anyone selling a schema type as the key to AI search is selling something Google has explicitly said does not exist.

Schema still helps, for two grounded reasons. It gives entity clarity, helping machines understand who and what a page is about, and it earns rich-result eligibility in classic Search, which sits in the same result as the AI answer. What it does not do is buy you a place in the AI response. For the full picture of what technical readiness for AI search actually involves, our guide to AI Overviews and technical SEO covers it, and the short version is that being crawlable, indexed and genuinely useful does far more than any markup.

Common mistakes

Most structured-data problems come down to a handful of repeated errors:

• Marking up content that is not visible on the page. The worst of them, and the one that earns manual actions. If it is not on the page for users, it does not belong in the markup.
• Declaring review or rating schema without genuine on-page reviews. An AggregateRating with no visible reviews is the classic mismatch, and a common trigger for a structured-data manual action.
• Leaving plugin-generated schema unchecked. Many SEO plugins generate schema automatically from fields written for keywords, not accuracy, and that markup can conflict with or duplicate manually added blocks. Two contradictory schema blocks on one page is a problem nobody put there on purpose.
• Schema injected by JavaScript after load. If the markup is added by a script that runs late, Google can index the page before the schema appears, so the test passes but the indexed page has no schema at all.

FAQs

What is schema markup?

Schema markup is structured data, usually written in JSON-LD, that labels your content for search engines: this is an article, this is the author, this is a product, this is its price and rating. It does not change what a person sees on the page. It gives machines an unambiguous description of what the page contains, which can make the page eligible for enhanced search listings called rich results.

Does structured data improve rankings?

No, not directly. Schema is not a ranking factor, and Google has said so repeatedly. Adding markup will not lift a page up the results on its own. What it does is make a page eligible for rich results, the enhanced listings with star ratings, prices and breadcrumbs, which can improve click-through. The benefit is presentation and eligibility, not a ranking boost, so treat any advice that sells schema as a ranking lever with suspicion.

What is the difference between schema and rich results?

Schema is the markup, the machine-readable label in your page’s code. A rich result is the enhanced search listing Google can show because of that markup, like a star rating or a breadcrumb trail under your title. They are not the same thing: schema is comprehension, the rich result is presentation. Valid schema makes you eligible for a rich result, but Google decides whether to actually display one, so you can have correct markup and still see a plain listing.

Can broken schema get my site penalised?

Yes. Markup that declares something the page does not show, a rating with no visible reviews, a price not in the copy, breaks Google’s rule that structured data must match visible content. The mild outcome is that Google ignores the markup. The serious one is a manual action for structured-data misuse, which removes rich-result eligibility for the affected URLs even if the markup is technically valid. This is why mismatched schema is worse than no schema, and why the rule is simple: only mark up what a user can actually see.

Do I need special schema for AI Overviews?

No. Google has stated there is no special schema required to appear in AI Overviews or AI Mode, and no AI-specific markup type. Pages are cited in AI answers with no FAQ schema and no special structured data at all. Schema still helps indirectly, through entity clarity and rich-result eligibility in classic Search, but it is not a key into the AI answer, and any tactic sold on that basis is selling something Google has said does not exist.

How do I check if my structured data is valid?

Use Google’s Rich Results Test to check validity and which rich results the page is eligible for, and the Schema.org validator to check generic syntax. For errors across the whole site, use the enhancement reports in Search Console. Remember the limit: these tools check structure and eligibility, not whether the markup matches your visible content, which is the mismatch that causes manual actions. So validate the code with the tools, then confirm with your own eyes that every claim in the markup appears on the page. Validate on publish and re-check after template or plugin changes.

Last reviewed: June 2026

This article provides general information about structured data and is not affiliated with or endorsed by Google. Supported rich-result types, validation tools and policies change, and which schema types earn rich results is narrowed by Google over time. Verify the current supported types and guidance against Google’s own structured data documentation before relying on it.