Schema markup validator tools and the 8 most common mistakes

You test schema markup with two free tools: the Schema Markup Validator for general schema.org correctness and Google’s Rich Results Test for the question of whether Google can turn it into a rich result. Most markup does not fail because of something exotic, but because of a handful of recurring mistakes that let your code exist while making it invisible to search engines and AI. In this guide you will walk through the right testing order and the 8 mistakes that quietly kill your markup, including how to spot and fix each one.

This is a practical test-and-debug layer underneath the GEO guide. Because schema markup is no longer just an SEO trick: it is one of the ways you tell an AI model, in machine-readable terms, who you are and what a page means.

Why should you test schema markup at all?

Testing schema markup is necessary because a mistake in your structured data is usually silent: no error message appears, the markup is simply ignored. Unlike a broken link or a 404, you see nothing in the browser. The code sits neatly in the page source, but Google and AI systems skip it because it is not parsable or not trustworthy.

That makes schema treacherous. You think your page is passing along extra signals, while in practice those signals never arrive. Testing is the only way to see the difference between “markup present” and “markup understood”. And that difference is big, because correctly validated structured data is one of the foundations of how AI determines what is true.

Which tools do you use to test schema markup?

To test schema markup you use three tools in a fixed order: first the Schema Markup Validator, then the Rich Results Test, and after that Google Search Console for your live pages. Each answers a different question.

• Schema Markup Validator (validator.schema.org). This is the official schema.org tool, the successor since 2021 to the old Structured Data Testing Tool. It checks your markup against the full schema.org vocabulary, without Google-specific rules. Use it to track down syntax errors and non-existent types or properties. It tells you whether your markup is correct, not whether Google does anything with it.
• Rich Results Test (search.google.com/test/rich-results). This tool checks whether your page qualifies for a Google rich result. It is stricter and more specific: it only knows the types Google supports and requires the mandatory properties per type. Use it as the second step, once the syntax is right.
• Google Search Console. The two tools above test a snapshot or a URL. Search Console shows the structured data on your actually indexed pages, at scale and over time. The Enhancement reports are where you see the errors that only surface on the real, crawled version of your site.

The most reliable approach is a chained check: first validate schema.org correctness, then check Google eligibility, and finally monitor the live situation in Search Console. Anyone relying on a single tool misses a layer every time. You will find more about the broader arsenal in AI visibility tools.

The 8 most common schema markup mistakes

1. Invalid JSON-LD syntax

The most common mistake is a broken JSON structure: a missing comma, a stray quotation mark or an unclosed brace. JSON-LD is unforgiving. One character out of place and the entire block is ignored, not just the line with the error. The Schema Markup Validator points to the exact line. Paste your markup into it before you put it live and you catch this in seconds.

2. Markup that does not match the visible content

Your structured data has to describe what is really on the page. Mark up a price, review or event that the visitor cannot see and you are breaking Google’s guidelines. In the best case it gets ignored, in the worst case a manual action follows. Do not test this only technically: put the markup next to the page and check that every field has a visible counterpart.

3. Invented or misspelled types and properties

Schema.org has a fixed vocabulary. Write Organisation instead of Organization, or invent a property that does not exist, and it will be ignored. The validator flags unknown terms as a warning, but those warnings are often clicked away. Check every type and property name against the schema.org documentation, certainly with hand-written markup.

4. Missing required properties

Every rich result type has required fields. A Product without price or availability, an Article without a headline or date: the markup is then syntactically correct but incomplete. The Rich Results Test is your best friend here, because it names exactly which required property is missing per type. The validator often still lets it through, the Rich Results Test does not.

5. Markup that is not rendered or crawled

Sometimes your code is perfect, but it never reaches the search engine. That happens when schema is only loaded via JavaScript and the crawler does not see the rendered version, or when the script sits in a block that is accidentally blocked. So always test the live URL, not just your standalone code snippet. The Rich Results Test shows the rendered HTML, so you can see what Google actually receives.

6. Loose entities without any connection

This is the mistake that weighs heaviest in an AI era. Many sites scatter loose blocks of markup around without connecting them: an Organization here, a Person there, a Product elsewhere, not a single reference between those entities. With @id references you link them into one coherent graph, so a model understands that the author, company and product are about the same entities. Without that coherence, an AI gets loose facts instead of a clear picture. This is the core of entity consistency for AI visibility.

7. Wrong @context or case sensitivity

Schema.org markup is case-sensitive and context-dependent. A @context that points to the wrong address, or a type with the wrong capitalisation, breaks the interpretation. Always use https://schema.org as the context and respect the exact spelling (FAQPage, not Faqpage). Small deviation, big consequences.

8. Relying on deprecated rich result types

This is where many sites go wrong without noticing. Google discontinued HowTo rich results in 2023 and phased out FAQ rich results step by step, first limiting them to certain authoritative sites and then retiring them entirely. Anyone still counting on those search appearances is debugging something that will not show up in the SERP anyway. Important: the underlying markup is not “broken”. FAQPage remains a valid schema.org type and is still read to understand your page. The value has shifted from a visual SERP element to machine readability for AI systems, and that is exactly why you should not simply strip it out.

What does a valid validation actually mean?

A green message in the validator means your syntax is right, not that you get a rich result or that AI cites you. This is the biggest misconception around schema markup testing.

At Customer Impact we therefore do not steer towards a “green checkmark” as a goal in itself, but towards the underlying effect: are you understood correctly and does that contribute to visibility that produces leads? Schema is a means, not a vanity metric. To be fair: correct markup increases your chance of being found and cited, but nobody can guarantee a rich result or an AI mention. Anyone who promises that is overselling. How schema fits into a broader, extraction-friendly setup is covered in content architecture for AI extraction.

How do you keep your schema markup error-free at scale?

Keeping schema error-free is not a one-off check but a maintenance routine, because most mistakes creep in during changes. Three habits make the difference:

1. Test every template, not every page. On a well-built B2B site, schema comes from templates. Test one representative page per template thoroughly, because a mistake in the template repeats across hundreds of pages.
2. Retest after every change. A redesign, a new CMS field or a plugin update can quietly break markup. Make a retest a fixed step in your release process.
3. Monitor in Search Console. Put the Enhancement reports into your regular rhythm. That way you see errors on the live, indexed version of your site, even when a standalone test looked green.

This discipline is the same one you apply to the rest of your structured data and is part of how we approach GEO optimisation for B2B companies.

The short summary

Schema markup testing comes down to the right order and avoiding a handful of recurring mistakes. First validate the schema.org syntax in the Schema Markup Validator, then check Google eligibility in the Rich Results Test, and monitor your live pages in Search Console. Watch out above all for the silent mistakes: invalid JSON, markup that does not match the visible content, invented properties, missing required fields, unrendered code, loose entities without connections, the wrong context, and reliance on deprecated rich result types. And remember that a valid validation guarantees neither a rich result nor an AI citation: it is a necessary foundation, not an endpoint.

Want to know whether your markup really helps you get found by AI, instead of just being technically correct? Book your free intake and we will look together at where your schema stands.