How to know if your schema is working
- The two free tools that confirm whether your schema is working
- How to use each one in plain steps
- What to do when something is wrong
Why this lesson matters
Adding structured data to a website is one thing. Knowing whether it’s working is another — and most websites that have schema in place have never actually verified it. Plugins claim to add it. Themes promise it’s included. Developers tick it off as done. None of these assurances are the same as confirmation.
Confirmation matters because broken schema is genuinely common. A typo in a field. A property that points to nothing. A type that’s been deprecated. A plugin that’s outputting markup the search engines no longer accept. Any of these can mean that the structured data you assume is helping you is doing nothing — or, occasionally, doing harm.
The good news is that confirming it works takes about two minutes and uses tools that are free, public, and reliable.
The two tools you need
There are two tools that, between them, cover essentially everything a website owner needs to check.
Google’s Rich Results Test. Available at search.google.com/test/rich-results. This is the tool that tells you whether your schema is valid in Google’s eyes and whether any of it qualifies for the “rich results” — the enhanced search listings with stars, FAQs, images, and so on. If something is wrong from Google’s perspective, this tool finds it.
The schema.org Validator. Available at validator.schema.org. This is the tool that tells you whether your schema is valid against the formal schema.org specification — which is the same specification that AI systems read. It catches errors that Google’s tool ignores and shows you the full structure of what you’ve published. It’s the more thorough of the two.
You use both because they catch different things. Google’s tool reflects what Google rewards. The schema.org validator reflects what’s actually correct. A small percentage of pages pass one and fail the other, and knowing which is which is useful.
How to use them
The process is the same for both:
Step 1. Open the tool in your browser.
Step 2. Paste the URL of the page you want to check into the field at the top. Click “Test URL” (Rich Results Test) or “Run Test” (Validator).
Step 3. Wait a few seconds. The tool will crawl the page, find the structured data, and display what it found.
Step 4. Read the result.
That’s the whole flow. There’s no signup, no integration, no command line. If you can paste a URL into a form, you can test schema.
Reading the results
Both tools give you broadly the same kinds of feedback. The headlines to look for:
Green ticks or “valid” messages. Mean the structured data was found and is well-formed. The type the tool detected (Person, Article, FAQPage, etc.) is listed clearly. This is what you want to see.
Warnings. Mean the schema is technically valid but has missing or weak fields — for example, a Person schema with no sameAs links, or an Article schema with no author specified. Warnings are not errors; the schema still works. But they’re worth fixing where you can, because each warning is a missed opportunity to give AI systems more clarity.
Errors. Mean something is genuinely broken. Maybe a required field is missing. Maybe a value is the wrong type (a date where a name should be, or vice versa). Maybe the schema points to a URL that doesn’t exist. Errors usually mean the schema isn’t doing what you think it’s doing, and they’re the ones to fix first.
No structured data detected. Means the tool found nothing — either there’s no schema on the page at all, or what’s there is broken so completely that the tool can’t parse it. Either way, this is the result that says “your schema isn’t working.”
Read the result, identify the category, and act on what’s there. If everything is green, you’re done. If there are warnings, decide whether they’re worth fixing. If there are errors, fix them. If there’s no schema at all, you’ve found a gap.
What to do when something is wrong
If you find errors or missing schema, you have three options.
Fix it yourself if you’re using a plugin. Most schema plugins — Rank Math, Yoast, Schema Pro — have settings panels where you can correct missing fields, change types, or add sameAs links. The validator’s error messages usually tell you exactly which field is missing, and the plugin’s interface usually has a field with that exact name.
Brief a developer. If the schema is hand-coded or embedded in the theme, this is a developer fix. Send them the URL, a screenshot of the validator’s error message, and a brief description of what you expected the schema to say. That’s enough to get the fix made cleanly.
Add it from scratch. If there’s no schema at all, you need either a plugin (for most CMSes) or a developer (for custom builds). Either route is valid; which one to take depends on how much custom work the rest of your site already has.
You don’t need to memorise the codes or learn the formats. You just need to know that the tools exist, that the results are readable, and that “no schema detected” means there’s work to do.
A note on validation frequency
Schema isn’t a “set it and forget it” job, but it doesn’t need constant attention either. A reasonable rhythm:
At launch. Test every important page — homepage, About, services, contact, and a sample of blog posts. Make sure each one returns what you expect.
After any major site change. Theme updates, plugin migrations, redesigns, and CMS upgrades can all silently break schema. Test the key pages after each.
Quarterly, lightly. Run the homepage and one or two key pages through both tools every few months. Plugins update, schema.org evolves, and Google occasionally changes what counts as valid. A quick check catches issues before they accumulate.
When something feels wrong. If you notice your business being described inaccurately by an AI tool, or your rich results disappearing from search, schema is one of the first things to check. The tools are quick enough that there’s no reason not to.
A useful mindset
Schema that hasn’t been validated is schema you’re hoping works. Schema that’s been validated is schema you know works. Hope is not a strategy.
If you can confirm — yourself, in two minutes, with free tools — that the structured data on your key pages is valid and complete, you’ve done the verification work that almost no other website does. That alone is a significant advantage.
Coming up in the next module: The new technical signals. We leave structured data behind and look at the technical layer that’s genuinely specific to AI — llms.txt, AI crawlers, and the decisions you need to make about which AI systems can access your content. Three lessons, all practical, all decision-focused.