SEO

Checking your schema markup: is it valid, and does it match the page?

Schema markup that passes a validator can still be wrong. A test tells you the syntax parses and the required fields are there. It doesn’t tell you the markup describes what the visitor sees. This post is about checking both, with an example of what I found on my own site, and it deliberately doesn’t repeat which schema types are worth adding, which I cover in which schema types still matter.

What a validator checks and what it doesn’t

A validator checks syntax and whether the fields a feature needs are present. It can’t tell you the values are true, current, or the same as the visible page. A product with a price in the markup that differs from the price on screen can pass every test.

Google’s Rich Results Test is its public tool for checking whether a page’s markup can produce rich results. A Search Central post says that if you don’t see the structured data in those tools, your markup isn’t valid [1]. The Schema Markup Validator at schema.org checks general syntax against the vocabulary. They answer slightly different questions, and I’d run the page through the Rich Results Test first for anything Google uses.

Google’s documentation also says structured data lets a feature be present but doesn’t guarantee it will appear [2]. A clean test result makes a page eligible for a feature and doesn’t mean it will get one.

The rule that catches most real faults

Google’s general guidelines say structured data must reflect what users can see on the page. Don’t mark up content that readers can’t see, mark up all the visible items you describe, and keep the markup a true representation of the page [2]. Merchant Center guidance goes further for products, saying structured data must match the values shown to the user [3].

So the check beyond validation is a comparison. Open the page, open the markup, and read the two against each other. Do the price, name, dates, author, rating, FAQ answers and opening hours in the JSON-LD match the text on screen? If the answer is yes for every field, the markup is doing its job.

An example from my own site

When I went back through the structured data on this site on 1 October 2026, I found that a placeholder had been published. The datePublished and dateModified fields on 22 posts and 22 pages contained the literal text [publish date], and five newer posts had a BlogPosting with no date at all. The pages looked fine, and the markup was valid JSON, so nothing flagged it.

The cause was a template that carried a placeholder, and nothing in the upload process replaced it. I fixed it across all 44 pages and the five posts, and the upload script now stamps the real date into the markup on every publish. That’s the pattern I’d look for on other sites. A template with a placeholder in it, a plugin that outputs a default, or a field that was filled in once and never updated.

Common mismatches

Mismatch Example Check
Placeholder left in [publish date], “Lorem ipsum”, a default author Search the JSON-LD for brackets and template text
Out-of-date values A price or opening hours that changed Compare to the visible page
Markup for hidden content FAQ or review text not shown on the page Read the page as a visitor would
Duplicated markup A plugin and the theme both output the same type View source and count the blocks
Wrong type Article markup on a product page Match the type to what the page is

Our post on FAQ schema is a good example of the third row: the markup is only appropriate if the questions and answers are visible on the page.

How to test at scale

Testing one page is easy. Testing a template across hundreds of URLs is where problems hide. A crawler that extracts structured data can export it for every URL, so you can filter for blanks, brackets and duplicates, then read a sample against the live pages.

For search appearance, Search Console reports on the rich result types it has detected, and any errors it finds. Use it to see which pages Google has read the markup on, and to catch a template fault that affects a whole group.

What I’d do

Test a handful of pages from each template in the Rich Results Test. Then export the markup for the whole site and search it for placeholder text and empty fields. Compare a sample of values against the visible pages. Fix at the template, and re-crawl.

If schema is part of a larger set of checks, the technical SEO audit covers it, and the schema markup service page describes what I’d set up. For the question of whether markup helps with AI answers, structured data in the age of AI answers looks at the evidence.

Checking your schema markup

FAQ

How do I test my schema markup?

Run the page through Google’s Rich Results Test for features Google supports, and the Schema Markup Validator for general syntax. Then read the markup against the visible page. The tools check structure and required fields, and they can’t tell you whether the values are correct.

Can valid schema still be wrong?

Yes. Valid means the syntax and required fields are fine. It doesn’t mean the values match the page. Placeholders, stale prices, hidden content and the wrong type all pass validation, which is why a manual comparison is part of the check.

Does schema markup guarantee rich results?

No. Google says structured data lets a feature be present but doesn’t guarantee it will appear. Eligibility, quality and the current policy for that feature all play a part. Treat a clean test as the starting point.

Should schema match the visible content?

Yes. Google’s guidelines say not to mark up content readers can’t see, and the markup should be a true representation of the page. If the markup says one thing and the page another, that is a problem worth fixing.

Sources

  1. Google Search Central Blog, Rich results and Search Console Webmaster Conference lightning talk, August 2020.
  2. Google Search Central, General structured data guidelines, accessed 2 November 2026.
  3. Google Merchant Center Help, Set up structured data for Merchant Center, accessed 2 November 2026.

Get started

Proof, not promises.

Tell me what's going on and I'll come back with what I'd do first. No obligation, and no report you need a translator for.

Over 925,000 impressions and 13,795 clicks in a year, from people asking one awkward question: what size do I need?condoms.uk, past 12 months

Get in touch WhatsApp

Brands and sites I have worked on

Brands worked on across nearly twenty years, in agency roles and directly.