Say only what the page says

What does this finding mean?

The report identifies this check as schema-content-match in Structured data. The category carries up to 16 of the 100 rubric points. This is a category weight, not an automatic deduction for this finding. Read the scan detail for the measured state, overlapping signals and any score cap.

Compare the schema headline, FAQ text, prices and dates with what readers actually see. Correct or remove unsupported values at their source.

Why does this matter for AI retrieval?

Valid markup labels the entities and content already visible on a page. Parsers can use those explicit relationships. No schema type guarantees selection by ChatGPT, Claude, Perplexity or Gemini, and this score is our rubric rather than a vendor ranking formula.

How do I check and resolve it?

1. Inspect every JSON-LD block and any microdata or RDFa on the route. Check theme and plugin output before adding another source.
2. Use the narrowest accurate schema type and stable absolute @id values. Describe the real publisher or Person, then link author and publisher references.
3. Validate JSON syntax and schema values. Compare every factual field with visible page content. Remove duplicated entities or conflicting plugin output.

Adapt the example to the reported issue and the actual route. Placeholder names, URLs and dates must be replaced with facts from your site.

Copy code

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@graph": [
    {"@type":"Organization","@id":"https://example.com/#publisher",
     "name":"Actual publisher","url":"https://example.com/",
     "description":"What this publisher actually does"},
    {"@type":"WebSite","@id":"https://example.com/#website",
     "url":"https://example.com/","name":"Actual site name",
     "publisher":{"@id":"https://example.com/#publisher"}}
  ]
}
</script>

Verify the result

Inspect the raw response after deployment and run Schema.org Validator. Use Google Rich Results Test only for supported Google features. Rescan to check the reported principle.

Open the related tool or technical guide, then run a new scan of the affected URL. Keep unknown checks marked unknown until you have a complete response.

Copy a prompt for your coding agent

Paste this into Codex, Claude Code or Cursor from your own repository. Include the scan's exact URL and detail so the agent can verify the observation.

Copy prompt for coding agent

Investigate the scanner finding "Say only what the page says" (schema-content-match) in this repository.
Observation: Some structured-data facts do not match the visible page content.
Required change: Compare the schema headline, FAQ text, prices and dates with what readers actually see. Correct or remove unsupported values at their source.
Inspect the existing route, shared layouts and configuration first. Confirm the finding against the actual response before editing. Preserve intentional access restrictions, existing content and unrelated changes. Use only factual identity and date values. Make the smallest appropriate fix, verify the deployed HTML or resource response and run the relevant project checks. Explain any measurement that remains unknown. Do not promise search rankings or AI citations.
Reference: https://letthebots.in/fixes/schema-content-match/