Contents 4
Window 1 is the HTML after a cache-bypassed request – the document a fetcher that never runs JavaScript can read. Window 2 is the document after load. They are reported apart and never summed.
What does the tool measure?
JSON-LD blocks, Microdata itemscope, RDFa typeof, declared types and sameAs values. Two counts. Not traffic, not rank, not citation.
How do you run it?
- Check this page – same-origin, complete comparison. Request with cache reload, then read the live document.
- Check delivered HTML for another URL – the browser will often block the read. A blocked fetch is a blocked fetch, not “no schema.”
- Open two windows – live URL in one browser window, parsed delivered HTML in the other.
- Bookmarklet on the target origin, in a private window. That is the cache deletion.
How do you read the verdict?
- Same types in both windows: the graph arrived with the response.
- Schema only after load: validators that render will report a graph; a fetcher that stops at the response will not.
- Nothing in either state: no Schema.org graph was found in the markup we could read. That is all it says.
What are the limits?
- One document, two states. Nothing about how any system uses the difference.
- Consent banners, personalisation and A/B tests change documents that are otherwise sound.
- The sibling measurement for words, images and tables is the client-side rendering check.
Next in this set Rich Results · Entity Recognition
Elsewhere on this site Entity grounding · Semantic search · Named entity recognition · Client-side rendering check · Large language models · Answer engine optimization · Generative engine optimization · Evidence · E-E-A-T