What engines actually take from schema
Less than the vendors selling schema plugins say, and more than nothing. Structured data does not put you in an answer. What it reliably does is remove ambiguity: which company this is, where it lives, what it sells, what it charges. An engine that has to guess those from prose guesses wrong some of the time, and a wrong guess is a lost citation.
This checker is not a validator. Google's Rich Results Test tells you whether markup qualifies for a search feature; this tells you what an answer engine can learn about you from the page.
The four types that matter for answers
Organization
Name, URL, logo, and sameAs links to your profiles elsewhere. sameAs is the part sites skip and the part that lets an engine tie your site to your LinkedIn, Crunchbase or Wikipedia entry.
Service or Product, with an Offer
What is sold and what it costs. On a services business this is a Service with an AggregateOffer; on software, a SoftwareApplication with Offer.
FAQPage
Questions and answers that are also on the page, word for word. Google ignores FAQ markup whose text is not visible.
WebSite
Names the site and its publisher. Small, and one line.
The mistakes we see in scans
- Schema on the home page only, while the product pages that answer buyer questions have none.
- An Organization with no sameAs.
- FAQPage questions that were edited on the page and not in the markup, so they no longer match.
- Prices in Offer that differ from the pricing page.
- A block that does not parse because of a trailing comma, which an engine skips without telling anyone.
Questions about schema
Does FAQ schema help ChatGPT?
It makes a question-and-answer pair easy to lift, and the answers we record do quote FAQ text more often than paragraph text. It is not a switch that puts you in an answer; the question has to be one buyers ask.
Microdata or JSON-LD?
JSON-LD. One block in the head, separate from the markup, is what Google recommends and what every parser reads first. This checker reads JSON-LD only.
One block or several?
Either. Several small blocks or one @graph both work. What matters is that the Organization, the Service or Product, and any FAQ are all present and consistent with the visible page.
Does wrong schema get you penalised?
Google acts on schema that misrepresents the page, for example FAQ questions not on it or reviews that do not exist. Honest, incomplete schema is simply ignored.
Readable is the first step
The full report checks all thirteen technical items and then whether the engines actually name you.