Structured data attracts two opposite mistakes. One treats it as a ranking switch that every page should have flipped. The other dismisses it as decoration. Google’s documentation sits between the two, and reading it closely makes the right expectations clearer. This article sets out what the documentation says structured data is for, the guidelines that decide whether markup counts, what the supported gallery covers, how Organization markup relates to facts about a business, and what Google says about AI features. Attributions are to pages we read for this article; our own readings are labelled.
What structured data is for, as Google describes it
Google’s introduction to structured data describes it as a standardised format for providing information about a page and classifying the page content. The aim is to give Google explicit clues about meaning. Google says it uses the markup to understand the page and also to gather information about the web in general, such as the people, books or companies that appear in it.
The introduction gives the practical benefit as rich results: search results that can be more engaging and might encourage people to interact with a site. It links case studies from named sites that reported higher click-through; treat those as Google’s chosen examples, not a forecast. The same page suggests a before-and-after test on a few stable pages, because page views vary for many reasons.
What the pages we read do not do is present structured data as a ranking factor. They describe understanding and eligibility for search features. Our reading, not Google’s wording: it is best thought of as a way to remove ambiguity and qualify for specific appearances, not as a general score boost.
On vocabulary, most Search structured data uses schema.org, but Google says to rely on its own Search Central documentation as definitive for Google Search behaviour, because schema.org holds more types and properties than Google requires. Google supports JSON-LD, Microdata and RDFa, and recommends JSON-LD where a site’s set-up allows it. The schema.org getting started guide is the vocabulary’s own introduction; it focuses on Microdata, and says the basic ideas apply to the other formats.
The guidelines that decide whether markup counts
Google’s general structured data guidelines apply to all markup. To be eligible for a rich result, markup should not violate the content policies, which include the spam policies. Within the quality guidelines, the points that matter most day to day are these:
- Do not mark up content that is not visible to readers. If the markup describes a performer, the page body must describe that performer.
- Do not mark up irrelevant or misleading content, such as fake reviews, and do not impersonate a person or organisation.
- Provide up-to-date information; time-sensitive content that is no longer relevant will not get a rich result.
- Specify all required properties for the feature, or the item is not eligible.
- Use the most specific applicable schema.org type, and put the markup on the page it describes.
- Do not block pages carrying structured data with robots.txt, noindex or other access controls, and make sure image URLs in the markup are crawlable and indexable.
Where a page holds several related items, such as a recipe and a video, Google says to link them with @id if they are more helpful together. Otherwise it may not know the video belongs to the recipe.
The page is firm that markup is not a promise. Using structured data enables a feature to be present; it does not guarantee that it will be, even if the Rich Results Test passes. Google’s reasons include that it may judge a plain text result or another feature better for a particular search. A structured data issue can also result in a manual action, which removes the page’s rich result eligibility but, as the page states, does not affect how it ranks in web search.
What the supported gallery covers
The search gallery lists the structured data features Google Search supports, each with its own guide. It can be filtered by category, such as ecommerce, organisations, jobs, news, entertainment, food and drink, and education and science. Entries include Article, Breadcrumb, Event, Job posting, Local business, Organization, Product, Profile page, Review snippet and Video, among others. The page notes that the actual appearance in results might differ from what is shown.
Only markup that matches a listed feature, with its required properties, makes a page eligible for that feature. Our reading: the gallery is the practical starting point for choosing what to implement, because it ties each type to something Google says it can do with it.
Organization markup and entity facts
Google’s Organization documentation shows the closest link between markup and facts about a business. It says organisation markup on the home page can help Google understand an organisation’s administrative details and disambiguate it from other organisations in search results. Some properties work behind the scenes for disambiguation, while others can influence visual elements such as the logo shown in results and in a knowledge panel.
No properties are required; Google recommends adding as many relevant ones as apply. Its example includes the URL, the name, a description, a logo, contact details, an address and sameAs links to profiles elsewhere. The general guidelines still govern it, so each fact should be true and match what the site visibly says.
Validating before and after release
Google’s Organization page sets out a routine that applies to any markup: add the properties that apply, follow the guidelines, validate with the Rich Results Test and fix critical errors, then deploy a few pages and use the URL Inspection tool to check how Google sees them, making sure they are accessible and not blocked by robots.txt, noindex or a login. It also advises allowing time for recrawling.
The Rich Results Test and URL Inspection catch most technical errors, but the quality guidelines, such as relevance and visibility, cannot be fully tested by a tool. That is the gap a human review has to cover.
No special markup for AI features
Google’s AI features guidance says you do not need to create new machine-readable files, AI text files or markup to appear in these features, and that there is no special schema.org structured data to add. It lists structured data that matches the visible text among the existing fundamentals, alongside crawlability and indexing. We cover the wider guidance in our article on what Google says about AI Overviews. Google’s pages describe Google Search; they do not tell you how other AI tools use markup, and we do not claim they do.
Where this fits at Cultured Digital: markup designed with the content
Structured data and entities are part of our Technical SEO work: schema design and validation, entity understanding, machine-readable site information, and how those feed AI search visibility. Our structured data and entities guide covers that scope in more detail. On the AI side, our machine understanding page covers entity and brand understanding and structured data with consistent facts.
When we build sites, structured data is listed with the things designed alongside the content model. Our home page also shows illustrative markup under “What the site tells machines”, labelled as illustrative, as a reminder that these signals are written for crawlers and answer engines as well as people. We do not promise rich results or rankings from markup.