Understanding schema types

A type is a noun

Schema.org is a shared vocabulary that search engines and answer engines read. At its heart is a list of types, and each type is a kind of thing: an Organization, a Person, an Article, a Product, an Event. When your page says "this is an Article", a crawler knows what it is looking at before it reads a word of the copy.

Each type comes with properties, the facts you can state about it. An Article has a headline, a datePublished and an author. A Product has a brand, an offers and a review. The type decides which properties make sense.

Types sit in a family tree

Types inherit from each other. NewsArticle and BlogPosting are both kinds of Article, so they carry every Article property plus a few of their own. InsuranceAgency is a kind of FinancialService, which is a kind of LocalBusiness, which is a kind of Organization.

This matters when you choose. A more specific type tells a crawler more, as long as it is true. Sirius Insurance is an InsuranceAgency, so that type says more than Organization does. A page about the company's history is still about the company, so it is still an InsuranceAgency, not an Article.

⤡ Expand

How to choose

Ask three questions about the page, in this order.

1. What is the page about? Not what it contains, what it is about. A page that lists travel insurance products is about products. A page that explains how travel insurance works is an article.
2. Is there a more specific type that is true? Start with the general type, then walk down the tree while every step stays true. Article, then TechArticle if it is technical. Organization, then LocalBusiness if customers visit a place, then the specific business type if one fits.
3. Does Google or an answer engine do something with it? Some types unlock rich results or feed answer engines directly: Article, Product, FAQPage, HowTo, JobPosting, Event, LocalBusiness, BreadcrumbList. If your page is one of those, use that type. If it is not, a true type is still worth more than no type.

Mistakes we see most

• Marking up the container, not the thing. A page that shows one product is a Product, not an ItemList or a WebPage. Use WebPage for pages that are about nothing else, such as a privacy policy.
• Choosing the type you wish you were. A company blog post is a BlogPosting, not a NewsArticle, however newsworthy it feels. Crawlers check.
• One type for every page. The home page, the product pages and the articles are different things. The value of schema grows with the number of content types you can tell apart.
• Forgetting the things on the page. A page has one main type, but it can mention others. An Article has an author who is a Person and a publisher that is an Organization. Those are the connections that let a crawler join your pages into one picture.

How this works in the app

In JSON-LD Schema Manager, each type is a block. Around 30 come ready to use, covering the types enterprise sites need most, and the whole schema.org catalogue sits behind them. You pick the block that matches your content type, and the graph does the rest. See Blocks vs graphs: where values live for what happens next.

⤡ Expand