Schema.org

Schema.org is the shared vocabulary announced by Google, Bing and Yahoo in summer 2011 and supported by Yandex. It describes entities and their properties through attributes written directly into the HTML — no separate export file, no feed to maintain.

How markup is written

Two steps. Wrap the entity in a container declaring its type, then label each value with a property attribute:

<span itemprop="streetAddress">Lev Tolstoy, 16</span>

An unmarked block of contact details is text to a machine. The same block with name, streetAddress, postalCode, addressLocality, telephone, faxNumber and email labelled is a record that can be extracted and displayed. The vocabulary runs to several hundred types and is extended continuously.

What Yandex actually does with it

Markup feeds specific snippet types, and only these are supported:

• Products and prices — special snippets for product pages.
• Software — applications, programs, games.
• Recipes.
• Film descriptions.
• Questions and answers — for mobile search, where Yandex can highlight the best answer.
• Student research and similar works.
• Organisation information and address — used by Yandex Business as well as Search.
• Images — improves their appearance in Yandex Images.
• Videos.

Marking up a type outside this list does no harm and produces nothing in Yandex.

Checking it

The structured data validator in the console is the authority on whether the markup parses. For common content management systems there are usually plugins; the platform documentation is the place to look before commissioning custom work.

The clause worth reading twice

Yandex states that information marked up according to Schema.org becomes publicly available and can be extracted or used by any service. That is the intended purpose, but it is also a decision: pricing, stock levels and contact details in machine-readable form are trivially harvested by competitors and aggregators. Marking up is publishing, not merely describing.