Cloaking is showing different content to users than to indexing bots. Yandex’s definition is broader than the classic case and includes content that changes according to how the user arrived.
What qualifies
- The site does not open via a direct link but does open for users arriving from search results or other links.
- What the visitor sees differs significantly from what the search result described — including the inverse case, where a direct link returns
404and the same URL from search results works perfectly. - One variant of the page carries block elements meant to manipulate ranking: a text block, an extra section or a tag cloud shown to the robot and hidden from users. Yandex is explicit that advertising units are not what it means here.
What is explicitly permitted
- The site being temporarily down, or some function temporarily unavailable.
- Responsive layout — different presentation on mobile and desktop with the same content.
- Regional content differences — showing a visitor prices and stock for their region.
- A/B testing, where users and search engines see the same content in different designs.
The distinction running through the whole list is content versus presentation. Changing how something looks is fine. Changing what it says depending on who is asking is not.
Why sites do this accidentally
Deliberate cloaking is rare among ordinary businesses. Accidental cloaking is not. A paywall or registration gate that opens for search referrals and blocks direct visitors is cloaking by this definition. So is a geo-redirect that sends a robot in one location to a different page than the user in another, and a bot-detection layer that serves a stripped page to anything it does not recognise as a browser.
How to check
Compare three retrievals of the same URL: as a browser via a direct link, as a browser arriving from a search result, and as the Yandex bot through the server response check tool, which shows the page content the robot actually received. If the three differ in substance rather than in layout, you have found the problem before Yandex does.
Comment nous l'appliquons
The accidental cloaking we find most often is a bot filter or a caching layer serving robots a stripped version of the page — no prices, no availability, sometimes no body text at all. It never shows up in a browser, so it survives every internal review. The server response check’s “page content” panel settles it in thirty seconds, and we run it on any site where the indexed snippet does not match what we see on screen.
Services associés