How links are counted
Content and navigation links, redirects, canonicals and pagination.
Where a link sits
Every link is placed in a region of its page, read from the page's HTML: the <nav>, <header>, <footer>, <aside> and <main> elements, their ARIA roles, and class names such as “menu”, “breadcrumb” or “sidebar”. The regions shown in the app are Content, Navigation, Header, Footer, Sidebar and Other.
A link is sitewide when the same target is linked from the same region on a large part of the site: at least 20 pages and at least 10% of all pages. The main menu and the footer are sitewide; a “related articles” box that repeats everywhere can be too.
Content links
A content link sits in the page's main content and is not sitewide. These are the links an editor places, and most of LinkTaste is about them. A page's details sort its links into three groups that never overlap:
- Content
- Links written into the page content. They carry the most weight.
- Other
- Links in a navigation, header, footer or sidebar block that does not repeat across the site.
- Sitewide
- Links from a block that repeats on most pages, such as the main navigation or the footer.
Links that page through a listing (“next”, “previous”, page numbers) are counted as navigation, even when the theme places them inside the main content.
What is counted
- Inbound and outbound links count different pages, not link occurrences: three links from one page to another count once. A page's links to itself are left out.
- Redirects are followed, up to five hops. A link to an address that redirects counts for the page it lands on, and is flagged so you can link the final address directly.
- Canonical tags are reported, not followed. A page whose canonical names another page stays a page of its own, and the Canonicalised check lists it.
- Nofollow links (
nofollow,sponsoredorugc, or a page with meta robots nofollow) pass no value. They still take their share of the page's value, as they do with Google. - Pages in the link graph are the URLs that answered with an HTML page. Links to errors, files and uncrawled URLs are kept, and show up in Issues and in the leaks in Authority.
Links built with JavaScript
The crawler reads the HTML your server sends and does not run JavaScript. On a site that renders its menu or content in the browser, those links are not seen. When more than half the pages have no links in their HTML, the overview warns about it.