Internal linking is the skeleton of a website. It decides which pages search engines find, how relevance is distributed within a topic and whether visitors get from a general question to the specific answer. On small sites it can be maintained by hand. With hundreds or thousands of pages it needs rules — and a check that ensures they are followed.
Hub and cluster
The basic pattern is one hub per topic linking to all pages of the cluster, and cluster pages linking back to the hub. On innopulse.io these are, for example, a foundational article per topic area, the overview pages of glossary, checklists and decision guides, and the service pages leading to the related offers. That gives each topic a recognisable centre. The underlying idea is described in our article on topical authority.
Meaningful cross-links
Within a cluster, pages link to each other where it makes sense: a glossary term to related terms, an expert article to the matching checklist, a cost page to the offer and the decision guide. What matters is the benefit to the reader. A link set only for SEO reasons is usually a weak signal for search engines too.
Link sibling pages in rotation
With many similar pages, such as glossary terms in one category, a problem arises: if all link to the same three neighbours, the same pages always benefit. Better is a rotating selection in which each page links to different siblings, so all receive inbound links.
Anchor text
The linked text describes the target: “high-risk classification checklist” rather than “here”. Variation is natural and welcome; identical anchor text across hundreds of pages looks mechanical. For glossary terms, the term itself is the best anchor.
Click depth
Every important page should be reachable from the home page in a few clicks. Pages deep in the structure are crawled less often and considered less important. On innopulse.io the maximum click depth is currently three — with over 1,300 indexable pages.
At least two inbound links
A page without inbound internal links is orphaned: it exists but is found only through the sitemap. A page with exactly one inbound link hangs by a thread. Our rule: every indexable page has at least two inbound links from thematically related pages. On average, pages on this website currently receive just over thirty.
Check automatically
These rules only work if they are checked. After every build, an audit computes the link graph from the HTML actually built: orphaned pages, pages with only one inbound link, pages unreachable from the home page and pages with excessive click depth. If any of these numbers is not zero, the batch does not go live. That is exactly how we recently caught a new glossary category that was built but not linked from the overview — and fixed it before release.
Common mistakes
- New pages are added to the sitemap but linked nowhere.
- Navigation and footer link everything, but no content points to specific pages.
- Related pages are selected at random rather than by topic.
- Internal links to redirected or non-indexable pages.
- No post-build check, so errors surface in Search Console only months later.
We offer building such structures in our topical authority cluster offer; how high speed combines with these rules is described in our article on content velocity.