Internal linking is the highest-leverage technical SEO work available to most sites, and the most consistently neglected, because it does not feel like a discrete project. It is not a task you complete. It is a property of your site's structure.
The model
Your site is a directed graph: pages are nodes, links are edges. Authority — from external links, however a search engine currently computes it — enters at certain nodes and distributes through the edges.
The original PageRank formulation distributes a page's value across its outbound links and applies a damping factor at each hop. Google's actual system has evolved far beyond that paper, and the specifics are not public. But the structural intuition survives and is directly useful: pages that receive many internal links from well-linked pages accumulate more authority than pages that do not, and pages nobody links to accumulate nearly none regardless of quality.
Four failure modes
Orphans. Pages with zero internal inbound links. They exist in the sitemap and possibly the index, and they receive nothing. Every CMS produces these eventually — a page unlinked when navigation was redesigned, a landing page built for a campaign that ended.
Find them by crawling your own site and diffing the discovered URL set against your sitemap. Anything in the sitemap but absent from the crawl is orphaned.
Depth. Click distance from the homepage. Pages more than three or four clicks deep receive materially less crawl attention and less authority. Deep is not always wrong — a large e-commerce catalog cannot be flat — but important pages being deep is always wrong.
Uniform distribution. A site where every page links to every other page through a giant footer distributes authority evenly, which means it distributes it to nowhere in particular. Concentration is the point. If everything is emphasized, nothing is.
Anchor text waste. "Click here," "learn more," and "read more" carry no topical signal. Anchor text is one of the clearest relevance signals available for internal links and it is free.
The hub-and-spoke structure
The pattern that works for most content sites: a hub page covering a topic broadly, spokes covering sub-topics in depth, spokes linking back to the hub, and the hub linking out to every spoke.
This does two things simultaneously. It concentrates authority on the hub, which is usually the page targeting the competitive head term. And it creates a dense, topically coherent cluster, which helps both conventional relevance signals and semantic retrieval systems recognize the site as covering the subject rather than mentioning it.
The common failure is building the spokes and forgetting the hub-to-spoke links, or building both and never linking spokes to each other where they are genuinely related.
Practical rules
Link from strong pages to pages that need help. Your highest-authority pages are usually your homepage, your most-linked blog post, and your main service pages. A link from any of those is worth more than a link from a page nobody visits. Identify your top pages by external links, then check what they point at.
Link contextually, in the body. In-content links carry more weight than navigation and footer links, which appear site-wide and are discounted accordingly. A relevant link inside a paragraph is worth several in a footer.
Use descriptive anchor text, varied naturally. Describe the destination. Vary the phrasing across links because natural language varies, not because you are avoiding a penalty.
Link new content from old content. New pages start with no authority. The fastest way to give them some is a link from an established page that already gets crawled regularly — which also gets them discovered faster.
Audit outbound link counts on hub pages. A page linking to 200 destinations distributes very little to each. If a category page links to everything, consider whether it should link to the most important things.
Measuring it
You do not need enterprise tooling. A crawl of your own site produces the edge list. From there:
- Count inbound internal links per URL. Sort ascending — the bottom of that list is your problem set.
- Compute click depth from the homepage via breadth-first search.
- Cross-reference against Search Console clicks. Pages with meaningful impressions, poor position, and few internal links are the highest-return fixes on the site.
That last cross-reference is the whole exercise in one query. Pages Google already shows to people, that rank just below where clicks happen, that your own site barely links to — fixing those is fast, cheap, and produces movement more reliably than almost anything else in technical SEO.
Why it stays neglected
Internal linking has no owner. It is not a content task, not a development task, and not a design task, so it falls between them. The result is that most sites' link graphs are an accident of how the site grew rather than a reflection of what matters now.
Setting aside two hours a quarter to look at the graph deliberately — orphans, depth, the inbound-link-count list — catches nearly all of the damage before it compounds.