Internal linking: the free lever most sites ignore

An internal link is a link from one page on your site to another, and it does two quite separate jobs that are worth separating because they suggest different decisions. The first is distribution: links are how importance moves around a site. A page linked from many places, particularly from pages that are themselves well linked, is being pointed at repeatedly by its own site, and a search engine reads that as the site's own statement about which of its pages matter. The second is meaning: the words in the link tell both the reader and the engine what the destination is about.

The second job is the one businesses underuse. The visible text of a link is the strongest hint anyone gets about the page at the other end, and it is entirely within your control — unlike the text of an external link, which somebody else writes. A site whose internal links all say 'read more' and 'click here' has thrown that away across every page at once, and has also made the site harder to use for anyone navigating by screen reader, since a list of links reading 'click here, click here, read more' describes nothing.

Most levers in search require something you do not control. Earning links from other sites depends on other people's decisions. Competitive data costs money. Rankings are somebody else's system. Internal links are different in kind: they are entirely yours, you can change them this afternoon, and you need nobody's permission or budget.

That also makes them the thing most often left in whatever state the site's template produced. A typical small business site links according to its navigation and nothing else, which means every page receives exactly the links the menu happened to give it and no more. A page's importance to the business — the treatment that pays the bills, the product with the margin — is nowhere reflected in how the site links to it, because nobody ever made a decision about linking as distinct from a decision about the menu. The gap between the pages that matter commercially and the pages the site actually points at repeatedly is usually the largest easily-fixed problem on a small site.

An orphan page is one nothing on the site links to. It exists, it may be perfectly good, and it can only be reached by knowing the URL. Nobody creates one deliberately; they are byproducts. A campaign landing page was built for a specific push, linked only from an advert, and stayed after the campaign stopped. A navigation menu was simplified and three pages fell out of it without being deleted. A product went out of stock and was quietly removed from the category listing while the product page remained. A blog series was linked from a sidebar widget that a redesign removed.

The consequence is not just weak ranking. A page nothing links to receives none of the distribution described above, which leaves it competing on almost nothing, and a search engine may struggle to discover it at all beyond a sitemap entry. A sitemap can get it crawled and does not make up for the absence of links, because a sitemap says 'this exists' where a link says 'this exists and here is what my site thinks of it'. The page ends up in the worst position available: indexed, unsupported, and invisible to anyone browsing.

A repeatable method, using only your own site. Start with a list of every page — from your sitemap, your CMS or your platform's page list. In a second column, write what the page is for, commercially: is it something you sell, something that explains what you sell, or background. In a third column, count how many other pages on your site link to it, which you can do by hand on a small site and by searching your own content for the URL on a larger one.

Now sort by the second column and read the third. The pages that are commercially important and have one or two internal links are the finding, and there are usually several. The fix is not to add links mechanically to every page; it is to find the places where a link genuinely helps a reader who is already reading about the subject. An article explaining how a treatment works should link to the page where the treatment is booked, in words that say what it leads to. That link helps the reader, carries meaning, and moves distribution to the page you care about, all at once — which is why editorial links inside body text tend to outperform another entry in a footer nobody reads.

The first is vague link text, covered above and worth repeating because it is so common and so easy to fix. Write what the link leads to. The second is linking everything to everything: a page with forty internal links in its body spreads its distribution thinly and reads like a directory, and the reader has no idea which link was the important one. Fewer, better-chosen links do more.

The third is the same anchor text pointing at different destinations, which makes both pages harder to characterise — if 'dental implants' links sometimes to the service page and sometimes to an article, the site is giving contradictory signals about which page that phrase describes. The fourth is depth: a page reachable only after five clicks from the homepage is being told, structurally, that the site does not consider it important. If something matters commercially, it should be close to the front. The fifth is treating a footer with every page in it as an internal linking strategy; a link that appears in the same place on every page carries very little information about any of them.

Almost everything in this article is verifiable from your own site for nothing, which is unusual in search. You can read every page's links out of its own markup. You can identify vague anchor text by looking at it. You can count how many pages link to a given URL by searching your own content. You can measure click depth by walking your own navigation. None of it needs a subscription, and you can re-check any of it the moment you change something.

Two honest limits. Finding every orphan page across a large site properly needs a full crawl of it, which is a piece of tooling rather than a purchase of data — the reason it is not trivial is engineering effort, not a vendor licence, and on a small site you can substitute the page-list method above. And whether a linking change improved anything is not answerable from your markup; that needs impression and click data from a verified Search Console property, or nothing. So do this work because the structure is wrong in ways you can see and describe — a commercially critical page with one link pointing at it is a defect on its own terms — and treat any measured improvement as confirmation rather than as the justification.