Internal Links: The Free Lever Almost Every Site Leaves Unused

There is an odd asymmetry in how most teams spend their SEO effort. Enormous energy goes into acquiring links from other websites — outreach, content campaigns, relationships, budget — while the links on their own site, which they control entirely and can change this afternoon, are whatever the template happened to produce three years ago.

This article is about correcting that using the Semalt platform to see the structure you actually have: finding orphan pages, measuring click depth, deciding where links should point, and turning the whole thing into a habit rather than a one-off project.

What internal links actually do

Three separate jobs, and conflating them leads to bad decisions.

They make pages discoverable. A crawler finds pages by following links. A page with no incoming internal links is reachable only if it happens to be in the sitemap, and being in a sitemap is a much weaker signal than being linked from a real page. In practice, orphan pages get crawled infrequently and treated as unimportant, because structurally the site is saying they are.

They distribute authority. Whatever value the site has accumulated flows through internal links. A page linked from the homepage and from twelve relevant articles is being described as important. A page linked from nowhere is being described as an afterthought, regardless of how good it is.

They describe destinations. The anchor text tells both readers and crawlers what to expect on the other side. "Read more" describes nothing. "How we price technical audits" describes a page precisely, and it costs nothing extra to write.

See also: Images and Video.

Working parameters from the method described below.

Finding the pages nobody links to

Orphans accumulate silently. A page is created for a campaign and linked from a banner that later disappears. A service is added and never added to the menu. A blog category is retired and its articles stay published, now unreachable. Nobody notices, because the pages still load perfectly when you type the address.

The crawl finds them by comparing what it can reach by following links against everything that exists in the sitemap and in the site's own records. The gap is your orphan list, and on a site older than three years it is nearly always longer than anyone expects.

What to do with each depends on the page, and there are only three answers:

Click depth and why it matters

Click depth is the number of clicks from the homepage to a page along the shortest path. It is a rough proxy for how the site itself ranks its own content, and search engines read it that way.

The practical rule: anything you want to rank should be reachable in three clicks or fewer. Beyond four, crawl frequency drops noticeably and pages start behaving as though they are peripheral — because in the site's own structure, they are.

Deep pages usually result from one of three architectural patterns, and each has a straightforward fix.

1. Deep category nesting

   Category, subcategory, sub-subcategory, then the page. Flatten by linking important leaf pages directly from higher levels, rather than restructuring the whole taxonomy.

2. Pagination burial

   Older items reachable only through page fourteen of a listing. Add hub pages that group items thematically, so nothing depends on someone paging through.

3. Menu-only navigation

   If the only internal links on a site are in the menu and footer, every page is described identically. In-content links are what create meaningful structure, and templates cannot produce them.

Anchor text without overthinking it

Internal anchor text is one of the few places where you can describe a page in your own words and have it count. Two rules cover almost everything.

The second item on the right deserves a note, because it is the classic overcorrection. Linking to a page fifty times with the identical exact-match phrase looks engineered and reads badly. Natural variation — the service name, a description of the problem it solves, a question it answers — is both safer and more useful to readers.

Building hubs that hold a topic together

The structural pattern that consistently works is simple: a hub page covering a topic broadly, linking to detailed pages beneath it, each of which links back to the hub and sideways to its siblings.

This does three things at once. It gives the search engine a clear statement of which pages belong together and which is the primary one. It gives readers a route through a subject rather than a series of dead ends. And it means a new page added to the cluster inherits relevance immediately, because it arrives linked rather than isolated.

The failure mode is a hub that is only a list of links. If the hub itself has nothing to say, it will not rank, and a hub that does not rank passes on very little. Write the hub as a page worth reading on its own, then link from it.

We cover this in detail in Structured Data in Practice.

Structure on a shop, where the rules bite hardest

Everything above applies more sharply to an online store, because a shop generates structure automatically and the automation has no opinion about what matters commercially.

The recurring pattern is that the category with the highest margin sits at the same depth as the category with four products, because the navigation was built alphabetically. Products that sell nothing receive the same internal linking as the ones paying the rent. Nobody decided this; the template did.

Three adjustments correct most of it without restructuring anything. Link your best categories from the homepage explicitly, rather than relying on a dropdown that appears on every page equally. Add contextual links between products that genuinely relate — the accessory that fits, the replacement part, the next size up — which is useful to the shopper and creates real structure at the same time. And make sure category pages link down to their subcategories in the body content, not only in the sidebar filter, since filter markup is often invisible as navigation.

One warning specific to shops. Faceted filters generate enormous numbers of internal links to URLs that should not be indexed at all, and those links dilute the signal reaching your real category pages. If the filters are already blocked from indexing, make sure the links to them are not spraying crawl attention across combinations nobody searches for. That single check often explains why a well-built shop has category pages that never get crawled deeply.

Making it a habit rather than a project

Internal linking degrades continuously as a site grows. New pages arrive unlinked, old pages lose links when templates change, and after two years the structure bears little resemblance to the plan.

Two habits keep it stable at almost no cost. First, at publication: before anything goes live, decide the two or three existing pages that should link to it and add those links. Two minutes, and it prevents the orphan problem entirely at source. Second, quarterly: run the crawl, check the orphan list and the depth distribution, and fix what drifted. Half an hour every three months keeps a site's structure honest indefinitely.

A note for smaller sites

Most Portuguese business sites are between thirty and three hundred pages, which is small enough that this work is genuinely finishable — and that changes the calculation. On a fifty-thousand-page site, internal linking is a governance problem requiring rules and automation. On a hundred-page site, one person can map the entire structure in an afternoon and fix it in a day.

That is worth saying plainly, because the published advice on this topic is written for large sites and reads as intimidating. If your site has a hundred pages, you can achieve a genuinely good internal structure this week, and it will keep working with quarterly maintenance measured in minutes.

Where to start

Run a crawl and look at three things: the orphan list, the click depth of your commercially important pages, and how many internal links point at each of them. Those three numbers describe your structure completely enough to act on.

Then fix in this order: orphans that matter, commercial pages buried deeper than three clicks, and pages with only one incoming link. That sequence puts the effort where the return is, and it is usually a day of work rather than a project.

If you have never looked at your own internal link structure, that crawl is the place to begin. Open the dashboard, crawl the site, and check how many pages have exactly one internal link pointing at them.