Playbook

An internal linking strategy that actually moves rankings

Internal linking moves rankings only when you control where link equity flows. Here is how to fix orphan pages, write anchors and build hub structure.

Key takeaways
  • Find orphan pages by diffing your sitemap and search data against a crawl, then give each one two to four contextual links from related pages.
  • Write each anchor as the phrase a reader would search for the destination, and vary the wording rather than repeating one exact-match phrase across every link.
  • Start with the pages that other sites already link to, and add contextual links from them to the pages you want to rank, since that is where the most authority sits.

Internal linking is the most underused lever in SEO, mostly because it looks too simple to matter. Add a link, move on. But on most sites the pages that deserve to rank are the ones with the fewest links pointing at them, while a few old posts collect internal links by accident. The fix is not more links. It is deciding which page each link is for, then placing it where both a crawler and a reader will find it and understand why it is there.

Here is the counter-intuitive part. The best internal linking opportunity on most sites is already sitting in your analytics: a handful of pages that other sites link to, and that link out to almost nothing useful. Those pages hold the most earned authority you have, and they often spend it on a changelog or an author bio. Pointing that flow at the page you actually want to rank is frequently a bigger move than publishing another batch of posts.

What internal linking actually does for rankings

Every internal link does three jobs when a crawler follows it.

It discovers the destination. Googlebot finds pages by following links. A page with no internal links depends on your sitemap to be found at all, and a sitemap tells Google that a URL exists. It does not tell Google that the URL matters.

It passes signal. Authority earned from the rest of the web, and from your own site, moves along links. A link from a page that other sites cite is a vote with weight behind it. A link from a page nobody visits is a much weaker one.

It describes. The anchor text and the sentence around the link tell Google what the destination covers. This is the job most people skip, and it is where internal links differ from external ones: you control both ends of the description.

Google’s link guidance keeps returning to the same basics. Links should be crawlable, meaning a real <a> element with an href rather than a click handler. Anchor text should describe the destination. And the link should be useful to a reader. Google does not publish a weighting formula for internal links, so anyone quoting you one is guessing. What you can work with is the mechanism, and the mechanism is enough to make good decisions.

One limit to keep in mind: internal linking redistributes authority, it does not create it. A site with no external links and thin pages will not gain anything by reshuffling its internal links. You are amplifying whatever signal already exists, so the first job is finding where that signal sits.

How link equity flows through your site

The model that works in practice is a budget. Each page has a share of signal to hand out, and each outgoing link takes a slice. The original PageRank paper described a page’s score being divided across its outgoing links. Google’s systems have moved well past that simple version, but the dilution logic still holds as a working heuristic. A page with twelve contextual links to focused, relevant destinations concentrates its signal. A page with 180 links, counting the footer, the mega-menu and a “related posts” widget that shows the same six articles everywhere, spreads it thin.

Three consequences follow.

Your homepage is a strong donor, but a blunt one. It usually has the most external links, so any page it links to gets a real boost. The problem is that navigation links are identical on every page and say little about any specific topic. The homepage links that help are the ones you curate in body copy or a featured section, tied to a subject.

Donor pages beat navigation. A page that has earned external backlinks and covers a subject close to a target page is the best place to put an internal link. On most sites that list is short, often just a handful of posts that other sites have cited. Those posts should be your first stop in any audit, because a link from them carries more weight than the same link from a page nobody else points to.

Click depth affects crawling. Pages that sit closer to the homepage, and that have more links pointing at them, tend to get crawled more often. The “three clicks from the homepage” rule circulates widely as a best practice. Google has not published it as a rule, but it is a useful target: if an important page takes six clicks to reach, it is probably being crawled less often than its value deserves.

Find and fix orphan pages first

An orphan page has no internal links pointing to it. It can still sit in your sitemap, stay indexed and even pull a few impressions. But it competes without any signal from the rest of the site, and orphans are one of the most common reasons a well-written page never gains traction.

Finding them is a set-difference problem with three lists:

  1. Crawled URLs. Run a crawler from the homepage, following only internal links. Anything it reaches is reachable through your site structure.
  2. Known URLs. Your XML sitemap, plus every URL with a Search Console impression or an analytics session in the last 90 days.
  3. Inlink counts. For each crawled URL, how many internal links point to it.

An orphan is a URL in list 2 that is missing from list 1. Screaming Frog can run this comparison directly when you connect it to a sitemap and an analytics export. If you are doing it by hand, export both lists and diff them. The 90-day window matters: a sitemap can hold years of old URLs, and the ones still earning impressions are the ones worth saving.

Once you have the list, each orphan gets one of three decisions:

  • Good page, no links. Add two to four contextual links from related pages. Spread them across different source pages rather than stacking them on one. Include at least one donor if a relevant one exists.
  • Weak page, a query you don’t want. Consolidate it into the stronger page that covers the same intent, and redirect the old URL there. The redirect carries over whatever signal the old URL had, and removing the dead page from the sitemap keeps the index honest.
  • Near-duplicate. Pick the better version, redirect or canonicalize the other, and link to the survivor.

Here is the arithmetic with invented numbers, to show the shape of the decision. Say a site has 212 indexable URLs and 31 orphans. Twelve of the orphans are decent posts still getting impressions. Giving each of those three contextual links from related posts adds 36 links to pages that had none. Of the other nineteen, eleven are near-duplicates to consolidate and eight are old pages that answer no question anyone still asks. Those get retired with redirects. The site ends with fewer URLs and more signal on each one that remains.

Write anchor text that describes the destination

Anchor text is the most visible signal in internal linking, and the most abused. The working test is simple: the anchor should be the phrase a reader would type if they wanted the page you are linking to. If the anchor could describe any of a dozen pages, it describes none of them.

Anchor type Example What it tells Google Verdict
Generic “click here”, “read more” Nothing about the destination Never
Bare URL “/blog/fix-content-decay” Little beyond the slug Rarely
Descriptive “how to find posts that stopped ranking” The destination’s topic Default
Exact-match, repeated “internal linking” on 40 posts A strong topic signal, but a visible pattern On the hub only
Branded “AutoSEO Stack” Entity identity Navigation and bios

The exact-match row needs more explanation, because it is where well-meaning teams go wrong. If every link to your internal linking guide uses the same phrase, verbatim, across forty posts, that pattern looks templated, and templated patterns are what content reviews flag. Vary the anchor the way a person writing a paragraph would, while keeping the topic words in it. “Our guide to internal linking” and “how to audit links across a large site” can both point at the same hub and still read naturally.

Two more details matter. First, the sentence around the link does real work. A link placed inside a sentence that discusses the subject is more useful to readers and crawlers than one dropped at the end of a paragraph as a “see also.” Second, linked images use their alt text as the anchor, so a logo or screenshot that links internally needs alt text describing the destination, not the file name.

Build hub and spoke structure on purpose

Hub and spoke is a topic structure. One broad hub page covers a head topic at a useful depth, and spoke pages each go deep on one subtopic. The linking pattern is what makes it work, and it has three rules: the hub links to every spoke, every spoke links back up to the hub, and spokes link to each other only where a reader would plausibly go next.

Hub: project planningGantt chart templateCritical path methodResource levelingTask dependencieslaterallateralSolid arrows: links both ways. Dashed: lateral links between neighbors.
Hub and spoke linking. The hub links down to every spoke, each spoke links back up, and lateral links connect neighboring spokes only where a reader would go next.

The reason this structure consolidates ranking is that the hub collects links from every spoke, so it becomes the most credible page on the topic and the one to rank for the broad term. Spokes then compete for long-tail variants instead of competing with each other. Without a hub, sibling articles tend to split the same signal and sometimes fight over the same query, which is cannibalization in its plainest form.

Here is a worked example. A project management product targets “project planning.” Its hub covers the topic at a survey level and links to nine spokes: a Gantt chart template, the critical path method, resource leveling, task dependencies and so on. Each spoke links up to the hub in its first few paragraphs. Two pairs of spokes link to each other, because the critical path article and the task dependency article genuinely overlap and a reader working through one needs the other. The spokes do not link to each other through a block of six “related” posts at the bottom of every page, because that pattern tells Google nothing about how the topics connect.

Two warnings about hubs. A hub that is only a list of links, with no substantive explanation of the topic, is thin and will not rank for anything. It needs its own real content. And a hub should not become a dumping ground. If a spoke does not belong to the topic, leave it out, even when a link would be convenient.

For the full build process, including how to pick the subtopics that justify a spoke, read pillar pages and content clusters, done right. For the wider argument about why depth across a subject tends to outperform a scattered library, how to build topical authority that ranks makes the case in detail.

The repeatable internal linking audit

Run this quarterly on a site of a few hundred pages, and monthly if you publish heavily. Each step produces a list you can act on, which matters more than the tool you use to produce it.

  1. Crawl from the homepage. Export every URL with its status code, indexability, click depth and inlink count. Flag any internal link that points to a non-200 URL, any indexable page deeper than four clicks, and any indexable page with zero inlinks.
  2. Pull what should exist. Export the sitemap and the Search Console pages with impressions over the last 90 days. These are the pages you are accountable for.
  3. Diff for orphans and underlinked pages. Anything in step 2 that is missing from step 1 is an orphan. Anything with impressions and fewer than two internal links is underlinked. Sort the list by impressions so the valuable pages come first.
  4. Identify donors. Use the Links report in Search Console to see which pages have external links. Those are your donors. List which of their outgoing links point at pages that are weak, and which weak pages sit in topics the donors already cover.
  5. Assign every page a home. Map each indexable page to one hub, or mark it standalone and decide whether it should exist. Pages with no cluster home are the most likely orphans, because nothing in the structure calls for them.
  6. Fix in priority order. Orphans with impressions go first, then donor links into weak clusters, then anchor cleanup on pages that are already linked.
  7. Remove what wastes equity. Update links that point at redirects or 404s to their final URLs, remove nofollow from internal links, and cut sitewide modules that repeat the same links on every template.
  8. Recrawl and measure. Confirm the new links appear in a second crawl. Then watch each target page’s impressions and average position over six to eight weeks, and log every change so you can see what preceded any movement.

Check the donor first: a contextual link from a page that other sites already cite does more than the same link from a page nobody links to. Pull the Links report before you spread links across the whole site.

Failure modes that quietly cancel the work

The silent failure: sitewide modules. A "related posts" block showing the same six articles on every page gives those six pages a fixed set of links from your whole site, dilutes the outgoing links every other page spends, and tells Google nothing about how your topics connect.

Most of the damage comes from a short list of habits:

  • Keyword auto-linking. Plugins that link every mention of a term to a single page look efficient and read badly. The result is many irrelevant links with identical anchors, which is the exact pattern you are trying to avoid.
  • Links to redirects. An internal link to a URL that redirects makes the crawler take an extra hop on every visit, and it keeps the link pointed at an address you no longer serve. Update links to final URLs whenever you consolidate.
  • Nofollow on your own pages. Marking internal links nofollow withholds nothing useful from your own site and only muddies the structure. Remove it.
  • Crawl traps. Faceted filters, calendar archives and endlessly paginated tag pages generate huge numbers of internal links to near-duplicate URLs. They soak up crawl activity that should go to the pages you want ranked. Canonicalize or block the parameters that multiply, and keep filter URLs out of your navigation.
  • Deleting without redirecting. Removing a page leaves its inbound links pointing at a 404 and drops whatever authority it had. Redirect it to the closest surviving page first, then update the links that pointed to it.

What to expect, and an honest wrap-up

Internal link changes are slower to show than most teams expect. A new link only counts once a crawler has found it, and how quickly that happens depends on the site’s size, its crawl history and its authority. There is no fixed timeline. A reasonable window to judge a change is six to eight weeks, and anything you check at ten days is mostly noise.

Judge the change page by page, not site-wide. Track each target’s impressions, its average position for the queries it is meant to rank for, and its inlink count. A single internal link rarely produces a dramatic jump. Internal linking is compounding maintenance: many small, correct links, applied on a schedule, that add up over quarters.

It will not rescue a page nobody should read, and it cannot substitute for content that answers the query better than the alternatives. What it does is make a good page findable, tell Google what that page covers, and send signal to it from pages that already carry some. Most of the gain comes from the unglamorous parts: an audit that runs on a calendar, anchors a human would write, and a hub with real content behind it.

That discipline is what AutoSEO Stack is built around. It researches what buyers actually search, writes articles in a human voice and publishes them on a schedule, so the per-page quality bar is built in rather than bolted on after the fact. If you want to see that running on your own site, start with the features.