When I audit a site, internal linking is usually where the cheapest wins are hiding. Not because the team doesn't know links matter, but because nobody owns them. Developers own templates, writers own articles, and the links between them happen by accident. This part of my technical SEO guide is about owning them on purpose.
What internal links actually do
An internal link does three jobs at once.
Discovery. Googlebot finds pages by following links. A page with no internal links is an orphan: it may be in your sitemap, but it gets crawled rarely and indexed reluctantly. I covered the diagnosis side of this in crawlability and indexability.
Signal distribution. Google still uses link-based authority, and internal links are how it flows through your site. Your homepage and a few strong pages collect most external links; where they point internally decides which pages inherit that strength. Most sites pour it into the "About" page and privacy policy via the footer, and trickle almost nothing to the category or guide pages that should be ranking.
Context. The anchor text and surrounding sentence tell Google what the destination is about. A hundred internal links saying "kurta sets for women" is a louder, more reliable signal than any amount of on-page optimisation on the target page itself.
External links are something you earn. Internal links are something you decide.
The five types of internal link
Each type has a different job. Problems start when one type is asked to do another's work.
Navigation links
Header and primary menu. They define the top of the hierarchy and appear on every page, so they carry enormous weight. Keep them to the sections that matter most; every item you add dilutes the rest.
Contextual links
Links inside body copy, in a sentence that explains why you'd click. These carry the richest context and are the ones most sites under-use. They're also the hardest to scale, which is why they need rules (below).
Hub links
A hub (category, pillar, topic page) links out to every page in its cluster, and every page links back up. This is the backbone of topical authority and the structure I use for every content programme.
Related links
"You might also like" modules. Useful when they're actually related (same cluster, same intent stage), useless when they're random or the same six items sitewide.
Breadcrumb links
Home › Category › Subcategory › Page. They reinforce hierarchy, give every page a link up to its parent, and show in search results with the right schema.
Footer and sidebar links
Technically a sixth type, and the one that most often does damage. Sitewide boilerplate links are heavily discounted and, in bulk, drain equity from the pages that matter.
Rules for anchor text
I bake a short set of rules into the editorial guidelines so they survive without me.
Describe the destination, not the action
"See my guide to site architecture" beats "click here" or "read more". The anchor should make sense if it were the only text on the page.
Vary naturally, but stay on topic
Ten links to the same page can use "site architecture", "how to structure a website", "architecture guide". They should not all be the identical exact-match phrase, and they should not wander into unrelated wording either.
One destination per concept per page
If a page mentions canonical tags four times, link the first meaningful mention. Google generally counts the first anchor to a given URL; the rest add noise and look like stuffing.
Link from the sentence where the reader has the question
The best contextual link appears at the moment the reader would want to know more. That's where the click rate is highest and the context is clearest.
No generic anchors to important pages
If your best-earning category page only ever gets linked as "here" and "products", you've thrown away the strongest signal you had.
Depth and click distance
Click depth is the minimum number of clicks from the homepage to a page. Google's crawl and the flow of authority both weaken with each hop. Pages at depth two or three are fine. Pages at depth six or beyond tend to be crawled infrequently and rank weakly no matter how good they are.
On ecommerce sites, the usual villain is pagination: product 400 sits at Home › Category › Page 17 › Product. On publishers, it's the archive: a three-year-old article is reachable only by paging back through the blog. On SaaS sites, it's documentation and integration pages nobody linked from the marketing site.
The fix is structural rather than manual: subcategories so each list is shorter, hub pages linking directly to every item in a cluster, related modules that cross-link within the cluster, and links from strong evergreen pages into deep ones. I cover the architecture choices in the site architecture guide.
Where link equity gets wasted
Every page has a finite amount of signal to pass on, split across all its links. The more links on a page, the less each one carries. This is where most sites bleed.
Footer mega-lists
A footer with eighty links to every city page, category and policy document appears on every page of the site. Google discounts boilerplate heavily, but the dilution on the page itself is real: a product page with 30 meaningful links and 80 footer links spends most of its equity on the footer. Keep footers short. If city or category links need to exist sitewide, put them on a hub page and link to the hub.
Boilerplate "related" modules
If every page shows the same six "popular posts", those six collect thousands of links and nothing else does. Related modules should be computed per page (same cluster, same category) or they are just a second footer.
Links to pages that can't rank
Login, cart, compare, sort and filter URLs, tag pages with one item. Each is a link that could have gone somewhere useful. On large ecommerce sites, filter and sort links alone can outnumber real links many times over. Links to redirecting or dead URLs belong in the same bucket, and are the easiest fix: point them at the final URL.
Scaling internal links with templates and rules
Hand-adding links works for a fifty-page site, not five thousand. At scale, internal linking has to be a system: rules that templates execute automatically, plus a small editorial layer on top.
The templated layer is straightforward once the taxonomy is clean. Every page links up to its hub; every hub links down to every member; breadcrumbs reflect the hierarchy; related modules pull from the same category. On programmatic sites, the same logic applies to entity relationships: a "SEO services in Delhi" page links to neighbouring locations, the parent services page and relevant guides, all generated from data rather than typed.
The editorial layer is for contextual links in body copy. The scalable version: maintain a map of target pages and preferred anchors, have writers include two or three per article via the content brief, and periodically crawl for pages that mention a target topic without linking to it.
I tested whether rule-based linking at this scale actually moves indexing and rankings in the internal linking at scale experiment. The full method, including how to pick targets and set rules, is in my internal linking framework.
Auditing internal links with a crawler
A site crawler (Screaming Frog, Sitebulb or similar) shows the shape of your linking. The reports I always pull:
| Report | What it reveals | What I do with it |
|---|---|---|
| Inlinks per URL, sorted ascending | Pages with one or two internal links, and orphans (zero) when combined with sitemap data | Add hub and contextual links to anything valuable in the bottom tail |
| Crawl depth distribution | How many pages sit at depth 4, 5, 6+ | Restructure categories or add hubs to flatten the deep sections |
| Anchor text per target URL | Whether important pages are linked with descriptive anchors or "click here" | Fix templates and editorial guidelines, not individual links |
| Outlinks per page | Pages with hundreds of links (usually footer and filter bloat) | Cut boilerplate, nofollow or remove links to unrankable URLs |
| Internal links to 3xx / 4xx | Equity leaking through redirects and dead links | Update links to final URLs; it's a batch find-and-replace in most CMSs |
| Link score / internal PageRank | Which pages your structure is actually promoting | Compare against the pages you want promoted; the gap is your to-do list |
The last row changes conversations with stakeholders. When the crawler's link score shows the privacy policy outranking the flagship category page, nobody argues about trimming the footer.
Frequently asked questions
How many internal links should a page have?
There is no fixed number. Every link should be one a user might plausibly click. On a long guide, five to fifteen contextual links plus navigation and breadcrumbs is typical. If a page has hundreds of links, most are boilerplate and are diluting the rest.
Do internal links with the same anchor text hurt?
Not in the way exact-match external anchors can. Internal links are under your control, and Google expects consistency. I avoid the identical phrase on every link because it reads as mechanical and loses the chance to describe the page in different useful ways.
Should I use nofollow on internal links?
Rarely. Nofollow on internal links doesn't redirect equity elsewhere; it drops it. It can stop Googlebot following filter or sort links into a crawl trap, but the cleaner solution is not to render those as crawlable links at all.
Is it worth linking from old blog posts to new pages?
Yes, and it's one of the fastest wins available. Old posts that have earned external links and still get traffic are ideal places to link to newer pages in the same cluster. A crawler's "mentions X but doesn't link to X" check finds the candidates in minutes.
How soon do internal linking changes show results?
Crawling and indexing changes can show within days to a few weeks once Google recrawls the linking pages. Ranking movement typically follows over several weeks, and depends on everything else about the target page being sound.
Where this connects
This guide sits under the technical SEO hub. The repeatable method is my internal linking framework; the controlled test of rule-based linking on a large site is the internal linking at scale experiment; the technical SEO and content architecture case study shows linking changes as part of a wider rebuild; and the implementation itself is part of my technical SEO services.