An internal linking strategy is your plan for which pages link to which, and with what words, so people and Google reach your important pages. On a site with thousands of pages, hand-linking fails. You set it in templates. Hubs link down to every child page, breadcrumbs link up, and related-link blocks link sideways by rule. Anchors come from page data. Then you measure orphan pages, click depth and links per page type every month, and fix the template that's failing.
I'm Brenden House. I run Searchmaxxed, an SEO and AI SEO agency for enterprise, B2B SaaS, venture-backed, manufacturing and public-sector companies.
In summary
- Write rules, not links. A large site needs one link rule per page template, rather than ten thousand editorial decisions.
- Every page needs a hub. Each page you care about should sit on a crawlable hub page and in a breadcrumb trail. Google says "every page you care about should have a link from at least one other page on your site."
- Shallow isn't enough. A page two clicks from the homepage with one link in still looks unimportant to a crawler.
- Pick related links by rule. "Popular posts" blocks pile links onto the same few pages. A rule based on the same parent, shared attributes, or a ring gives every page an equal share.
- Measure with fresh data. Track orphans, depth and inlinks per template, and check that your crawler isn't reading cached pages.
What is an internal linking strategy?
An internal linking strategy decides how pages on one website link to each other, so Google can find, understand and rank the important ones. On a small site, that's an editor adding links in copy, while on a large site it's a set of rules inside page templates.
A few terms come up in every conversation about it:
- Internal link: a link from one page to another page on the same domain.
- Anchor text: the visible, clickable words of a link. Google calls it link text.
- Hub page: a page whose job is to list and link to every page in one group, such as a category, topic or directory.
- Click depth (crawl depth): the fewest clicks it takes to reach a page from the homepage.
- Orphan page: a page with no internal links pointing to it, so crawlers find it through a sitemap, if at all.
- Link equity: the shorthand SEOs use for the value one page passes to another through links. Google's name for its system is PageRank.
- Template-level linking: links your page templates generate from data, such as breadcrumbs, hub lists and related-link blocks.
The top guides on this topic explain the link types and then tell you to add links by hand. For a few hundred pages, fine. Past that, the template decides where links go, so that's where this guide spends its time.
Why do internal links matter for SEO and AI search?
Internal links are how Google finds pages and judges which ones matter. Google's link best practices say it "uses links as a signal when determining the relevancy of pages and to find new pages to crawl". That's the whole game. Its ranking systems guide adds that PageRank "continues to be part of our core ranking systems".
Three jobs follow from that:
- Discovery. Googlebot follows
<a href>links. Google's own SEO Starter Guide says it "primarily finds pages through links from other pages it already crawled." - Importance. A page that 300 pages link to looks more important than one with a single link from page 14 of a paginated archive.
- Meaning. The anchor text and the words around it tell Google what the target page is about.
AI search runs on the same plumbing. Google's guide to AI features says a page "must be indexed" to show as a supporting link in AI Overviews or AI Mode. Its list of SEO basics for those features includes "making your content easily findable through internal links". Internal links alone won't make an AI answer name you, but they do decide whether your best page is in the running at all.
At Searchmaxxed, internal linking sits inside our technical SEO service, next to crawl, indexing and schema work. Same root cause: a page Google can't reach or can't place.
What types of internal links does a large site have?
A large site has two kinds of internal links: links templates generate from data, and links editors write into copy. On a site with 5,000 pages, the templates produce most of them, so they set where your link value goes.
| Link type | Who controls it | What it's good for | How it fails at scale |
|---|---|---|---|
| Main navigation | Template | Top hubs and money pages | Mega menus with 150 links flatten every signal |
| Breadcrumbs | Template | Linking each page up to its parents | Missing on some templates; no visible <a> links |
| Hub and category lists | Template | Linking down to every child page | Paginated with "next" links only, so deep pages sink |
| Related-link blocks | Template | Sideways links between siblings | "Popular" logic sends every link to the same few pages |
| Contextual body links | Editor | Specific, descriptive anchors | Don't scale; old articles never get updated |
| Footer | Template | Legal, company and key hub pages | Sitewide keyword anchors to every service |
| Pagination | Template | Reaching later pages of a long list | Page 30 sits 30 clicks deep |
| Faceted navigation | Template | Filtering products, docs or listings | Endless parameter URLs that eat crawl time |
| HTML sitemap | Template | A safety net for discovery | Thousands of links on one page, no context |
Google's faceted navigation guidance warns that parameter-based filters "can generate infinite URL spaces". The fix it gives is to block crawling of filter URLs you don't need indexed, using robots.txt.
How do you build internal linking at template level?
You build it by writing one link rule per template, shipping those rules in code, and checking the rendered HTML. Here's the order we work in.
- List every template and count its pages. Pull URLs from your sitemaps, CMS export and a crawl, then group them by template.
- Give each template a job. Hubs rank for broad searches and pass links down, while detail pages rank for the specific searches underneath them. Articles answer questions and pass links to commercial pages.
- Choose a parent for every page. Every page belongs to exactly one hub, and that hub sits within two clicks of the homepage.
- Write the rule for each template. Specify links up, down and sideways, how many, how to pick them, and which data field supplies the anchor.
- Cap the template-generated links. Leave room for editors' contextual links, and remove duplicate links to the same URL on one page.
- Check the rendered page. Make sure every link is a real
<a href>with a full URL, in the HTML Googlebot gets.
A worked example: a B2B SaaS site with about 6,000 pages
Take a B2B SaaS company with eight solution areas, 40 use-case pages, 450 integration pages, 300 comparison pages, 1,200 articles and about 4,000 docs pages. These are example figures for illustration, so swap in your own counts.
| Template | Pages | Links up | Links down or sideways | Anchor comes from |
|---|---|---|---|---|
| Homepage | 1 | None | 8 solution hubs, integrations hub, comparison hub, resources hub | Hub short names |
| Solution hub | 8 | Breadcrumb to home | Every use case in its group, its top 10 integrations, 3 case studies | Use-case name field |
| Use-case page | 40 | Breadcrumb to its solution hub | 4 sibling use cases, 6 tagged integrations, 2 articles | Page H1 or short name |
| Integration category | 12 | Breadcrumb to integrations hub | Every integration in the category, about 38 each | Integration name |
| Integration page | 450 | Breadcrumb to its category | 5 same-category siblings by ring rule, use cases it supports, its setup doc | Integration name |
| Comparison page | 300 | Breadcrumb to comparison hub | Product page, 3 related comparisons with a shared competitor | "X vs Y" title field |
| Article | 1,200 | Breadcrumb to its topic hub | 3 to 5 editor links, 4 same-topic articles, one commercial page | Article title |
| Docs page | 4,000 | Breadcrumb and sidebar tree | Previous and next page, parent section | Docs nav label |
Look at the integrations. With 12 category pages of about 38 integrations each, every integration sits three clicks from the homepage: home, integrations hub, category, integration. List all 450 on one hub instead, 20 per page with only a "next" link, and the last ones sit 23 pages deep. Same pages and the same content, yet one rule decides whether they're reachable.
How should hubs and topic clusters work on a large site?
A hub page should link to every page in its group in plain HTML, and it should split into sub-hubs before the list gets unmanageable. Topic clusters are the same idea for content: a pillar page links down to every article on the topic, and each article links back up.
Depth comes down to arithmetic. If your homepage links to 10 hubs and each hub links to 100 pages, 1,000 pages sit two clicks from the homepage. Add one more level of 100-link hubs and you reach 100,000 pages in three clicks. Shallow is cheap.
| Links per hub level | Pages 2 clicks deep | Pages 3 clicks deep |
|---|---|---|
| 10 hubs, then 50 links each | 500 | 25,000 |
| 10 hubs, then 100 links each | 1,000 | 100,000 |
| 20 hubs, then 100 links each | 2,000 | 200,000 |
My arithmetic, assuming each level links to new pages only.
What makes a good hub:
- It lists every child page, with a sentence or two on each, so the hub can rank on its own.
- It splits by a real attribute.
- Its pagination links to page numbers as well as "next", so page 9 is one click away.
- It sits in the main navigation or one click below it.
On sites we build, each page has a distinct job, and we connect related pages through internal links as they ship. Our programmatic SEO service runs on the same rules, because a page family of thousands breaks first at the hub.
Do breadcrumbs still matter for SEO?
Yes. Breadcrumbs give every page a crawlable link up to each of its parents, and a large site needs that whether or not Google shows them. Since January 2025, Google no longer shows breadcrumbs on mobile results; it still shows them on desktop and still supports the markup.
Two parts do two jobs:
- The visible trail is a set of normal
<a href>links, and that's the part that passes link value up through the site. - BreadcrumbList structured data tells Google the page's place in the hierarchy. Google's breadcrumb documentation says a trail "indicates the page's position in the site hierarchy".
Generate both from one parent field in your CMS. Then the trail and the markup can't disagree, however many pages you add. If a template shows breadcrumbs only to logged-in users, or renders them as buttons, crawlers get nothing.
How should related-link blocks choose their links?
Related links should be chosen by a rule that spreads links evenly across siblings, not by popularity. A "most popular" block sends every link to the pages that already have the most, and the rest of the group gets none.
Take the 38 integrations in one category. If each page shows the same five "popular" integrations, those five get 37 links each from their siblings, and the other 33 get none. With a ring rule, each integration links to the next five in an alphabetical list, wrapping at the end. That's five links in, for every page in the category.
| How the block picks links | What each page gets | Use it for |
|---|---|---|
| Most popular or newest | A few pages get dozens; most get none | Nothing you want indexed |
| Same parent, shared tags | Even, if tags are complete | Use cases, articles, comparisons |
| Ring (next N in a sorted list) | Exactly N links in, every page | Large flat families: integrations, locations, templates |
| Editor-picked | Strongest anchors, uneven coverage | Money pages and pillar articles |
Mix the methods. A block of four links can take two by shared tag and two by ring, so relevance and coverage both hold. Fix the order in code, too. A block that reshuffles on every load makes Googlebot see different links on every visit.
What anchor text should internal links use?
Internal anchors should describe the target page in the words people search, kept short. Google's guidance asks for anchor text that's "descriptive, reasonably concise, and relevant", and warns against cramming keywords into it.
At template level, anchors come from data fields, so set them once:
- Use a short-name field for each page ("Salesforce integration", "SOC 2 compliance software") and pull anchors from it.
- Let the target page's main search phrase appear in some anchors, and use natural variations elsewhere.
- Ban "learn more", "click here" and "read more" in templates; card links can wrap the card title instead, a real anchor for free.
- Give each search phrase one target page. Two templates linking with the same anchor to two different URLs tell Google you have two answers.
The same failure turns up on any templated page family: country pages, industry pages, "X for Y" pages. When three templates target one search, internal links can't save them. Merge the families or give each a distinct search, then link to the winner.
How do you find and fix orphan pages?
Compare three URL lists. One is what your crawler reaches from the homepage by following links. The others are your XML sitemap and the URLs in Search Console or your server logs, which show what Google already knows about. A URL that's in the sitemap or logs but missing from the crawl is an orphan.
| In crawl? | In sitemap? | In logs or Search Console? | What it is | Fix |
|---|---|---|---|---|
| No | Yes | Yes | A live page no template links to | Add it to its hub and to sibling related blocks |
| No | No | Yes | An old URL Google still requests | 301 to the closest live page, or return 410 |
| Yes | No | Yes or no | A linked page missing from the sitemap | Add it, or noindex it if it shouldn't rank |
| Yes | Yes | No | A page Googlebot ignores | Check depth, inlinks and duplicate content |
Server logs show what Googlebot requested, which a crawler can't. Our guide to log file analysis for SEO walks through getting the logs and verifying the bots.
Orphans on large sites nearly always come from templates. A new page type launches without its hub, or a CMS filter hides expired items from the list while their pages stay live. Or a migration drops a category. Fix the template, and the whole family's orphans go with it.
What is crawl depth, and how deep is too deep?
Crawl depth is the fewest clicks from the homepage to a page, and I aim to keep every revenue page within three of them. That's my working rule, not a Google number, and it holds up on large sites because hubs make it cheap to meet.
Depth matters most when Google has more URLs than it wants to crawl. Google's crawl budget guide says it's written for sites with "1 million+ unique pages" changing weekly, or "10,000+ unique pages" changing daily. It also covers sites with many URLs in Search Console's "Discovered - currently not indexed" state. In those groups, depth costs you directly.
Below that size, depth still tells Google how much you value a page, and deep pages tend to come from the same few template habits:
- pagination with only "next" links
- hubs that show only the 20 newest items
- pages you can reach only through on-site search or filters, which crawlers don't use
- links rendered by JavaScript without an
<a href>
That last one catches design systems. Google's link guide says that, generally, it "can only crawl your link if it's an <a> HTML element" with an href attribute. A card that navigates with a click handler works for a person and dead-ends a crawler, so check the rendered HTML, not the design file. The rest of these checks sit in our technical SEO checklist.
How many internal links should a page have?
Google publishes no limit on internal links per page. Count unique links to useful pages and leave room for the ones you care about, because navigation and footer links add up fast.
My rule of thumb on a large site:
- Navigation and footer: 30 to 60 links, mostly hubs.
- Breadcrumb: 2 to 4 links.
- Related block: 4 to 8 links.
- Body copy: 3 to 5 contextual links in a normal article, more on a pillar page.
Our own pages carry a median of 46 internal links, from our navigation, footer and page body, in our 27 October crawl. More links per page isn't the goal. A link should send the reader somewhere they'd plausibly want to go next.
How do you measure whether internal linking is working?
Measure it by template, before and after each change: orphan count, click depth, inlinks per page and the share of pages Google indexes. Then compare matched 28-day periods of impressions and clicks for the templates you changed.
| Metric | Where to get it | What good looks like |
|---|---|---|
| Orphan pages | Crawl versus sitemap versus logs | Zero live indexable pages |
| Pages deeper than 3 clicks | Crawler's depth report | No revenue pages |
| Inlinks per page, by template | Crawler's inlinks export | No page in a family below about 3 |
| Pages with one inlink | Same export | Few, and none that should rank |
| Indexed share, by template | Search Console Page indexing, filtered by sitemap | Rising after the change |
| Googlebot requests, by template | Server logs | Requests reach the deep pages |
| Impressions and clicks, by template | Search Console, matched 28-day windows | Up for the pages you linked |
One tip that saves hours: give each template its own XML sitemap. Search Console's Page indexing report filters by sitemap, so you see indexation per template instead of one site-wide number.
Where does Searchmaxxed fit?
Searchmaxxed is an SEO and AI SEO agency for enterprise, B2B SaaS, venture-backed, manufacturing and public-sector companies. Internal linking is part of our technical SEO work, and we treat it as engineering. We map your templates, write the link rules, build them into your site, and measure the result by template.
Here's what that looks like in practice:
- We crawl the site the way engines do and compare it with your sitemaps, Search Console and server logs. Problems show up by template.
- We write a link rule for every template, covering hubs, breadcrumbs, related blocks and anchor fields. Then we rank the fixes.
- We build the rules into the site and check the rendered HTML and live pages after release.
- We measure inlinks, depth, indexation and clicks per template in matched periods.
Here's how we do it. We build on a website you own, our Agentic Website. AI agents pull your Search Console and GA4 data, crawl the site and read the Google results for your searches. That feeds the Managed Search Loop, which shows which page, fix or link should bring in the most enquiries next. Off your site, the Off-Page Source Layer builds the links, digital PR and mentions that internal links then pass through to the right pages.
| Package | Who it suits | What it includes |
|---|---|---|
| Growth System | A company growing in one or two markets | 30 page releases a month, landing-page work on up to 4 existing pages, AI visibility tracking |
| Agentic Search Engineering | A competitive or national market | 50 page releases a month, comparison, industry and integration page families, digital PR and link campaigns |
| Custom & Enterprise | Several brands or markets, or a huge site | Multiple markets, dedicated engineering, custom reporting and original first-party research |
In the Growth System, technical, schema and internal-link work doesn't reduce the page-release count. Every managed package includes ongoing audits, a 90-day plan in weekly releases and the build work itself. You own the site, repository, content and data.
Searchmaxxed's work has generated more than A$10 million in revenue through search for its clients.
Is Searchmaxxed right for you?
We're a good fit if you run a large or fast-growing site and want the linking fixed in the templates. Built and measured. We're the wrong call if you need one page relinked or want an audit to hand to your own developers and nothing more.
| Searchmaxxed is a good fit if | Look elsewhere if |
|---|---|
| Your site has thousands of pages across several templates | Your site has 50 pages and an editor can link them by hand |
| You want link rules built into the site, not a spreadsheet of suggestions | You want a one-off crawl report |
| You want technical SEO, content and AI search from one team | You only need a plugin to suggest links in WordPress |
| You want results measured per template, against enquiries | Your developers can't ship template changes for months |
| You want to own your site, repository, content and data | You want the cheapest retainer you can find |
Right-hand column? A crawler licence and a good in-house editor will serve you better.
Frequently asked questions
What is internal linking in SEO?
Internal linking in SEO means linking from one page of your website to another page on the same site. It's free. Google uses those links to find pages, judge which ones matter and work out what each covers. Cheap lever.
What is the difference between internal and external links?
An internal link points to a page on your own domain, while an external link points to another website. You control every internal link, so they're the fastest way to direct crawlers and link value to your most important pages.
What is the three-click rule?
The three-click rule says important pages should be reachable within three clicks of the homepage. It's a working rule, not a Google limit. Hub pages that list every child page make it easy to meet on a large site.
How do I find orphan pages?
Compare three URL lists: pages your crawler reaches from the homepage, your XML sitemap URLs, and URLs from Search Console or server logs. Anything in the sitemap or logs but missing from the crawl is an orphan.
Should internal links use exact-match anchor text?
Some can. Use the target page's main search phrase in some anchors and natural variations in others. Keep anchors short and descriptive, and never point the same anchor at two different pages.
Can internal linking be automated?
Yes, and on a large site it should be. Breadcrumbs, hub lists and related-link blocks are automated by template rules. Keep editors on contextual links to money pages, and check automated blocks spread links evenly.
Do internal links help with AI search?
Indirectly. AI answers draw on pages that get crawled and indexed, and internal links decide how easily crawlers reach yours. They won't make ChatGPT name you on their own. Still, an unlinked page might never be in the running, because nothing points crawlers to it.
What should you do next?
Crawl your site from the homepage today, export inlinks per page, and group them by template. Find the family with the most pages on one link or none, and fix that template's rule before you touch a single article by hand. If you'd like us to scope it for your site, let's talk pricing.