Internal Linking: Architecture That Ranks
Internal linking is the one ranking lever you control completely. You need nobody's approval, no editor and no budget – and yet it is the most neglected part of technical optimisation. This guide shows how to set up architecture, click depth and link placement so that the right pages get strength and crawl priority.
1. Why internal linking gets underrated
External links are hard to get, so they get attention. Internal links are free, so they get none. That is the wrong priority, because internal links determine three things:
- What gets crawled. Crawl priority follows linking. A page with no internal link sits at the back of the queue – see crawl budget.
- What gets indexed. "Crawled – currently not indexed" is markedly rarer for well-linked pages. Details under indexing.
- What a page stands for. The anchor text of internal links is your explicit statement about each target page's topic.
On top of that: external links almost always land on the homepage and a handful of top pages. Whether that strength reaches the pages that actually generate revenue is decided solely by internal structure.
2. The four jobs of an internal link
| Job | What gets transferred | What it depends on |
|---|---|---|
| Discovery | Knowledge that the URL exists | An <a href> present in the rendered DOM |
| Crawl priority | Fetch frequency | Number and strength of inbound internal links |
| Link equity | Share of authority | Source page strength divided by its link count |
| Relevance | Topical attribution | Anchor text and surrounding context |
The four are independent of one another. A footer link does jobs 1 and 2 well, job 3 weakly and job 4 barely. A contextual link in the main body does all four.
3. Architecture models compared
| Model | Principle | Suited to | Weakness |
|---|---|---|---|
| Flat structure | Everything within two clicks of the homepage | Small sites up to ~200 URLs | Does not scale; navigation becomes unmanageable |
| Silo / topic tree | Strict separation of subject areas | Shops with clear category logic | Cross-connections missing; feels artificial |
| Hub and spoke | A central overview page links all details | Knowledge bases, guides | The hub itself must be well linked |
| Topic cluster | Hub plus reciprocal detail-to-detail linking | Building topical authority | Upkeep needed with every new piece |
In practice the topic cluster is the strongest variant, because it combines two things: a flat hierarchy via the hub and a dense graph via the cross-references. This wiki is built exactly that way – the index page is the hub, and each guide points to its topical neighbours.
4. Measuring and reducing click depth
Click depth is the number of clicks from the homepage to a target page along the shortest path. It is the most informative single metric in internal linking, because it is measurable and directly actionable.
| Click depth | Typical crawl behaviour | Assessment |
|---|---|---|
| 0–1 | Very frequent fetches | Homepage and main categories |
| 2–3 | Regular fetches | The target zone for all important pages |
| 4–5 | Less frequent fetches | Acceptable for long-tail content |
| 6 and beyond | Markedly rarer, sometimes never | Needs attention |
Four ways to reduce depth:
- Introduce hub pages. One overview page per topic shortens many paths to two clicks at once.
- Limit pagination. A category with 40 pages creates depth 41 for the last one. More items per page, or finer subcategories, solve it.
- Cross-references from the content. A contextual link can attach a deep page directly to a shallow one.
- A "related content" module. Effective when the selection is computed topically – not at random.
An important clarification: click depth is not the same as
URL depth. A URL like /a/b/c/d/product can be
reachable in one click from the homepage. Google evaluates the click path, not the number of
slashes.
5. Link placement in the layout
| Placement | Weight | Appears on | Use for |
|---|---|---|---|
| Main navigation | High | every page | Only your most important areas – each entry raises the divisor everywhere |
| Contextual link in the body | Highest | one page | Targeted strengthening with a precise anchor |
| Breadcrumb | Medium | deeper pages | Explaining hierarchy, strengthening parent pages |
| Related content | Medium | detail pages | Densifying clusters |
| Sidebar | Low to medium | page types | Secondary navigation |
| Footer | Low | every page | Required pages, a few core areas |
The decisive difference is repetition: a navigation or footer link exists on every page. It helps discovery, but it raises the outbound link count on every page and thereby shrinks the share of every other link. A contextual link exists once – and works hardest exactly where it sits.
6. Building topic clusters in practice
A cluster consists of a hub page and several detail pages. Five steps to build one:
- Define the scope. A cluster covers one topic for which a category-level query exists.
- Build the hub page. It explains the topic at overview level and links every detail page with descriptive anchor text.
- Link the details back. Every detail page links to the hub – ideally from the body text, not only via the breadcrumb.
- Add cross-references between details. Only where a genuine relationship exists. Indiscriminate linking creates noise.
- Maintain it as you expand. Every new piece needs a link from the hub and two or three cross-references from existing pieces. This step is the one most often forgotten.
The connection to entities is direct: if each detail page covers exactly one term and every mention of that term links there, you are building a machine-readable topic graph – see entity SEO.
7. Finding orphan pages
An orphan page is reachable but linked from no other page. It receives no link equity and is rarely crawled. How to detect them:
- Crawl the site through internal links only – without using a sitemap as the starting point.
- Compile every known URL: XML sitemaps, the Search Console performance report, server logs, a CMS export.
- Take the difference. Anything in list 2 but not in list 1 is orphaned.
- Decide: link it, merge it, redirect it or delete it.
Common causes: deleted categories whose posts remain; campaign landing pages; pages reachable only through a search function; content behind JavaScript interactions that produce no real link in the rendered DOM.
8. How many internal links per page?
There is no limit – Google removed the old recommendation of at most 100 links from its guidelines. The mathematical effect remains, though: more links mean a smaller share per link.
Workable orientation from practice:
- Homepage: 40 to 80 links. Beyond that, prioritisation dissolves.
- Category page: navigation plus the items on the page plus pagination.
- Guide article: 3 to 10 contextual links in the body, depending on length.
- Footer: under 25 links.
A better yardstick than any number: does the content justify the link? A link a reader finds useful at that point is right. A link that is there only for search engines is noise.
9. Internal links as crawl control
Internal linking is also a steering instrument. Four rules:
- Do not link to blocked URLs. Links to paths blocked by robots.txt waste their share.
- Do not link to redirects. Internal links should point straight at the final target. After any relaunch this is the most important clean-up step.
- Link the canonical variant. If internal links point at
?sort=pricevariants while the canonical points at the clean URL, your signals contradict each other. - No internal
nofollow. The share is not redistributed, it evaporates – see link attributes.
10. Common mistakes
- A mega menu containing everything. 250 links on every page means no page is prioritised.
- Using a sitemap instead of linking. A sitemap makes URLs known but transfers no strength.
- Only linking top-down. Category to product, but never product to product or product back to the matching category.
- Links that are not links.
<div onclick>and<button>instead of<a href>– see rendering. - Internal links to
404. Pure waste and a user experience problem. - Generic anchor text. "Learn more" throws away the relevance signal.
- Automated keyword linking. Plugins that link every occurrence produce unreadable copy and hundreds of identical anchors.
- Not maintaining clusters. Last week's new article still has no inbound link.
11. Measuring internal linking
- Search Console, Links report → internal links: shows which URLs are most linked internally – Google's view of your prioritisation. Compare that list against your actual revenue and traffic goals. The gaps are your concrete to-do list.
- Your own crawl: provides click depth, inbound internal links per URL, anchor texts and internal links pointing at errors or redirects.
- Log files: a URL's crawl frequency is a good proxy for how important Google considers it.
- Sitemap versus crawl: the standard method for detecting orphan pages.
12. Checklist
- Measure click depth for every revenue and traffic target; aim for three clicks maximum.
- Compare the internal links report against your own priority list.
- Identify orphan pages via sitemap-versus-crawl.
- Fix internal links pointing at
404targets. - Repoint internal links that hit redirects to the final target.
- Correct internal links pointing at non-canonical variants.
- Remove internal
nofollow. - Replace generic anchor text on important pages.
- Reduce the number of navigation and footer links.
- Create a hub page per core topic and link it from the navigation.
- Give every new article a hub link and two cross-references at publication.
- Check that all links in the rendered DOM are real
<a href>elements. - Re-measure the crawl frequency of the strengthened pages after four weeks.
Pro tip: Rank-O-Saur shows every link on a page with its anchor text, rel attribute and status code – separated into internal and external. That lets you check on the spot whether a page carries too many outbound links and whether internal links are hitting redirects or error pages.
13. Frequently asked questions
How many internal links should a page have?
There is no fixed limit; Google removed the old 100-link recommendation from its guidelines. As orientation: 40 to 80 links on the homepage, 3 to 10 contextual links in a guide article, under 25 in the footer. What matters is whether the content justifies the link.
What is the difference between click depth and URL depth?
Click depth counts clicks from the homepage along the shortest path; URL depth counts directory levels in the address. Google evaluates the click path. A URL with five levels can be one click away and is then unproblematic.
Are internal links as important as external ones?
They work differently. External links bring in new authority from outside; internal links distribute what you have and define what each page is responsible for. Since internal linking is entirely in your hands, it is usually the faster lever to act on.
Where should internal links sit?
Contextual links in the main body work hardest, because they are singular and carry topical context. Navigation and footer links help discovery but raise the outbound link count on every page, diluting all the others.
How do I find orphan pages?
Crawl the site exclusively through internal links, without a sitemap as the starting point. Then compare every known URL from sitemaps, Search Console and logs. Anything known but absent from the crawl is orphaned.
Do navigation and footer links count at all?
Yes, but more weakly than links in the main content. Google recognises recurring boilerplate areas and weights them lower. For discovery and crawl priority they remain effective.
Should I use a related posts module?
Yes, if the selection is computed topically. Random or purely chronological modules create noise rather than structure. The most effective option is manually curated cross-references within a topic cluster.
Does internal linking help with indexing problems?
Yes, for "Crawled – currently not indexed" it is often the most effective step. In practice a link from a frequently crawled, topically related page works faster than an indexing request in Search Console.
Read next
