The short answer
There is no correct number of internal links per page, and anyone selling you one — five, ten, a hundred — is selling a rule that was never measured. Link where a reader would genuinely want to go next, make sure every important page is reachable from the homepage within three clicks, and make sure no page is reachable from nowhere at all. Those three tests cover everything a number would pretend to cover.
Internal links do three jobs at once, and the confusion starts because people count links as if only one job mattered. They help readers navigate, they tell crawlers which pages exist and how they relate, and they distribute the authority a site has earned across its pages. A count optimises none of the three. A reader who finds the next step, a crawler that can reach every page, and a new page that inherits links from established ones — that is what good internal linking looks like, and it has no fixed total.
Why no number exists
Google has never published a limit, and the old guideline of "no more than 100 links per page" was withdrawn years ago. It was a crawling concern from an era when pages were simpler and crawlers weaker, not a quality judgement, and it has not been relevant for a long time.
The deeper reason is that pages differ. A short answer page may need two links: one to the detailed guide, one back to the section it belongs to. A long reference page may need thirty, because it touches thirty topics a reader might want to explore. A navigation page may need hundreds, because its entire job is to list destinations. Judging all three by one number would mean the short page is over-linked, the reference page is under-linked, and the navigation page is broken — and all three judgements would be wrong.
What matters instead is whether each link earns its place. A link earns its place when a reader following it learns something the current page does not say, or reaches a tool that does something the current page only describes. A link that points at a page the reader has already read, or at a page with nothing new to add, is clutter — and clutter is the only real excess in internal linking.
The three-click rule
If one number is worth keeping, it is not a count of links but a count of clicks: any important page should be reachable from the homepage within three clicks. Homepage to section page is one, section page to article is two, article to related detail is three. Beyond three, a page is deep enough that both readers and crawlers reach it rarely, and rarely-visited pages are rarely indexed, rarely linked, and rarely read.
Three is a working rule, not a law. A large site with thousands of pages will have some pages four clicks deep, and that is fine if those pages are genuinely minor. The rule is a diagnostic: if a page you care about is five clicks from the homepage, that is a structural problem, and the fix is not to add links to that page alone but to reconsider the section it lives in. Deep pages are usually a symptom of a section that grew without a plan.
Note the direction of the fix. When a page is too deep, people add a link from the homepage to it, which technically satisfies the rule and helps nobody: a homepage with a hundred direct links is not navigation, it is a sitemap pasted into the header. The right fix is to give the page a proper parent — a section page that lists it alongside its siblings — so the path is short and the journey makes sense.
Orphan pages are the real problem
An orphan page is one that nothing links to: no navigation entry, no link from another article, no mention anywhere on the site. The only way a crawler can find it is through the sitemap, and the only way a reader can find it is by knowing the URL. Our guide to why pages are not indexed lists this as one of the most common causes of a technically perfect page never being indexed, and it is the internal-linking failure that actually costs traffic.
Orphans happen quietly. A page is published, the author means to link to it from the overview article, the overview is never updated, and the page sits alone. A redesign drops a section page, and the ten articles under it lose their only parent. A draft is published from a preview link and nobody adds it to the index. None of these are dramatic, which is why they accumulate.
Finding them is mechanical. List every URL in your sitemap, list every URL that receives at least one internal link, and compare the two sets. Any URL in the first list but not the second is an orphan. Fix it by linking to it from the page a reader would most naturally arrive from — usually the overview or the article that mentions its topic — and add that link to your publishing checklist so the next page does not repeat it.
Anchor text: say where the link goes
The clickable words of a link should describe the destination. "Read our guide to meta description length" tells a reader what they will get; "click here" tells them nothing, and "read more" tells them slightly less than nothing. This matters for readers first — people scan links before clicking them — and for crawlers second, because anchor text is one of the signals used to understand what the destination page is about.
Two habits to avoid. The first is using the same anchor text for different destinations, which tells both readers and crawlers that two different pages are the same thing. The second is forcing the exact same keyword phrase into every link to a page, which reads as manipulation and makes the text worse. Describe the destination in the words that fit the sentence you are writing, and vary them naturally.
One more point worth stating: link from the body of the page, not only from navigation and footers. A link inside a paragraph, at the point where the reader is thinking about the topic, is worth more than the same link in a sidebar — to the reader, because it arrives when it is wanted, and to the crawler, because body links carry more weight than boilerplate.
Can a page have too many links?
Yes, but the threshold is far higher than most people assume, and it is about usefulness rather than count. A page with two hundred links that a reader can scan — a directory, an index, a sitemap page — is fine. A page with twenty links where every second word is blue is not, because the text becomes unreadable and no single link stands out.
The genuine excess case is sitewide boilerplate: the same fifty links in the footer of every page. Those links count as internal links, technically, but they teach a crawler nothing about the relationship between two specific pages, because they appear on all of them. If your link audit shows a page is "well linked" only because of the footer, it is not well linked at all. Count body links separately from template links, and judge the page by the body links.
There is also a crawl-budget consideration on very large sites: every link is a URL a crawler may follow, and a site with millions of internal links can spend its crawl allowance re-walking the same navigation instead of reaching new content. For a site under a few thousand pages this is not a practical concern. For a large one, the fix is to stop linking to filtered, sorted, and paginated variants from every page, not to ration links on articles.
What to do
For a new page, link to at least two existing pages from the body, and link to the new page from at least one existing page — ideally the overview or article that a reader would arrive from. That is the whole rule for publishing, and it is also step seven of our pre-publish checklist. Two links out, one link in, every time, and orphans stop happening.
For an existing site, run the orphan check described above, then check the click-depth of your ten most important pages. If any of them is more than three clicks from the homepage, restructure the section rather than bolting on a homepage link. Then read the five pages that earn the most traffic and ask of each link whether a reader would want it: remove the ones that point nowhere useful, and add the ones a reader would have wanted but did not get.
Repeat the orphan check quarterly. Internal linking decays: pages get moved, sections get dropped, and links rot. A fifteen-minute check four times a year is the entire maintenance cost, and it catches the failure that no amount of content work can compensate for — a good page that nobody can find.
Limitations
The three-click rule is a heuristic from practice, not a measured threshold. Google has never stated that pages beyond three clicks are penalised, and deep pages do get indexed and do rank. The rule is useful because depth correlates with neglect — deep pages tend to be the ones nobody linked to, updated, or read — not because depth itself is a penalty.
Likewise, the advice here assumes a site small enough that a human can hold its structure in mind. On a site with hundreds of thousands of pages, internal linking is a data problem solved with crawl data and templates, not a writing problem solved with checklists. The principles are the same; the tooling is not.