How many internal links per page? Fewer rules, better reasons.

“How many internal links should a page have?” is one of the most searched questions in internal linking, and most of the answers you will find give a number. Five or six per page. Twenty to forty for a blog post. Three to five per thousand words. They disagree with each other because none of them comes from anywhere in particular.

This guide gives you something more useful than a number: where the old limit came from, how search engines are thought to treat many links on one page, what sensible ranges look like for three very different pages, and a rule of thumb you can apply to any post in about a minute. It is a narrower deep dive under our internal linking strategy for blogs, which covers what to link and in what order.

The short answer: there is no magic number

Google’s own documentation answers the question directly. In its guide to making links crawlable, under internal links, it says there is “no magical ideal number of links a given page should contain”, followed by: “if you think it’s too much, then it probably is.” The same section asks that every page you care about has a link from at least one other page on your site.

That gives you two real constraints and a lot of room in between:

• A floor. Every page that matters needs links pointing at it, or it is an orphan page that readers and crawlers only reach by accident.
• A ceiling set by readers. Past a certain density, links stop helping people decide where to go next and start getting in the way of the text.

The published rules of thumb sit all over that range. One plugin vendor suggests around five or six internal links per page; another suggests 20 to 40 for a blog post. Neither cites data for its figure. Treat any fixed number as somebody’s habit, not a finding.

Where the “100 links” rule came from, and why it went

The number that still haunts SEO audits is 100. It was real: Google’s guidelines used to say it. But the history is shorter, and the reason more mundane, than its reputation suggests. Archived copies of Google’s guidelines show the wording change over time.

Why 100 in the first place? Matt Cutts, then a Google engineer, explained it in March 2009: Google used to index only about the first 100 kilobytes of a page, so a page with hundreds of links risked being cut off. By 2009 that limit had gone, and he said the remaining reason to stay under roughly 100 was user experience. He also wrote that pages with lots of links “are not automatically considered spammy”. In February 2011 Google published a Webmaster Help video titled “Does Google still recommend 100 links or fewer per page?”, and by that June the number had gone from the guidelines.

How link value is thought to split across a page

The other reason people worry about link counts is link equity: the idea that a page has a fixed amount of value to pass on, shared between its links. That idea comes straight from the original PageRank formula.

In the 1998 paper that introduced Google, Sergey Brin and Lawrence Page define a page’s PageRank using, for each page linking to it, that page’s own PageRank divided by the number of links going out of it. Ten links out, and each carries a tenth of what the page passes on. A hundred links, a hundredth.

Two things complicate that tidy picture, and an honest answer has to include both.

• Hiding links does not save value. In his 2009 post on PageRank sculpting, Cutts explained that Google had changed how nofollow links were counted: they still count towards the number of links on the page, but pass nothing. A page with ten links, five of them nofollowed, passes the same share through each followed link as if all ten were followed. Nofollowing your own menu or footer links does not concentrate anything on the rest. He described the classic model as a simplification, and advised letting PageRank flow freely within a site.
• Links are probably not weighted equally. A Google patent, US7716225B1, “Ranking documents based on user behavior and/or feature data”, describes weighting each link by how likely a reader is to follow it, using features such as the font size of the anchor text and the link’s position in a list. This is often called the reasonable surfer model. A patent shows what Google has worked on, not what it runs today, so treat it as a strong hint rather than a fact.

Google has also said internal and external links are not treated the same. Asked in 2017 whether the calculations differ, John Mueller of Google replied with one word, “definitely”, without saying how.

So the practical reading is cautious. Every extra link on a page probably shares out a little more thinly what the page passes on, and links readers are likely to use probably count for more than ones they are not. That is a reason to make each link earn its place. It is not a reason to strip a page down to three links, and it says nothing about a specific number.

Outbound and inbound: two different counts

“Links per page” mixes up two numbers that do different jobs.

• Outbound links are the links on a page pointing elsewhere on your site. They help readers move on and pass a share of the page’s value to its targets.
• Inbound internal links are the links on other pages of your site pointing at this one. They are how readers and crawlers find the page, and they are your own vote on how important it is.

If you only have time to improve one, improve inbound. Google’s floor is about inbound links (“a link from at least one other page”), and the most cited study of the subject measured inbound links too. In its study of 23 million internal links across 1,800 websites, Zyppy found that URLs receiving 40 to 44 internal links got about four times the Google clicks of URLs receiving 0 to 4, and that the trend reversed after about 45 to 50 links, which the author attributes mostly to sitewide navigation links pointing at lower-traffic pages. The author is clear that it is mostly a correlation study: pages that matter more tend to get both links and clicks, so neither proves the other.

On a real site the two numbers are often badly out of balance. A personal-finance blog with 200 posts might find its 2,500-word emergency-fund guide, the page it most wants to rank, has three inbound links from other posts, while a 2019 post about a since-discontinued savings account has thirty, because it sat in a “popular posts” widget for years. Neither page’s outbound count tells you that.

Seeing both numbers in WordPress

• Yoast SEO adds two columns to Posts → All Posts: internal links in each post, and internal links pointing at it. You can sort by either, and if the columns are missing they are switched on under Screen Options. Yoast’s guide to the text link counter covers this, and Yoast staff have pointed to Tools → Count links in your texts in the Yoast SEO menu to count posts written before the plugin was installed.
• Rank Math shows internal, external and incoming link counts in the Posts and Pages lists, according to its link counter documentation. Seeing the individual links, with each source post and its anchor, needs the Link Genius module in Rank Math PRO.
• Google Search Console has a Links report with a table of your most internally linked pages. It is capped at 1,000 rows, which is plenty for most blogs.

If you are choosing between those plugins’ link features and a dedicated tool, our Silotree and Yoast comparison puts the two side by side.

Realistic ranges by post length

Length changes the number of places a link can sit naturally, so it is the most useful single input. Here are three real-looking pages. The ranges are judgement calls for links written into the body, not findings from a study.

Three things to notice:

• The ceiling is rarely the target. The tent guide could carry 30 links, but it only has about 13 distinct pages worth sending a reader to. Linking the same review three times does not help anybody, so the real number is set by how many useful destinations exist, not by length.
• Short pages still need their few links. The help article is short because the task is. Its two or three links are the most valuable thing on the page, because a reader stuck at step one needs the step before, and a reader who finished needs the step after.
• The template does not count towards this. All three pages also carry a menu, and the help centre has a sidebar listing thirty articles in the same category. Those are decided once, in the theme. The table is about the links you choose post by post.

For context on how other sites behave, LinkStorm’s analysis of 2.5 million in-content internal links across 1,700 sites describes how its customers link, such as anchor length and how deep the targets sit, but it does not report links per page or per thousand words. That gap is telling: nobody has published good evidence for a per-page number.

A rule of thumb you can actually use

You still need something to go on when you are editing a post. This is the rule of thumb we use. It is a heuristic, not a measurement of anything Google does.

1. Ceiling: about one body link per 100 words

   A 1,200-word post tops out around 12. Most posts land well under it, and that is fine.

2. Floor: two or three, whatever the length

   One up to the hub or category the post belongs to, and one or two to the obvious next question.

3. One link per destination

   Pick the sentence where it helps most and link it there once.

4. At most one new link per paragraph

   A paragraph with three links asks the reader to make three decisions at once. Spread them out, or drop the weakest.

5. Inbound before outbound

   For each page you care about, aim for at least three links from related posts. When you add a post, add links to it as well as from it.

If you would rather not do the arithmetic, use the internal links per page calculator to work out a range for the post you are editing.

Silotree’s own link budget follows the same pattern for the links it suggests: about one per 100 words, no more than one new link in any paragraph, and no more than six of its links in a single post. That is a product choice made for readers, not a claim about a ranking threshold.

Which words to put those links on is its own subject. Our guide to choosing anchor text for internal links covers it, and the deep dive on how much exact-match anchor text is too much covers the one anchor question that worries people most.

Menus, footers and other sitewide links

Most of the links on a typical page are not in the post at all. They are navigational links in the header, sidebar and footer, repeated on every page. A small travel blog can easily ship a mega menu with 60 destination links, a sidebar of 15 recent posts and a footer of 20 more: nearly a hundred links before the first word of the post.

Template links are not bad. They are just a different tool:

• They are good for reach. A hub in the menu is one click from every page. Nothing written into a single post can do that.
• They say little about any one page. A link that appears on every page tells a reader nothing about why this page links there. That is why contextual links, written into a relevant sentence, carry the context template links lack. Our guide to contextual and navigational links goes into the difference.
• They dilute, if the classic model holds. On the right-hand page above, each body link is one of 60 rather than one of 9. Zyppy’s study saw clicks fall above about 50 inbound links and put that down to sitewide navigation, though it is a correlation and the author says so.

The cheapest fix is usually in the theme, not in the posts. Cut a mega menu down to the hubs readers actually use, drop “recent posts” from the sidebar if your home page already lists them, and keep the footer to the pages people look for there: contact, about, policies. On a big archive, a good menu also keeps old posts within reach; see how click depth works for why that matters.

Signs a page has too many links

Since there is no number, use symptoms. If you see several of these on one page, it has too many links, whatever the count says.

• Whole paragraphs turn blue. If two or three links share a sentence, the reader can no longer tell which one matters.
• The same destination is linked more than once. Usually a sign links were added in a hurry, or by an automatic keyword rule.
• Links point at weak or off-topic pages because they happened to share a word: a sourdough post linking “starter” to a post about car batteries.
• Every occurrence of a phrase is linked. This is the classic output of an auto-linking plugin left running.
• Links in headings, buttons or image captions that nobody would expect to be clickable.
• Nobody clicks them. If your analytics show a link-heavy post sending almost no readers onward, the links are decoration.

The fix is subtraction. Keep the link in each paragraph that a reader is most likely to want, and remove the rest. You lose nothing: the target pages keep the links that were doing the work.

What to do this week

A short plan you can finish in an afternoon:

1. List your ten most important pages: the ones that earn money, sign-ups or most of your search traffic.
2. Check their inbound counts in Yoast, Rank Math or Search Console. Any with fewer than three links from other posts go to the top of the list.
3. Add two or three inbound links to each, from related posts where a sentence already mentions the subject.
4. Open your three longest posts and apply the rule of thumb: no paragraph with more than one new link, no destination linked twice.
5. Look at your template once. Count the links in your menu, sidebar and footer. If it is more than about fifty, decide which ones readers really use.

Where Silotree fits

The counting is easy; finding the right sentence in the right post is the slow part. The free link checker crawls your published pages, with no account, and lists orphan pages, dead ends, broken internal links and click depth, which tells you where inbound links are missing. A Silotree account then suggests links from sentences you already wrote, within the budget above, and you approve or skip each one; the first 10 links are written free on every plan, and any batch can be undone.

Questions people ask