Internal Linking: How the Pages of a Website Strengthen Each Other

Why do so few links lead to the most important page?

Count where the links on your website lead. On many business websites, almost every link points to the homepage, the contact form, the legal notice and the privacy policy. The service page with which the business earns its money, by contrast, can only be reached via one menu item. Articles stand side by side without referring to each other. Anyone who has read an article leaves the website because nothing shows them the next step.

Internal linking sounds technical. In fact it is a task of organization that every business can solve itself. The central question, then: How do you connect the pages of your website so that visitors find the way to the matching service and search engines recognize which pages are important? To answer it, we first clarify the terms and two models, one from the early days of search engines and one from research on information seeking. Then we look at what a study shows about structure and link texts, and end with a rule for every page and a check every six months.

A page that no link points to exists for visitors and search engines only on paper.

Terms: internal link, link text, orphan page

An internal link leads from one page of your own website to another page of the same website. An external link leads to another website; an incoming link from another website is called a backlink. The link text (anchor text) is the visible, clickable text of a link. An orphan page is a page that no internal link points to. It is published but can only be reached via the sitemap or a direct address.

Two kinds of internal links are also distinguished. Navigation links sit in the menu, the footer or the breadcrumb trail and are the same on many pages. Text links sit in the body text and point to where a thought suggests going deeper or taking a next step. Navigation links show the structure of the website; text links show connections between content.

The thinking model: links as paths and as cues

The first model comes from the early days of Google. Brin and Page (1998) described their search engine as a system that makes heavy use of the structure of hypertext, meaning the links between pages. A page that many others point to is considered more important, and the text of a link says something about what the linked page is about. The method has developed considerably since then, and Google today names many other signals. The basic idea also applies within a website, however: internal links show the search engine which pages the business itself considers important and how they relate.

The second model describes visitors. The article The FAQ page presents the information foraging theory of Pirolli and Card. According to it, people weigh how much useful information a path promises and how much effort it costs. The cue they use to estimate the benefit is what the theory calls information scent. This is mainly the text of a link or a heading and how much it sounds like the user’s goal. For internal linking this means: a link is only as good as its text.

Both models lead to the same conclusion. For the search engine, the link is a connection with meaning; for the visitor, a path with a sign. In both cases, the link text decides whether the connection is understood.

Two views: structure or link text?

The first view says that good navigation begins with a structure that matches the way visitors think. Whoever groups the content correctly need not worry much about individual links. The second view says that visitors do not follow a structure but words, and every link must be understandable on its own. One study tends to support the second view, as the next section shows. I nevertheless consider both necessary: the structure decides whether a path exists, the link text whether it is taken.

What a study shows about structure and link texts

Schmettow and Sommer (2016) set out to test whether websites are easier to use when their structure matches users’ mental model. First, they had 27 students at the University of Twente sort terms from municipal websites into groups, a method called card sorting. This produced an average mental model. Then they reconstructed the navigation of five Dutch cities, including Amsterdam, Utrecht and The Hague, and calculated for each item how far its position in the menu deviated from the mental model. Finally, 50 participants solved search tasks on these websites.

The result contradicted the authors’ expectations: the match with the mental model had no recognizable influence on how quickly and in how many steps an item was found. The length of the shortest path to the goal, by contrast, had a strong effect. And the side observations showed what the authors interpret as the importance of expressive link labels: on one website, an item on population figures was well placed under the menu item “News”. Only 2 of 8 participants found it within four and a half minutes. The authors conclude that a well-built structure fails when the wording is not right.

The authors name the limits themselves: students as participants, government websites instead of business websites, and only steps and time measured, not satisfaction. They deliberately did not consider cross-links in the text, only the main navigation. Two things can nevertheless be derived for a business. Short paths to the important page matter, and link texts must sound like the destination, not like the internal order.

The four-way rule: which links every page needs

The article SEO for SMEs describes the pillar-cluster principle: one overview page per main topic, several in-depth articles, all linked to each other and to the overview page. From this I derive a rule for every single page, the four-way rule. Every page offers four paths, where they exist.

1. Up

   A link to the overview page of the topic area, in the text and in the breadcrumb trail. That way the visitor knows where they are, and the search engine knows what the page belongs to.

2. Across

   One to three links to articles that answer a follow-up question, placed in the text where the question arises.

3. Down

   From the overview page to every in-depth article, with a link text that names its question.

4. To the service

   From every article, a link to the matching service page, with a sentence on why it fits here. This is the path to an inquiry.

The fourth path is the one most often forgotten. An article on the cost of a flat roof renovation that does not point to the service page for flat roof renovation answers the question and lets the reader go. The article Understanding search intent calls this a guide without a connection to the matching service. How service pages themselves are built so that the path continues there is described in the article The service page.

Good link texts: what goes into the link

Google recommends in its documentation that link texts be short and specific. Generic phrases such as “click here” tell neither readers nor Google what the target page is about. A simple test: read the link text on its own. If it makes no sense without the surrounding sentence, it is too generic. Google also advises against stuffing link texts with keywords or placing many links right next to each other (Google Search Central).

• Name the destination. The link text describes the target page, not the act of clicking.
• Build it naturally into the sentence. The link sits where the thought suggests going deeper, not collected at the end.
• Vary it. Different pages can link to the same target page with different fitting words.
• Describe images. If an image is linked, Google uses its alt text as the link text.
• Use real links. Search engines reliably follow only ordinary links with an address, not buttons that reach the destination only via a script.

The six-monthly check: finding orphan pages and weak paths

Google writes in its documentation that every page you care about should have a link from at least one other page on your own website. The article Understanding Google Search Console describes the four stages of visibility: found, indexed, shown, clicked. Internal links affect the first stage. Google often does not find pages that are not linked anywhere.

1. Open the Links report

   In Search Console, the “Links” report shows which pages are most often linked internally. Note where your most important service pages stand.

2. Look for orphan pages

   Compare the list of all pages from the sitemap with the internally linked ones. Pages without an incoming link get at least one fitting text link or are merged.

3. Check paths to the service

   For the five most visited articles, check whether each one points to the matching service.

4. Read the link texts

   Go through the links on the most important pages and replace every link text that makes no sense on its own.

5. Repair dead links

   Links to deleted or moved pages are switched to the new address.

The check becomes particularly important after a rebuild. When addresses change, old internal links run into nothing or take detours. Redirects catch this, but internal links should point directly to the new address. The article Planning a website relaunch describes the redirect list this requires.

The bigger picture

Internal linking shows whether a business thinks of its content as a connected whole or as a collection. A website with many well-written articles that do not refer to each other is like a library without a catalog. Whoever asks with every article where the reader wants to go next, by contrast, builds paths that visitors and search engines use alike. How specialist content is organized overall, as a blog or as a knowledge section, is described in the article Blog or knowledge section.

The limits are clear. How strongly Google weights internal links in detail today is not public. The model by Brin and Page describes the state of 1998, and the study by Schmettow and Sommer concerns government websites and students. What is robust is the common direction: short paths, clear link texts and no orphan pages help visitors, and what helps visitors is also understood by a search engine.

What remains and how to start

We set out to ask
to connect the pages of your website so that visitors find the matching service and search engines recognize what is important. The answer: by having every page link up, across, down and to the service according to the four-way rule, by making every link text name its destination and by leaving no important page without an incoming link. The two models explain why this works: for the search engine, links are connections with meaning; for visitors, paths with a sign; and in both cases the text decides.

• Identify the three most important service pages.
• Check in the Search Console Links report how often they are linked internally.
• Add a link to the matching service to the five most visited articles.
• Replace every link text such as “here” or “learn more” with the destination.
• Look for pages without an incoming internal link and connect or merge them.
• Add both linking steps to the workflow for new articles.

A website does not become stronger because it has more pages. It becomes stronger when its pages know each other.