Internal linking for SEO: how to build silos, fix orphan pages and control crawl depth

Internal linking is not a list of links appended to the end of an article. It is the architecture of the site: how pages support each other, in what order they are discovered and how quickly a visitor reaches the answer they came for.

A site with strong articles but no coherent internal links looks to Google like a library where the books are scattered across the floor. Every page exists, but nobody knows how to get to it. Links between pages are the shelves and labels that turn a pile of URLs into a structure a crawler can walk and a person can navigate.

01. What internal linking actually does

Every internal link carries four things at once: a discovery route, an amount of authority, a topical context and a priority hint. When a link is missing, all of those signals go missing with it.

Crawl budget is not a fixed quota handed out by Google. It is the result of three factors: how fast the server responds, how much demand exists for the content and how clean the URLs are that the crawler encounters. If a large site links every page to every other page without discrimination, Googlebot burns resources on filters, sorting and parameter variations while the commercial pages wait at the back of the queue.

That is why internal linking has to be designed together with the URL structure and the indexing rules. A clean silo reduces the number of low-value pages the bot can reach and increases how often the important ones are revisited. If you want to see how these mechanisms are checked on a live site, we walked through the process in our technical SEO audit guide.

02. Silo architecture: hub, subpages, articles

A silo is a closed topical group in which pages link primarily to each other and only occasionally to the outside. The working model has three clear levels: a hub page that covers the broad subject, subpages that detail each branch and articles that support the topic with contextual information.

The practical rule is simple: every page has one clear topical parent, and sideways links exist only between pages that answer the same search intent. Commercial pages should sit no more than three clicks from the homepage. Articles can go three or four levels deep, as long as they are linked from their hub and from related material.

Breadcrumbs: visible hierarchy, not decoration

Breadcrumbs are generated automatically by the layout, but that does not make them correct. If the trail reads Home > Blog > Article, then the article must actually be linked from the blog index and from its topical hub. A breadcrumb that promises a hierarchy that does not exist is just a visual element. Check the BreadcrumbList JSON-LD too, because it has to mirror the same structure the visitor sees.

Why flat structures fail

A flat structure appears when every page is placed in the main menu and the footer, and every page links to every other page. At first glance it looks generous. In reality every page lands at the same depth, none of them receives a priority hint, and the menu becomes a wall of links that visitors ignore. A new article in a well-built silo is discovered because its parent is strong, not because it was squeezed into the footer.

03. Anchor text: descriptive, varied, natural

The anchor is the only part of a link a visitor reads before deciding whether to click. For Google it is also a hint about the subject of the target page. That is where the balance gets difficult: explicit enough to be useful, natural enough not to look written for an algorithm.

In practice, good anchors describe the destination in terms of the topic, not the click. "Technical SEO audit guide" tells you what you will find. "Click here" tells neither the visitor nor the crawler anything. And repeating the same exact phrase on every link to the same page turns a useful signal into a pattern that is easy to recognize.

One detail that is easy to overlook: menu and footer links almost always use generic anchors, repeated across the whole site. They are necessary for navigation, but they add nothing to contextual relevance. That is why in-body links carry more practical weight: they appear in a unique context, inside a sentence that explains the relationship between two subjects.

There is no magic number of links per page. There is, however, a limit to attention: if a paragraph has more links than ideas, the reader no longer knows what to choose. Two or three well-placed links in a medium-length article usually outperform ten links crammed into the last section.

04. Orphan pages: how to find them

An orphan page is an indexable page that receives no internal links. It may sit in the sitemap and may even earn traffic from external links, but inside the site there is no path to it. For Googlebot that page is a dead end: it gets visited once, forgotten quickly and never revisited when the content changes.

Orphan pages appear for mundane reasons: an article published without a link from the index, a service page built for a campaign and then forgotten, a category removed from the menu, a product taken out of a listing but left indexable. Nobody sees them in the navigation, so nobody fixes them.

Detection methods

• Compare the sitemap with the URLs found during a crawl. Anything that appears in the sitemap but not in the link graph is an immediate orphan candidate.
• Check the number of internal links received. Crawl tools show how many internal links point to each URL. Zero means orphan, regardless of its position in the sitemap.
• Read the server logs. Look for Googlebot requests to URLs with no internal referrer. A URL visited once, with no recurring traffic, deserves a closer look.
• Study the indexing report in Search Console. Indexed pages that earn no impressions, or pages stuck at "discovered but not indexed", often point to a linking problem rather than a content problem.

Fixing an orphan does not mean adding it to the footer. It means finding two or three relevant pages that would have a real reason to mention it and adding a contextual link there. If you cannot find a single page with a reason to link, the right question is whether the page deserves to exist at all.

05. Common mistakes that cancel out internal linking

Nofollow on internal links

Putting rel="nofollow" on internal links does not save authority and does not "reserve" anything for the important pages. In practice it only hides the relationship between pages and leaves the crawler without context. If you do not want a page indexed, use a robots tag or a robots.txt rule instead, not nofollow on the link pointing to it.

JavaScript-only navigation

A link has to be an <a> element with an href attribute. Buttons that change the page through onclick, or divs restyled to look like links, do not exist for the crawler when it parses the HTML. Google can render JavaScript, but rendering costs resources and is not guaranteed for every URL. Real links remain the safest discovery route.

Too many links and too little priority

There is no threshold of 100 or 150 links that blocks indexing, but dilution is real. When every page links to 80 other pages, each link passes less value and the reader has less chance of choosing. Keep the menu to the main sections and move secondary links into relevant contexts.

Duplicate anchors and decorative links

If the same phrase leads sometimes to one page and sometimes to another, the search engine can no longer infer the real destination. Likewise, a link to the contact page dropped into a paragraph about technical optimization adds noise, not context. Every link should have a defensible editorial reason to be there.

Linking everything to the homepage

It is the most widespread reflex: every mention becomes a link to the front page. The homepage piles up internal links it cannot use, while the pages that actually need authority get none. Link the specific page when one exists; the homepage is for orientation, not for every sentence.

Links to redirected or noindex pages

Every link to a URL that answers with a 301, 302 or 404 is a road with an extra stop. Authority leaks along the way and crawl resources are spent on redirects. Update internal links after every structural change and audit destinations with a crawl on a regular basis.

06. How to measure the results

Internal linking does not produce a dramatic chart in the first week. The effects appear as Googlebot revisits pages and recalculates signals. That is why measurement should happen at the level of page groups, not the whole site, and should be compared against a baseline of at least six to eight weeks.

• Crawl depth: the distribution of URLs by number of clicks from the homepage. The goal is to raise the share of commercial pages at three clicks or fewer.
• Search Console: impressions, clicks and average position for pages inside the silo, compared with the previous period. Watch the indexed page count in the indexing report as well.
• Analytics: events on internal link clicks, landing pages and the path toward conversion pages. A useful new link shows up in internal navigation, not just in total traffic.
• Server logs: how often Googlebot requests each directory and which status codes it meets. More requests on the right section confirms that discovery is fixed.

Do not confuse correlation with causation. A traffic increase after a linking change can also come from seasonality or an algorithm update. Note the date of the intervention, keep an untouched group of pages for comparison and track several signals at once.

A practical internal linking checklist

Run through these points before publishing a new article and review them after every structural change:

Where implementation usually gets stuck

On a simple brochure site, internal linking can be fixed by hand in a few hours. On an online store with thousands of products, categories, filters and brand pages create a graph you can no longer control from the editor. That is where code takes over: clear linking rules in templates, a correctly generated sitemap and a URL structure that does not spawn infinite variations.

If your current platform spends too much JavaScript on navigation, or does not let you control which links appear in listings, a rebuild with Laravel web development gives you direct control over the HTML you ship. You can also read about ecommerce architecture in Laravel if you want to see how a catalog structure connects to a content structure.