NOTE

The internal linking problems we keep finding on websites

While researching websites for SEO outreach, my team and I kept running into the same problem: sites with hundreds of articles and several service pages, but very little connecting them. Internal linking is not an advanced technique. It is basic site architecture, and it was repeatedly missing.

By Maharshi KushwahaUpdated 8 min read
On this page
  1. Five patterns we kept seeing
  2. What published industry data says
  3. Why it costs more than it looks
  4. Seven problems to check in an audit
  5. How to measure the problem
  6. What to fix first
  7. What we are doing next
  8. The lesson
  9. Sources and further reading

We were reviewing websites to understand their SEO condition, find weaknesses and identify businesses that could benefit from professional SEO work. Rather than looking only at keywords, rankings or traffic estimates, we examined site structure and the technical and on-page elements behind it. Internal linking became one of the issues that kept coming back, across very different kinds of website.

In some cases a site had plenty of content and almost no connection between it. This note sets out what we saw, puts it next to published industry data, and explains what I would check and fix first.

Five patterns we kept seeing

1. Related articles were not connected

Sites published several articles on closely related subjects with very few meaningful links between them. A site might have separate articles on technical SEO, website speed optimisation, crawl budget, XML sitemaps, robots.txt, website architecture and indexation problems, and treat each one as an independent piece. There was no clear content hierarchy and no consistent contextual linking connecting them.

2. Important service pages got little support from content

A digital marketing agency might have an SEO services page, a technical SEO page, a local SEO page, a content marketing page and dozens of informational blog posts, but the posts did not consistently link to the relevant service pages. A reader researching a technical SEO problem could naturally benefit from finding the agency's technical SEO service. That connection was frequently missing, which is a missed chance to connect informational content with commercial intent.

3. Links existed, but without a strategy

Some sites did have internal links, just not useful ones:

  • Repeated "click here" and "read more" anchors.
  • Links to loosely related articles.
  • Many articles linking to the same destination while other relevant pages were ignored.
  • Important pages receiving little contextual support.
  • Links added to raise the count rather than to help navigation.

Having internal links is not the same as having an internal linking strategy. Relevance and placement are what matter.

4. Topic clusters existed in the content but not in the architecture

Some sites had clearly invested in a subject. A site might have 30 articles about business accounting that do not consistently link to a central accounting guide or to the relevant supporting resources. The topical content is there; the internal architecture does not reflect the relationships between it.

5. Linking was treated as a one-off, not ongoing work

Internal links are considered at publication, if at all, and articles are rarely revisited for new linking opportunities. As a site grows, new articles are published, old pages go out of date and new commercial pages are added. Without maintenance, the linking structure drifts further apart.

Publishing more content does not automatically create a better-connected website. Content has to be connected on purpose.

What published industry data says

Our observations are consistent with a large technical SEO study Ahrefs published in August 2023, which examined 1,002,165 domains for common technical SEO issues. Two of its findings are directly relevant:

Issue in the Ahrefs studyShare of domains
At least one page with only one dofollow incoming internal link66.2%
At least one page linking to a redirect62.7%

Both numbers need reading carefully. The first does not mean 66.2% of all pages had only one internal link; it means the issue was found on at least one page on 66.2% of the domains. The second is a separate internal-link maintenance issue, not evidence that 62.7% of internal links point to redirects.

Even read conservatively, about two-thirds of the domains in the study had at least one under-linked page. That is independent, industry-level context for the pattern my team saw by hand.

Why it costs more than it looks

Internal links do more than help visitors move around:

  • Discovery. Search engines use links to find pages and explore a site's structure. A page with no internal links may still be found through an XML sitemap, a backlink or another source, but that does not replace a meaningful internal structure.
  • Relationships. Linking an article about crawl budget to a guide about website architecture tells readers and search engines the two belong together.
  • PageRank. Internal links influence how PageRank flows through a site. More links do not guarantee better rankings, though; relevance, accessibility, page quality and overall structure matter too.
  • Commercial pages. A local SEO service page should be connected to the articles that support it, such as Google Business Profile optimisation, local citations, local keyword research and local landing pages.

The mechanics are covered in depth in the internal linking guide.

Seven problems to check in an audit

ProblemWhat to look forKeep in mind
Orphan pagesPages that cannot be reached through the internal links found in a crawl. Common causes: removed category links, old campaign landing pages, disconnected blog posts, redesigns, discontinued products, pages published without links.They may still be indexed or discovered through other sources.
Pages with one incoming internal linkImportant service pages, category pages, pillar articles or landing pages that depend on a single link.Not automatically a problem for every page; importance and link type matter.
Excessive crawl depthHow many clicks a page is from the homepage. Semrush recommends keeping your most important content at a crawl depth of 3 or less.A practical architecture benchmark, not a Google ranking requirement.
Generic anchor text“Click here”, “read more”, “learn more”, “visit this page”.Not always wrong, but descriptive anchors tell readers where they are going.
Disconnected topic clustersRelated articles that do not link where the relationship would help, such as crawl budget, internal linking, XML sitemaps, canonical tags and indexation on a technical SEO site.Connect where useful, not everything to everything.
Broken and redirecting internal linksLinks to broken pages, and links that pass through a redirect before reaching the final URL.Link straight to the final destination.
Important pages without contextual linksPages linked only from the header, footer or navigation, with little support from related content.Separate template links from contextual links in the audit.

The step-by-step audit, including how to separate template links from contextual ones, is part of the internal linking framework.

How to measure the problem

To turn observations into evidence, the same metrics need to be collected for every site:

MetricFormula
Orphan page rateOrphan indexable pages ÷ total indexable pages × 100
Single-inlink page ratePages with one unique incoming link ÷ total indexable pages × 100
Internal link coverageImportant pages with contextual incoming links ÷ total important pages × 100
Broken internal link rateBroken internal links ÷ total internal links tested × 100
Redirecting link rateLinks pointing to redirects ÷ total internal links tested × 100
Deep page rateImportant pages beyond target depth ÷ total important pages × 100
Generic anchor rateGeneric contextual links ÷ contextual links analysed × 100

As an illustration only, a site with 1,000 indexable URLs, 80 orphan pages, 150 pages with a single incoming internal link and 25 of its 100 important pages beyond crawl depth 3 would have an orphan page rate of 8%, a single-inlink page rate of 15% and 25% of important pages beyond depth 3. These are hypothetical figures to show how the metrics work, not findings from our research.

What to fix first

PriorityProblemAction
CriticalImportant orphan pagesAdd relevant internal links
HighImportant pages with very few incoming linksFind contextual linking opportunities
HighImportant pages buried beyond depth 3Improve their accessibility
HighBroken internal linksRepair or remove broken destinations
MediumDisconnected topic clustersBuild relevant content connections
MediumGeneric contextual anchorsImprove anchor descriptions
MediumLinks pointing to redirectsUpdate links to final URLs

The goal is not to add as many links as possible. It is to make the site easier to navigate, make content relationships clear, and give important pages the internal support they need.

What we are doing next

The research started as prospecting, but the pattern deserves a proper study. The plan is a sample of 100 websites: 25 local service businesses, 25 ecommerce sites, 25 blogs and publishers and 25 professional service sites. For each one we will record the total indexable URLs, orphan URLs, pages with one unique incoming internal link, average and median incoming internal links, important pages beyond crawl depth 3, broken and redirecting internal links, generic contextual anchors, the share of important pages with contextual links, and the number of disconnected topic clusters, along with the site's category, approximate size, CMS and audit date.

One caveat will travel with the results: this is a prospect-based sample, not a random sample of the web. Sites selected because they may need SEO help are likely to show more problems than a random group.

The questions we want to answer:

  1. How many sites in our prospect database have orphan pages?
  2. What percentage of important service pages receive no contextual links?
  3. How does internal linking differ between local service sites and ecommerce sites?
  4. How many sites have important pages more than three clicks from the homepage?
  5. How often do sites publish related articles without linking them?
  6. Does improving internal linking lead to measurable changes in crawling, indexing, impressions or clicks?

The last question needs a controlled test rather than a survey, which is what the internal linking at scale experiment is designed to measure.

The lesson

The surprising part is not that some sites have technical SEO problems. It is how many invest in publishing content without paying attention to how it connects. A site can have hundreds of articles and still have a disconnected architecture, several service pages that no informational content points to, and excellent content that leaves important pages with almost no meaningful internal links.

Don't just publish more pages. Build meaningful connections between the pages you already have. Internal linking belongs in the content strategy, the site architecture and the ongoing SEO process, not added as an afterthought. It is the same argument I make about volume in publishing more is not the answer.

Sources and further reading

Work with me

Need help applying this?

Reading is the easy part. If you want this implemented on your site, tell me what you're working on.

Work With Me Follow the SEO Roadmap
Email — [email protected]
Based in — India · working remotely
Practising SEO — since 2018