Internal Linking: +7% Organic Traffic in a Controlled A/B Test

Internal linking is the most underestimated lever in on-page SEO. While external backlinks get a lot of attention, the structure of your internal links decides which pages Google finds in the first place, how link equity is distributed across the site and which content shows up in AI answers. This article positions internal linking as an on-page discipline, describes an actionable strategy for larger B2B sites in the DACH region and shows how to measure success. It relies exclusively on documented data from industry studies and official Google guidelines.

Why internal linking matters for B2B sites

Internal links are the connections between pages on the same website. They do three jobs at once: they help search engines and people find content, they distribute relevance and link equity across the site, and they provide topical context through the anchor text. Google says this directly in its official guidelines. According to Google Search Central, paying more attention to the anchor text of internal links helps both people and Google understand your site more easily and find other pages.

The sheer prevalence of internal links shows that they are the norm. According to the HTTP Archive Web Almanac 2024, the median is 41 internal links per page across all sites, while external links are far less common at just 6 per page. For the top 1,000 sites, the figure rises to 129 internal links per page versus 16 external links on desktop. Internal links are therefore the dominant type of link on the web, and large B2B sites with product, solution and knowledge sections in particular depend on their quality.

The economic context makes this more pressing. Most pages get no search traffic at all. According to an Ahrefs analysis of around 14 billion pages, 96.55 percent of these pages get zero traffic from Google. Internal links are one of the few levers that can pull pages out of this zero-traffic majority, because they redistribute link equity in a targeted way. The same analysis names exactly this mechanism: stronger sites have more PageRank, which they can pass to pages through internal links so that those pages rank.

For Austria, the market structure adds another factor. Google dominates the search engine market with 81.87 percent market share, ahead of Bing at 9.01 percent. If you optimize internal linking in Austria, you are effectively optimizing for Google and its crawling, ranking and AI Overview logic.

How internal links control crawling, indexing and link equity

Internal linking works first at the technical level: it controls which URLs Google discovers and crawls. The format is what matters here. According to Google Search Central, Google can generally only crawl a link if it is an <a> HTML element with an href attribute. The guidelines explicitly list constructions such as a span with href, or navigation triggered by JavaScript events without href, as not recommended. In practice, only real href links present in the HTML count for discovery and link equity. Elaborately rendered menus without proper anchors are worthless from an SEO perspective.

On this basis, internal links control the crawl budget. The better an important page is linked internally, the sooner and more often it gets crawled. On smaller sites, crawl budget is rarely a bottleneck. On larger B2B sites with thousands of URLs, filters and pagination, however, the internal link architecture becomes the central control instrument for steering crawlers to business-critical pages instead of endless parameter variants.

The second level is the distribution of relevance and link equity. Every internal link passes on part of the linking page's authority. Important pages should therefore be linked from many relevant places, and unimportant or duplicate pages correspondingly less often. This deliberate control of internal link flows, often called PageRank sculpting, is today less a matter of technical tricks than of a clean topical architecture.

Strategy: anchor text, topic clusters and the sweet spot

An effective internal linking strategy rests on three pillars: descriptive anchor text, a topical hub-and-spoke architecture and a sensible link density.

Descriptive anchor text: Anchor text is the strongest on-page signal of an internal link. Google's official guidelines call for anchor text that is descriptive, reasonably concise and relevant to the target page. Generic anchors such as "click here" or "learn more" waste this signal. In practice, the picture is mixed. According to the HTTP Archive Web Almanac 2024, 84 percent of desktop pages and 91 percent of mobile pages pass the Lighthouse test for descriptive link text, which means roughly 9 to 16 percent fail with non-descriptive anchors.

Relevance between anchor and target matters more than description alone. An analysis of 2.5 million contextual internal links on 1,700 websites shows a widespread quality problem. According to LinkStorm, around 81 percent of anchors are keyword-rich and only about 15 percent are generic, yet 28 percent of anchors show no semantic similarity to the title of the target page. Descriptive anchor text alone is not enough; it has to match the topic of the target page.

Hub-and-spoke and topic clusters: The most viable architecture for entities and E-E-A-T is the topic cluster model. A pillar page covers a topic broadly, and several cluster articles go deeper into specific aspects. Cluster articles link contextually up to the pillar page and across to related pages. The result is a semantically dense network that signals topical authority to search engines and concentrates link equity on the central pages. The same LinkStorm analysis shows that 71 percent of contextual links appear within the first two hierarchy levels of a site, which underlines the importance of a flat, clearly structured architecture.

The sweet spot for link density: More links are not automatically better, but too few are a clear problem. The documented reference point for a content-rich B2B page is close to the median of 41 internal links per page that the HTTP Archive Web Almanac 2024 measures across all sites. Very large, authoritative sites commonly reach much higher values, up to the 129 internal links of the top 1,000 sites. The practical consequence: a central page with only a handful of internal links is under-linked. A sensible number of contextually relevant links, embedded in a clean navigation and content structure, is the norm on good sites.

Best practices for implementation

The data and Google's guidelines point to a concrete approach.

Limit click depth: Important pages should be reachable within three clicks from the homepage. The deeper a page is buried, the less often it gets crawled and the less link equity it receives. A flat architecture with clear hubs keeps business-critical pages close to the surface.

Place contextual links in body text: Links within editorial content carry more topical weight than pure navigation or footer links, because descriptive text surrounds them. Place several contextual links per article to relevant pillar and cluster pages.

Use only crawlable href links: Make sure every relevant internal link is a real <a> element with href, as Google's guidelines on crawlable links require. Avoid JavaScript-only navigation without href for SEO-relevant paths.

Use descriptive, varied anchors: Write anchor text that describes the target page, and avoid the 28 percent of anchors without a semantic connection to the target that LinkStorm documents. Vary the wording instead of mechanically repeating the exact same anchor.

Link bidirectionally and by topic: Link from cluster to pillar and from the pillar page back to the most important cluster articles. This creates a consistent topical network that strengthens the entity.

A controlled test shows that targeted internal linking has a measurable effect. In an SEO A/B test by SearchPilot on a site with around 8,000 regional pages, adding internal links to the geographically closest regions led to 7 percent more organic traffic to the linked target pages. The effect was measured against a forecast baseline with a confidence interval, which makes it independent of seasonality, algorithm updates and other external influences.

Common internal linking mistakes

Many visibility problems come from avoidable mistakes in the internal link structure.

Orphan pages: A page without a single internal link receives no link equity and is barely crawled. Orphan pages feed directly into the zero-traffic majority where, according to Ahrefs, 96.55 percent of the pages studied end up. Every important page needs at least one relevant internal link, preferably several.

Generic anchors: If you consistently link with "click here" or "learn more", you waste the strongest on-page signal. The 9 to 16 percent of pages with non-descriptive anchors in the Web Almanac 2024 show that this mistake is still common.

Anchors without topical relevance: A descriptive but off-topic anchor is almost as harmful as a generic one. The 28 percent of anchors without semantic similarity to the target page in the LinkStorm study show how often anchor and target diverge.

Click paths that are too deep: Pages that can only be reached after many clicks get crawled less often and rank worse. A hierarchy that is too deep buries important content.

Non-crawlable links: Navigation that only works through JavaScript without href is, according to Google, not a crawlable link, and it passes neither discovery nor link equity signals.

Internal linking for featured snippets, GEO and schema

Internal linking increasingly affects AI answers and structured search results as well. Generative Engine Optimization (GEO) aims to get content cited in the answers of generative search engines. A study accepted at KDD 2024 quantifies the potential. According to the GEO study by Aggarwal et al., GEO can increase visibility in the answers of generative search engines by up to 40 percent. The most effective methods in the study are adding statistics, quotations and sources to structured answer paragraphs. Internal links to such evidence-based content make these sources discoverable for crawlers and AI systems and thereby support citability.

For featured snippets and AI Overviews, a similar logic applies: contextual internal links in clearly structured answer paragraphs lead users and machines alike to in-depth content and strengthen the topical cohesion of a cluster. If a pillar page with precise definitions and documented statistics has strong internal links, its chance of being selected as a source goes up.

Structured data reinforces this effect. BreadcrumbList markup reflects the internal hierarchy in machine-readable form and improves how pages appear in search. With sameAs and entity connections, you can also link internal content to established entities. Google shows in several case studies that structured data has a measurable effect on click-through rate. According to Google Search Central, Rotten Tomatoes added structured data to 100,000 pages and achieved a 25 percent higher click-through rate, Nestlé measured an 82 percent higher click-through rate for pages with rich results compared with non-rich results, and Food Network recorded 35 percent more visits after the switch. Internal linking and schema markup belong to the same on-page strategy.

Metrics, workflows and measurement on large sites

Internal linking should be managed based on data, especially when you scale to thousands or tens of thousands of URLs.

Orphan pages and internal links per URL: The most important baseline metric is the number of internal links pointing to each URL. A crawl of the entire site uncovers orphan pages and under-linked important pages. The goal is that no business-critical page is left without relevant internal links, so you do not worsen the zero-traffic rate of 96.55 percent with self-inflicted errors.

Click depth: Measure how many clicks each important page is from the homepage. Pages beyond three clicks are candidates for better internal linking.

Anchor quality: Check the share of descriptive anchors and compare it against the benchmarks. The Web Almanac figure of 84 percent of desktop pages with descriptive link text and the LinkStorm finding of 28 percent mismatched anchors are useful reference points.

Prove impact with A/B tests: On large sites, you can cleanly isolate the effect of internal linking. The SearchPilot test with 7 percent more organic traffic from additional internal links is the model: variant against control, measured against a forecast baseline with a confidence interval. This separates the contribution of internal linking from external influences.

For the workflow: manual, editorial linking delivers the highest quality for central pages but does not scale to tens of thousands of URLs. Automated or tool-assisted linking covers the breadth but needs clear rules for anchor relevance and link density, so that it does not produce masses of generic or mismatched anchors. Regular link audits keep the structure consistent.

Next steps and conclusion

Internal linking is one of the few on-page disciplines that influences crawling, ranking and AI visibility at the same time and is entirely under your own control. The sequence for a B2B site in the DACH region is clear: first set up the architecture as hub-and-spoke with pillar pages and cluster articles, then make every important page reachable within three clicks and through several relevant internal links, and finally set descriptive and topically relevant anchors and use only crawlable href links.

Three points are decisive. First, no important page should be left orphaned, or it adds to the zero-traffic majority of 96.55 percent. Second, the anchor has to fit the target page, not just sound descriptive, to avoid the LinkStorm problem of 28 percent mismatched anchors. Third, it pays to combine internal linking with statistics, quotations and schema markup, because according to the GEO study this can bring up to 40 percent more visibility in generative answers and, according to Google, up to 82 percent higher click-through rates through rich results. Sensible next steps are a full crawl to identify orphan and under-linked pages, an audit of anchor quality and a first controlled A/B test on a scalable page template.