Internal Linking Strategies for Topical Authority

The on-page signal most sites under-invest in by an order of magnitude

Enric Ramos · · 8 min read
A wooden table topped with scrabble tiles spelling the word link

Internal linking is the on-page optimization most sites treat as an afterthought — and the one with the largest unrealized upside. External backlinks are hard (someone else has to link to you). Internal links are entirely under your control. Every article you publish is an opportunity to reinforce the topical connections in your site. Most sites use 20% of this opportunity.

This article covers what makes internal linking actually effective for rankings, how to decide which articles link to which, and the weekly habit that compounds more than any single audit.

Why internal linking matters so much

Internal links signal three things to Google:

  1. Topical relevance of the target URL. When many articles on your site link to your "crawl budget" pillar with related anchor text, Google concludes that page is your authoritative crawl budget resource.
  2. PageRank flow. Links pass authority from source to target. Pages linked by many high-authority pages accumulate signal — this is the mechanic behind page authority in third-party metrics like Moz's PA.
  3. Crawl priority and discoverability. Heavily-linked URLs get crawled more frequently. Orphan URLs (no internal links) get crawled rarely.

The third point is particularly underrated. Internal linking directly affects crawl budget allocation — Google prioritizes pages that are obviously important to your site, and internal link count is one of the strongest "obviously important" signals.

Three data points:

  • 3-6 internal links in a 2,000-word article is the natural range for content-heavy articles.
  • 10-20 links on a category/hub page is fine because that's the page's job — surfacing children.
  • 50+ links on a page dilutes individual link signal. Each link has a fraction of the page's authority; with too many, no target gets enough.

Scale roughly linearly with article length. A 5,000-word pillar article naturally has 8-15 links; a 1,000-word supporting article has 3-5.

Below the minimum (under 3): the page is under-linked, leaving ranking signal on the table.

Above the maximum (10+ in a 2,000-word article): over-linking. Readers get distracted, Google's signal dilutes.

Priority order for what an article should link to:

  1. The pillar of its cluster. Every supporting article should link to its pillar. This is non-negotiable and one of the single highest-leverage linking rules.
  2. Relevant glossary terms. Terms genuinely discussed in the article, linked at first substantive mention. Reinforces topical vocabulary.
  3. Sibling articles in the same cluster. "If you found this useful, see also..." — but inline in the content, not just at the bottom.
  4. Related pillars in other clusters. Where the relationship is genuinely relevant. E-commerce SEO pillar naturally links to Technical SEO pillar (e.g., when discussing faceted nav + crawl budget).
  5. High-conversion product/commercial pages. Where the topic naturally warrants it, link to product or service pages for transactional intent.

What not to link to:

  • Your homepage, from every page. Homepage already has authority; adding more links to it wastes yours.
  • About/Contact pages. Irrelevant to most content; dilutes signal.
  • "Click here" on random unrelated pages. Misleading.
  • URLs that redirect (301/302 chains). Pass equity through the chain; just link to the final URL.

Anchor text patterns that work

Anchor text is the descriptor Google uses to understand what the target URL is about. Good anchor text is:

  • Descriptive of the target — not "click here" or "this article."
  • Natural in context — reads as prose, not as a keyword insertion.
  • Varied across links — don't always link to your crawl budget article with "crawl budget optimization." Mix: "crawl budget optimization," "how crawl budget actually works," "crawl budget for large sites."

Anchor text style examples:

Example Quality
"crawl budget" Good — direct and relevant
"how crawl budget works" Good — descriptive phrase
"crawl budget optimization framework" Good — longer descriptive phrase
"see here for more on crawl budget" OK — "see here" is generic but context compensates
"click here" Weak — no descriptive value
"crawl budget optimization for large sites with Googlebot" Over-engineered — natural prose beats stuffed anchors

For internal links, unlike backlinks, exact-match anchors are fine — you control the link graph, and some keyword overlap signals your editorial intent. Just don't stuff; use it where natural.

Silo vs mesh: what works in 2026

Silo approach

Each pillar + its supporting articles form a tight cluster. Cross-cluster links are sparse.

     [Pillar: Technical SEO]
     /   |   |   |    \
   TS1  TS2 TS3 TS4  TS5
   
     [Pillar: On-page SEO]
     /   |   |   |    \
   OP1 OP2 OP3 OP4  OP5

Pros: Clear topical clustering. Google sees concentrated signal per subject area. Strong authority-building within a cluster.

Cons: Misses cross-cluster connections that genuinely exist. Rigid; hard to maintain at scale.

Mesh approach

All content cross-links freely based on contextual relevance. Clusters emerge from content gravity, not from enforcement.

Pros: Reflects how humans actually read and connect ideas. Natural cross-topic links.

Cons: Can dilute topical signal. Every article linked to every other weakens each connection.

The hybrid (what works best)

Primary structure is siloed — pillars and supporting articles form tight clusters. But editorial cross-cluster links are encouraged when genuinely relevant.

Concrete rule: 70-80% of a supporting article's internal links go within its cluster (1 to pillar, 2-3 to siblings). 20-30% can cross clusters where topic warrants.

Example: a "faceted navigation for SEO" article in the E-commerce cluster naturally links to the crawl budget article in the Technical SEO cluster, because the two topics genuinely connect. That cross-cluster link strengthens both articles.

The mistake is being rigid either way. Pure silos miss real connections; pure mesh dilutes.

Internal linking is a compounding investment. Each new article you publish is an opportunity to:

  1. Link out to 3-6 existing articles — strengthens those articles.
  2. Link to glossary terms — reinforces terminology.
  3. Update 2-3 older articles to link to the new one — distributes signal to the new article.

The third point is the tactical habit that matters. Most sites publish new articles without updating the old ones. Result: new articles take weeks to build internal authority; older articles stagnate.

The weekly habit:

Every Friday, for each new article published that week:

  • Identify 2-3 older articles where the new article is genuinely relevant.
  • Edit those older articles to include a link to the new one.
  • 15 minutes total per new article.

After 6 months of this habit, your internal linking graph looks different from sites that don't do it. Over 2 years, the difference in topical authority is visible in rankings.

Detecting and fixing orphan pages

An orphan page has zero internal links pointing at it. It's invisible to Googlebot discovery via internal linking and gets crawled rarely via sitemap or backlinks only.

Audit workflow:

  1. Crawl your site starting from the homepage, following only internal links. Export the full list of discovered URLs.
  2. Compare against the full URL list (sitemap + GSC + database query).
  3. URLs in the full list but not reached by the internal link crawl are orphans.
  4. Prioritize by backlink count and historical traffic — orphans with external links are higher priority to reconnect.

Fixes:

  • For orphans worth keeping: identify 3-5 related articles and add links.
  • For orphans not worth keeping: 301 redirect to the most relevant live URL, or 410 if truly dead.
  • Prevention: ensure every new article gets linked from at least 2 existing articles on publish day.

How to audit internal linking at scale

Tools:

  • Screaming Frog — crawl + export inlinks and outlinks per URL. Reports "pages with few internal links" and "pages with no inbound internal links."
  • Sitebulb — cleaner visualizations of the link graph.
  • ahrefs Site Audit — integrates internal linking with authority metrics.
  • Custom: export internal links to a graph database (Neo4j) for complex queries — "show me all articles in cluster A that don't link to pillar A."

What to look for in an audit:

  • Articles with fewer than 3 inbound internal links.
  • Articles with fewer than 3 outbound internal links.
  • Pillar pages with fewer than 8 inbound links from their cluster.
  • Cluster articles not linking to their pillar.
  • Orphan pages.

Fix in batches. Prioritize pages with traffic potential or existing rankings that could benefit from an internal linking boost.

Common mistakes

Linking to the homepage from every article. Homepage already has plenty of authority. Link to relevant specific pages instead.

All links going to the same 3 "money pages." Concentrates authority but leaves many URLs under-linked. Distribute.

Nofollowing internal links to "sculpt PageRank." Setting rel="nofollow" on internal links was a valid technique in 2008; Google patched it in 2009. Nofollow internal links just waste the flow. Don't do it.

Links added only at the article's end. Users rarely reach the bottom. Inline links in the body get more clicks and do more for topic reinforcement.

Never updating old articles when new related content ships. Covered above — the single biggest missed opportunity.

Using identical anchor text across all links to one target. Google's systems notice unnatural anchor patterns. Vary naturally.

Frequently asked questions

Homepages naturally have many — navigation, featured content, footer. 30-60 is typical. What matters: the linked-to pages should be the ones you want to rank. Don't waste homepage link slots on pages that don't deserve the authority.

No. Link the first occurrence on the page; leave subsequent mentions as plain text. Over-linking the same term on one page dilutes the signal and annoys readers.

Rarely, and not for the reasons people think. PageRank sculpting doesn't work. The only legitimate use is on links to low-trust user-generated content (comment authors' profile links), but those are few.

How long before internal linking changes affect rankings?

Typically 4-8 weeks for Google to recrawl and reindex the updated link graph. Aggressive interlinking changes on high-traffic sites can show effect within 2-3 weeks.

External links to authoritative, relevant sources (not just competitors — any authority) is a quality signal. Linking to competitors when they're genuinely the best source for a sub-topic is fine. Not linking out at all looks suspicious to quality systems.

Related articles

Screenshot of the medium website search results page

Featured Snippet Optimization: Formatting for Zero-Click Wins

Featured snippets occupy prime SERP real estate — position zero, above organic results. Winning them requires structural formatting that Google's snippet extractor can reliably parse. Here's which formats work, how to structure content, and the queries worth targeting.

· 8 min read