The web was built on pages linking to each other. But publishers increasingly link inward — to their own articles, their own apps, their own gardens — rather than outward to sources, other sites, or the wider web. This study measures that shift: what percentage of links on a typical page point to a different website, and how has that changed over the past decade?
Data: Common Crawl (306K sites) + detailed per-site analysis (1,000 top sites, via web archive), 2014 → 2020 → 2024. About
About this study
Data sources: Common Crawl WAT files (306K domain-year pairs across 2014, 2020, 2024) + Wayback Machine deep-dive (1,000 top sites with per-site link analysis).
Method: For each page, all <a href> links in the raw HTML are classified as internal (same registrable domain) or external (different domain). External links are further split into social media (Twitter, Facebook, etc.) and genuine editorial links (excluding same-organisation cross-domain links like bbc.com ↔ bbc.co.uk). Domain classification uses the Public Suffix List (tldextract).
SPA detection: Pages with <300 characters of visible text in the raw HTML are flagged as potential JavaScript-rendered shells (links may exist but are invisible to crawlers).
Limitations: Web archive captures are point-in-time snapshots; results may vary by capture date. The random sample does not cover all websites. JS-rendered pages are underrepresented in link counts.
This site loads a 14MB SQLite database (1,000 sites × link analysis) entirely in your browser. Download once — cached for future visits.
Median External Link Percentage
What % of links on a typical page point to a different website. Lower = fewer external links = more insular. Purple = all external links. Amber = excluding same-organisation links (e.g. bbc.com linking to bbc.co.uk counts as external but is the same org).
Distribution of External Link Percentage
What % of all sites fall into each bucket. Each bar = the share of sites whose external-link ratio is in that range. Purple = 2014, Amber = 2020, Green = 2024. The left-shift over time means more sites now have fewer external links.
Per-site verdicts (1,000 top sites, detailed analysis) 🔗 Share this view