38% Of The Web Is Already Gone

When an old link fails, most people assume something dramatic happened, a company folded, a server went dark. Usually the site is perfectly healthy and one page inside it was quietly deleted, and that ordinary process has removed 38% of the web that existed in 2013. This file opens the story of link rot: what the Pew Research Center found when it tried to open a decade of webpages, why news and government sites are worse than average, how much the Internet Archive manages to rescue, and the one number nobody can ever measure, the pages that vanished without anybody having written the address down. I'm the RiddleWanderer. Archivist of mysteries. My consultant is a rubber duck named Archimede. He has never once been helpful. Subscribe, and we'll open the next file together. Chapters: 0:00 You have done this 0:34 Consulting the expert 0:46 What actually dies 1:47 Somebody measured it 3:06 Where it hurts 4:20 The number about numbers 5:01 The people who tried 6:08 What can't be counted 6:54 Why I like this one 8:06 Close #riddlewanderer #lostmedia #internethistory Sources Pew Research Center (2024), "When Online Content Disappears: Link Rot and Digital Decay on Government, News and Other Webpages", Chapekis, Bestvater, Remy & Rivero, 17 May 2024 - the 38%, 25%, 8%, 23% and 21% figures, and the tweet finding (X/Twitter specifically, not social media generally) Internet Archive (2026), "Gone but Not Forgotten: Recovering the Dead Web", blog.archive.org, 23 April 2026 - the Wayback Machine recovery rates against the Pew dataset Zittrain (2021), "The Internet Is Rotting", The Atlantic - the analysis of around two million New York Times links Ahrefs (2024), link decay analysis, originally published April 2022 - a much higher figure from a different sample, showing how much the result depends on what you sample Web Science and Digital Libraries Research Group, Old Dominion University (2024), "Some URLs Are Immortal, Most Are Ephemeral" - on the 44, 75 and 100-day figures dating from 1996–2003 research Credits Screen capture and diagrams: original Photography: pexels.com · Library of Congress, loc.gov/free-to-use Music: YouTube Audio Library