"The average web page only lasts 100 days" is one of those statistics that's been repeated so often it's basically folklore. It's also from 2003. Before that, a 2001 study put the number at 75 days, and before that, a 1997 Scientific American piece claimed 44. Those figures describe the web of the dial-up era — GeoCities pages, personal sites hand-edited in Notepad, a completely different internet than the one anything you're reading now lives on. So what does current research actually say, and does "average" even mean anything useful here?
The most rigorous recent answer comes from "Not Your Parents' Web," a 2024 collaboration between the Internet Archive, Old Dominion University's Web Science and Digital Libraries Research Group, and the Filecoin Foundation, analyzing 27.3 million archived URLs captured by the Wayback Machine between 1996 and 2021. Their finding: the median lifespan of a web page is about 2.3 years — half of everything that ever existed on the web is gone within that window. The average, pulled upward by a long tail of pages that survive for decades, comes out higher, around 5.1 years. That gap between the median and the average is the whole story: most pages die young, and a small minority live an extremely long time, which is exactly why a single "average lifespan" number is misleading on its own.
Overall, the study found roughly 60% of the archived URLs it examined are no longer accessible on the live web at all — gone to a 404, a DNS failure, or a domain that stopped resolving entirely, not just moved or redesigned.
The study's most useful finding for anyone deciding what to actually save is the split between root URLs (a site's homepage or top-level domain) and deep links (the specific article, thread, or page several clicks in). Root URLs are dramatically more durable — only about 10% vanish in a given year. Deep links have a median lifespan of just 1.3 years, roughly 18 months. In other words, the specific thing you actually wanted — the article, the answer, the thread — is the least durable part of the URL you bookmarked. The homepage it lived under will probably still be there. What you needed usually won't be.
It's tempting to picture link decay as a slow fade measured in years, but Pew Research Center's May 2024 study of 5 million tweets found removal can start almost immediately: 1% of tweets were gone within an hour, 3% within a day, 10% within a week, and 15% within a month of being posted. Eighteen percent were no longer publicly visible by the end of their tracking window. Pages don't announce when they're about to disappear, and "I'll come back and save it properly later" assumes a timeline the content itself isn't guaranteed to honor.
The same Pew study looked at longer-lived content too: of pages that existed at any point between 2013 and 2023, a quarter were inaccessible as of October 2023 — 38% of 2013-era pages were gone, versus 8% of 2023-era pages, a pattern that lines up closely with the half-life curve above: the older it is, the more of the decay has already happened, but the decay itself started from day one.
A page doesn't wait until it's old to start dying. It starts the moment it's published — some pages just take longer to finish the job.
None of this requires guessing which pages will be the unlucky ones. The only strategy that works against a decay curve that starts on day one is saving the actual page — not the link to it — the moment it's worth keeping.
The Starchive saves a full, faithfully rendered copy the moment you archive a page — not a link that inherits the original's decay curve. Start a free 14-day trial, no credit card required.