Why links break over time (link rot) and how archives help

Click an old link and there’s a fair chance you’ll land on an error page, a site’s home page, or something that has nothing to do with what the link promised. This is called link rot, and it’s one of the quieter problems of the web: nothing announces it, but references stop working all the time.

Two ways a link goes bad

Researchers usually separate two problems:

• Link rot: the address no longer leads anywhere useful. You get a “not found” error, the domain has expired, or the server doesn’t answer.
• Content drift (sometimes called reference rot): the link still works, but the page has changed. The article was rewritten, the figures were updated, or the address now points to something else entirely.

Content drift is harder to spot, because nothing looks broken. You only notice when the page doesn’t say what someone quoted.

How common it is

Two studies give a sense of the scale.

In 2024, the Pew Research Center looked at samples of web pages collected each year from 2013 to 2023. It found that 38% of pages that existed in 2013 were no longer accessible when checked in late 2023, and that a quarter of all the pages it sampled across the decade were gone. It also found broken links inside pages that are meant to be reliable: 23% of news pages and 21% of government pages had at least one, and 54% of Wikipedia articles had a reference link pointing to a page that no longer exists.

Ten years earlier, a study by researchers at Harvard Law School, published in 2014, examined links cited in legal writing. It found that more than 70% of the links in the law journals it studied, and about half of the links in U.S. Supreme Court opinions, no longer led to the information originally cited, either because the page was gone or because it had changed.

The details vary by study and method, but the pattern is consistent: the older a link is, the less likely it still shows what it once did.

Why pages disappear

There’s rarely a single cause. The usual ones:

• Redesigns and new systems. When a site moves to a new platform, old addresses often aren’t carried over or redirected.
• Expired domains. If nobody renews a domain, the whole site goes, and someone else may later buy the name.
• Deliberate removal. Pages get taken down for legal reasons, because they’re out of date, or because an organisation would rather they weren’t found.
• Reorganisation. Content is merged, split or moved into new sections with new addresses.
• Ordinary edits. Pages are updated in place, and the old version is simply overwritten.
• Closures. Companies, projects and publications shut down, and their sites go with them.

How archives help

Web archives exist largely because of this problem. Services like the Wayback Machine and archive.today keep copies of pages as they were on a particular date, so a reference can still be followed after the original is gone or changed.

Archives can’t catch everything. They only have a page if someone or something saved it before it changed, and they can only see what a logged-out visitor sees. But for most public pages, there’s a good chance at least one copy exists.

If you run into a broken link, paste it into byepaywall and try the archives listed there. They open one after another from the same page. How to view the archived version of a page explains how to read what you find, including how to pick a good date in the Wayback Machine’s calendar.

Habits that keep your own references working

You can’t stop other people’s pages from changing, but you can make sure the version you relied on survives:

• Save a copy when you cite something, not later. A public archived copy gives you a stable link anyone can open. How to save a copy of a page shows how.
• Include the archived link alongside the original in references, reports and footnotes, with the date you accessed the page. How to cite an archived page covers the format.
• Prefer stable sources where they exist, such as published reports, document repositories and identifiers like DOIs for academic papers.
• Keep a private copy of anything you might need to prove later, such as terms you agreed to or a receipt, as a PDF on your own device.

The bigger picture

The web was built to be changeable, which is part of what makes it useful. The cost is that it has no memory unless someone keeps one. Archives do a lot of that work, largely as non-profit or volunteer efforts. If you rely on them, it’s worth knowing who runs them and supporting them where you can.

To learn why we built byepaywall around these services, read about us.