Link rot, by the numbers: how fast the web forgets your links

Ask anyone who has run a website for more than a few years and they'll tell you the same thing: links die. Not all at once, and rarely with any warning — they just quietly stop working. A page moves, a domain lapses, a company gets acquired and folds its old URLs into a redirect that eventually gets switched off. The link that worked when you published it now goes nowhere.

This isn't a hunch. People have measured it, and the numbers are worse than most publishers assume.

A quarter of the web from ten years ago is already gone

The most-cited recent study comes from the Pew Research Center, which in 2024 looked at a large sample of web pages and the links inside them. They found that 38% of pages that existed in 2013 were no longer accessible a decade later. For pages from 2023 — just one year old at the time of the study — 8% had already vanished.

It compounds year on year. A link you added in 2019 has had five or six years to break. The academic literature calls this "link rot," and the numbers are remarkably consistent across studies:

| Source | What they measured | Rot rate | | --- | --- | --- | | Pew Research (2024) | Pages online in 2013, checked 2023 | 38% gone | | Harvard Law / Zittrain (2014) | Links in academic papers | ~50% dead | | Columbia Journalism (various) | Links in news articles | ~25% within a few years |

The exact figure moves around depending on what's being measured, but the direction never does. Left alone, a body of links loses something like a quarter to a half of its members over a decade. The web does not maintain itself.

Why links rot

There's no single cause. In practice it's a handful of failure modes that each chip away at your links:

• Domains lapse. A small retailer stops paying for its domain, or a brand consolidates and retires an old one. Every link pointing at it dies at once.
• Sites reorganise. A CMS migration, a redesign, or a move to a new platform renumbers every URL. Sometimes there are redirects. Often there aren't, or they get removed a year later to "clean things up."
• Products get discontinued. This one is specific to anyone linking to things for sale. A product page that returns a clean 404 — or worse, silently redirects to a category page or the homepage — is dead as far as your reader is concerned.
• Acquisitions and shutdowns. Companies get bought, merge, or close. The new owner rarely cares about preserving inbound links to the old catalogue.

None of these announce themselves. Your page keeps rendering. The link is still right there in the text, blue and underlined and completely broken.

Affiliate links rot faster than the rest

If you earn from affiliate links, you have a worse version of this problem, for two reasons.

First, you're linking to commerce, which churns more than the rest of the web. Product pages come and go constantly. A gadget that was in stock when you wrote the review is discontinued eighteen months later. The link doesn't 404 — it redirects to a "similar products" page, which looks fine and earns you nothing.

Second, and this is the part that catches people out: an affiliate link can fail while still loading perfectly. The classic case is a stripped tracking tag. A plugin update or a caching layer quietly removes the ?tag= parameter that connects a click to your account. The link still resolves. The product page still loads. The reader can still buy. You just don't get paid. Every ordinary check — does the link work? does the page open? — passes, while the money silently stops.

So the real affiliate rot rate is higher than the headline "38% of pages gone," because it includes a whole category of links that are technically alive but financially dead. We know because we found 25 of them on our own site.

What you can actually do about it

You can't stop the web from changing. What you can do is stop finding out months late.

The manual version is exactly what you'd expect: periodically open your highest-value pages, click every outbound link, and check that each one lands where it should — and, for affiliate links, that your tag survived the trip. It works. Nobody keeps it up, because it's tedious and the failures are invisible until you go looking.

The automated version is the whole reason LinkWatch exists: crawl your live pages on a schedule, follow every link the way a reader would, and flag the ones that break, redirect somewhere unexpected, or lose their tag. You find out in days, not at the end of a quarter when the earnings report finally looks wrong.

Link rot is not a risk you can eliminate. It's a maintenance cost you either pay deliberately, a little at a time, or pay all at once when you realise how much of your archive quietly stopped working.