The Real Reason Most Web Pages Don't Survive

Bookmarking a page on the internet is, statistically speaking, a short-term commitment. Jeff Huang recently found this out the hard way while sifting through his own saved links, discovering that a large share of the pages he'd marked for later had already rotted away. The Internet Archive captures much of what vanishes, but preservation can't rely on a single backstop.

Huang's own recommendations for making a page last are straightforward, if familiar: stick to plain HTML and CSS, avoid unnecessary minification, keep content consolidated on single pages, forbid hotlinking, limit yourself to a small set of web-safe fonts, compress images relentlessly, and never let a URL become a dead end.

Not All Advice Is Equal

Some of that list is easier to defend than others. The ban on hotlinking has clear practical value for long-term survival. The font advice, on the other hand, is harder to take seriously—web fonts fall back gracefully, so what you specify has little bearing on whether a page will still render a decade from now.

Even the calls for vanilla HTML and unminified source code only make sense when you look past the literal advice. Both suggestions are really arguments against complex build pipelines. Tools and processes age poorly: they stop compiling, dependencies break, and before long the source of a site is unusable. You don't need to abandon build tools—many developers couldn't work without them—but you should recognize them as a form of technical debt. That point was made well by Bastian Allgeier in his own piece on simplicity.

The Static Answer

If the question is purely technological—what architecture makes a site most likely to persist—then the answer is prerendering. Build everything into static files ahead of time, serve them from a host, and depend on nothing else. No server-side logic to patch, no third-party APIs to watch. It's a small, self-contained unit of web that's easy to keep alive.

The URL question is just as important. Few things kill a site's longevity faster than a URL reorg. A 404 is terminal, and there's no shortage of sites that sent every old address to the graveyard in a single redesign.

What Actually Keeps a Page Alive

But there's a deeper, non-technical ingredient that no amount of careful architecture can replace: ownership. A page lasts when someone has a reason to keep it alive. That means caring enough not to let URLs rot, being willing to fix what breaks, and having some ongoing benefit from the page's existence. Technology can help, but the incentive to maintain is what really decides whether a page will still be there next year—or a decade from now.