What Is a 404 Error? Meaning, Causes, and How to Fix It

August 23, 2026 · 7 min read · Broken Link Checker team

A 404 Not Found error is the HTTP status code a server returns when it received your request just fine — but has no page at the address you asked for. The connection worked, the server is healthy; the specific resource simply isn't there. That distinction matters: a 404 is not a server crash, not a network problem, and not always a mistake. It becomes a problem when real links — yours or other people's — point at it.

What 404 means, technically

Every HTTP response starts with a three-digit status code. The first digit tells you who's "at fault": 4xx codes mean the request couldn't be satisfied (client side), while 5xx codes mean the server failed. 404 is the 4xx code for "the thing you asked for doesn't exist here."

So when a browser shows a 404 page, this happened: DNS resolved the domain, a connection was made, the server parsed the request — and then looked up /that/exact/path and found nothing. Everything except the last step succeeded.

Myth check: 404 is not named after a room at CERN where the web's first server lived. The number is just HTTP's numbering scheme — 4 for client errors, 04 as the sequence within that class. (Room 404's neighbors, 403 Forbidden and 405 Method Not Allowed, would need their own rooms.)

404 vs 410 vs soft 404

Three related states get mixed up constantly:

404 Not Found — "nothing here." Deliberately noncommittal: maybe the page never existed, maybe it moved, maybe it'll be back. Search engines retry 404 URLs for a while before dropping them from the index.
410 Gone — "this existed and was removed on purpose." A stronger signal; crawlers tend to deindex 410 URLs faster. Use it when you genuinely retired content.
Soft 404 — the impostor. The page says "not found" (or is empty, or redirects everything to the homepage) but returns status 200 OK. Google detects the mismatch, labels it a soft 404 in Search Console, and link checkers that trust status codes will call the link "valid" while your visitors see an error. Soft 404s are worth fixing at the server level so the code tells the truth.

What actually causes 404 errors

Deleted content — a product retired, a post unpublished, a user profile removed. The links pointing at it stay behind.
Changed URLs without redirects — the most common cause on real sites. A slug edit (/blog/best-tools → /blog/best-tools-2026), a CMS migration, a new URL scheme — every old address 404s unless a redirect is set.
Typos in links — a missing letter, a duplicated https://, a trailing parenthesis or period swallowed from surrounding text when the link was pasted.
Case and trailing-slash mismatches — many servers treat /Page and /page as different URLs.
The other site died — for outbound links: domains expire, startups shut down, platforms fold. This is link rot, and it accounts for a huge share of broken links on older pages.

Why 404s matter (and when they don't)

A 404 URL sitting in isolation — nothing links to it, nobody visits it — is harmless. Google has said repeatedly that 404s are a normal part of the web and don't hurt the rest of your site. The cost appears when links point at the 404:

Visitors hit a wall. Someone clicked expecting content and got an error page. Many leave the site entirely.
Internal link paths break. Crawlers discover your pages by following links; a broken internal link severs that path and wastes crawl requests.
Inbound link value evaporates. If another site links to a URL that now 404s, the traffic and authority that link would pass go nowhere — until you add a redirect. (More on the SEO mechanics in do broken links hurt SEO?)

How to find 404s before your visitors do

You find them from two directions. On your pages: scan the links you publish. A browser extension like Broken Link Checker tests every link on the page you're viewing in one click and highlights the dead ones — handy right before hitting publish. Pointing at your site: Google Search Console's Indexing → Pages report lists URLs Google crawled and found missing, which catches broken links on other people's sites that you'd never see otherwise. A full comparison of the approaches is in how to find broken links.

How to fix a 404

Match the fix to the cause:

The link has a typo → correct the link where it's written. No redirect needed.
The content moved → set a 301 redirect from the old URL to the new one, and update your own internal links to point directly at the new address.
The content is gone for good → let it return 404 (or 410), and make the error page useful: a short apology, a search box, links to relevant sections. Don't blanket-redirect everything to the homepage — Google treats that as a soft 404, and visitors find it disorienting.
An external page you linked to died → replace the link with a live equivalent or an archived copy. Step-by-step options are in how to fix broken links.

FAQ

Does the number 404 come from a room at CERN?

No — that's a popular myth. The number simply follows HTTP's scheme: the first digit 4 marks a client-side error, and 04 is the sequence number assigned to "Not Found" in the specification.

Are 404 errors bad for SEO?

A 404 by itself doesn't penalize your site — Google calls them a normal part of the web. The damage is indirect: broken internal links waste crawl paths and frustrate visitors, and inbound links pointing at a 404 pass no value until you redirect them.

Should I redirect every 404 to my homepage?

No. Google treats blanket redirects to the homepage as soft 404s because the destination doesn't match what was requested. Redirect a dead URL to the closest relevant page, and let genuinely gone content return a real 404 or 410 with a helpful error page.