What Is 404 Error?

How a 404 Error Works

Every time a browser or crawler requests a URL, the web server answers with an HTTP status code in the response header. That code, not the visible page, is the machine-readable verdict on what happened. A 404 means the server understood the request perfectly well but holds no resource that matches the URL. In the words of MDN, the 404 Not Found status code is “a client error response status code [that] indicates that the server cannot find the requested resource.” It is standardized in RFC 9110, the current specification for HTTP semantics.

The defining trait of a 404 is what it leaves unsaid. It reports that a resource is missing but makes no claim about why or for how long. MDN is explicit: “A 404 status code only indicates that the resource is missing without indicating if this is temporary or permanent. If a resource is permanently removed, servers should send the 410 Gone status instead.” So a 404 is the correct answer when the server genuinely doesn’t know — a mistyped URL, a page that may return, a resource that never existed.

What matters for search is that the status code travels in the header, invisible to the human reading the page. A well-designed site pairs the 404 code with a friendly HTML page (“Sorry, we couldn’t find that”), but the friendly page is decoration. Software — including Googlebot — keys off the numeric code. This is the seam where soft 404s are born, which we return to below.

How Google Handles a 404

For crawling and indexing, Google’s behavior is documented and blunt. Its guidance states that “all 4xx errors, except 429, are treated the same: Google crawlers inform the next processing system that the content doesn’t exist.” The downstream effect: “the indexing pipeline removes the URL from the index if it was previously indexed. Newly encountered 404 pages aren’t processed. The crawling frequency gradually decreases.”

Two practical consequences follow. First, a 404 is not a penalty. Google drops that one URL and eventually stops re-requesting it, but it does not dock the rest of your site for having dead pages — 404s are a normal, expected part of any site that changes over time. Second, because Google treats 404 and 410 almost identically for crawling, the choice between them is about semantics and speed of cleanup, not about avoiding harm. If you know a page is gone for good, a 410 is the more honest signal; if you’re unsure, 404 is right.

Hard 404 vs Soft 404

There are two ways a “not found” experience can be delivered, and only one is a true 404:

• A hard (real) 404 returns the 404 status code in the HTTP header. The header and the page agree: this URL has no content.
• A soft 404 returns a success code — usually 200 OK — while the page body says the content is missing. The header and the page contradict each other. Google flags these because it can no longer trust the status code to tell it what to index.

The goal is always the hard 404. A soft 404 wastes crawl budget on a page Google has to evaluate, then guess about, rather than drop.

Example of a 404 Error

Consider the lifecycle of a deleted product page — a clearly illustrative walkthrough built on the documented behavior in Google’s and MDN’s references. A store removes /products/old-widget. A visitor who bookmarked it makes a request, and the server replies with a status line reading HTTP/1.1 404 Not Found. MDN’s own reference response for this code even attaches caching headers such as Cache-Control: max-age=604800, showing that a 404 is a cacheable response a server can legitimately ask clients to remember for a week.

Now Googlebot re-requests the same URL during a routine crawl. It receives the 404, and per Google’s documented pipeline it “inform[s] the next processing system that the content doesn’t exist.” Because /products/old-widget was previously indexed, the indexing pipeline removes it from the index. Over subsequent crawls, Google requests that URL less and less often — the “crawling frequency gradually decreases” — until it effectively stops checking a page that keeps answering “gone.”

The point of the example is what does not happen. The rest of the store’s catalog is untouched. There is no sitewide demotion, no manual action, no lingering ghost — provided the page returns an honest 404. The only real cleanup left is internal: any menu link, related-products widget, or sitemap entry still pointing at /products/old-widget should be updated or removed, because those broken links waste crawl requests and send users into a wall. The status code did its job; the maintenance is on the links that lead to it.

404 Error vs Soft 404

A 404 is a signal working as designed; a soft 404 is that same signal broken. The fix for a soft 404 is almost always to make the server return a real 404 or 410 for the missing URL — see the full write-up on soft 404s.

Frequently Asked Questions

The Bottom Line

A 404 error is the web’s standard way of saying "there is nothing here." It is a client-error status code that a server sends when a URL points to no resource, and it deliberately leaves open whether the gap is temporary. Treated correctly, it is healthy plumbing — the honest answer that lets an index drop a dead page. The failures around 404 come not from the code but from hiding it: masking it behind a 200 or a homepage redirect turns a clean signal into a soft 404.