404, 410 and Soft 404: Handling Removed Pages Correctly

A 404 is the server's way of saying "there is nothing here". Two myths follow this code around in SEO: first, that 404s "penalise" a site; second, that redirecting all of them to the home page is the fix. Both are wrong.

The reality is simpler. A 404 is a normal, useful response. The real problem appears when the code and the content disagree — the page does not exist, yet the server returns 200. That is a soft 404, and it quietly eats your crawl budget.

What a 404 is and how Google behaves

404 (Not Found) is the best known member of the client error family (4xx). The server is saying the request was valid but there is nothing at that address.

Google's documentation describes the behaviour precisely:

• The content of URLs returning 4xx is not used.
• If the URL was previously indexed, it is removed from the index over time.
• Newly encountered 404 pages are not processed at all.
• The crawling frequency of that URL gradually decreases.

The crawl budget guide adds a direct recommendation: return 404 or 410 for permanently removed pages. The reasoning is practical — a 404 is a strong signal to the crawler not to visit that URL again, whereas blocked URLs linger in the crawl queue much longer.

Do 404s harm rankings

The short answer: not in themselves. A 404 is not a penalty mechanism — Google simply removes the URL from the index.

That does not mean you should ignore them. The real losses lie elsewhere:

• Lost traffic. If the page brought visitors from search, that traffic disappears.
• Lost external links. If other sites link to that page, the value of those links is wasted. This is the most common leak in backlink work.
• User experience. When links inside the site return 404, visitors leave.
• Crawl budget. At scale, 404s keep crawlers busy for nothing on large sites.

So the question is not "there are 404s, is that dangerous?" but "was there value behind this particular 404?"

404 versus 410

410 (Gone) means "this page has been permanently removed" — a more definitive statement than 404.

For SEO the practical difference is close to zero. Google's documentation groups the two together and the outcome is identical: the URL is removed from the index and crawl frequency drops. The crawl budget guide recommends either one.

Practical advice: use 410 where you are certain, 404 everywhere else. It is not worth rebuilding a system just to serve 410s — the outcome is almost the same.

Soft 404 — the fake page that returns 200

A soft 404 is a page that returns status code 200 while its content says "this does not exist". It is not a code the server sends; it is a label Google applies after comparing the code with the content.

Google's crawl budget guide for large sites states it plainly: "soft 404 pages will continue to be crawled, and waste your budget."

The usual sources:

• JavaScript-built sites. The server returns 200 for a route that does not exist and JavaScript paints a "page not found" message.
• Filtered catalogues. "No results found" pages served with 200, generating thousands of variations.
• Deleted products. A product page turns into "this item is no longer available" while the code stays 200.
• Empty categories. Category pages with no items in them.

The fix: align the code with reality. If the page does not exist, return 404 at the server level rather than merely changing the text.

Redirect or keep the 404

This is the most frequently asked practical question. The decision tree is simple:

When you do decide to redirect, follow the rules in the 301 redirect guide: direct, chain-free and relevant in content.

How to build a 404 page

A good 404 page keeps the visitor on the site. It should contain:

• A clear message. "The page you were looking for could not be found" — without technical jargon.
• A search box. So the visitor can type what they were after.
• Links to the main sections. Services, blog, contact.
• Popular or recent content. A few useful destinations.
• The same design as the rest of the site. It should feel like part of the site, not a bare server page.

The critical technical condition: this page must return a 404 status code. A beautifully designed 404 page that returns 200 is exactly what a soft 404 is.

To check:

curl -I https://example.com/this-page-does-not-exist
# expected: HTTP/2 404

How to find your 404s

The most valuable combination: Search Console plus server logs. The first gives you Google's view, the second the real traffic.

Which 404s are worth fixing

Fixing every 404 is neither possible nor necessary. In priority order:

1. URLs with external links. If a page that earned links from other sites was removed, that is a direct loss — 301 it to the closest match.

2. URLs still receiving traffic. Removed pages that still show impressions in analytics or Search Console.

3. URLs linked from within the site. These are the easiest fix: update the link and no redirect is needed at all.

4. URLs still in the sitemap. A removed page has no business being there.

Not worth fixing: invented URLs that never existed, addresses probed by spam bots (/wp-admin, /.env and the like), and old test pages. For those, 404 is precisely the right answer.

Common mistakes

1. Redirecting every 404 to the home page. The most repeated mistake of all. Google frequently classifies irrelevant redirects as soft 404s.

2. Serving the 404 page with a 200 code. Right design, wrong code — the result is a soft 404.

3. Keeping removed pages alive with noindex. The crawl budget guide advises against it: the page still has to be crawled and then discarded, which wastes crawling time. For a removed page, 404/410 is cleaner.

4. Never cleaning the sitemap. If removed URLs stay in it, Google keeps rechecking them.

5. Leaving broken internal links in place. Fixing the link is better than building a redirect around it.