Are 404 errors bad for SEO? The short answer is no, not in the way most audit tools imply. A crawler report showing hundreds of 404s looks alarming, and the instinct is to redirect every one of them to the homepage. That instinct usually makes things worse. A 404 is a status code doing its job: telling Google a URL has nothing on it. What matters is which URLs return it, and whether they should.
What Google actually does with a 404
Google’s documentation on how HTTP status codes, and network and DNS errors affect Google Search is plain about it. URLs that are already indexed and return a 4xx status code “are removed from the index”, and “the crawling frequency gradually decreases.”
Read those two sentences carefully, because they describe a contained effect. The URL that returns 404 drops out. Google checks it less often. Nothing in that description says the rest of the site is penalised, demoted or crawled less. A site that deletes old, thin or expired pages and lets them 404 is behaving exactly as Google expects.
That is why the Page indexing report in Search Console lists “Not found (404)” as a reason a page is not indexed rather than as a problem to fix. Most of the URLs in that bucket are old, mistyped or never existed. Our guide to using Google Search Console covers where to find the report and how to read it.
How status codes compare
The 404 is easier to judge when you set it beside the other responses Google documents. The difference between them is mostly about what happens to URLs already in the index.
| Response | What Google does | When it is the right answer |
|---|---|---|
| 404 / 410 (4xx) | Indexed URLs are removed; crawling of them gradually decreases | The page is gone and has no close replacement |
| Soft 404 | Success code, but content suggests an error; flagged as a soft 404 in Search Console | Never. Fix it to a real 404, real content or a redirect |
| 5xx and 429 | Crawlers temporarily slow down; indexed URLs are preserved, but eventually dropped | Only as a short-lived server problem |
The row that surprises people is the third. A server error is treated more gently at first than a 404, because Google assumes it might be temporary. But if it persists, the pages go anyway. A site throwing intermittent 500s is in a worse position than one with a tidy set of 404s, even if the crawl report makes the 404s look bigger.
Soft 404s are the problem worth fixing
A soft 404 is a page that returns a success code while the content suggests an error. Google’s examples are an empty page or an error message. Search Console flags these separately. They are worse than honest 404s because the server is telling Google one thing and the page another, and the URL sits in a grey zone until someone notices.
Common sources are internal search pages with no results, category pages with every product out of stock, and custom “page not found” templates that forget to send a 404 header. Our guide to crawl budget covers why large sites in particular should care: soft 404 URLs keep being crawled and waste requests that could go to pages you want indexed.
The fix depends on what the page should be. Google’s guidance gives three options:
- Return a real 404 or 410 if the content is gone and is not coming back.
- Add real content if the page should exist but is currently thin or empty.
- Redirect to a relevant page if a genuine replacement exists.
Note what is missing from that list: redirecting everything to the homepage. A redirect to a page that has nothing to do with the old one is not a relevant replacement, and it hides the real signal from your own reports, so you can no longer see which content disappeared or why.
When 404 errors are bad for SEO
The cost of a 404 is not the status code. It is what was pointing at the URL. Three situations are worth acting on.
The page has backlinks. External links to a URL that now 404s pass nothing to your site. If the content moved or has a close equivalent, a 301 recovers that value. Our guide to 301 vs 302 redirects covers which to use, and link reclamation covers finding the broken links in the first place.
The page still gets traffic. If people are arriving from search, bookmarks or old emails, a dead end loses them. Check analytics for landing pages that return 404 before deciding they are unimportant.
Your own internal links point at it. This one is entirely within your control. Internal links to 404s waste crawl requests and send users to dead ends. Fix the link at the source rather than adding a redirect to patch it.
For pages that are gone and have none of the above, leave the 404 alone. A useful custom 404 template with search and links to main sections helps the people who land there, as long as it still returns the 404 status code.
A triage order for a 404 report
When a crawl or Search Console export lands with a long list of not-found URLs, work through it in this order rather than top to bottom.
- Filter to URLs with external backlinks. Redirect those with a relevant destination.
- Filter to URLs with recent landing-page traffic. Restore or redirect them.
- Find internal links pointing at 404s and update them.
- Check for soft 404s in Search Console and fix each by type.
- Leave the remainder. Old, mistyped and junk URLs returning 404 are fine.
After a site migration or redesign, run this list before launch as well as after. Most avoidable 404 damage comes from URL changes made without a redirect map. Our guide to website redesign SEO covers building one.
The limitation is that Google’s documentation describes behaviour, not timing. It says crawling of 404 URLs “gradually decreases” and that 5xx URLs are “eventually dropped”, without saying how gradual or how eventual. You will see 404 URLs in reports for months after deletion, and that is not evidence of a problem. It is Google rechecking a URL it once knew about.
The practical rule is to judge each 404 by what it is losing you. If the answer is nothing, the 404 is the correct response and the report entry is just a record.
Frequently asked questions
Do 404 errors hurt SEO for the whole site?
Google’s documentation says indexed URLs returning 4xx are removed and crawled less often. It does not describe any site-wide penalty for having 404s.
What is a soft 404 error in SEO?
A page that returns a success code while its content suggests an error, such as an empty page or an error message. Search Console reports it as a soft 404.
How do I fix a soft 404 in Google Search Console?
Return a real 404 or 410 if the page is gone, add real content if it should exist, or redirect it to a genuinely relevant page.
Should I redirect every 404 to the homepage?
No. Google’s guidance is to redirect to a relevant page. Where none exists, a real 404 or 410 is the honest answer.
The takeaway A 404 on a dead page is correct. Spend your time on 404s with backlinks or traffic, broken internal links, soft 404s and persistent server errors.
