A 404 error means the server could not find the page at that URL. For SEO, a 404 is normal and does not penalize your site. Google drops the URL from its index and crawls it less over time. You only need to act when a page that should exist returns a 404.
What does a 404 error mean?
Every request to a web server gets a three-digit HTTP status code back. A 200 means "here is the page". A 404 means "nothing lives at this address". The visitor usually sees a "page not found" message, but the status code is what a search crawler reads.
A 404 says nothing about why the page is missing. It might have been deleted on purpose, moved without a redirect, or never existed because someone mistyped a link. The code only reports the result.
Do 404 errors hurt SEO?
Not by themselves. Google has said this plainly for years. Its Search team wrote that 404s on URLs that do not exist on your site do not hurt you or count against you in search results. That post is old, but the current crawler documentation still describes the same behavior.
That documentation says Google treats all 4xx errors except 429 the same way. In practice this means:
- A URL that is already indexed and starts returning 404 is removed from the index.
- A newly discovered 404 URL is not processed.
- Google crawls the URL less often as time passes.
- Any content received from a 404 URL is ignored.
So the real cost of a 404 is not a penalty. It is the loss of the page itself. If a page that ranked or earned links now returns 404, the traffic that page brought in goes with it.
Which 404s matter and which can you ignore?
A site of any age collects 404s. Most of them are harmless. Sort them by one question: did this URL ever matter?
| Kind of 404 | Example | What to do |
|---|---|---|
| Page you deleted on purpose, no replacement | A discontinued service page | Leave it as a 404, or return 410 |
| Page that moved | A blog post under a new slug | 301 redirect to the new URL |
| Page that should exist but is broken | A template error took down a product page | Fix the page and restore the URL |
| Typo in someone else's link | /blog/seo-gude | 301 redirect only if the link comes from a source worth keeping |
| URL that never existed | Random strings from bots and scrapers | Ignore |
| Broken internal link | Your own menu points to a deleted page | Update the link |
Google's advice matches this table. It recommends investigating 404s on pages you care about, redirecting moved content with a 301, and returning a 404 or 410 when you remove content with no replacement.
How do you find 404 errors on your site?
- Search Console. Open the Page indexing report and look for "Not found (404)" in the list of reasons pages are not indexed. Click it to see the affected URLs.
- Your own links. Crawl your site with any link checker and list every internal link that returns a 404. These are the ones you fully control.
- Server logs or analytics. Look for 404 responses that get repeat visits. A URL hit fifty times is worth a redirect. A URL hit once by a bot is not.
- Your sitemap. A sitemap should list only live pages. If a listed URL returns 404, remove it. Our guide on submitting a sitemap to Google covers keeping that file clean.
How do you fix a 404 error?
Work through the cases from the table above.
If the page should exist
Restore it. Check whether a deploy, a CMS change or a slug edit broke the URL. If the content moved, put it back or redirect the old address.
If the page moved
Use a 301 redirect from the old URL to the new one. Keep it one hop. Chains of redirects slow crawling, and we cover that in how to fix redirect chains. If you are unsure between a 301 and a 302, this comparison explains when each fits.
If the page is gone for good
Leave the 404 in place, or return a 410 (Gone). Google's documentation groups the two and does not describe a separate effect for 410, so do not spend time agonizing over the choice. Then remove internal links that point to it.
If your error page returns 200
This is the one mistake that does cause trouble. A custom "not found" page that sends a 200 status is a soft 404. Google may flag it and leave pages out of the index. See what soft 404 errors are and how to fix them.
Should you redirect every 404 to the home page?
No. Google's advice is to avoid redirecting unknown URLs to your home page because it can confuse how Google understands your site. A visitor who clicked a link to a specific article and lands on your home page is lost, and Google may treat the redirect as a soft 404.
Redirect a missing page to the closest relevant page. If there is no close match, a 404 is the honest answer.
What should a good 404 page do?
Your error page cannot rescue the SEO of a deleted URL, but it can keep a visitor on your site. Keep it simple:
- Say plainly that the page was not found.
- Link to your home page, your main sections and your search box.
- Return a real 404 status code, not a 200.
- Keep the site header so people can navigate.
Pages with no internal links pointing to them are a related problem, because crawlers may never find them. See what an orphan page is if a page seems to be missing for that reason.
Frequently Asked Questions
Is a 404 error bad for SEO?
No, not by itself. Google removes a 404 URL from its index and crawls it less, but it does not mark your whole site down. The loss is the page, not a penalty.
How long does Google keep a 404 page in its index?
Google says an indexed URL that starts returning a 4xx status is removed from the index. It does not give a fixed number of days, so check the URL Inspection tool if you want to confirm.
Should I use 404 or 410?
Either works for content you removed on purpose. Google's documentation handles both the same way, so pick whichever your server makes easier.
What is the difference between a 404 and a soft 404?
A 404 sends the correct status code. A soft 404 shows a "not found" message but sends 200, which tells Google the page is fine. Fix the status code so the two agree.
Do I need to fix every 404 in Search Console?
No. Fix the ones on pages that should exist, pages with traffic or links, and your own broken internal links. Ignore URLs that never existed.