404 errors are the response that tells you the server could not find the requested page, and they appear on every site sooner or later. Most site owners panic on seeing a 404 and think "my SEO is broken"; yet the situation is usually much calmer than that. What really needs attention is important pages mistakenly returning a 404, and pages that should return a 404 instead returning 200, creating a soft 404. In this article we explain which 404s are a problem, which are not, and how to handle them.
What is a 404 error?
A 404 (Not Found) is the HTTP status code that tells you the server understood the request but could not find a resource at that address. The causes are usually these:
- The page was deleted or its address (slug) changed, but no redirect was added.
- An internal or external link address was typed incorrectly.
- Some pages were not matched during a site migration.
- The user typed the address incorrectly by hand.
Do 404 errors hurt SEO?
Google's official position is clear: returning a 404 for pages that do not exist is normal and does not harm your site's overall ranking performance. It is expected behavior for a deleted page to eventually drop out of the index. So seeing a few hundred "Not found (404)" entries in Search Console is not, on its own, a penalty signal.
The problem starts in these situations:
- If a page that receives traffic or links returns a 404, that value is lost.
- If addresses listed in the sitemap or shown through internal links return a 404, Google wastes time and users become frustrated.
- If important pages that were working by mistake (products, categories, services) start returning a 404, you lose traffic directly.
Important distinction: The goal is not to try to reset every 404. The goal is to find the 404s that should not exist: addresses that carry value, receive links, or are displayed from within your own site.
What is a soft 404?
A soft 404 is when a page shows the user a "not found" message or empty content, yet the server still returns a 200 (success) code. Google notices this and keeps the page out of the index, but because the signal is confusing, it also consumes your crawl budget. Typical examples:
- Your custom 404 page says "Page not found", but the status code is 200.
- A category with emptied content, or a search results page that says "no results found", returns 200.
- Thousands of deleted pages are redirected in bulk to the home page.
- Thin pages of a few sentences that carry no meaning.
To check on your own site, act like a visitor: type a non-existent address and read the response code in the Network tab of your browser's developer tools; if you see 200 instead of 404, this is the issue. The "Pages" report in Search Console also lists these cases under the Soft 404 heading. For the other statuses in that report, see our Google indexing issues guide.
404, 410 or redirect: which one?
| Situation | Correct response | Explanation |
|---|---|---|
| Page was deleted by mistake | Restore it (200) | Best is to republish the content at its old address |
| Address changed, same content in a new place | 301 redirect | Exact mapping from old address to new; you can find the differences in our 301 and 302 article |
| Content removed, a very close alternative exists | 301 redirect | Redirect only to a genuinely relevant page |
| Content removed, no equivalent exists | 404 or 410 | This is the honest response; do not redirect everything in bulk to the home page |
| Content deliberately and permanently removed | 410 (Gone) | Says "it no longer exists"; in practice Google's handling is very close to 404 |
The practical difference between 404 and 410 is small; both tell Google the page no longer exists. A 410 expresses it a little more definitively. The main point to watch is returning the right code and not blurring the signal with irrelevant redirects.
How do you find 404 errors?
- Search Console → Pages: Open the "Not found (404)" and "Soft 404" reasons, and check the sample addresses and the "first detected" date.
- Crawling tool: A desktop crawling tool walks through your site and lists the internal links that return 404 together with the page they appear on. This is the most useful output for fixing.
- Server logs: They show which addresses get the most 404s and which referrers they come from; this is where you catch incorrectly typed external links.
- Analytics: In GA4, pageviews filtered by the title of your custom 404 page reveal which addresses show an error to users.
- Backlink reports: They find addresses that receive links but return a 404; the most valuable fixes are often here. For the value of links, see our backlink guide.
Gather the list you find into a single table: address, source (internal link, external link, sitemap), the number of visits or links it receives, and your decision (restore, 301, leave as 404). That way it does not get lost among hundreds of addresses, and it is clear within the team who is doing what.
Fixing steps
- Prioritize the list. Put addresses that receive external links, brought traffic, or appear in the sitemap at the top.
- Fix internal links. Every link from within your own site that leads to a 404 is a fixing opportunity; point it to the new correct address or remove the link. Our internal linking strategy article is a good guide for this.
- Redirect those with a counterpart using a 301. Map the old address directly to the most relevant new address.
- Leave those with no counterpart. Pages that were genuinely deleted and have no equivalent should remain 404 (or 410).
- Clean up the sitemap. Addresses returning 404 should be removed from the sitemap.
- Fix soft 404s. Either fill empty pages with real content, or change their response to a 404 code.
- Verify in Search Console. Use the "Validate fix" option on the issue group you corrected.
Practical scenarios
The same 404 code requires different decisions in different situations. Three common examples:
- Sold-out product page (e-commerce): If the product is temporarily out of stock, keep the page open with a 200, showing the "out of stock" information and similar products. If the product has been permanently removed and a very close new model exists, a 301 is appropriate; otherwise a 404/410 fits.
- Old blog post: Updating and strengthening an outdated post is often more rewarding than deleting it. If you must delete it, give a 301 to the relevant current post.
- Campaign page: If a finished campaign page recurs every year, keep the address and have the page ready for the next period. If it was a one-off, redirecting to a similar category or returning a 404 makes sense.
The common thread in these examples: when deciding, ask "is this address still receiving visitors, links, or search demand?" If the answer is yes, offer users something useful instead of abandoning the address; if no, a 404 is not a problem. Another practical tip: when publishing a new page, think about the address from the start, because changing it later always creates a redirect obligation.
What should a good 404 page look like?
A 404 page that keeps users from leaving is more about user experience than SEO, but it indirectly protects your site. Pay attention to the following:
- The status code must be 404 (or 410); no matter how nice the page text is, it must not be a 200.
- Keep your site's design and top menu so visitors understand they are still on the site.
- Write a short, clear explanation: "The page you are looking for could not be found. It may have been moved or deleted."
- Provide a search box, a link to the home page, and a few links to popular pages.
- Do not automatically send visitors who land on the page to the home page; this can be counted as a soft 404.
Preventing new 404s
Preventing new ones is just as important as fixing existing problems:
- Before changing a page's address, prepare the redirect and publish it at the same time as the change.
- Before deleting content, check that page's external link and traffic value.
- On large sites, add a "broken link scan" step to the publishing process; after a theme, CMS or site migration, crawl the entire site again.
- On multilingual sites, also check the reciprocal links between language versions; see our hreflang guide for details.
- Monitor the trend of the 404 count and soft 404s in Search Console each month; a sudden increase usually points to a publishing error.
Finally, also look at which device or browser the 404s appear on. Errors produced only by a specific template on mobile point to a problem caused by the theme or a plugin, and a single fix can repair hundreds of addresses at once.
Frequently asked questions
Do I need to fix the 404s in Search Console one by one?
No. It is normal for deleted pages with no value to return a 404. Focus, in order of importance, on addresses that carry value and are shown through internal links.
Is the value of links pointing to a 404 page lost completely?
If a link goes to that page and the page does not exist, the value is lost at that point. A 301 to a similar page largely recovers that value.
Should I set my 404 page to noindex?
There is no need. The correct status code (404) is enough; Google does not index these pages anyway.
How long does Search Console take to update after I fix an error?
The report updates as Google recrawls the addresses; this can take anywhere from a few days to a few weeks.
Conclusion
404s are not a penalty but a natural part of a site; what matters is distinguishing which ones create a real loss. Restore the addresses that carry value or redirect them with a 301, honestly return a 404/410 for the rest, eliminate soft 404s, and keep internal links clean. To make this routine part of your technical audit, you can use our technical SEO checklist.