Google Search Console 101: Fixing Not Found 404 Errors
- Harold Bell
- 3 days ago
- 13 min read
Key takeaways
|
In more than 16 years of B2B content and demand generation work, no Search Console status has kicked off more pointless cleanup projects than Not found (404).
A team sees a four-figure number next to it in the Page indexing report and assumes the site is falling apart. It usually isn't. Google expects 404s. They are a normal, healthy part of how the web works, and the overwhelming majority of URLs in that report need nothing from you.
The problem is the handful that do. Buried in that list are the URLs that used to rank, the ones with external links pointing at them, and the ones an answer engine has already cited. Those are quietly costing you traffic and authority while you spend the afternoon redirecting scraper junk to your homepage.
Here's how to tell them apart.
What does not found 404 mean in Google Search Console
Not found (404) means Googlebot requested a URL on your site and your server replied that the URL does not exist. Google records it in the Page indexing report under the reasons pages are not indexed. It is a status, not a penalty. Google does not index 404 URLs because there is nothing on them to index. |
You'll find it in Search Console under Indexing, then Pages, then scrolling to the section listing why pages aren't indexed. Click the row and you get the affected URLs plus an export.
Before you touch anything, understand that there are two different 404 statuses in that report and they do not mean the same thing. If you want the wider tour of the report, start with using Google Search Console for SEO and the URL Inspection tool.
Not found (404) means Google discovered the URL on its own, through an internal link, an external link, or an old crawl record, and the URL returns a 404. Google found it. You never promised it existed.
Submitted URL not found (404) means the URL is sitting in a sitemap you handed to Google. You told Google the page exists and matters, and then it didn't. That one is always worth fixing, because it is a direct contradiction between what you claimed and what your server does.
The second status is entirely within your control, which makes it the faster win.
Why does Google report 404s you never created
Most 404s in your report were never pages you published. Google accumulates URLs from external sites with broken or truncated links, from scrapers, from old crawl records, and from JavaScript that generated paths which never resolved. It keeps re-checking them long after they stop existing. |
The usual sources, in roughly the order I find them:
A page was deleted on purpose and nothing was put in its place
A URL slug changed and no redirect was created
Another site linked to you with a typo, a truncated URL, or trailing punctuation glued onto the end
A sitemap still lists URLs that were removed months ago
Your own internal links point at an old path
Scrapers and bots requested URLs that never existed, and Google picked up the pattern
A platform migration changed the URL structure without a redirect map
On this site I renamed a blog post slug because it was colliding with a service page path. The rename was correct. The old URL started returning 404 within a day and appeared in the Page indexing report the following week. That is the system working exactly as designed. What mattered was not the 404 showing up. It was whether I redirected the old URL to the new one before the equity attached to it evaporated.
Do 404 errors hurt SEO
Not directly. Google has been consistent for years that 404s do not trigger a penalty and do not damage the rankings of your other pages. What costs you is what a specific 404 destroys: link equity from external sites pointing at that URL, an entry point that used to convert, and crawl budget spent re-checking dead paths at scale. |
This distinction matters because it changes what you do on Monday morning. If 404s were a penalty, the right response would be to eliminate every one of them. They are not, so the right response is triage.
Three things genuinely cost you money:
Lost link equity. If another site links to a URL that now returns 404, that authority stops flowing anywhere. A 301 to the closest equivalent recovers most of it. Doing nothing recovers none of it. Pull the list from checking backlinks in Google Search Console before you decide what to abandon, and read it alongside your backlink strategy.
Lost entry points. A URL that used to bring in qualified traffic still has demand attached to it. Old search results, bookmarks, newsletters and Slack messages keep sending people there, and they arrive at nothing. That is a conversion problem before it is an SEO problem.
Crawl waste at scale. A few hundred 404s is noise. Tens of thousands, usually from a broken template or a bad migration, is a real crawl efficiency problem on a large site. Google spends requests confirming URLs are still dead instead of finding your new content, which is the same underlying constraint behind discovered currently not indexed.
None of that is a penalty. All of it is leakage.
How do you decide which 404s to fix
Fix a 404 when the URL earned something you would lose by abandoning it. If it has external links, historical traffic, or a live equivalent elsewhere on the site, redirect it. If it was never a real page, or the content is genuinely gone with no replacement, leave it. A 404 is the honest answer to a request for something that does not exist. |
Run every URL in the export through this table. It takes about an hour for a few hundred URLs and it is the whole job.
What you are looking at | What it means | What to do |
Had backlinks or historical clicks, and a close equivalent exists | Real value, recoverable | 301 to the closest equivalent page |
Had backlinks or clicks, but no equivalent content exists | Real value, nowhere to send it | Rebuild the page, or 301 to the most relevant hub |
Deleted on purpose, no links, no traffic | Working as intended | Leave the 404, or serve 410 |
Sitting in your sitemap and returning 404 | Your own contradiction | Remove from the sitemap in this pass |
Typo, truncation, or scraper artifact | Was never real | Ignore, unless the broken link sits on a site worth contacting |
Your own internal links point at it | Self inflicted | Fix the internal links first, then decide on the URL |
How do you fix not found 404 errors
Export the URL list, score each URL for historical traffic and external links, then 301 the ones with real value to their closest live equivalent. Remove dead URLs from your sitemap, repoint your own internal links, and use Validate Fix to trigger a recheck. URLs that were never real pages need no action at all. |
Export the full URL list. Open Not found (404) in the Page indexing report and export rather than working from the sample view on screen.
Score each URL for value. Cross-reference against Search Console performance data for historical clicks, and against the links report for external links pointing at the dead URL.
Map redirects one to one. Each dead URL goes to its closest live equivalent. Not a category page, not your homepage. If nothing close exists, that is a signal to leave the 404 rather than force a match.
Implement the redirects. On Wix that's Marketing and SEO, then SEO Tools, then the URL Redirect Manager. Every platform has an equivalent. Use 301 for permanent moves.
Clean the sitemap. Any URL returning 404 comes out of the sitemap in the same pass, so you stop contradicting yourself.
Repoint your own internal links. A redirect handles the outside world. Your own links should point at the live URL directly with no hop.
Validate the fix. Click Validate Fix in the report to queue a recheck, and spot-check individual URLs with URL Inspection. Expect days to weeks, not hours.
Repeat in Bing. Bing runs its own crawler and its own index, so cleaning up in Google does nothing for Bing.
One sequencing rule, because it is the mistake I see most often: get the redirect and the internal links live before you ask for a recrawl. The crawl should arrive to find everything already in place. Requesting indexing first just books an appointment for the engine to confirm the problem still exists. The same principle governs internal linking strategy generally, and it is why getting your website indexed by Google and Bing is a sequencing discipline rather than a button.
When should you use 410 instead of 404
Use 410 Gone when you have deliberately removed content and want the engine to stop asking. A 404 says the URL is not here. A 410 says it is not here and is not coming back. Google treats both as removal signals and drops 410s marginally faster, but the difference is small enough that it is not worth re-architecting your stack for. |
410 earns its keep in three situations: bulk removal of programmatic pages, expired listings such as jobs, events and inventory, and retired products where you are certain nothing is returning. If there is any chance the URL comes back, use 404 and keep your options open.
On most Wix, HubSpot and Webflow sites you cannot serve a 410 without custom work, and that is fine. The gap between the two statuses is not worth an engineering ticket. Spend the effort on the redirect map instead, then push the changes through IndexNow so Bing hears about them immediately.
What is a soft 404 and why is it worse
A soft 404 is a URL that returns a 200 OK status while the page has no real content, or literally tells the visitor the page was not found. It is worse than a hard 404 because the status code lies. The engine cannot tell the page is dead, so it keeps crawling it, may index it, and thin URLs pile up in your index and your crawl budget. |
The usual causes:
A custom 404 page that returns 200 instead of 404, which is a template configuration problem rather than a content problem
Empty search results, empty filter pages, and category pages with nothing in them
A failed database call or a missing include that renders the template shell with no content inside it
Mass redirecting removed URLs to the homepage. Google treats a redirect to an irrelevant page as a soft 404, which is exactly why the homepage catch-all is a bad habit rather than a shortcut
Fixing it means one of three things: return the correct status code, put real content on the page, or redirect somewhere genuinely relevant.
If you only have budget to chase one status this quarter, chase soft 404 rather than Not found (404). Not found is usually the system working. Soft 404 is always something misconfigured, and it sits in the same family of problems as crawled currently not indexed, where the page technically resolves and still earns nothing.
How should your 404 page be built
A good 404 page returns a genuine 404 status code, says plainly that the page is gone, and gives the visitor somewhere useful to go. Full site navigation, a search box, links to your best content, and your primary call to action. What it must never do is return 200, auto-redirect to the homepage, or dead end. |
Keep your header and navigation. A 404 without navigation sends the visitor back to search results, which is the outcome you are trying to avoid
Add site search and three or four links to your highest-value pages
Keep your brand voice. A dead end is the wrong place to sound like a server log
Do not add a noindex tag. The 404 status code already handles exclusion, and noindex on a page that returns 404 does nothing
Do not put the 404 page in your sitemap
Track it in analytics with the referrer captured. The referrer data from your 404 page is the fastest way to find broken internal links on your own site
That last point is worth more than it sounds. Search Console tells you which URLs 404. Your own analytics tells you which of them people are actually hitting, which is a completely different and much shorter list. That is the list you fix first, and it usually overlaps with whatever came out of your last SEO gap analysis.
Does a 404 hurt your AI search visibility
Yes, and this is the version of the problem that actually costs you. Answer engines cite URLs. When a cited URL starts returning 404, the citation breaks. The link in the AI answer goes nowhere, the retrieval layer drops the page on its next crawl, and the authority that page accumulated transfers nowhere unless you redirect it. |
Across SEO, AEO, GEO and LLMO, a 404 is a Citation lever failure caused by a Distribution problem. The content was good enough to get cited. The URL stopped resolving. Nothing about the content quality changed, which is why teams working on AI search visibility keep rewriting pages when the actual defect is a redirect that was never built.
There is a second-order problem too. Answer engines do not re-crawl on your schedule. An AI answer can keep pointing at a dead URL for weeks after you delete it, so every person who follows that citation lands on nothing and forms the impression that your links do not work. That is the opposite of what you were trying to build when you set out to get cited by AI in the first place, and it undoes the structural work behind answer engine optimization.
Bing matters more here than most teams realize, because ChatGPT search retrieval leans heavily on Bing's index. If a URL is dead in Bing and you only cleaned up in Google, you have not fixed it where it counts. Run the same pass through the Bing Webmaster Tools guide and confirm the redirect landed in Bing indexing.
So here is my rule, and it is not negotiable on client sites: never delete a page that has ever been cited. Redirect it. The redirect preserves the path for humans following an older citation and passes the accumulated authority to whatever replaced it. Deletion throws both away for the sake of a tidier sitemap.
The bottom line
Most of what is sitting in your Not found (404) report is fine and always was. Some of it is not, and the difference is worth about an hour with a spreadsheet and the triage table above. The URLs that had links, the URLs that had traffic, and the URLs that had citations are the entire job. Everything else is noise you are allowed to ignore, which frees you up to spend the time on optimizing content for SEO instead.
If you would rather have someone run the triage, map the redirects, and repair the citation paths across both indexes, book 30 minutes with MQL Magnet.
Frequently asked questions
What does not found 404 mean in Google Search Console?
It means Googlebot requested a URL on your site and your server responded that the URL does not exist. Google lists it in the Page indexing report under the reasons pages are not indexed. It is a status, not a penalty. Google does not index 404 URLs because there is no content on them to index.
Do 404 errors hurt SEO?
Not directly. Google has been consistent that 404s are a normal part of the web and do not trigger a ranking penalty or damage your other pages. The cost is indirect: link equity from external sites pointing at the dead URL stops flowing, an entry point that used to convert disappears, and at large scale Google wastes crawl requests re-checking dead paths.
How do I fix not found 404 errors?
Export the URL list from the Page indexing report, score each URL for historical traffic and external links, then 301 redirect the ones with real value to their closest live equivalent. Remove dead URLs from your sitemap, repoint your own internal links, and use Validate Fix to trigger a recheck. URLs that were never real pages need no action at all.
What is the difference between not found 404 and submitted URL not found 404?
Not found (404) means Google discovered the URL on its own, through an internal link, an external link, or an old crawl record. Submitted URL not found (404) means the URL is in a sitemap you gave Google. The second one is always worth fixing because you told Google the page exists and your server contradicts you.
Should I redirect all my 404 pages to the homepage?
No. Google treats a redirect to an irrelevant page as a soft 404, so a homepage catch-all converts a clean status into a misconfiguration. Redirect only to a genuinely equivalent page. If no close equivalent exists, letting the URL return 404 is the honest and correct answer.
What is a soft 404?
A soft 404 is a URL that returns a 200 OK status code while the page has no real content or tells the visitor the page was not found. The status code lies, so the engine keeps crawling and may index a page that should have been removed. Soft 404 is a genuine misconfiguration and deserves attention before Not found (404) does.
When should I use a 410 instead of a 404?
Use 410 Gone when you have deliberately removed content and are certain it is not coming back, such as expired listings, retired products, or bulk-removed programmatic pages. A 404 says the URL is not here. A 410 says it is not here and will not return. Google drops 410s marginally faster, but the practical difference is small.
How long does it take for 404 errors to disappear from Search Console?
Expect days to weeks rather than hours. After you implement redirects and clean your sitemap, click Validate Fix to queue a recheck. Google revalidates in batches and continues re-testing known dead URLs for a long time, so some legitimately removed URLs stay in the report indefinitely without indicating a problem.
Do I need a custom 404 page?
Yes. A custom 404 page keeps visitors on your site instead of sending them back to search. It should return a genuine 404 status code, state plainly that the page is gone, and offer full navigation, site search, links to your best content, and your primary call to action. It should never return 200 or auto-redirect.
Can I remove 404 URLs from Google's index?
A URL returning 404 is not in the index, so there is nothing to remove. If a dead URL is still appearing in search results, use the Removals tool in Search Console for a temporary block while the 404 or 410 propagates. The status code is what permanently removes it.
Does Bing report 404 errors too?
Yes. Bing Webmaster Tools surfaces 404s through its crawl information and site scan reports, and Bing runs its own crawler and its own index. Cleaning up 404s in Google does nothing for Bing, and because ChatGPT search retrieval leans heavily on Bing's index, Bing is where broken URLs quietly cost you AI citations.
Do 404 errors affect AI search visibility and ChatGPT citations?
Yes, and this is where 404s actually cost you. Answer engines cite URLs. When a cited URL starts returning 404, the citation link goes nowhere and the retrieval layer drops the page on its next crawl. Redirecting preserves the path for anyone following an older citation and passes accumulated authority to the replacement page.
Comments