What Google does with a link to a missing page
A new URL that Google finds already returning 404 is not processed. Known URLs that keep returning a 4xx code other than 429 get crawled less and less over time. Google treats all 4xx codes except 429 the same, so Googlebot handles 410 like 404: the content is reported as not existing. RFC 9110 reserves 410 Gone for a resource that is no longer on the server, probably for good.
Google's position is that 404s in general do not hurt a site in search: 404s on some URLs do not affect how the site's other (200) URLs perform.
Soft 404s are a separate case. There, the page tells visitors it does not exist, but the server still answers 200 (success). Google says redirecting a missing page to another page such as the homepage, instead of returning 404, can cause problems (soft 404).
Fix the href, redirect, or let it 404
The real cost falls on readers and on the pages you need people to reach. MDN advises keeping the number of broken links, internal and external, to a minimum because they frustrate readers. Google advises investigating 404s on pages you actually care about.
Fix the href when the link itself is wrong.
Redirect. If the page moved or has a close substitute, Google's guidance is a 301 to that page instead of an error. Google recommends mapping commonly misspelled URL variants to the correct page with redirects. For an inbound link with a typo in the URL, Google suggests a 301 redirect to the correct URL to keep that traffic.
Let it 404. When nothing replaces the removed page, Google wants a 404 or 410 response.
A content inventory that lists dependencies, such as linked forms or campaigns, helps prevent broken links. The W3C Link Checker checks that every link in a page can be followed to its target, including fragment anchors. It also reports HTTP redirects, including directory redirects. Other crawl tools are compared in best tools for a successful website migration.
How Pagepro keeps migrated redirects from breaking
On a migration, Pagepro checks where each imported redirect points, not only where it starts. A redirect whose target has no route on the new site ends in a 404, even when the redirect table looks fine. For the full move, see website migration services.
Sources
- 404 Not Found - HTTP | MDNMozilla (MDN) · developer.mozilla.org
- How HTTP status codes, and network and DNS errors affect Google SearchGoogle · developers.google.com
- RFC 9110: HTTP SemanticsIETF · rfc-editor.org
- 404 (Page Not Found) errors - Search Console HelpGoogle · support.google.com
- Do 404 errors hurt my site? (Google Search Central Blog, 2011)Google · developers.google.com
- Troubleshoot Google Search crawling errorsGoogle · developers.google.com
- W3C Link Checker DocumentationW3C · validator.w3.org
- Content inventory and audit (Web Style Guide)Government of British Columbia · www2.gov.bc.ca