How does Google really deindex an expired site or one that’s globally 404?

Why doesn’t Google remove all pages at once?

Google operates with a distributed crawl budget based on the perceived priority of each URL. When a site switches to a global 404 or expires, Googlebot does not revisit all indexed pages simultaneously.

Daily crawled URLs — homepage, main categories, frequently updated pages — are revisited quickly and noted as inaccessible. They drop out of search results within a few days. Orphaned, old, or rarely explored URLs remain cached for several weeks, even months, without validation of their status.

What does “reducing visibility in normal search results” really mean?

Google applies a relevance filter to pages it suspects belong to a dead site. Even if technically indexed, they no longer rank for regular queries. You can still find them via site:example.com, but they no longer rank for their usual keywords.

This is a functional rather than a technical deindexing: the engine considers this content obsolete but does not immediately purge its database. This latency explains why some expired domains retain a visible trace in Google for several weeks after expiration.

What is the difference between a global 404 and domain expiration?

A global 404 occurs when all pages of an active site return a 404 code — often due to server configuration errors or a faulty .htaccess. The domain is still registered, the server responds, but all URLs return Not Found.

Domain expiration involves the end of renewal: the registrar reclaims possession, the DNS points to a parking page, or resolves no longer. The server no longer responds or returns entirely different generic content.

In both cases, Google detects unavailability, but the speed of response varies according to the nature of the signal. A 404 is a clear HTTP signal; expiration may first generate timeouts or redirects, complicating immediate detection.

• Gradual deindexing based on the crawl frequency of each URL
• Visibility filter applied before final removal from the index
• Site query: continues to display cached pages during the transition period
• Variable delay depending on the historical priority of each page (homepage vs orphaned content)
• Technical distinction between HTTP 404 error and total domain unavailability

Does this statement align with field observations?

Yes, completely. We regularly observe expired domains that remain partially visible in Google for 4 to 8 weeks after the end of registration. Homepages and categories disappear in 7-10 days, while orphaned blog articles may persist for up to 2 months.

This latency actually creates an angle of attack for black-hat SEOs who buy expired domains: as long as Google hasn’t purged the index, the domain retains a trace of authority. But this window is shrinking — Google has visibly accelerated its detection since mid-2023. [To verify] whether this acceleration applies to all TLDs or just prioritized .com/.net domains.

What nuances should be added to this rule?

Mueller speaks of reduced visibility in normal results, but doesn’t specify whether Google applies a declining freshness score or a binary filter. Experience shows it’s more of a filter: pages suddenly drop from position N to outside the top 100, without a gradual decline.

Another point: pages with strong external backlinks seem to enjoy a longer grace period. If an orphaned URL still receives referral traffic, Googlebot occasionally revisits it even after detecting the global 404. This is not officially documented but is consistent with the crawling budget logic oriented towards demand.

Finally, the site: query is not a reliable indicator of real indexing. Google repeats: site: shows the cache, not necessarily the active index used for ranking. A site can display 500 URLs in site: and rank none on an actual query.

When does this rule not apply?

If a site temporarily switches to a global 404 and then recovers within 48-72 hours, Google typically does not start the deindexing process. The engine tolerates brief outages, especially for sites crawled daily.

Soft-404s — pages that display empty content but return a 200 code — complicate matters. Google takes longer to detect that a site no longer exists if the server responds with OK. The algorithm must analyze the content to identify the anomaly, which delays deindexing by several additional weeks.

What should I do if my site accidentally goes global 404?

Restore availability immediately — every hour counts. Once the site is back up, force a re-crawl of priority URLs through Google Search Console (URL inspection → request indexing) to speed up recognition of the return to normalcy.

Monitor server logs to identify which pages Googlebot has revisited post-incident. If entire sections have not been re-crawled within 10 days, manually submit an updated XML sitemap with a recent <lastmod> to signal content freshness.

How can I anticipate the deindexing of a site I plan to shut down voluntarily?

If you plan to permanently close a site, set up 301 redirects to a destination domain before expiration or server shutdown. Even if the content is not identical, redirecting to a homepage or a similar category retains some of the signal.

Keep redirects active for at least 6 months — the time it takes for Google to crawl all indexed URLs and transfer equity. If you cut off abruptly without redirecting, you permanently lose accumulated SEO juice.

What mistakes should be avoided during a migration or redesign?

Never assume Google will instantly detect a status change. A poorly executed migration can lead to a temporary global 404 long enough to trigger a short-term irreversible partial deindexing.

Test the new infrastructure in pre-prod with a crawlable subdomain, ensure all critical URLs respond with 200, then switch DNS. Keep the old server active for 48 hours in parallel if possible, to absorb the last requests from Googlebot before the complete update of its DNS cache.

• Check Google Search Console daily for any unusual spike in 4xx errors
• Maintain an updated XML sitemap with differentiated priorities (homepage/categories at 1.0, secondary content at 0.5)
• Set up monitoring alerts for returned HTTP codes (Uptimerobot, Pingdom, or custom script)
• Document the crawl frequency of your strategic pages (Search Console → Crawl Stats) to anticipate deindexing timelines
• In the event of an accidental global 404, force manual re-crawl of the 20-30 most strategic URLs as soon as recovery occurs
• Plan for a minimum of 6 months of active 301 redirects during a voluntary site closure