How can 404 errors after a server migration destroy your organic traffic?

What actually happens when a URL migrates without a redirect?

When you change the structure of your URLs during a server migration, Google sees the new URLs as completely new pages. Old URLs continue to exist in the index for a while, but now return 404 errors.

The crawler visits these old URLs, detects the 404 error, and eventually completely de-indexes them. The ranking signals (PageRank transmitted by backlinks, performance history, user signals) remain attached to the old URL that no longer exists. The new URL starts from scratch, as if your content had just been published for the first time.

How do 301 redirects preserve ranking signals?

A 301 redirect tells Google that a page has moved permanently to a new address. The engine then transfers most of the ranking signals from the old URL to the new one. This transfer includes the PageRank from external backlinks, accumulated authority, and part of the performance history.

According to Google’s official statements, a 301 redirect transmits almost 100% of PageRank. In practice, the new URL quickly inherits the position of the old one in the SERPs, sometimes with a slight temporary fluctuation as Google reevaluates the page in its new context.

How long does Google tolerate 404 errors before de-indexing?

Google does not immediately de-index a URL that returns a 404. The engine re-crawls the page several times over a period of a few weeks to verify that the error is indeed permanent and not a temporary technical glitch.

If the 404 error persists, the page eventually disappears from the index. The exact timeframe varies depending on the authority of the site and the usual crawl frequency. For a site crawled daily, expect 2 to 4 weeks. For a less prioritized site, it could take several months. But right from the first crawls that return errors, the position in the SERPs begins to drop.

• Migration without redirects = complete loss of ranking signals accumulated by the old URLs
• Permanent 301 redirects = almost complete transfer of PageRank and authority to the new URLs
• Variable de-indexing timeframe (2 weeks to several months depending on crawl frequency), but immediate impact on ranking
• External backlinks pointing to old URLs no longer transmit PageRank if they return 404 instead of redirecting
• Google makes no distinction between server migration and URL overhaul: the redirect rule applies in all cases

Does this statement align with on-ground observations?

Absolutely. The migration data I have observed across dozens of projects confirms that the lack of a complete 301 redirect plan leads to traffic collapses of 40 to 80% in the weeks following the migration. Pages that are not redirected lose their position in the SERPs in just a few days.

What is less evident is that even with redirects in place, some sites experience temporary losses of 10 to 20%. The reason: Google reevaluates the content in its new URL context, and if the new structure breaks internal linking elements or changes page depth, ranking may fluctuate. Redirects preserve signals but do not guarantee absolute stability if the SEO architecture changes.

What nuances should be made to this general rule?

The first crucial point: not all 404s are disasters. If you deliberately remove outdated or low-quality content, allowing these pages to return 404s is perfectly acceptable. Google has confirmed this multiple times: a healthy site has 404s. The problem arises when high-traffic pages or pages with backlinks disappear without a redirect.

The second nuance: redirect chains dilute PageRank. If you redirect A to B, then B to C, some PageRank is lost at each jump. The ideal is to redirect directly from A to C. In migration, map the final URLs before setting up the redirects, especially if your new structure alters multiple levels of hierarchy.

[To be checked] Google claims that 301 redirects transfer "almost 100%" of PageRank but refuses to quantify the exact residual loss. SEO tests indicate a retention of 90 to 95%, but it is challenging to establish this with certainty given the numerous variables (page context, domain authority, content quality). Consider that there is always a micro-loss, even if minimal.

In what cases does this rule not apply or become secondary?

If your site does not depend on organic traffic (for example, a web application behind a login, or a purely advertising site), you can ignore redirects without consequence. Similarly, very recent pages without backlinks or ranking history do not need a redirect: they have not accumulated any signal worth preserving.

Another borderline case: sites undergoing a complete content overhaul. If you retain no content from the old site, redirecting to unrelated pages may create a discrepancy between the anchor of the backlinks and the destination content. Google may then partially ignore these redirects. In this rare scenario, sometimes it is best to leave them as 404s and start fresh rather than force absurd redirects.

What should you do concretely before a server or URL migration?

Thoroughly map all indexed URLs before migration. Export the list from the Search Console (Indexing > Pages section) and cross-reference it with a complete site crawl using Screaming Frog or Oncrawl. Also identify the URLs that receive backlinks through Ahrefs, Majestic, or your preferred tool. These are the URLs that must be redirected.

Then create a mapping table that links each old URL to its new destination. For deleted pages, decide whether they merit a redirect to a category page or whether it’s better to leave them as 404. This mapping will serve as a basis for configuring 301 redirects in your .htaccess file, nginx.conf, or via a plugin if you are on WordPress.

What mistakes should be avoided when setting up redirects?

The first classic mistake: redirecting all old URLs to the homepage. Google detects this practice ("soft 404") and may ignore these redirects if the destination content has no relation to the original URL. Each old URL must point to the most semantically relevant new URL. If no match exists, leave it as 404 rather than forcing an inconsistent redirect.

The second frequent pitfall: forgetting URL parameters. If your old URLs include parameters (e.g., /product?id=123), ensure that your redirects capture them as well. A redirect that ignores parameters will create 404s for those variants. Test your redirects with all the URL patterns used on the site, including versions with and without trailing slashes.

How can you verify that the migration occurred without signal loss?

Monitor three key indicators in the Search Console for 4 to 6 weeks post-migration: the number of indexed pages (Indexing > Pages section), 404 errors (Pages > Unindexed section), and organic traffic (Performance). A sharp drop in the number of indexed pages or traffic indicates a redirect issue.

Also check that external backlinks are properly transmitting their PageRank to the new URLs. Use a backlink tracking tool to confirm that the links pointing to the old URLs are now following the 301 redirects. If some backlinks point to 404s, contact the webmasters to request updating the links directly to the new URLs, which eliminates the redirect jump.

• Export the complete list of indexed URLs and URLs with backlinks before migration
• Create a 1:1 mapping table between old and new URLs
• Configure 301 redirects avoiding chains and generic redirects to the homepage
• Test the redirects on a representative sample of URLs (with/without www, http/https, trailing slash, parameters)
• Submit the new XML sitemap and monitor indexing via the Search Console for 6 weeks
• Keep redirects active for at least 12 months after migration