You migrate a WordPress site without losing rankings by mapping every old URL to a new one, serving permanent 301 redirects for each, and leaving those redirects in place for months while Google re-crawls. The ranking loss people blame on migrations almost never comes from the move itself — it comes from URLs that were never mapped, redirects that were removed too early, or a database search-and-replace that broke the site's internal links.
The work splits cleanly into two very different jobs. Changing hosts while keeping the same URLs is a DNS and verification exercise. Changing the URLs — a new domain, HTTPS, a different permalink structure — is a mapping exercise, and that is where the risk lives.
Table of Contents
- Redirects do not cost you ranking signals
- Build the URL map before you touch anything
- Getting the WordPress side right
- Search Console, the Change of Address tool, and its limits
- How long it actually takes, and when you can stop
- A host change with the same URLs
- Frequently Asked Questions
Redirects do not cost you ranking signals
The single most common fear is misplaced. Google's site-move documentation states it directly: "Don't worry about link credit. 301 and other permanent redirects don't cause a loss in PageRank." A URL-changing migration preserves link equity as long as every old URL resolves to its real new equivalent. The qualifier is doing all the work.
Equity flows through a redirect that points at the matching page. It does not flow usefully through a redirect that points at a category archive, a search page, or the homepage because nobody bothered to find the right target. So the goal is not "set up redirects." The goal is that an old URL's redirect lands on the page a visitor was actually looking for. Anything less is a redirect that technically returns 301 and still loses the traffic.
Build the URL map before you touch anything
Google's instruction is to prepare the inventory first: "Prepare a URL mapping from the current URLs to their corresponding new format," and only then configure server redirects from old to new. A migration that starts with the new site and works backwards is how archives turn into mass 404s.
Pull the old URL list from more than one place, because no single source is complete: Then decide a destination for every row — including the rows with no destination. For content you are not carrying over, Google says the server must return HTTP 404 or 410 rather than redirecting everything to the homepage, because a blanket redirect to one page is treated as a soft 404. Retiring a page honestly is a legitimate outcome; disguising it as a move is not.
- The existing XML sitemap, for everything WordPress currently knows about
- Search Console's performance and index-coverage reports, for URLs that earn impressions
- Server access logs, for old URLs still being requested that WordPress no longer lists
- Analytics landing-page reports, for pages with inbound links you forgot about
Getting the WordPress side right
WordPress stores its own address in two options, `siteurl` and `home`, and the official migration documentation treats changing those as step one — either in the database or by setting `WP_HOME` and `WP_SITEURL` in `wp-config.php`. Those two values control what WordPress thinks it is. They do not touch anything already written into your content. That is why a separate search-and-replace is required for URLs embedded in post bodies, widget text, and theme settings. And this is the step that quietly destroys sites: the WordPress docs warn that a database-wide find-and-replace corrupts serialized data — PHP's packed array format, which records the byte length of every string it contains — unless you use a serialization-aware tool such as WP-CLI.
A plain SQL `REPLACE` changes the string and leaves the stale length, and the option stops unpacking. Theme customizer settings and widget configurations are exactly where serialized data lives. One column stays untouched: the docs say the GUID must never be changed. It looks like a URL, but feed readers use it as a permanent identifier. Rewrite it and every subscriber sees your entire back catalogue republished as new.
Search Console, the Change of Address tool, and its limits
Before the move, check that your verification survives it. Google's guidance is explicit — "Make sure your Search Console verification will continue to work after the site move" — because tokens can be tied to the URL. Lose verification and you lose the Change of Address tool and your post-move diagnostics at the exact moment you need them. The tool itself is narrower than most people assume.
Per Google's Change of Address help page, it only works on domain-level properties: "You cannot move properties at the path level." It will not move subdomains beneath the named domain, so moving `example.com` does not move `www.` or `m.`. It runs blocking pre-move checks, must be used after redirects are already live, and should be skipped entirely for an HTTP→HTTPS move, which Google detects on its own. Since a June 17 documentation update, Google asks owners to verify and submit a change-of-address request for every variant of the old domain — all subdomains plus both www and non-www, even unused ones — as reported by Search Engine Land. Worth confirming the current wording on Search Central before you plan around it. Google also warns against chaining moves: going A→B and then immediately B→C is a known way to confuse the process.
How long it actually takes, and when you can stop
A move is not finished when the redirects go live. It is finished when Google has crawled it, and Google is blunt about the timeline: "Googlebot will have to visit every URL on your old and new site at least once. There are no fixed crawl frequencies." Speed scales with site size, so a WordPress archive of tens of thousands of posts settles over weeks.
Keep the redirects for at least 180 days, and Google says longer "if you still see any traffic to them from Google Search." A change-of-address request itself may take up to 180 days to complete, and can be cancelled inside that window. Treat half a year as the floor, not the plan — if old URLs are still sending visitors, the redirects are still earning their keep. Google's other pacing advice is to phase changes "one after the other, not everything at the same time." Combining a domain change with a permalink restructure, a theme replacement, and a host move makes diagnosis impossible when traffic dips, because every candidate cause landed on the same day.
A host change with the same URLs
If the URLs are not changing, there is nothing to map and nothing to redirect — the risk is a crawl interruption during DNS propagation. Google's hosting-change guidance is to lower your DNS TTL "to a conservative low value (for example, a few hours) at least a week in advance," so resolvers pick up the new address quickly instead of holding a day-old cached answer. Expect a dip.
Google says to anticipate "a temporary drop in Googlebot's crawl rate immediately after the launch," with recovery over the following days. That dip is normal and not a signal that something broke. Keep the old host running until you have confirmed the new one is serving Googlebot. Google's advice is to use the URL Inspection tool to verify access before shutting the old server down — propagation is uneven, and some resolvers will still be pointing at the old IP after yours has switched.
Frequently Asked Questions
Do I need a redirect plugin, or should redirects live at the server level?
Google's instruction is to configure server redirects from old to new. Server-level rules run before WordPress loads, so they survive plugin changes and are faster, which matters when Googlebot has to re-crawl every URL.
Can I use the Change of Address tool to move from HTTP to HTTPS?
No. Google's help page says to skip it for an HTTP-to-HTTPS move because Google detects that change itself.
What should happen to posts I am deleting rather than moving?
Serve 404 or 410 for those URLs. Google treats a mass redirect of dropped content to the homepage as a soft 404, so it gains you nothing.




