Will a new website lose my Google rankings? How to plan the move so you keep them

• 301 and 308: permanent. Google's crawler documentation treats 301 as a strong signal that the target should be canonical, and 308 as equivalent to 301.
• 302, 303 and 307: temporary. Google treats 302 as a weak signal, and its indexing pipeline doesn't use a temporary redirect as a signal that the target should be canonical.
• Instant meta refresh: read as permanent. A delayed meta refresh is read as temporary.
• JavaScript redirects: only if server-side or meta refresh redirects aren't possible, because rendering can fail.

Within each pair, the HTTP standard (RFC 9110) lets a client change a POST to a GET on a 301 or 302, but not on a 307 or 308. So 308 is the choice when a form posts to a moved address. For ordinary pages, 301 and 308 behave the same in Search.

Two server examples in the form Google's redirects page gives them (the page also covers mod_rewrite, NGINX rewrite and PHP):

# Apache, mod_alias
Redirect permanent "/old" "https://example.com/new"

# NGINX
location = /service {
  return 301 $scheme://example.com/about/service;
}

URL Inspection suits single addresses; for a full map, script it.

# One old URL: the status and where it points
curl -sI https://www.example.co.uk/boiler-servicing

# Every hop in the chain
curl -sIL http://example.co.uk/boiler-servicing | grep -iE '^(HTTP|location)'

# Final status, number of hops, final URL (a GET, for servers that answer HEAD differently)
curl -s -o /dev/null -L -w '%{response_code} %{num_redirects} %{url_effective}\n' http://example.co.uk/boiler-servicing

# The whole map. map.csv: old,new on each line, both as full addresses starting https://
while IFS=, read -r old new || [ -n "$old" ]; do
  new=${new%$'\r'}
  read code hops final < <(curl -s -o /dev/null -L -w '%{response_code} %{num_redirects} %{url_effective}' "$old")
  [ "$final" = "$new" ] && [ "$code" = "200" ] && [ "$hops" -le 1 ] && echo "OK   $old" || echo "FAIL $old -> $final ($code, $hops hops)"
done < map.csv

Good: a first response of 301 or 308, one hop, and a final 200 on the intended page.

Red flags: a 302 or 307 on a permanent move; a final 404; the homepage as the destination for a page that had its own equivalent; more than one hop; a loop, which shows up as a very high hop count, or as a "Maximum redirects followed" error if you add -S to the curl command.

Chains build up quietly. An old http:// address on the bare domain can go to https, then to www, then to the new page: three hops. Googlebot follows up to 10, but the site-move guide advises redirecting to the final destination directly, and keeping any chain ideally to no more than 3 and fewer than 5.

• Internal links: change them from the old URLs to the new ones.
• Canonicals: each new URL should carry a self-referencing rel="canonical".
• hreflang: on multi-language sites, update every annotation to the new URLs. Each version has to reference all the alternates.
• Sitemaps: submit both the old-URL and new-URL sitemaps. Indexed pages from the old one should fall towards zero over time, and then it can go.
• robots.txt: make sure the new file blocks only what you mean to block, and remove anything that was only there for the build.
• Capacity: Google crawls a newly moved site more heavily than usual. A server that buckles returns 5xx errors, and those make Google temporarily slow its crawling.

Check for a leftover noindex in the header and the HTML:

curl -sI https://www.example.co.uk/ | grep -i x-robots-tag
curl -s https://www.example.co.uk/ | grep -iE "name=[\"']?(robots|googlebot)"

Since WordPress 5.3, the "Discourage search engines" box outputs a robots meta tag; on current versions it reads <meta name='robots' content='noindex, nofollow' />. A noindex is only seen if robots.txt lets Google fetch the page, so don't block and noindex the same URL.

In the Crawl stats report (under Settings), a large share of "Moved permanently (301)" is expected after a move. A large share of "Moved temporarily (302)" means the redirects are the wrong type.

Change of Address: open it in a domain-level property. It doesn't move subdomains, including www, so run it for every version of the old domain (www and non-www, and any subdomains), even ones you don't use, with each one verified. Its effects continue for 180 days from the start, and Cancel Move withdraws it. Keep the redirects for at least a year regardless.

New host, same URLs: lower the DNS TTL to a conservative value, such as a few hours, at least a week before the switch. Remove any crawl blocks from the new copy when the move starts. A temporary drop in Googlebot's crawl rate straight after launch is normal, followed by a steady increase over the next few days.

# Which address does the name point at, before and after the switch?
dig +short www.example.co.uk

# Read the TTL as set, from the domain's own nameserver (the number after the name, in seconds)
dig +noall +answer www.example.co.uk @$(dig +short NS example.co.uk | head -1)