Skip to content

Blog

Website Migration SEO Checklist, With an Example

Website migration SEO comes down to one rule: every old URL that people or search engines might still use must lead, in one permanent redirect, to the closest page on the new site. Find those URLs before launch, map each one, redirect it, and test the whole list automatically, so nothing the old site earned is lost by accident.

This checklist is for marketing leads at B2B companies moving to a new site, a new platform, or a new URL structure. It follows Google's own guidance on site moves with URL changes, and uses our own rebuild of webika.net as the worked example: a site with about 20 years of history, moved to a new build with every old URL kept working.

Why website migration SEO goes wrong

Most migrations that lose traffic lose it for dull reasons:

  • Old URLs nobody listed. The new site launches, the old pages return 404, and the links other sites gave them stop counting.
  • Everything redirected to the home page. Google's guidance warns that redirecting many old URLs to one irrelevant page, such as the home page, can be treated as a soft 404.
  • Redirect chains. Old URL to a staging path to the final page, each hop slower and less certain.
  • The staging site left blocked. A noindex tag or a robots.txt block copied from staging to production.
  • No test. Redirects checked by hand on launch day, then never again.

None of these needs special tools to avoid. They need a list, a map, and a check.

The website migration checklist

1. List every old URL, from more than one source

The old site's own sitemap isn't enough: it lists what the old site knew about, not what the web still links to. Use several sources:

  • Your analytics and Search Console, for pages that had traffic or impressions.
  • A links export from Search Console or an SEO tool, for pages other sites link to.
  • The Wayback Machine, for pages that existed years ago and may still carry links.
  • Bing Webmaster Tools, which sometimes holds URLs the others don't.

What we did: webika.net's history goes back about 20 years. We pulled the Wayback Machine's capture index for the domain and its subdomains on 18 October 2026, which listed 200 URLs. We then filtered out what was never a page: 60 404s and crawler probes, 23 asset files, 27 admin, login, and system pages, 13 URLs from a 2022 parked domain, and 2 template queries. That left 75 URLs worth keeping, including old pages, a PDF brochure, and pages from an Arabic version of the site.

Six days later, Bing Webmaster Tools showed one more URL, /lander, that the Wayback Machine had never captured. It was most likely a parked-domain landing page, and we added it to the list. A second source is worth having for exactly this reason.

2. Keep the list as a file that only grows

Write the list down as a file in the site's code repository, with where each URL came from and when it was first seen. Treat it as append-only: rows are added, never removed. If a URL ever existed and might still be linked, it stays on the list.

What we did: the list lives in redirects/legacy.csv, with a source and a first-seen date for each URL. Our local CI script, which runs before every release, compares it with the last released version and fails if any row has been removed.

3. Map each old URL to its closest new page

For each URL, decide one of two things: it redirects to a new page, or a page already exists at the same address. Pick the closest match by topic, not the most important page.

What we did: a second file, redirects/map.csv, holds one row per old path, with its action, its target, and a note saying what the old page was. Old SEO pages go to our SEO service page, old web design and ecommerce pages to the web design page, hosting pages to development, the old portfolio to case studies, and the old blog and news posts to the blog. Arabic pages go to their closest equivalent. Old home pages from 2008 to 2010 go to the new home page, and so do a few URLs with no closer match, such as an old photo gallery and the PDF brochure: six old paths in all, out of 69 in the map.

4. Use one permanent redirect per URL

Google's guidance asks for server-side permanent redirects, such as 301 or 308, straight from each old URL to its final page, with no chains. It also says to keep the redirects for at least a year, and ideally longer.

What we did: we chose 301 for every redirect. The new site runs on Next.js, whose own permanent redirect setting sends a 308, and whose automatic trailing-slash redirect ran before ours, so an old URL ending in "/" took two hops. We turned that off, sent a single 301 for each old path with and without a trailing slash, and added one site-wide 301 that removes a trailing slash from any other path. The old site matched URLs regardless of case, so the new one does too. Because the list only grows, the redirects stay.

5. Test every URL automatically

A redirect map is only as good as its last test. The test should request every URL on the list against a real build of the new site, and fail if any of them doesn't answer properly.

What we did: a script requests every row of legacy.csv against a running production build. Each must have a row in map.csv, and must answer 200, or a 301 in one hop to a page that answers 200. Unit tests also check that every redirect target is a page that exists, that no target is itself redirected, and that no URL appears twice. The local CI script runs the check before every release, so a change that breaks an old URL is caught before it reaches production. When we checked production on 20 October 2026, all 75 URLs on the list at the time passed.

6. Make sure the new site can be indexed, and the staging site can't

Check that production has no leftover noindex and no robots.txt block, and that staging does say noindex. Google's guidance on blocking indexing notes that a page blocked in robots.txt can't be seen to carry noindex, so it can still end up indexed from links.

What we did: the site decides from its own address whether it's production. Anywhere else, every response says noindex, while robots.txt still allows crawling so the noindex can be read. A misconfigured environment fails closed: it isn't indexed.

7. Pick one host and redirect the others

Decide whether the site lives at www or the bare domain, and on HTTPS, and send every other variant there with a 301.

What we did: www.webika.net redirects to webika.net with a 301, and the hosting platform redirects HTTP to HTTPS.

8. Update everything that points at old URLs

Internal links, canonical tags, and the sitemap should all use new URLs. Google suggests submitting the new sitemap, keeping the old one available for a while, and asking sites that link to you to update their links where you can. Use Search Console's Change of Address tool only if the domain itself changes.

SEO migration checklist for launch day

On the day, work through this list:

  • Run the redirect check against production, not only staging.
  • Submit the new sitemap in Google Search Console and Bing Webmaster Tools.
  • Request a few old URLs by hand and confirm one 301 to the right page.
  • Confirm production pages have no noindex.
  • Check the canonical tag on a sample of pages points at the page itself.
  • Watch the server logs for a rise in 404s over the first days, and add any new old URLs you find to the list.

Website redesign SEO: when the URLs stay the same

A redesign that keeps every URL is lower risk, but not free of it. Templates that drop headings, product specs moved into images or scripts, internal links that disappear from menus, and pages that load more slowly can all change how search engines read the site. Compare a sample of key pages before and after: their titles, headings, text, structured data, and speed.

Our website redesign service keeps URLs where they work and redirects the rest, and how much a website costs explains why redirect work belongs in the quote.

What to expect afterwards

Google's guidance is plain that rankings can fluctuate for weeks or longer while it recrawls and reindexes a moved site. A clean migration keeps what the old site earned; it doesn't promise anything on top.

That's where we are with our own site. The new webika.net has only just been indexed: in a check on 14 November 2026, Google showed about 26 of the 41 URLs in our sitemap. We're watching Search Console as the rest are crawled, and we'll report what we see rather than predict it.

Where Webika fits

Every site we build or rebuild gets this process: the old URLs found, mapped, redirected in one 301, and tested before launch. If your migration is already done and you're worried about what it lost, our B2B SEO service starts with exactly that audit. Book a free call to talk it through.

Talk it through first

30 minutes, booked online. Free, with no obligation.

Book a call