Moving a website without losing visitors

A new domain, a new CMS, or a tidy-up of page addresses all change URLs. Every old address is in someone’s bookmarks, an old email, a printed leaflet and Google’s index; without a redirect each becomes “page not found”, and the search ranking the old page had is lost. These tools plan the redirects and check the new site, in your browser.

1. List every old address

Before anything is switched off, collect the old addresses: the site’s sitemap.xml, an export from the CMS, Google Search Console’s Pages report, and the most visited pages in your analytics. Pasting the whole sitemap.xml into the Redirect Map Builder lists its pages one a line, ready for each new address after the comma. For a long list, or one gathered from several places, a spreadsheet is easier: old addresses in one column, the new address for each in the next. Old addresses with tracking or session parameters can be cleaned up in the URL Inspector, which splits a link into its parts so you can see which parameters actually choose the page (such as ?id=5) and which don’t (utm_source, fbclid).

Point every old page at its closest new page, not all at the home page: search engines treat a mass redirect to the home page as “not found”.

2. Turn the list into redirect rules

Copy the two columns and paste them into the Redirect Map Builder. Choose For: Apache (.htaccess), nginx, Netlify or Cloudflare Pages (_redirects), Vercel (vercel.json) or Caddy. Leave the status on 301 Moved permanently, which passes the old page’s search standing to the new one. A whole section that moved can be one line: /blog/*,/news/*.

Fix what it lists above the rules: an address redirected twice to different places, loops that send visitors round in circles, and chains (/a to /b to /c), which Point chains straight at their end rewrites so each old address goes straight to its final page. Copy the rules or Save as a file and add them to the new site.

3. Know the status codes

When you test old addresses in the browser’s developer tools (Network), each one should answer 301 then the new page 200. A 302 or 307 means a temporary redirect, which search engines don’t treat as a move; a 404 means a page was missed; a 410 says it’s gone for good on purpose. HTTP Codes & MIME Types explains any other code you see.

4. Don’t block the new site

Sites under construction often carry Disallow: / in robots.txt to keep search engines out, and it’s easily left there at launch. Paste the new site’s robots.txt into the robots.txt Tester with a few of its addresses: each should say Allowed for Googlebot, and its Sitemap: line should point at the new domain. Then submit that sitemap in Google Search Console, and use its Change of Address tool if the domain changed.

Before submitting it, open the new sitemap in the Sitemap Inspector and the old one as the Earlier version to compare. Removed lists every old address the new site no longer lists: each needs a line in the redirect map. Pages whose address only changed from http:// to https://, or gained or lost a final /, show as removed and added too, so they need redirects as well. If any address is still on the old domain, the checks above the list say the sitemap uses two hosts.

5. Check how pages look when shared and in search

The move can break the pictures and titles shown when a page is shared, especially if og:image still points at the old domain. Paste a new page’s HTML into the Social Card Preview to see how it will look on the main platforms and which tags are missing.

A new theme or platform often drops the structured data that gave search results their prices, dates or breadcrumbs, or leaves the old domain in its addresses. Paste an old page and its new version into the Structured Data Viewer in turn and compare the items: the same types should be there, with url, @id and item addresses on the new domain. Getting a website ready to launch covers the rest of the launch: favicons, security headers and the certificate.