How to redesign a website without losing your Google rankings

Redesigning a website without losing Google rankings comes down to one rule: every page that Google already knows about must keep working, at its old address or through a permanent redirect to its new one, and the pages that rank must keep the content that earned the ranking. Most traffic lost after a redesign is lost to URLs, not to the new design.

Rebuilding sites that already rank is the part of my work that carries the most risk, so this is the checklist I work from. It applies whether you are changing the design, the platform, or both.

Why redesigns lose traffic

The same few causes come up again and again:

• Old URLs return a 404 because the new platform builds its addresses differently.
• Every old page is redirected to the homepage, which Google treats much like a missing page.
• The text on pages that ranked is rewritten or cut during the redesign.
• Page titles, descriptions, and structured data that the old templates produced never make it across.
• The staging site's "do not index" setting goes live with the new site.
• An old address of the site, such as a staging or hosting subdomain, stays reachable and competes with the real one.

Each step below closes one of those gaps.

1. Record what ranks before anything changes

Export your data before the first design meeting, because once the old site is gone you cannot measure it again.

• From Google Search Console, export the Pages and Queries reports for the longest period available. These show which addresses get impressions and clicks, and for which searches.
• From your analytics, export the landing pages that bring in enquiries or sales.
• Note any pages other websites link to. Links point at addresses, and a broken address wastes the link.

This list is your baseline, and it decides which pages get the most care.

2. Crawl every URL on the current site

Search Console only shows what gets impressions. A crawler finds everything else: old campaign pages, PDFs, image files, tag and category pages, and paginated archives. On WordPress sites, image files under /wp-content/uploads are often indexed and linked to from other sites, and they are easy to forget because nobody thinks of them as pages.

3. Decide what happens to each URL

Every old address gets one of three decisions, written down with a reason:

• Keep. The page stays at the same address. Where you can, keep the addresses of pages that rank. It is the safest redirect there is.
• Move or merge. The page gets a new address, or its content joins another page, and the old address redirects to the closest match.
• Retire. The page is removed. Check first that it has no links pointing at it and no traffic worth keeping. If it has either, redirect it instead.

4. Write the redirect map

The redirect map is a spreadsheet with one row per old URL: the old address, the new address, the status code (a permanent 301 or 308), and the canonical address the new page declares. A few rules keep it clean:

• One hop only. An old URL should land on its final page directly, not through a chain of redirects.
• No blanket rule that sends everything to the homepage.
• Match each old page to the most relevant new page, not just a page in the same section.
• Include the variants: with and without www, http and https, trailing slashes, and old query strings.

Write the map before the build starts, not after. It often changes the new site's structure, and that is far cheaper to learn early.

5. Keep what Google already understands

For the pages that rank, carry across the page title, the main heading, the body text, the internal links pointing at them, image alt text, and any structured data. A new design can change how a page looks without changing what it says. If the content also needs rewriting, consider doing it after the new site has settled, so that if rankings move you know which change caused it.

6. Keep staging out of Google, and out of the live site

The staging copy of the new site should be blocked from indexing, with a password or a robots tag that says not to index it. Then make sure none of that goes live: check the robots.txt file, the robots meta tags, and that canonical tags point at the real domain rather than the staging address.

7. Redirect every old address of the site, not just the old pages

This is the step people forget, and it happened on my own site. Before idreesworks.tech, the site ran on a Vercel subdomain, muhammadidrees-dev.vercel.app. When I checked Search Console, three pages on the new domain (the about page and two articles) were not indexed. The reason given was "Duplicate, Google chose different canonical than user": Google had picked the old Vercel address as the original and treated the real pages as copies, even though their canonical tags pointed at idreesworks.tech.

The old address now sends a permanent redirect to the same page on the new domain, and a live test of one of those pages in Search Console shows it as available to Google. What they need is a recrawl, which you can ask for with Request Indexing in the URL Inspection tool.

The lesson applies to every host your site has ever answered on: staging subdomains, hosting addresses on vercel.app or netlify.app, the version without www, and plain http. All of them should redirect permanently to the one address you want indexed, from the day the new site launches.

8. On launch day

1. Run every row of the redirect map against the live site and confirm each one lands on the right page in one hop.
2. Check robots.txt, and confirm no page you want in Google carries a robots tag telling search engines not to index it.
3. Check the canonical tags on each template point at the live domain.
4. Submit the new sitemap in Search Console.
5. Use URL Inspection on your most important pages and request indexing for them.
6. Launch early in the week, so you are around for the first few days if something needs fixing.

9. After launch: what to check, and when

I check three times:

• After 48 hours. The redirects resolve, the important pages are being crawled, and nothing unexpected is returning errors.
• After two weeks.Search Console's Page indexing report, and the reasons it gives for anything not indexed. A dip in traffic during this period is normal while Google recrawls.
• After six weeks. Rankings and clicks for the searches in your baseline. A drop that is still there at six weeks is a fault to find and fix, not something to wait out.

If you have already lost traffic

It is usually recoverable. Find the old addresses that used to bring in traffic, from Search Console's Pages report, your analytics, or the Internet Archive's snapshots of the old site. Then redirect each one to its closest new page and restore any content that ranked. The sooner the redirects go in, the more of the old ranking signals carry over.

Getting help with a redesign

If your site brings in real enquiries through search, the crawl and the redirect map are worth having done properly. You can run the free migration risk check on your current site to see what a rebuild would trip over, without giving an email. If you are moving from WordPress, the WordPress to Next.js cost guide covers the platform-specific parts.

To have me plan and carry out the move, see website migration without losing rankings, or send me your current site.