Website Redesign SEO: How to Redesign Without Losing Your Google Rankings

Most owners come to a redesign worried about the wrong thing. They worry the new site will look worse. The real risk is quieter: the new site goes live, looks great, and the phone slowly stops ringing because Google no longer knows where anything went. Website redesign SEO is the work that prevents that, and almost all of it happens before launch.

I have built more than 282 websites over twelve years. The redesigns that hold on to their rankings tend to share the same unglamorous habits, and none of them are complicated. This is the version I would give a friend: what to record first, what to keep, how to redirect, and what to check in the weeks after launch. Along the way I will show you how it played out on real redesigns my team handled for agency partners, including one launch we had to stop.

Google is clear about this in its own documentation. When a site changes its URLs, Google says to expect temporary fluctuation in site ranking during the move while it recrawls and reindexes the pages. Temporary is the important word. The permanent losses come from pages that disappear without a redirect, content that gets cut because it looked old, and launch-day mistakes that tell Google not to index the new site at all.

A purely visual refresh, where the URLs and the words on each page stay the same, is low risk. A redesign that also moves to a new platform, rewrites the copy and reorganizes the menu is really a site migration, and it deserves migration-level care. Most small business redesigns are the second kind, even when everyone calls it the first.

You cannot protect rankings you never wrote down, and you cannot prove the new site is doing better without a before picture. Here is the snapshot I take:

• Every live URL. Crawl the current site or pull the URL list from its sitemap. This becomes the left column of your redirect map.
• Top pages by clicks. Export the last 12 months of pages from Google Search Console. These are the pages you cannot afford to lose.
• Top search queries. Export the queries too, so you know which phrases each page actually ranks for.
• Pages other sites link to. Search Console’s Links report shows which of your pages earn links from elsewhere. Those links are worth keeping alive.
• Conversions. Note which pages generate calls, form fills and bookings. Rankings are only useful if they turn into customers.

Save it somewhere that will still exist in six months. When someone asks whether the new site is working, this is the file that answers them.

The most expensive redesign mistake is cleaning house. An old blog post from years ago can quietly be the page Google sends the most visitors to. It gets cut because nobody on the redesign team recognizes it, and the traffic goes with it. Your baseline export stops that from happening, because it shows you which “old” pages are actually doing the work.

Sometimes the problem is bigger than one post. Gladwell Orthodontics, a practice with seven offices around the Triangle, had a redesign built by another firm that was days from launch when GO Media asked me to review it. The new site did not keep the old site’s addresses, and 413 pages from the old site had no equivalent on the new one. If it had launched that way, most of the pages Google showed for the practice would have returned errors. I sent an email with the subject line “DO NOT LAUNCH,” and GO Media pushed the launch back so it could be fixed properly.

On big sites, the page list is the project. When we rebuilt Cleggs, a pest control company, for MOD Marketing, the old site had 1,634 URLs, nearly a thousand of them individual review pages. We worked from a shared spreadsheet with one tab for every URL we were keeping and another for the pages other websites linked to, so nothing with value was cut by accident.

For every page on the current site, decide one of three things:

1. Keep it. Same topic, same URL, improved design and content. This is the default for anything with traffic.
2. Merge it. Two thin pages about the same service become one stronger page. The old URL redirects to the new one.
3. Retire it. Truly dead pages with no traffic, no links and no purpose. Even these get a redirect to the closest relevant page, so nobody lands on an error.

Content is where most of your rankings live. A new layout around the same words is a small change in Google’s eyes. New words on the same layout can be a big one, so when you do rewrite a page, keep the phrases people actually search for.

Every URL change asks Google to transfer what it knew about the old address to a new one. Redirects make that work, but the safest redirect is the one you never needed. Good reasons to change a URL include moving to a new domain, cleaning up addresses that are genuinely broken (long strings of numbers, duplicate versions of the same page), or merging pages. “The new platform does it differently by default” is usually not a good reason, because most platforms, WordPress included, let you set the URL to match the old one.

If you are switching platforms, match the old URL structure wherever you can. It is a little more setup work and it removes most of the risk.

That is what we did for Phoenix Cleaning Solutions, a carpet and tile cleaning company we rebuilt for Heather Pierson Media. The old site had about 120 published pages. Every city page, every service and town page and all 13 blog posts kept their exact old addresses on the new site, and we kept closing the gap between the old page list and the new build until only a handful of pages were left to decide on before launch.

Google’s guidance is reassuring here: 301 and other permanent redirects don’t cause a loss in PageRank. The losses come from redirects that are missing, pointed at the wrong place, or chained together. Here is how I build the redirect map:

1. List every old URL from your baseline crawl in one column.
2. Match each one to the most relevant new URL in the next column. A page about roof repair redirects to the new roof repair page, not to the homepage.
3. Avoid chains. If page A already redirects to B, and B now moves to C, point A straight to C.
4. Use permanent redirects (301 or 308) for anything that has moved for good. Google’s redirect documentation explains the differences between the types.
5. Test the whole list on launch day. Every old URL should land on a live page, not an error.
6. Leave them in place. Google recommends keeping redirects generally at least 1 year. There is rarely a good reason to remove them after that, either.

The shortcut to avoid is pointing every old URL at the homepage. Google specifically warns against redirecting many old URLs to one unrelated page, because it might be treated as a soft 404 error, which means the old page’s ranking does not carry over.

Here is what that looked like in practice. On Gladwell’s launch day we put 354 permanent redirects in place, mapping old URLs to the new structure, then caught 9 more old location pages still erroring that afternoon and fixed those too. For Cleggs we created 278 redirects, sorted by type: 141 blog posts, 61 pest pages that moved into the site’s new pest library, 31 location pages and 16 pest categories. Before launch we checked all 1,638 old URLs against the new build, so we knew where every one would land.

A new theme or page builder often resets the parts of a page that visitors never look at. The design review passes because everything looks right, and the SEO quietly disappears underneath. For each page you are keeping, check that the new version still has:

• the same (or a better) title tag and meta description
• a single, clear H1 and logical subheadings
• the internal links that pointed to it from other pages, updated to the new URLs
• alt text on the images that matter
• any structured data the old page had, such as business details, reviews or FAQ markup
• the same core wording for the phrases it ranked for

Internal links deserve extra attention. They are how Google understands which pages matter most on your site, and a new menu or a new homepage layout can change them more than people expect.

Two examples from our own projects. On Phoenix, we moved the page titles and meta descriptions over from the old SEO plugin, so 121 of the 124 published pages and posts launched with a proper title and description, and where the old title was better than the new draft, the old one won. On Cleggs, a later quality pass found 630 internal links still pointing at old addresses and passing through redirects, so we pointed each one straight at its final page.

Most new sites are built on a staging copy that is deliberately hidden from Google. In WordPress that is the “Discourage search engines from indexing this site” box. If that setting, or a noindex tag, makes it to the live site, Google will start dropping your pages no matter how good they look. Here is my launch-day list:

1. Indexing is on. No noindex tags, and the robots.txt file is not blocking the site.
2. The sitemap lists the new URLs and has been submitted in Search Console.
3. Redirects pass the test from the step above.
4. The site still works on a phone, on a normal cell connection, not just on office wifi.
5. Speed has not gone backwards. Run your main pages through PageSpeed Insights and compare them with the old site.
6. Analytics and Search Console are connected, so the first weeks of data are not lost.
7. Forms and phone links work. Submit every contact form yourself and confirm the email arrives.

That last one is not strictly SEO, but it matters more than anything else on this list. Rankings that send people to a broken contact form are worth nothing. It is also the most common way a site loses business without anyone noticing, which I wrote about in why websites stop getting leads.

I have seen this one up close. During the Gladwell review we found the contact forms were sending patient inquiries to the builder’s inbox instead of the practice. The form said “thank you,” and the lead went nowhere. On launch day the ad agency’s tracking also went missing in the build, and we reinstalled it the same afternoon. For Cleggs, the launch-day crawl confirmed all 441 published pages loaded without an error.

The weeks after launch are when a redesign either settles in or slowly leaks. I check these on a schedule:

• Page indexing report in Search Console, for new “not found” errors and pages Google has excluded.
• Top pages versus the baseline, so a page that has dropped out shows up within days rather than months.
• Search queries, to confirm the phrases you ranked for are still showing up.
• 404 errors from old URLs that slipped past the redirect map. Add redirects as they appear.
• Leads, because calls and form fills are the real scoreboard.

Expect some movement. Pages may wobble for a while before they settle. What you are looking for is a page that drops and keeps dropping, because that almost always points to a specific cause you can fix: a missing redirect, lost content, or a technical setting.

A month after Phoenix launched, we pulled every URL Google had shown for the site over the previous 16 months and checked each one against the new site. Only 4 returned a real error, none of them had earned a click, and we added redirects anyway. That kind of check takes an afternoon and tells you for certain whether anything slipped through.

Google’s own estimate is that for medium-sized websites, it can take a few weeks or more for search results to start showing the new URLs instead of the old ones, and longer for larger sites. A redesign that keeps the same URLs and content usually settles faster than one that changes both. If rankings are still sliding a month or two in, treat that as a problem to diagnose rather than something to wait out.

Done carefully, a redesign can be close to invisible in the numbers. Phoenix’s new site went live on August 25, 2026. Its clicks from Google search were 261 in June, 263 in July, 286 in August and 261 in September, so the relaunch landed without a dent. Two days after Gladwell’s launch, Ellie Baldini, GO Media’s VP of Marketing, wrote: “We are not seeing a negative SEO impact (trendlines are still going up and to the right).” By the August audit, branded searches for the practice were up 77% year over year and local searches were up 13%.

Most of the sites my team builds launch in 0 to 2 months. That timeline only works because the SEO protection is part of the plan from the first week, not a scramble the night before launch. The order I work in looks like this:

1. Take the baseline before touching anything.
2. Decide keep, merge or retire for every page.
3. Build the redirect map while the new site is still on staging.
4. Carry over titles, descriptions, headings and structured data page by page.
5. Run the launch-day checks, then submit the sitemap.
6. Compare against the baseline weekly for the first month.

If you take one thing from this, make it the baseline. Almost every redesign that lost its rankings could have been saved by someone writing down what the old site was doing before it was replaced.

And if you are still deciding whether a redesign is worth it, it helps to know what a small business website should cost before you get quotes. If you would rather have someone handle the whole thing, send me a note. The first reply comes from me.