A website redesign doesn't lose you rankings if the URLs that bring in traffic stay the same, or each one has a 301 redirect to the equivalent page, and if the titles, copy and internal links on those pages don't change without a reason. When addresses do change, some fluctuation arrives even when the work is done well: Google's guide to site moves says a medium-sized site can take "a few weeks or more" before Google starts showing the new URLs.
In the UK "website redesign seo" is searched 170 times a month on average, with peaks of 720 searches in September 2025 and 390 in May 2026 (Google Ads data via DataForSEO, October 2026). "Website redesign services" gets 1,300.
And people searching mostly find the same reassurance.
We read the eight pages on Google UK's first page for "website redesign seo", plus the one agency service page ranking for "website redesign". Three use language that reads like a promise: "Rankings Kept, Design Renewed", "relaunch your site without losing rankings", "it will safeguard your rankings". Only one says plainly that nobody can guarantee a position on Google.
And checks after launch? Where a time frame is given, it runs from "a few weeks" to six months, and the agency service page watches the site for 30 days.
How long does a site really take to get back to where it was?
For domain migrations, the heaviest case, a SALT.agency study of 1,052 cases gives a median of 304 days. You'll find it further down, with all the limits of the data.
The mistakes we see most often, when a site comes to us after a redesign, are always the same:
- URLs change "because the new theme generates them that way", and nobody writes the redirect map;
- pages that looked pointless to the designer disappear, including the ones that brought in quote requests;
- the noindex from the staging site goes live along with the new site.
None of the three shows up when the owner looks at the home page on their phone. And that's how almost every redesign gets signed off.
Why should you listen to us?
Because we measure what a redesign puts at stake every month. Macropix, a Milan-based maker of LED walls, took five years to climb from position 88 to 2 on Google Italy for "monitor pubblicitario" (advertising display), and today has 79 of 135 tracked keywords on page one. One old URL without a redirect can throw part of that away in an afternoon.
Sunscape, an Italian solar farm developer, shows the other side. In March 2026 the site was getting 329 organic impressions a month; we built an architecture of pages by region and by August it was getting 5,809. There was nothing to defend there, only things to build. On a site like Macropix's, by contrast, every URL has a history.
In this article:
- what Google sees in a redesign, in John Mueller's words;
- how long the drop lasts according to two studies of 892 and 1,052 migrations;
- the redirect map and the three rules Google reads it by;
- our test on 14 sites ranking on Google UK for website redesign searches;
- what to check after launch, and for how long.
This guide is about a redesign on the same domain. If you're also changing the site's name and address, the route to follow is the one in our rebranding checklist.
Let's start.
What Google sees in a redesign (and why even the layout alone moves rankings)
To Google a redesign isn't an event: there are URLs that change, content that changes and code that changes. If all three stay the same, only what the user sees changes, and the SEO risk is low.
So if I keep the same URLs, I'm safe?
Almost.
In 2020 someone asked Google's John Mueller why their blog had lost rankings after moving from WordPress to Hugo, with the same content and the same URLs. The answer, reported by Search Engine Journal: "Changing the layout of your pages can affect your search results". Changing the layout can move results, because the use of headings, the internal links and the context around each article all change.
A year later Mueller gave the other half of the picture (SEJ, July 2021): nothing on Google's side forces rankings to change after a redesign, but it's best to keep URLs the same as far as possible. If URLs, layout and content all change at once, Google may treat the site as "a new website".
Here's how much each type of change weighs.
| What changes | What Google sees | Risk | What to check |
|---|---|---|---|
| Colours, fonts, decorative images | Almost nothing, apart from page weight | Low | Core Web Vitals of the pages with the most clicks |
| Layout and templates (menu, headings, footer) | A different heading hierarchy and different internal links | Medium | H1, H2 and menu links on the main pages |
| Copy rewritten or shortened | New content to reassess | Medium to high | Queries per page in Search Console, before and after |
| New URLs (slugs, folders, language) | A site move with URL changes | High | 301 redirect map, one by one |
| New CMS or technology | Different code and rendering, often new URLs too | High | HTML served without JavaScript, canonicals, robots |
| Pages deleted or merged | Addresses that disappear | High | 301 to the equivalent, otherwise 404 or 410 |
The row people underestimate most is the second. A redesigned menu that drops ten items drops ten internal links from every page on the site. For Google that's a change of site architecture, even if no URL has moved, and the pages that lose those internal links get crawled and assessed differently.
How long the drop lasts after a redesign (the numbers nobody quotes)
Google doesn't give a figure for redesigns. For moves with URL changes it talks about ranking fluctuations while the site is recrawled and reindexed, and a few weeks or more for a medium-sized site. The broadest data available covers domain migrations, the worst case, and comes from two studies by Dan Taylor.
| Time since migration | Sites back to their previous organic traffic |
|---|---|
| 30 days | 5.1% |
| 90 days | 22.8% |
| 180 days | 35.4% |
| 1 year | 59.5% |
| 2 years | 83.5% |
| 3 years | 86.1% |
Median: 304 days. After three years, 13.9% of sites still hadn't returned to their starting level.
The same author's earlier study, published in Search Engine Journal in January 2025 and based on 892 migrations measured with Ahrefs traffic estimates, gave an average of 523 days, with 17% of sites never getting back to their previous traffic after 1,000 days.
But I'm not changing domain, I'm just rebuilding the site.
Right. And that's why these numbers should be read as a ceiling, not a forecast. A redesign on the same domain doesn't have to transfer authority to a new address, and if the URLs don't change there's no move to process at all. Both studies use international samples; we haven't found one limited to UK sites.
One thing does apply to you, though: 30 days of monitoring isn't much. In the SALT sample, 5.1% of sites were back to their previous traffic by day 30. If your redesign quote includes a month of checks, at the end of that month you still know very little.
Before the redesign: a snapshot of the site you have (which you can't take afterwards)
The "before" snapshot is a list of every URL that exists today, with how much traffic and how many links each one brings in. You can't rebuild it after launch, because the old site no longer exists.
- Crawl the current site with a crawler. The free version of Screaming Frog handles up to 500 URLs per crawl, enough for many company sites. Save PDFs and images that get visits too.
- Export the Performance report from Search Console by page and by query, over the longest period available.
- Pull your backlinks with Ahrefs, Semrush or the Links report in Search Console, and mark the pages linked from the most domains.
- Note the title, H1 and meta description of the pages that bring in clicks: they're the first things a new theme rewrites on its own.
- Save the sitemap with the old URLs. On launch day it becomes the list to check one by one.
On Macropix the query "ledwall" (LED wall) alone collected 67,560 impressions in twelve months. Knowing which page collects them, with what title and which inbound links, is the difference between a redesign that protects them and one that only notices when enquiries drop.
The snapshot is useful for something else too. In the SEO audit of that site we found duplicates created by WPML and spikes of pages crawled but not indexed. Faults like these do no harm while the site stands still, but in a redesign they're carried over to the new site exactly as they are if nobody has put them on the list. We covered them in our guide to common URL problems.
The redirect map: one row for every URL (not one rule that sends everything to the home page)
The redirect map is a sheet with two columns: the old URL and the equivalent new URL. It's the part of a redesign that Google's documentation describes most precisely, and there are three rules.
- Every old URL is mapped to its equivalent. A service page to a service page, an article to the rewritten article.
- No mass redirects to the home page. Google asks you not to send lots of old URLs to a single irrelevant destination: it confuses users and may be treated as a soft 404, in other words as an error page.
- Pages with no equivalent return a 404 or 410. A 404 error on a deleted page is the correct response, not a problem to hide.
The rest is execution. The redirects are 301s, configured on the server, and stay in place as long as possible: Google says "generally at least 1 year". Every old URL points straight to the final address, because Googlebot generally follows up to 10 redirects and every extra hop is wasted crawling. The internal links on the new site point to the new URLs, not the old ones: the same guide asks for this.
What if a redirect is wrong and I fix it later?
The browser may already have cached it. Until the map is stable, it's worth sending the 301s with a Cache-Control of a few hours: it's a detail that almost never appears in migration checklists.
Got a redesign coming up?
An old URL without a redirect is lost along with its whole ranking history, and you usually find out months later, when enquiries drop. Before launch we compare the redirect map with the pages that bring you clicks and enquiries today, one by one.
Book a strategy consultationOur test on 14 sites ranking on Google UK for website redesign (3 take two hops to reach the home page)
On 10 October 2026 we took the sites on Google UK's first page, on mobile, for "website redesign" and "website redesign seo": 16 unique domains after leaving out YouTube and Reddit. Eleven are agencies and web designers that sell redesigns; the other five are software companies, including Webflow and Shopify. On each one we made two requests from our server: the home page at the http address without www, following the redirects one at a time, and a page that doesn't exist.
| Check | Result |
|---|---|
| Domains on page one for the two queries (excluding YouTube and Reddit) | 16 |
| Not checkable (return 403 to automated requests) | 2 |
| Sites checked | 14 |
| Home page reached with a single redirect | 11 of 14 |
| Home page reached with two chained redirects (http, then https, then www) | 3 of 14 |
| Non-existent page returning a 404 straight away | 10 of 14 |
| Non-existent page redirected first (trailing slash removed), then 404 | 4 of 14 |
Two hops aren't a disaster: Googlebot follows up to ten. But they show how chains are born. The http-to-https rule and the non-www-to-www rule were added at different times, one on top of the other, without being rewritten as a single rule. It's the same mechanism that, after two redesigns, produces chains of three or four hops on the old URLs of inner pages, where nobody looks. One site in the sample also used a 302, a temporary redirect, to move visitors from http to https.
None of the 14 sends a missing page to the home page, the case Google describes as a possible soft 404. Four add a redirect before the 404, to strip the trailing slash.
So are they all to be avoided?
No. All fourteen end up returning a 404 and none has serious errors. The point is a different one: these checks take ten minutes with a crawler in list mode, and they need doing on your site before and after launch, whoever rebuilds it.
One change at a time (the rule in Google's own guide)
Google's guide to site moves says it plainly: "Plan your changes to your site one after the other, not everything at the same time". A typical redesign does the opposite: new CMS, new URLs, rewritten copy and new design, all on the same day.
If traffic drops, which of the four changes caused it?
You don't know. And you don't know if it goes up either.
Mueller added a detail that makes it even trickier: sometimes a redesign goes live in the very days of an algorithm update, and at that point separating the causes becomes almost impossible. The order we follow:
- Technology first (CMS, hosting, theme) with the same URLs and the same content. The new design can go out at this step: it's the change that weighs least.
- Then the URLs, only if they really have to change, with the redirect map already tested.
- Then the content, in groups of pages, starting with those that bring in the least traffic.
It's the same principle we applied on Sunscape, in a different area: tracking first, then the page architecture, then the budget. Reversing the order would have produced traffic on pages that couldn't be measured.
A note on the staging site. It needs to be locked behind a password or a noindex, and that noindex must not reach production. It sounds obvious. On its own it's enough to make a site disappear from Google, and we come across it more often than you'd think.
After launch: what to check on day one, in the first month and at 90 days
Monitoring after launch is there to find mistakes in the map while they're few and recent. It splits into three moments.
Launch day.
- robots.txt and meta robots don't block anything;
- the old URLs saved from the sitemap, run through the crawler in list mode, return a 301 in a single hop to the right page, or a 404 or 410;
- the sitemap with the new URLs has been submitted to Search Console;
- every page has a canonical pointing to itself.
The first four weeks. In the Pages report in Search Console, look at three entries: "Not found (404)", "Page with redirect" and "Crawled - currently not indexed". Compare clicks per page with the same period before launch, and fix the map wherever an important page has lost traffic for no reason.
At 90 days. The comparison is done query by query, on the pages that used to bring in the most clicks. If one page loses rankings on a query and the others don't, the cause is on that page: title, H1, copy, internal links. If the whole site drops together, the diagnosis follows a different order, the one in our guide to a drop in website traffic.
This is the work we do on website development and SEO consultancy projects: SEO joins the project when the wireframes are drawn and stays until sign-off, rather than turning up the day after launch to count the damage.
Website redesign and SEO FAQs
It depends. If the URLs stay the same and the pages that bring in traffic keep their titles, copy and internal links, the risk is low, although Google confirms that even a change of layout can move rankings. If the URLs change, a few weeks of fluctuation is normal even with redirects done properly. Lasting losses usually come from old URLs without redirects or from pages deleted without an equivalent.
No. A redesign changes the website: design, structure, sometimes the technology. A rebrand changes the company name and often the domain, and on the SEO side it also means keeping hold of the old name, which people go on searching for years.
It depends on the traffic and links the old URLs still receive, but Google says at least one year and recommends keeping them as long as possible. An old URL with backlinks from other sites is best never removed.
No, unless the current ones have a structural problem, such as a date in the address or parameters instead of words. Every changed URL is one more redirect to maintain and one more move for Google to process.
It depends how much changes, but a month is too short: in the SALT study of 1,052 domain migrations, only 5.1% of sites were back to their previous traffic after 30 days. On a redesign with new URLs, 90 days of page-by-page comparison is the minimum to tell whether the redirect map has held.
Website redesign and SEO in brief
If URLs don't change the risk is low, but the layout alone can move rankings.
Every old URL gets a 301 to its equivalent, never one rule that sends everything to the home page.
For domain migrations the median recovery is 304 days: a month of checks isn't enough.
One change at a time: technology first, then URLs, then content.
In our test, 3 of the 14 sites ranking for website redesign take two redirect hops to reach their home page.
Something the industry rarely says: the SEO of a redesign is decided in the contract, not in the design. If the quote describes pages, templates and rounds of revisions but doesn't say who writes the redirect map, who checks it and when the work counts as finished, nobody does that part. The acceptance test fits in one line: every URL that has had at least one click from Google in the last twelve months returns a 200, or a 301 to an equivalent page. Put it in the contract before you sign.
