What a redesign does to your rankings, and how to keep them
A redesign changes your URLs, your internal links and your page templates in the same week. Here is what carries across, what quietly breaks, and the checks that decide whether traffic dips for a fortnight or for a year.
Every redesign is a controlled demolition of the thing search engines had already worked out. New URLs, new templates, new internal linking, new copy, usually all in the same week. Google has to re-learn a site it already understood.
Done properly that costs you a fortnight of noise. Done badly it costs a year. The difference is almost never the design. It is whether the old addresses still resolve, whether the new pages say what the old ones said, and whether anyone checked before the DNS changed.
This is the checklist we run on a rebuild, and the failure modes we see when it gets left until after launch. It is the same work as a platform migration, minus the part where the domain changes.

What a redesign actually changes
Redesign is one word for four separate changes, and search treats each of them differently.
Four things move at once:
URL structure: Every page either keeps its address or gets a new one. This is the change that decides how bad the rest of it gets.
Templates and markup: Headings, internal links, image alt text and structured data are all regenerated by the new build. None of it travels on its own.
Content: Copy gets rewritten, pages get merged, thin pages get deleted. Rankings follow the words on the page, not the layout around them.
Rendering: A new platform serves pages differently. Server-rendered, client-rendered, or a mixture, and that determines what a crawler actually receives.
You can change any one of these on its own and recover in days. Changing all four in one launch is what turns a redesign into a six-month recovery, because once traffic moves you have no way of telling which change moved it.
If you take one thing from this page: change the design or change the URLs, but try not to change both in the same week.
What carries across and what breaks
Most redesign advice is written in the optimistic tense. Here is the honest version, using the same four states we use on the migration guides, because a redesign is a migration that happens to keep the domain.
| What you own | What happens to it | Status |
|---|---|---|
| Domain authority | It sits with the domain, not with the design. Nothing about a rebuild resets it, as long as the domain itself does not change. | Carries over |
| Page URLs | They survive only if you keep them, or map every old one to its closest new equivalent. Unmapped URLs are the single biggest cause of traffic loss after a launch. | Rebuilt |
| Backlink value | Other people's links point at your old addresses and nobody is going to update them. A 301 passes their value along. A 404 throws it away. | Rebuilt |
| Internal links and anchor text | Regenerated by the new navigation and the new body copy. The old link paths are gone unless the new build reproduces them deliberately. | Rebuilt |
| Plugins, embeds and integrations | There is rarely a like-for-like version on a new platform. You get a native feature that does the same job with different markup, which means the structured data around it has to be rewritten. | Replaced |
| Rankings for pages you delete | A page removed with no equivalent to redirect to loses its positions and the links pointing at it. Sometimes that is the right call. It is never a free one. | Doesn't move |
Read that table as a work list. Everything marked rebuilt is something a person has to do before launch, and everything marked replaced is a decision somebody has to make on purpose rather than discover afterwards.
Not sure which of your pages actually earn anything?
Send us the URL. We will pull the pages bringing you visits and links, and tell you which ones the rebuild is not allowed to weaken.
Someone opens your site and records what they find. No automated score, no templated PDF with your logo dropped into it.
Conversion leaks, search and AI visibility, speed, design and trust, copy and positioning, and the order we would fix them in.
You get the walkthrough whether or not we ever work together. No pitch, no obligation, and it lands within 2 business days.
Sample site audit
62/100
18 issues foundPerformance score
Top findings
- Largest Contentful Paint over 4s
- Missing meta descriptions on 9 pages
- No structured data for search
Why traffic drops after a launch
Three causes account for nearly all of it, and none of them are visible in a design review.
The URL changed and nothing said so
A page that ranked at one address now lives at another, and the old address returns a 404. Every link pointing at it stops counting, and the position it held goes to whoever is second. This is entirely preventable and it is still the most common cause we see.
The new page says less than the old one
Design-led rebuilds cut copy. That is usually the right instinct for a homepage and the wrong one for the service page carrying half your organic traffic. A page that ranked on nine hundred words of detail, and now carries two hundred words and a large photograph, has stopped answering the question it ranked for.
The content moved behind JavaScript
If the new build renders your body copy on the client, a crawler may fetch a shell and come back for the content later, or not at all. Answer engines are stricter than Google here: several read the HTML they are served and nothing more. Check what your page contains with JavaScript disabled. If the copy is not there, it does not reliably exist.
Lost positions: The old URL 404s, so its rankings and the links pointing at it stop counting on the day you launch.
Split intent: A rebuild splits one strong page into two weaker ones, and search picks the wrong one, or neither.
Wasted crawl: An old sitemap left in place sends crawlers to hundreds of dead addresses, so the pages that do exist get looked at less often.
Redirects do most of the work
A 301 tells Google the page moved permanently and that the new address should inherit what the old one earned. Google's own documentation on redirects is worth ten minutes of anyone's time before a launch. It is the hour of work that decides most of the outcome, and it is usually handed to whoever has a free afternoon on launch day.
Map every URL that has ever earned a visit, a link or an impression. Not the pages you remember building. The pages Search Console says exist.
The map itself is two columns, old path and new path. On Wix Studio you upload it as a CSV in the redirect manager rather than editing server files, so what you are producing looks like this:
Old URL,New URL
/webdesign,/services/wix-studio-website-design
/marketing,/services/seo-aeo-ppc
/about-us,/about
/blog/2019/07/why-we-rebranded,/blog/why-we-rebranded
/shop/product/blue-widget-2,/shop/blue-widgetFour rules decide whether that file does its job:
- Point each old URL at its closest equivalent, not at the homepage. An irrelevant redirect gets treated as a soft 404 and passes nothing.
- One hop, never three. Chains slow crawling and lose value at each step, so send the old address straight to the final one.
- Keep them permanently. A redirect you tidy away after six months is a 404 with a delay on it.
- Let genuinely dead pages die. A page with no traffic, no links and no equivalent should return a 404 and fall out of the index. That is a clean signal, not a mistake.
Redirecting every old URL to the homepage is not a redirect strategy. Google reads it as a soft 404, which means you have done the work and kept none of the benefit.
The pre-launch checklist
All of this happens before the DNS change. Afterwards, half of it turns into forensics.
- Export every URL Search Console has an impression for, plus everything in your current sitemap. That is your inventory, and it will be longer than you expect.
- Pull the top fifty pages by clicks and the top fifty by referring domains. Those are the pages the rebuild is not allowed to weaken.
- Write the redirect map. Every old URL gets a destination or an explicit decision to let it 404. No blanks.
- Check content parity on those hundred pages. If the old version answered something the new design dropped, put it back before launch, not after the ranking goes.
- Keep the title tags and H1s that already work. A rebuild is the worst possible moment to rewrite the title of your best-performing page for stylistic reasons.
- Rebuild the structured data. Organization, Breadcrumb, FAQ and Article markup do not travel with the copy, and nobody notices they are missing until the rich results stop appearing.
- Confirm the staging site is blocked and the live site is not. A noindex left on at launch is a week of nothing, and it is the easiest disaster on this list to cause.
- Generate the new sitemap, submit it, and leave the old one in place until the redirects have been crawled.
Rebuilding on Wix Studio?
We write the redirect map in week one, from Search Console data, before anyone opens a design tool. It is the difference between a fortnight of noise and a quarter of recovery.
The first 30 days
Launch day is the start of the measurement, not the end of the project.
Watch four things, in this order:
Coverage: In Search Console, watch the Pages report for a spike in Not found or Page with redirect. A 404 spike on day two means the map has holes in it.
Crawl rate: Crawl stats should climb as Google re-reads the site, then settle. A flat line means it has not noticed anything changed, which is its own problem.
Impressions before clicks: Impressions recover first. If they come back and clicks do not, the titles and descriptions got worse, not the pages.
The hundred pages you protected: Track them individually rather than as a site average. A site-wide number will happily hide one important page falling off a cliff.
Then change nothing for two weeks unless something is plainly broken. A rebuild is already a large enough variable. Stacking fixes on top of it means you will never learn which one worked.
Reading the drop
Not every dip is a problem, and the shape of the curve tells you which kind you have.
| What you see | What it usually means | What to do |
|---|---|---|
| Down 10 to 20% for a fortnight | Ordinary re-crawl noise while Google re-evaluates pages it already knew. | Nothing. Wait it out. |
| A cliff on day one, then flat | Something is blocking indexing: a noindex, a robots rule, or a sitemap pointing at the wrong host. | Check robots.txt and the index status of your top pages today, not next week. |
| A steady slide over six weeks | Content loss. The new pages say less than the old ones did. | Compare old and new versions of your top twenty pages, word count included. |
| One page collapses, the rest hold | That URL changed without a redirect, or two new pages now compete for its query. | Find the old address, confirm it 301s, and merge whichever pages are competing. |
| Impressions hold, clicks fall | Titles and meta descriptions were rewritten during the rebuild. | Put the old ones back on the pages that lost the most, and leave the rest alone. |
Three questions come up on every one of these calls, so they are worth answering here.
Two to four weeks for a clean rebuild on the same domain. Six to twelve if the URLs changed and the redirect map was incomplete. Longer than that is not recovery, it is a different problem wearing a redesign costume.
All at once, provided the redirects are right. A staged launch leaves two versions of your site in the index competing with each other, which is a harder thing to fix than a single clean cutover.
No. Search engines index HTML, not the tool that produced it. What hurts is what the move changes along the way: the addresses, the copy, and how the page renders.
How we run a rebuild
The short version is that the search work happens before the design work, not after it.
- The URL inventory and the redirect map are written in week one, from Search Console data, before a single page is designed.
- Every page that earns traffic gets a parity check. The new version has to answer what the old version answered, in as many words as that takes.
- Titles, H1s and structured data are content the client signs off, not something the build emits while nobody is looking.
- Indexing is verified on launch day rather than assumed: staging blocked, live open, top twenty URLs resolving or redirecting.
- We watch Search Console for thirty days after launch and fix what actually moves, rather than everything at once.
It is the same process behind our platform moves, whether a site is coming off WordPress or Squarespace. A redesign is the easier version of it, because the domain stays where it is.
A redesign should make you more money, not cost you the traffic you already had. Those two things only pull against each other when the search work gets left until after the launch party.
Want this done on your own site?
We build and move websites on Wix Studio, and build custom when you outgrow it. Tell us what you're working with.
Questionspeople ask
If yours isn't here, ask it on a call. We answer in plain terms.
Not if the URLs are mapped and the content survives. Almost every measurable loss comes from unredirected pages or from a rebuild that cut the copy a page ranked for, rather than from the design itself.
Keeping them is the lowest-risk option and it is free. If you change them, every old URL needs a 301 to its closest equivalent, and you should plan for a few weeks of re-crawl noise while Google works through them.
Two to four weeks for a clean rebuild on the same domain. If URLs changed and the redirects were incomplete, six to twelve weeks is more realistic, and the gap between those two outcomes is the redirect map rather than the design.
You can, and Google provides a Change of Address tool for it, but it doubles the number of things that can go wrong and makes any drop much harder to diagnose. If you have the choice, move the domain first, let it settle, then rebuild.
That the live site is indexable, that robots.txt is not blocking anything it should not, that your top twenty URLs either resolve or redirect correctly, and that the new sitemap is submitted. Those four cover most launch-day disasters.
It can be, but it is rarely the main one. Fix indexing and content parity first, then speed. A fast page that Google cannot find is worth nothing, and the reverse is at least worth something.
A free audit that actually tells you something
Drop your site in and we'll review it by hand, then send a short video walkthrough of the highest-impact fixes. No auto-generated PDF, no obligation.
Someone opens your site and records what they find. No automated score, no templated PDF with your logo dropped into it.
Conversion leaks, search and AI visibility, speed, design and trust, copy and positioning, and the order we would fix them in.
You get the walkthrough whether or not we ever work together. No pitch, no obligation, and it lands within 2 business days.
Sample site audit
62/100
18 issues foundPerformance score
Top findings
- Largest Contentful Paint over 4s
- Missing meta descriptions on 9 pages
- No structured data for search
Keep reading
More on the same thing, or the next thing worth knowing.
Sources
Primary documentation behind the claims above, so you can check them rather than take our word for it.
Google Search Central: Redirects and Google Search
How Google treats permanent and temporary redirects, and why an irrelevant redirect is read as a soft 404.
Google Search Central: Site moves with URL changes
Google's own sequence for changing URLs, including keeping the old sitemap available while redirects are crawled.
Google Search Central: Consolidate duplicate URLs
Canonicalisation, and what happens when two new pages compete for what used to be one page's query.
Google Search Central: Introduction to robots.txt
What robots.txt can and cannot block, which is the first thing to check after a launch-day traffic cliff.
Google Search Central: Intro to structured data markup
Why Organization, Breadcrumb and FAQ markup has to be rebuilt by the new templates rather than carried over.
Wix Help Centre: Importing and exporting URL redirects with a CSV file
How the redirect map is actually uploaded on a Wix or Wix Studio site.


