There is a version of a website redesign that nobody plans for. The new site launches, everyone agrees it looks far better than the old one, and six weeks later the enquiries are down.
Nothing was done carelessly. The design is good. The problem is that the project treated the website as a design object. A working website is also an asset. It has a search history, a set of pages people already find, and links pointing at addresses that are about to stop existing.
This guide is about protecting that while you rebuild. It starts somewhere unusual for a design project: in your analytics.
What a website redesign actually changes
A website redesign changes the structure, content and functionality of a site, not only its appearance. New templates, new navigation, and often a new CMS underneath it.
That is worth stating plainly, because the same word gets used for two very different jobs. One of them carries almost no risk. The other can cost you a year of search progress.
The distinction is not academic. It changes the budget, the timeline, and how much of the project has to be spent on protection rather than design. It is also the first thing we establish before quoting any web design project.
Refresh, redesign, rebuild: three different jobs
| What you are doing | What changes | Search risk |
|---|---|---|
| Refresh | Colours, images, small layout edits. Same pages, same addresses | Almost none |
| Redesign | Templates, navigation, page structure. Some addresses change | Real, and manageable |
| Rebuild | New platform, new structure, new content. Most addresses change | High without a plan |
A website revamp that keeps every page where it is has nothing to protect. A rebuild that changes the address of every page has everything to protect.
Most projects sit in the middle and are scoped as though they sit at the top. That gap is where traffic goes missing.
Start with the pages that already work
This is the step almost every redesign guide leaves out, and it is the one that decides whether you keep your rankings.
Before anyone opens a design tool, find out what your current site already earns. You need two lists.
List one: what search sends you. Open Search Console, go to the Performance report, set the range to the last twelve months, and switch to the Pages view. Export it and sort by clicks. This is a website audit in its most useful form: not a score out of a hundred, but a ranked list of the pages Google currently rewards.
List two: what actually produces enquiries. Your analytics will tell you which pages people were on before they filled in a form, called, or messaged. These are not always the same pages as list one. A blog post can bring the traffic while a service page closes it.
Then combine them. Anything appearing in either list is now a protected page.
That combined list is your content audit, and it is more useful than any tool score. A tool tells you what is technically imperfect. This tells you what is commercially valuable, which is a different question and the one that matters when someone is about to delete pages.
Most sites are surprised here. A handful of pages usually carry the majority of the value, and they are rarely the pages anyone talks about in a redesign meeting. The homepage gets all the attention. A four year old article answering one specific question often earns more.
Turn that list into a redesign brief
A protected page list is only useful if it becomes a rule the design has to respect. Write it into the brief in plain terms:
- These pages keep their address. No exceptions without a redirect plan attached.
- These pages keep their titles and headings. You can improve the wording. Do not replace the subject.
- These pages keep their depth. If a page answers twelve questions today, the new version cannot answer four and be considered finished.
- These pages keep their internal links. If eight pages link to your best guide today, the new navigation has to reach it too.
That last point matters more than it sounds. A cleaner, shorter menu is a common redesign goal, and it quietly cuts the paths that were feeding your best pages.
Everything else on the site is free to change however the design needs. That is the trade: total freedom on most of the site, protected rules on the small part that earns.
This is also the point where UI and UX design decisions and search decisions stop being separate conversations, because the same page structure serves both.
The seven reasons people redesign, and which hold up
Not every reason justifies a rebuild. These four usually do:
The structure cannot hold what you sell now. You added services, markets or product lines, and the navigation was built for a smaller company. This is the strongest reason of the seven, because no amount of editing fixes an information architecture that is wrong.
The platform is blocking your team. Every small change needs a developer. That is a tax you pay monthly, and a rebuild ends it.
It is slow in a way that cannot be patched. Sometimes the code, theme or plugin stack is the problem and there is no fix short of replacing it.
It does not work properly on a phone. Not just narrow, but genuinely awkward to use.
These three usually do not:
It looks dated. Sometimes true, and often a refresh solves it at a fraction of the cost and risk.
Somebody new arrived. A new marketing lead wanting a visible first project is a common cause of an unnecessary rebuild.
A competitor launched something. Their site had a problem. That is not evidence yours does.
When you do not need a redesign at all
Here is the part most agencies will not put in writing.
If the complaint is "we get traffic but no enquiries", that is usually not a redesign problem. It is a message, offer and page problem, and rebuilding the whole site is an expensive way to avoid diagnosing it. Our guide to conversion rate optimization covers the specific fixes, and most of them are edits rather than rebuilds.
The test is simple. Write down the three things you dislike about your website. If all three are about specific pages, wording or forms, fix those and keep your money. That work is conversion rate optimization, and it is a fraction of the cost of a rebuild.
If the three things are about structure, speed or the platform itself, you have a real case for a rebuild.
One honest note from experience: a redesign that launches without addressing those three complaints will not have addressed them. A new coat of paint on the same argument performs the same.
The redesign process, step by step
The web design process for a rebuild runs in this order. These website development steps are deliberately front loaded, because the cheap decisions all live at the start.
1. Audit what exists. The two lists above, plus a full crawl so you have every current address in a spreadsheet.
2. Agree the goal in one sentence. Not "modernise the site". Something you could measure.
3. Fix the structure. Sitemap and information architecture before any visual work. What pages exist, what sits under what, what the menu contains.
4. Wireframe the templates. A website wireframe is a layout with no styling, which is exactly why it is useful. You are agreeing what goes on the page and in what order, while changes are still free.
5. Design the templates, not the pages. Five to eight templates usually cover a site. Designing page by page is how projects overrun.
6. Write the content against the protected list. Not after the design is signed off. Alongside it.
7. Build on a staging site. Blocked from search, shared with the client, tested on real devices.
8. Map every redirect. Covered properly in the next two sections.
9. Launch, then watch. The work does not end at launch, and the first month is when problems are cheapest to fix.
If you would rather have this run for you, that is what our website redesign work is.
What breaks your traffic in a rebuild
Four things account for almost every post launch drop. None of them are about how the site looks.
Addresses changed and nothing catches the old ones. Every link anyone ever built points at an address that no longer resolves. A permanent redirect passes that value to the new page. No redirect means the value is simply gone, along with the ranking it held up. This is the single biggest cause, and it is the easiest to prevent.
Content got shorter. Redesigns love white space. A page that answered a question thoroughly, rewritten as three elegant lines, now answers less than the page that outranks it. Design improved. Usefulness dropped.
The same instinct removes other things quietly. A clear call to action gets replaced by something subtler. A frequently asked questions block gets cut because it looked untidy. Both were doing work.
Pages got merged. Four specific service pages become one beautiful overview. You have just replaced four things that could rank with one thing that ranks for the broadest term only.
The new build is slower. Bigger images, more fonts, heavier scripts. Core Web Vitals measure this, and a redesign is the easiest way to make all three of them worse in one afternoon. If the site is being rebuilt as a JavaScript application, how it renders matters as well, which is what JavaScript SEO covers.
All four are decided before launch day. Site redesign SEO is not a phase you add at the end. It is a constraint you design inside, and it belongs to whoever handles your technical SEO.
If the rebuild comes with a new name or a new domain, that is a bigger job again, and our guide to rebranding a business covers that version.
The pre-launch checklist
This is the site migration checklist for a redesign. Work through it before the new site goes live, not after.
Before launch
- Export every current address. From a crawl and from Search Console. You cannot redirect what you have not listed.
- Mark the protected pages. From your two lists. These are the ones that must not break.
- Map old to new, one by one. Closest matching page, never the homepage.
- Keep redirects to a single step. Old address to new address. Chains leak value and slow crawling.
- Check the content survived. Open your ten best pages side by side, old against new. If the new one says less, it will perform worse.
- Test on a staging site. Real devices, real forms, real payment flows if you have them.
Launch day
- Turn redirects on with the site, not after it.
- Unblock search. Staging sites are blocked on purpose, and forgetting to remove that is a genuinely common disaster.
- Test twenty important old addresses. Each should land on the right page in one step.
- Submit the new sitemap in Search Console.
After launch
- Watch coverage and crawl reports weekly for the first month. Rising errors mean a missing redirect.
- Re-check speed on your top landing pages against the old numbers.
- Watch your protected pages individually. Site totals can look fine while your best page quietly falls.
We ran exactly this process on a website migration for SignArtPro, which is the version of this job where the mapping work is most of the project.
What actually drives the cost
We do not publish a price list, because the honest answer is that four things decide it and they vary enormously.
Template count. Not page count. A site with fifty pages built from six templates is a smaller job than a site with twelve pages that all need bespoke layouts.
Content. Reusing existing copy is cheap. Rewriting the site is a separate project that often costs more than the design.
Platform change. Staying on your current platform is far cheaper than moving. Moving means migration, retraining and a longer test cycle.
Integrations. A brochure site is one thing. A site talking to a CRM, an inventory system and a payment gateway is another.
If a quote arrives before anyone has asked how many templates you need or looked at your current page count, it is a number attached to a guess.
Where redesigns actually stall
Almost never design. Almost never code.
Content. The client owes twelve pages of text and four rounds of photography, and it arrives in pieces over two months. This single item delays more projects than everything else combined.
Too many approvers. Five people with opinions and nobody with a decision.
Scope added late. A booking system mentioned in week six was not in the plan in week one.
The fix for the first one is the content inventory at the start. Knowing you owe forty pages of text in week one is a different conversation from discovering it in week seven.
What is different about this in the UAE
The process is the same everywhere. Three things change locally.
Bilingual is a structure decision, not a translation task. Arabic reads right to left, so a layout designed only for English often has no clean Arabic version. Decide at wireframe stage whether the site is bilingual, because retrofitting it later means redoing templates.
Phones dominate more than most teams assume. The mobile version is the site here, not a smaller copy of it. Design and test in that order.
Directory listings hold your old addresses. Businesses here appear across a long list of local portals, and each one holds a link to a page you may be about to move. Those links are exactly what your redirects protect.
How to judge it after launch
Record these before you go live, because afterwards everybody has an opinion and nobody has a baseline.
| What to record | Why |
|---|---|
| Clicks and impressions by page | The clearest signal that something broke |
| Your protected page list, individually | Site totals hide a single page collapsing |
| Core Web Vitals on top landing pages | Catches a heavier build early |
| Bounce rate and time on page | Tells you if the new layout actually reads better |
| Enquiries and conversion rate | The only number the business cares about |
Speed deserves its own note, because it is the drop people misdiagnose most often. Google treats page experience as part of how it assesses a page, and Interaction to Next Paint specifically measures whether an interface responds quickly when someone taps it. A design that feels smooth on a fast laptop can fail this on a phone.
Then give it a realistic clock. A few weeks of movement while Google reprocesses new addresses is normal. A drop that has not recovered after six weeks is not settling. It is a problem, and it is almost always a redirect.
If you also run organic alongside the rebuild, our guide to how long SEO takes sets out the same curve over a longer window.
The short version of all of this: a redesign is not a fresh start. It is a renovation of something that already has value, and the job is to keep that value while you change the building around it.
