A website redesign is where plenty of businesses quietly lose the traffic they spent years earning. This article is about not being one of them. The new site looks better, yet enquiries fall anyway. There was no hack and no penalty. Instead the addresses moved, a few pages vanished, and nobody told the search engines.
What protects you is dull work. You need an inventory, a redirect map, and a careful walk through the old addresses on launch day. None of it is glamorous. Still, it is cheaper than earning the same rankings a second time.
Why a website redesign loses traffic
In practice the damage usually comes from three things, and only one is technical in any interesting way.
First, addresses change. A page that ranked at one address now answers nothing, so the ranking has nowhere to land. Second, content shrinks. Designers turn long pages into short ones, and the words that made a page useful leave with the clutter. Third, the new build runs slower or blocks the crawler.
Notice what is absent from that list. The design itself rarely causes the drop. A prettier site with the same addresses, the same words and the same speed usually keeps what it had.
Measure what you have before anyone draws anything
You cannot protect what you never wrote down. So the first stage of a website redesign is a dull inventory, and it happens before anybody opens a design tool.
The pages that already earn
Pull a list of every page that brings visits from search. Then pull a second list of every page that brings enquiries. Those two lists rarely match. A blog post often brings the visits while a service page brings the work.
Mark both lists as untouchable. If the rebuild wants to drop one, that becomes a decision somebody makes on purpose. It is no longer a casualty of a tidier menu.
Every address you already own
Export your addresses from the sitemap, from analytics and from the content system itself. Then merge the three into one list. Old pages nobody links to internally can still hold links from other sites, so they still carry weight.
Keep that list somewhere obvious. It becomes the checklist you test against after launch, and without it you are guessing.
What a website redesign can change safely
Not every change carries the same risk. Sorting them early stops the whole project feeling dangerous.
| What the rebuild changes | What it does to your traffic |
|---|---|
| Colours, fonts and layout | Almost nothing on its own, so change freely |
| The words on a page | Depends, because the topic is what ranked |
| The menu and how pages sit together | Little, if the addresses stay put |
| The address of a page | Everything, unless you redirect it |
| Removing a page | Loses whatever that page earned |
| The platform underneath | Risky, since templates rename pages |
Treat the bottom three rows as the project's risk list. Every address you plan to change, every page you plan to drop and any platform move goes on it, with a name beside each one. The top three rows need no list at all, since appearance was never what ranked.
Keep the addresses, and map the ones you cannot
The safest website redesign keeps every address exactly as it stands, trailing slash included. New design, same doors. Nothing has to move just because the pictures did.
Sometimes an address genuinely has to change. When it does, redirect the old one to its closest replacement, one at a time. A permanent redirect is the clearest signal you can send that the page moved rather than vanished.
Avoid the shortcut of pointing everything at the homepage. A homepage does not replace a product page, so the redirect answers nobody's question. Readers who land in the wrong place simply leave. If there is truly no replacement, let the old address answer honestly as gone instead.
A worked example, with round numbers picked to make the arithmetic easy rather than to describe a typical site. Say the old site holds 200 addresses and only 40 of them bring any visits at all. Suppose ten of those 40 bring most of the visits. Redirect those ten correctly and get the rest wrong, and the damage is small. Get those ten wrong and the rest right, and you have lost most of your traffic. So order the list by value before anybody starts mapping.
The parts that quietly break
Rankings get all the attention during a website redesign. Meanwhile the smaller failures cost enquiries on day one.
Speed
New sites usually carry more scripts, bigger images and a heavier theme. A slow page loses people before it finishes loading. That costs enquiries whatever search engines make of it. There is more in our piece on what actually slows a site down.
Layout on a phone
Check your own analytics for the phone share rather than assuming it. Whatever the number says, designs get approved on a large monitor and used on a handset. So try the real thing on a real handset before launch. Our note on designing for every screen covers what usually goes wrong.
Tracking and forms
Analytics tags, conversion tracking and form handlers all live in the old templates. They rarely survive a rebuild untouched. So send a real enquiry through the new site and confirm it reaches a real inbox.
Planning a website redesign, stage by stage
The order matters more than the calendar. Each stage exists to make the next one cheaper to undo.
- Inventory first, so you know what you are protecting.
- Agree what the new site must do, in one page of plain sentences.
- Structure and addresses next, because everything else sits on them.
- Design after that, on the real page structure rather than on invented content.
- Build, then move the content across page by page.
- Finally, test on a staging copy that search engines cannot see.
That sequence is the one we set out in the stage by stage build process, applied to a site that already exists. The inventory at the front is the difference.
Scope is the hardest part to hold. A project that also adds a booking system, a customer login and an app is no longer a redesign, and our plain guide to building web and mobile products is the better starting point for that. Keep the two apart, so a delay in one does not hold the other hostage.
Choosing the platform without regret
A website redesign is often the moment a business changes platform. That can be reasonable, though it puts two hard jobs inside one project.
Ask who will edit the site next year. If the answer is a member of staff rather than a developer, that answer should decide the platform. Our guide to picking a system your team will use goes through the trade offs.
Whatever you pick, check how it builds addresses before you commit. Some platforms insist on their own pattern and quietly rename every page you import.
Launch day, in order
Launch is a checklist rather than an event. Work through it in this order.
- Publish, then immediately ask the search engines to crawl the important pages.
- Walk your old address list and confirm each one lands somewhere sensible.
- Submit the new sitemap, since the old one now describes a site that no longer exists.
- Confirm the site is visible to search engines at all, because staging blocks often survive the move.
- Send a test enquiry through every form and watch it arrive.
Do this on a quiet day, while the people who built the site are still around. Friday evening is a poor choice.
The first weeks after a website redesign
Expect movement. A dip while search engines recrawl and re-judge the new pages is ordinary.
What matters is the shape of the dip. A shallow wobble that recovers is normal. A steep fall that keeps falling means something broke, and the address list is where you look first.
Watch enquiries as well as visits. Traffic can hold while enquiries collapse, which usually points at a broken form rather than at search.
When a website redesign is the wrong answer
Plenty of sites do not need one. The complaint is usually that the site looks dated, yet the real problem often sits elsewhere.
If the site is slow, fix the speed and leave the design alone. When nobody finds the site, the gap is content and search rather than layout. Visitors who arrive and leave are telling you about the offer, so test the wording on one page before rebuilding the lot.
A rebuild is also a poor way to fix a business problem. If the enquiries are the wrong sort, a new theme will not change who arrives.
Frequently asked questions
How long does a website redesign take?
It depends far more on content than on design. Drawing pages is quick. Deciding what each page should say, gathering the words and moving them across is the slow part, and it needs somebody inside the business. The technical work usually finishes before the writing does. A useful way to shorten it is to freeze the page list early, then let design and content run alongside each other rather than one after the other.
Will a website redesign hurt my Google rankings?
Not by itself. Rankings attach to your addresses and to the words on the page, not to the colours, so a new look on the same addresses is close to neutral. What hurts is everything else a rebuild tends to carry along: renamed pages, deleted posts, a slower build, and a staging block left switched on. The more of those a website redesign changes at once, the harder it gets to tell which one caused a drop.
Should I keep my old page addresses?
Yes, wherever you can. An address that has ranked for years carries links and history, and keeping it costs nothing at all. Change one only when the old address is genuinely wrong, such as a name that no longer describes the page. When an address must move, redirect the old one straight to its closest replacement, one at a time. Never point a batch of unrelated pages at the homepage.
What should happen to old blog posts?
Keep the ones that earn visits or links, then decide on the rest one by one. Deleting a whole archive to make the new site feel tidy throws away the pages that often bring the most search traffic. If a post is out of date, update it and keep the address rather than removing it. Where two posts cover the same ground, merge them into one and redirect the weaker address to the survivor.
Can I change my CMS during a website redesign?
You can, though it puts two hard jobs inside one project. A platform move changes where pages live, how addresses form and how staff edit the site. Check the address pattern before you commit, because some platforms rename every page they import. If the current system works and only the design looks tired, keep it. When staff cannot update the site without help, the platform is the real problem and the design is a side issue.
How do I know my site needs rebuilding?
Look for a reason that is not about taste. Good reasons include a site staff cannot update, a layout that fails on a phone, a structure that cannot hold the services you now sell, and code nobody can maintain safely. Poor reasons include boredom and a competitor launching something shiny. If the real complaint is that nobody finds the site, that is a search and content problem, and a new theme will not solve it.
What is the biggest mistake in a website redesign?
Launching without a list of the old addresses. You can fix almost anything else afterwards, yet nobody can redirect pages they never recorded, and by then the old site has gone. Export every address before the rebuild starts, from the sitemap, from analytics and from the content system. Then use that same list as the launch day checklist. It is dull work, and it prevents the one failure that is genuinely hard to undo.
How do I check a website redesign worked?
Compare like with like, and give it time. Watch visits from search, enquiries and the pages that used to earn, then measure each against the same period before launch rather than against last week. Crawl the site and confirm no important page returns an error. Send a test enquiry through every form. A website redesign that keeps its traffic and its enquiries has done its job, whatever anybody thinks of the colours.
Can I rebuild a website in stages?
Often yes, and it is usually safer. Rebuilding section by section keeps most of the site stable while you learn how the new templates behave. Start with a part that earns little, such as a rarely visited section, then move to the pages that matter once the pattern works. The catch is that two designs live side by side for a while, so agree in advance how long that is acceptable.