Website Redesign, Rebuild or Staged Modernisation: Which Do You Need?
Redesign, rebuild or modernise a legacy website in stages? A decision table, a five-stage roadmap and how to handle URLs, PHP end of life and SEO risk.
Quick answer
A redesign changes a website's visual design and UX while generally keeping its technical foundation; a rebuild replaces that foundation, sometimes with a redesign alongside; incremental improvement makes smaller, targeted changes. For a legacy site whose platform is the problem but which still earns traffic or leads, there is a fourth option: modernise in stages. Stabilise it, upgrade in place, rebuild section by section behind the existing URLs, change URLs once if you must, then switch off the old system. The right approach depends on whether the problems are visual, architectural or isolated, how large the site is, and whether the current platform can still be kept secure. Whichever you choose, answer these 15 research and analytics questions first.
Key takeaways
- Diagnose first: visual and UX problems point to a redesign; platform and architecture problems point to a rebuild.
- Staged modernisation lowers risk for sites with real traffic and many integrations, but it runs two systems for a while.
- Stage the back end, not the URL change. Google recommends moving all URLs at once for small and medium sites.
- PHP 8.2 reaches end of life on 31 December 2026, which makes this a good time to check your runtime.
- Small, simple sites are often better rebuilt in one go.
Three distinct approaches
These terms get used loosely, which causes real confusion when planning a project. Being precise avoids overpaying for a full rebuild when a redesign would solve the problem, or underinvesting in a redesign when the real constraint is technical.
A redesign updates visual design, often UX and information architecture, and sometimes content, while the underlying technology, CMS and core architecture stay in place. It fits when the problems are about how the site looks and is used, not what it is built on.
A rebuild replaces the technical foundation: the framework, CMS, hosting architecture or all three. It may include a redesign or keep the existing look. It fits when the current foundation is itself the constraint.
Incremental improvement makes targeted changes, such as improving a specific page, fixing a performance bottleneck or updating one flow. It works when problems are isolated rather than pervasive.
The decision framework
| Symptom | Likely need |
|---|---|
| Outdated look, but the site works fine technically | Redesign |
| Poor UX or navigation, technical foundation is sound | Redesign |
| Slow performance rooted in the platform or architecture itself | Rebuild, possibly staged |
| CMS the team has clearly outgrown | Rebuild, possibly staged |
| Significant technical debt blocking new features | Rebuild, possibly staged |
| Unsupported runtime or CMS that can still be upgraded | Upgrade in place first, then decide |
| One or two specific pages or flows underperforming | Incremental improvement |
| Both an outdated look and a limiting technical foundation | Rebuild, often with a redesign included |
Design problems vs technical problems
Design and UX problems typically show up as usability complaints, weak conversion or a visually dated feel. Technical problems show up as slow performance the design cannot fix, difficulty adding features, or a CMS that has become a bottleneck for the team. Both are often present at once, which usually points towards a rebuild that includes a redesign. For the symptoms in more detail, see signs your website needs a redesign.
Technical debt, the shortcuts and workarounds accumulated over time, is a common reason a site feels hard to maintain even when it does not look broken. If every new feature takes disproportionately long, or the team routinely works around platform limits, that points to a rebuild rather than a surface redesign.
Modernising a legacy website in stages
Martin Fowler's strangler fig article describes replacing a legacy system gradually: new components take over pieces of the old one until it can be retired. His view is that "the reduced risk and earlier value from the gradual approach outweigh its costs". Our guide to the strangler fig pattern covers it for software systems; the roadmap below applies it to a business website.
The key adjustment for websites comes from Google's guidance on site moves. For small and medium sites, Google recommends "moving all URLs on your site simultaneously instead of moving one section at a time"; larger sites can move by section. So stage the back end, not the URL change: rebuild piece by piece behind the URLs you already have, and make any URL changes in one planned cutover.
| Stage | What happens | Done when |
|---|---|---|
| 1. Stabilise | Take full backups and test a restore. Inventory pages, URLs, forms, plugins, integrations and scheduled jobs. Record a baseline from analytics and Search Console. Check the PHP, CMS and plugin versions against their support dates and apply security updates. | You can restore the site, you know everything it does, and known vulnerabilities are patched |
| 2. Upgrade in place | Move to a supported PHP version and current CMS core. Remove dead plugins and replace abandoned ones. If changing hosts, keep URLs identical (a move without URL changes). | The site runs on supported software with no visible change to visitors |
| 3. Rebuild section by section | Put a routing layer (a reverse proxy, CDN or edge routing rules) in front of the site. Build one section, such as the blog or pricing pages, on the new stack and route its existing URLs to it. Repeat until the old system serves nothing important. | Each section is live on the new stack under its existing URLs, with analytics and forms working |
| 4. One URL cutover | If you must change URL structure or domain, do it once: map every old URL to its new one, use permanent (301 or 308) redirects, update internal links and sitemaps. | Every old URL returns a permanent redirect to the right page |
| 5. Decommission | Once the old system receives no traffic, switch it off. Remove its credentials, DNS records, plugins and licences, and archive a final backup. | Nothing depends on the old system and it is no longer reachable |
What each stage needs to get right
Stabilise. PHP 8.2 receives security fixes only until 31 December 2026; 8.3 is supported until 31 December 2027, 8.4 until 31 December 2028 and 8.5 until 31 December 2029. WordPress recommends PHP 8.3 or greater, and says older versions it still runs on "have reached their official End Of Life and may expose your site to security vulnerabilities". These dates can change, so check both pages when you plan.
Upgrade in place. Test on a staging copy first: older themes, plugins and custom code are the usual cause of breakage after a PHP upgrade. For a hosting move that keeps the same URLs, Google's guidance on moving a site without URL changes advises lowering DNS TTL in advance, keeping Search Console verification working and shutting down the old host only when its traffic has dropped to zero.
Rebuild section by section. The routing layer is the mechanism that lets two systems share one domain: requests for /blog go to the new stack, everything else to the old one. Keep a shared header, footer, cookie consent and analytics configuration so visitors and reports see one site. Start with a section that is valuable but self-contained, not checkout or login.
One URL cutover. Google says to keep redirects "for as long as possible, generally at least 1 year", to "expect temporary fluctuation in site ranking during the move", and that a small to medium site "can take a few weeks for most pages to move". The SEO-safe website migration guide covers the redirect map and launch checks in detail.
Decommission. Old systems left running are easy to forget and hard to secure. Switch them off deliberately and record what was removed.
Stage it or rebuild in one go?
| Factor | Favours staged modernisation | Favours a single rebuild |
|---|---|---|
| Site size | Many sections and templates | A handful of pages |
| Revenue or leads from the site | Significant and continuous | Low, or tolerant of a short disruption |
| Integrations | Several, each with its own risk | Few or none |
| Current platform | Can be upgraded and kept secure meanwhile | Cannot be patched or upgraded safely |
| Team capacity | Can run two systems and a routing layer | Small team, wants one system to look after |
| Design change | Mainly technical; the look can change later | A full brand and UX overhaul anyway |
Worth noting
Rebuild now, rather than stage, when the runtime or CMS cannot be upgraded to a supported version, when the CMS cannot model your content without workarounds, or when the team cannot deploy changes safely at all.
When staged modernisation is not the right call
Staging has real costs. You run and pay for two systems, the routing layer is another component to configure and monitor, and design consistency is harder while old and new sections sit side by side. For a small brochure site, those costs usually outweigh the benefit; a planned rebuild with one cutover is simpler. Staging also does not help if the old platform has an unpatched security problem that cannot wait; contain that first.
Finally, a staged plan can stall halfway, leaving you maintaining both systems indefinitely. Give each stage an owner and a date, and treat decommissioning as part of the project, not an optional extra.
Hypothetical example: a services firm on old WordPress
Hypothetical example: a professional services firm runs a WordPress site on PHP 8.2 with a custom theme, a booking plugin no longer maintained and a blog that brings in most of its enquiries. In stage 1 it backs up, inventories 400 URLs and finds three unused plugins. In stage 2 it upgrades PHP and WordPress on staging, replaces the abandoned booking plugin and goes live with no visible change. In stage 3 it rebuilds the service pages on a new stack behind the existing URLs, then the blog. Because the URL structure stays the same, there is no URL cutover. Once the old WordPress install serves nothing, it is switched off and its accounts removed.
Content, integrations and SEO in any rebuild
Whichever route you choose, a rebuild usually means migrating content and re-implementing integrations on the new architecture. Scope both during planning; the requirements document guide shows how to capture integrations and acceptance criteria. Any change of URLs or architecture carries SEO risk, which is manageable with a complete redirect map and careful checks during the migration. If the constraint is the platform itself, the website replatforming guide covers moving to a new CMS or platform.
Sources
Martin Fowler: Strangler Fig Application (22 August 2024); Google Search Central: site moves with URL changes (last updated 20 August 2026); Google Search Central: moving a site without URL changes; PHP: supported versions; WordPress: hosting requirements.
Accessed 14 October 2026. Support dates and recommendations change; check the current pages before planning.
Conclusion
Redesign, rebuild, incremental improvement and staged modernisation solve different problems. Diagnose whether your constraint is visual, technical or isolated; check your runtime and CMS against their support dates; and if the platform needs replacing on a site that earns its keep, consider rebuilding in stages behind your existing URLs, with any URL change made once. Next steps: run stage 1 (backups, inventory, version check) whatever you decide, because it informs every option. Our website development page explains how we approach rebuilds.
Get a staged modernisation plan
ZSpace Labs is a remote-first technology studio that works with clients globally on website rebuilds and modernisation. If useful, we can talk through whether your current site should be redesigned, upgraded in place, rebuilt in stages or replaced.
Common questions.
A redesign changes the visual design and often the UX, typically keeping the underlying technical architecture. A rebuild replaces the underlying technical foundation, sometimes with a visual redesign alongside it and sometimes without.
Yes. If the existing technical architecture is sound and the problems are mainly visual, UX or content-related, a redesign on the existing foundation can be the right, lower-cost choice.
When the underlying technical architecture itself is the constraint: significant technical debt, an outgrown CMS or platform, or performance and scalability limits that a visual redesign cannot fix. It is also worth considering when the site is small enough that staging the work would cost more than replacing it.
Not always. Incremental changes work well for isolated problems, but a site with many compounding issues can sometimes be addressed more efficiently with a coordinated redesign or rebuild than a long series of small fixes.
Any change of URLs or architecture carries some risk, and Google says to expect temporary ranking fluctuation during a site move. The risk is manageable with a complete redirect map and careful technical SEO during the migration, so it is not a reason to avoid a rebuild that is genuinely needed.
Design and UX problems typically show up as usability complaints, weak conversion or an outdated look. Technical problems show up as slow performance, difficulty adding features, or a CMS the team has outgrown. Often both are present.
Yes, deliberately. Content migration and redirect mapping should be planned explicitly, not treated as an afterthought once the new site is otherwise ready.
A rebuild often requires re-implementing integrations on the new architecture. Scope this explicitly as part of rebuild planning rather than discovering it mid-project.
No. The site will keep running, but PHP 8.2 stops receiving security fixes from the PHP project after 31 December 2026, according to php.net. Any newly found vulnerability in that version will then go unpatched upstream, so plan an upgrade to a supported version and check that your themes, plugins and custom code are compatible.
For small and medium sites, Google recommends moving all URLs at the same time rather than section by section. You can still rebuild the back end in stages while keeping the existing URLs, then make any URL changes in one planned cutover.