Website Replatforming: How to Tell If You've Outgrown Your Platform
A 10-point scorecard to tell whether your website has outgrown its platform, dated end-of-life checkpoints, and how to replatform without losing search traffic.
Quick answer
Replatforming means moving your website to a different technology stack, such as a new CMS, commerce platform or framework, along with its content, data and integrations. You have outgrown your platform when the platform itself is the constraint: its version or runtime is out of support with no realistic upgrade path, it cannot connect to the systems you need, it cannot model your content, and editors depend on developers for routine work. Score these with the 10-point diagnostic below. If most of the pain comes from how the site was built rather than what it was built on, a rebuild or upgrade on the same platform is usually cheaper and safer.
Key takeaways
- Diagnose the platform, not the design: slow pages and an awkward CMS are often build problems, not platform problems.
- End-of-life dates are deadlines for a decision, not automatically for a replatform. Check whether an in-place upgrade exists.
- Score ten platform-level questions with evidence, then let the total point you to optimise, modernise or replatform.
- Every replatform is also a migration: plan URLs, redirects, data and integrations from the start.
Replatforming, redesign and rebuild
These terms get mixed up, which leads to mis-scoped projects. If your concern is mainly how the site looks or how visitors move through it, start with the signs your website needs a redesign. If the code is the problem but the platform is fine, redesign vs rebuild covers rebuilding and staged modernisation. This guide is for the narrower case where the platform underneath is the constraint.
| Approach | What changes | Typical trigger |
|---|---|---|
| Redesign | Visual design and UX, usually same platform | Dated look, weak conversion |
| Rebuild | Code rewritten, platform may stay the same | Technical debt in the codebase |
| Replatform | Underlying platform changes | Platform limits the business |
Have you outgrown your platform? A 10-point diagnostic
Score each row 0 (fine), 1 (strained) or 2 (blocking), using evidence rather than opinion. Gather the evidence first: version numbers, a month of content-change requests, Search Console data and a list of integrations you need. Every row is about the platform, not the visual design.
| # | Question | Evidence to collect | Scores 2 when |
|---|---|---|---|
| 1 | Support status. Are the CMS, its runtime (for example PHP or Node.js) and key extensions in security support? | Your versions, checked against the official lifecycle pages in the table below | Already past end of life, or ending within six months with no upgrade path |
| 2 | Upgrade path. Can you move to the supported version in place? | A developer's assessment of custom code and extensions that block the upgrade | The upgrade is effectively a rebuild |
| 3 | Field performance. Do real users get good Core Web Vitals after standard fixes? | Search Console or CrUX data at the 75th percentile, mobile and desktop | LCP above 2.5 s, INP above 200 ms or CLS above 0.1, and fixes to images, scripts and hosting have not moved it |
| 4 | Editor independence. Can editors make routine changes alone? | Count a month of content requests that needed a developer | Most routine page, landing-page or content changes need a developer |
| 5 | Change lead time. How long does a small template or feature change take? | Dates from your ticket history over the past year | Small changes take weeks and are getting slower |
| 6 | Integration ceiling. Can the platform connect to the systems you need (CRM, ERP, PIM, SSO, search)? | A list of required integrations and how each works today | A needed integration is impossible, or runs only through a fragile workaround |
| 7 | Content model fit. Can the CMS model your content? | Requirements for structured content, languages, multiple sites or personalisation | You are forcing content into fields and pages that do not fit |
| 8 | Security and compliance burden. Can you keep it secure and compliant at reasonable effort? | Incident history, unpatched extensions, compliance requirements such as accessibility or payment card rules | Recurring incidents, or a requirement the platform makes very hard to meet |
| 9 | Cost curve. Are licence, hosting, extension and maintenance costs rising faster than the value? | Your own costs and maintenance hours over two or three years | Costs keep rising while the platform does less of what you need |
| 10 | Talent pool. Can you hire or contract developers for it? | Response to recent hiring or agency enquiries | Few people still work with it, and they are hard to book |
What your score means
The bands below are a ZSpace Labs heuristic to structure the decision, not an industry standard. Use them to decide what to investigate next, not as a verdict.
| Score (out of 20) | Likely conclusion | Next step |
|---|---|---|
| 0 to 5 | The platform is not the constraint | Optimise in place: see why a website is slow and the website maintenance guide |
| 6 to 11 | Strained, but fixable without moving | Targeted fixes, an in-place upgrade or staged modernisation: see redesign vs rebuild |
| 12 or more, or a 2 on row 1 with no upgrade path | The platform is likely the constraint | Build a replatforming business case and compare at least two target platforms against written requirements |
Key takeaway
Outgrown the platform or outgrown the build? If rows 1, 6, 7 and 10 score low but rows 3, 4 and 5 score high, the platform is probably fine and the implementation is the problem. A rebuild or refactor on the same platform usually fixes that for less risk than a move.
Dated end-of-life checkpoints
End of life means the project stops publishing security fixes, so new vulnerabilities stay open. The dates below were checked on the official pages on 15 October 2026. Schedules can change, so confirm on the linked page before you plan around them.
| Platform or runtime | Status as of 15 October 2026 | Source |
|---|---|---|
| Drupal 7 | End of life on 5 January 2025; no further security or compatibility updates | Drupal.org |
| Drupal 10 | Reaches end of life on 9 December 2026. Drupal 12 is scheduled for the week of 7 December 2026, though Drupal says release windows are not guaranteed | Drupal.org release schedule |
| PHP 8.2 | Security support ends 31 December 2026. PHP 8.3 has security support until 31 December 2027, 8.4 until 31 December 2028 and 8.5 until 31 December 2029 | PHP.net supported versions |
| Adobe Commerce 2.4.6 | Standard support ended 11 August 2026. 2.4.7 standard support ends 31 May 2027. The page covers Adobe Commerce; check Magento Open Source support separately | Adobe Experience League |
| WordPress | Only the latest version is officially supported; the security team backports some fixes to older versions as a courtesy | WordPress.org security |
| Magento 1 | Support ended on 30 June 2020 according to Adobe's announcements at the time. Treat any Magento 1 site as unsupported unless you pay a third party for patches | Adobe developer community post and contemporary coverage (primary announcement not retrieved) |
Worth noting
An end-of-life date often calls for an upgrade, not a move. Drupal 10 sites can move to Drupal 11, and PHP 8.2 sites to a newer PHP branch, if custom code and extensions allow it. Row 2 of the diagnostic is where that question gets answered. If your software comes from a vendor, ask for its support roadmap in writing; the UK's voluntary Software Security Code of Practice suggests vendors give at least a year's notice before support ends.
Hypothetical example: one site, two readings
This example is hypothetical. A B2B manufacturer runs its site on Drupal 10 with PHP 8.2. Its marketing team scores the diagnostic: row 1 scores 2 (both reach end of life by the end of 2026), row 4 scores 2 (every landing page needs a developer), rows 3 and 5 score 1, and the rest score 0. Total: 6.
A developer then checks row 2 and finds the custom modules are compatible with Drupal 11 after modest changes, so row 1 has an upgrade path. The editor problem turns out to be a build issue: the original agency hard-coded page layouts instead of giving editors reusable components. The sensible plan is an in-place upgrade plus a rebuild of the page templates, not a replatform. Had the custom modules blocked the upgrade and the CRM integration been impossible, the score would have pointed the other way.
Legacy technology problems
Older platforms often accumulate outdated dependencies, unsupported extensions and custom patches nobody fully understands. Each update becomes risky, which leads to updates being skipped, which increases security exposure. That cycle is a common reason businesses start considering a move, and it is why rows 1 and 2 of the diagnostic carry most weight in practice.
Performance, security and CMS limitations
Before blaming the platform, rule out fixable causes like heavy scripts, images or hosting, using the slow website guide. Google's "good" Core Web Vitals thresholds are LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less, measured at the 75th percentile of page loads. If problems remain rooted in the platform's architecture, security model or editing limits after those fixes, that's a genuine replatforming signal.
Not sure whether your platform is the real problem?
ZSpace Labs can look at your scorecard and current stack and tell you whether to optimise, upgrade, rebuild or replatform.
Integration limits and developer bottlenecks
When the business needs deeper connections to CRM, ERP or other systems than the platform supports, or when every change waits on scarce specialist developers, the platform is costing more than its licences. See common website integrations for what modern stacks typically need to support. Online stores have their own constraints; ecommerce replatforming covers how to tell whether a commerce platform is the limit.
Cost considerations
Budget for more than the new build: content and data migration, integration rework, redirects and SEO work, training, parallel running during transition, and post-launch fixes. Data migration and integration rework are the most commonly underestimated items. The cost breakdown guide explains the broader drivers.
Reducing replatforming risk
- Diagnose why the current platform fails before choosing a new one
- Choose the new platform against documented requirements
- Consider phased approaches, such as a new front end on the existing backend
- Plan content and data migration with validation steps
- Rebuild and test integrations early, not at the end
- Protect SEO with a full URL map and redirects
- Train editors before launch
- Monitor closely after launch
Planning a replatforming project?
Talk to ZSpace Labs about a phased, lower-risk path to a stack that fits where your business is going. Learn more about our Next.js and React development work.
A Replatforming Cutover Checklist That Protects Search Visibility
Replatforming differs from a redesign or a data migration: the technology changes underneath, often with new URLs, templates and integrations at the same time. Most search losses come from avoidable URL and content mistakes. Google's guidance on site moves is the reference point.
- Evaluate the platform against requirements: content model, editor experience, integrations, performance, accessibility, hosting and total cost, before falling for a demo.
- Inventory URLs: export every indexable URL with traffic, backlinks and conversions from Search Console and analytics.
- Keep URLs where possible; where they must change, map every old URL to its closest new equivalent and implement permanent (301 or 308) redirects. Avoid redirecting everything to the home page.
- Keep redirects in place long term; Google says generally at least a year, and longer is safer for valuable URLs.
- Migrate content and metadata: titles, meta descriptions, headings, structured data, canonical tags, alt text and internal links.
- Rebuild integrations (forms, CRM, analytics, payments) and test them end to end.
- Test on staging: crawl it, check redirects, page speed and accessibility, and block it from indexing.
- Cut over with a rollback plan, submit the new sitemap, and monitor Search Console, analytics and error logs daily for the first weeks. Some ranking fluctuation after a move is normal.
- Data: for moving records rather than pages, see data migration planning; for older systems, legacy software modernisation.
Worth noting
Australian businesses: also check hosting location and APP 8 if the new platform stores personal information overseas, keep GST-inclusive pricing and legal pages intact, and avoid launching just before peak trading periods.
When replatforming is the wrong answer
Do not replatform to fix a content problem, an unclear brand or a weak conversion path; those follow you to the new platform. Do not replatform because an agency prefers a different stack, or because one deadline is near when an in-place upgrade is available. And avoid replatforming during your busiest trading period or while the business is changing its offer, because you will be migrating a moving target.
Sources
Lifecycle and support: Drupal 7 end of life; Drupal core release schedule; PHP supported versions; Adobe Commerce lifecycle policy (updated 18 September 2026); WordPress security; Magento DevBlog: preparing for Magento 1 end of life.
Performance and migration: web.dev: Web Vitals; Google Search Central: site moves with URL changes.
Vendor support notice: UK Software Security Code of Practice.
Accessed 15 October 2026. Support dates and schedules change; check the official pages before planning around them.
Conclusion
Replatforming is justified when the platform itself limits the business, and it's risky enough to deserve careful diagnosis and planning. Score the ten rows with evidence, check the end-of-life dates that apply to you, rule out an in-place upgrade, then choose the new stack against clear requirements and treat migration as a project in its own right. The website migration guide covers the move step by step. For choosing who does the work, see how to choose a website development company.
For related guides, see legacy ecommerce migration and ecommerce migration checklist.
Common questions.
Moving a website from one technology platform to another, such as a different CMS, commerce platform or framework, usually along with its content, data and integrations.
A redesign changes how the site looks and works for users, often on the same platform. Replatforming changes the underlying technology, and may or may not include a redesign.
A rebuild rewrites the site, sometimes on the same technology. Replatforming specifically means moving to a different platform. In practice the terms overlap, and many projects are both.
Persistent performance problems the platform can't fix, security or support concerns with aging technology, integrations the platform can't handle, editors blocked by the CMS, and development work that's consistently slower than it should be.
Yes. Risks include SEO loss, data migration errors, broken integrations, and disruption to teams. Careful planning, phased approaches and a thorough migration process reduce them.
It depends on content volume, integrations, custom functionality and whether a redesign is included. Data migration and integration rework are often underestimated.
Often yes, for example moving the front end first while keeping the existing backend, or migrating sections of the site in stages.
Diagnose why the current platform is failing first. If the real issue is content, process or design, a new platform alone won't fix it.
Google's site move guidance says to keep redirects generally for at least a year, and longer for valuable URLs that still receive links or traffic. Removing them early can lose search visibility and break links customers rely on. Monitor Search Console before retiring any redirect.
Not necessarily. Often the right move is an upgrade to the supported version on the same platform, such as moving from Drupal 10 to Drupal 11. Replatform only if the upgrade is close to a rebuild anyway, or if the platform also scores badly on integrations, content model or editor independence.
It becomes progressively riskier. After end of life the project no longer publishes security fixes, so new vulnerabilities stay open unless you pay for third-party extended support. That can buy time for a planned move, but it is not a long-term answer.