Redesign when the foundation is healthy and the problem is mainly experience, content or presentation. Rebuild when architecture, CMS, performance, security, integrations or maintainability prevent the website from meeting its goals. Many projects need a controlled hybrid.
Key takeaways
- Do not decide from appearance alone.
- Audit content, URLs, analytics and integrations before choosing.
- Preserve valuable URLs and evidence even during a rebuild.
- A hybrid can replace high-risk layers while retaining useful assets.
Who this is for: Organizations with an outdated, slow or difficult-to-manage website deciding how much should change.
Three different interventions
A visual redesign changes presentation and interaction while retaining most of the platform. A structural refactor improves templates, components or performance without replacing everything. A rebuild creates a new implementation and migrates selected content, data and integrations.
| Option | Best when | Main risk |
|---|---|---|
| Redesign | CMS and code are healthy; UX and brand are weak | New visuals inherit old structural limits |
| Refactor | Specific technical layers cause measurable problems | Hidden dependencies expand scope |
| Rebuild | Foundation blocks growth, security or maintainability | Migration and launch risk if evidence is not inventoried |
Audit before deciding
A redesign decision should use evidence from analytics, search, content, technology and operations. A stakeholder saying the site looks old is a signal, not a complete diagnosis.
- Top landing pages and conversion paths
- Organic URLs, backlinks and redirects
- CMS editing pain and publishing workflow
- Performance and accessibility patterns
- Security, dependencies and hosting constraints
- Forms, CRM, payment and data integrations
- Content quality and ownership
- Mobile and browser behavior
Decision tree
Use the first decisive constraint, then validate downstream effects.
- 1If the platform is unsupported, insecure or cannot meet core requirements, evaluate a rebuild.
- 2If users cannot complete key tasks because information architecture is wrong, redesign templates and navigation; rebuild only if the platform prevents it.
- 3If performance problems come from isolated components, refactor before replacing the whole stack.
- 4If content and URLs perform well, preserve and migrate them even when the code is replaced.
- 5If the internal team cannot operate the current CMS, include governance and training in the decision.
Protect SEO and measurement
Inventory every indexable URL, metadata set, structured-data pattern, internal link and conversion event before launch. A new visual design does not excuse lost redirects or missing analytics.
old_url -> new_url
status -> 301
reason -> consolidated / renamed / retired
owner -> content or engineering
validated -> yes/no
When a hybrid is better
A hybrid replaces the risky foundation or highest-value templates first while retaining stable systems. It works when boundaries are explicit and temporary integrations do not become permanent debt.
Define the end state, migration sequence and retirement criteria before starting. Otherwise a hybrid becomes two overlapping websites with duplicated maintenance.
How to use this guide with your team
A redesign changes a live business asset, so treat it as a controlled improvement program rather than a visual replacement exercise. Preserve the evidence that already works while addressing the problems that can be demonstrated.
Marketing, content, design and engineering should review the same baseline. If each team uses a different definition of success, a visually successful launch can still create commercial or search losses.
After launch, compare results with the original baseline and document what changed. This creates a better foundation for the next iteration instead of beginning another redesign from opinion.
Practical next steps
Use the following actions as a short working session. Record decisions, owners and unresolved questions so the article becomes an implementation aid rather than passive reading.
- 1Capture the current performance and content baseline.
- 2Separate cosmetic, journey and platform problems.
- 3Map every high-value URL and integration.
- 4Define launch and rollback ownership.
- 5Measure the first 30 days against the baseline.
Frequently asked questions
Will a rebuild always improve performance?
No. Performance improves when requirements, budgets, architecture, media and third-party scripts are controlled and tested.
Can we keep the same URLs?
Often yes, and valuable URLs should normally remain stable. Changed URLs need deliberate one-to-one redirects and internal-link updates.
Sources and further reading
Last reviewed: September 1, 2026
