Redesign is attractive because it produces visible change. Conversion optimisation is attractive because it sounds smaller and more measurable. Neither is the right starting point until the business can explain where the current journey is failing and whether the existing site can support the required change.
Start with the commercial symptom, not the preferred project
Define what is not moving: relevant visitors do not understand the offer, trust is weak, enquiries are low quality, mobile friction is high, the content structure cannot support the buying journey, or leads disappear after submission. Different symptoms justify different interventions.
A redesign is not a diagnosis. Conversion optimisation is not automatically cheaper if the underlying structure prevents meaningful testing or improvement.
Optimise the current site when the foundation can carry the job
Keep the current platform and structure when the offer can be understood, the pages can be edited safely, analytics and forms work, the buyer journey is coherent enough to improve and the main friction sits in messaging, proof, hierarchy, calls to action or specific conversion steps.
In that situation, targeted changes create evidence faster than replacing the whole surface. Preserve what already works and change the smallest material constraint first.
Redesign when the structure itself blocks the buying journey
A rebuild becomes more defensible when information architecture is fundamentally wrong, the content model cannot express the offer, mobile behaviour is persistently broken, technical debt makes ordinary changes risky, ecommerce architecture no longer matches the catalogue or the site cannot connect reliably to the systems that receive the enquiry.
The business should be able to name the responsibility the new structure must carry. “It looks old” may justify a brand refresh, but it does not prove a commercial rebuild is necessary.
Inspect what happens after the conversion event
A site can improve form completion while the business still loses value through slow response, unclear qualification, missing ownership or no visible next action. If the expensive leak is downstream, redesigning the front end can produce a prettier version of the same commercial failure.
Follow one enquiry from page to inbox or CRM, owner, next action and outcome before deciding the website is the whole problem.
Migration risk belongs in the decision
A redesign can disturb rankings, URLs, analytics continuity, product data, integrations and proven content. Those are operating assets, not implementation trivia.
If the current site has search visibility or working commercial paths, define what must be preserved before committing to replacement. The burden of proof for a rebuild rises as the amount of useful existing equity rises.
Use a diagnosis to choose the scope
Map comprehension, trust, action, mobile experience, handoff and measurement. Then classify each issue as repairable inside the existing system, structurally constrained or outside the website entirely.
The decision should follow evidence: optimise what can be improved safely, redesign what cannot carry the required responsibility, and fix downstream ownership where the website is not the real constraint.
Choose redesign or optimisation from the constraint the website and commercial journey can actually demonstrate, not from aesthetic preference or a universal conversion checklist.
Apply the logic to a real process: Estimate the website responsibility →