Redesign or rebuild? A decision tree

“Our website needs a refresh” is how the conversation usually starts. But a refresh can mean two very different projects: new paint on sound foundations, or sound foundations under acceptable paint. Pick wrong in either direction and you’ll pay twice — a beautiful reskin of a rotten platform rots again in a year; an unnecessary rebuild burns budget the marketing needed. Here’s how we actually decide.

First, separate the two questions

A website is two things wearing one URL: the surface (design, copy, structure — what visitors experience) and the foundations (the code, platform and hosting underneath). Judge them separately, because they fail separately.

Signs the foundations are the problem (rebuild territory)

• Slow no matter what you optimise — the platform itself is the weight.
• Every small change needs a specialist, costs a surprising amount, or breaks something else.
• The stack is abandoned: an unsupported builder, plugins without updates, a PHP version your host keeps warning about.
• You can’t edit your own content without fear.
• Security patches have become a monthly drama.

Signs the surface is the problem (redesign territory)

• The site is technically fine but looks five years old next to competitors.
• The message no longer matches the business — services changed, positioning moved on, the copy didn’t.
• Traffic arrives but doesn’t convert: unclear headlines, buried calls to action, a journey that loses people.
• The brand evolved and the site didn’t come along.

The decision tree, in prose

Foundations rotten? Rebuild — and treat the redesign as included, because you’re touching everything anyway. Never reskin a condemned building. Foundations sound but surface failing? Redesign on the platform you have, and spend the difference on the marketing that fills it. Both failing? Rebuild once, properly, rather than two overlapping projects. Neither failing? You don’t have a website problem — look at traffic and attribution before spending anything on pixels.

One honest caveat: the answer needs a look under the bonnet, not just at the paint. A twenty-minute technical review settles it with evidence — which is exactly the sort of thing our Digital Scan conversation starts. Diagnose first. Then spend once, in the right place.