Almost every redesign conversation starts the same way: an owner is embarrassed by their site and wants to "start from scratch." Nine times out of ten, starting from scratch is the wrong move. A meaningful chunk of most existing sites — the content that already converts, the structure that already makes sense, the pages that already rank — should survive a redesign untouched. The job is not demolition. It is deciding, page by page, what actually needs to change.
The three-bucket test
Before any design work starts, we sort the existing site into three buckets, and this alone usually saves weeks of scope creep.
Keep. Pages and content that are structurally sound and already doing their job — ranking, converting, or both. These get a visual refresh to match the new design system, but the actual words and structure stay. Rewriting something that is already working is a common way redesigns quietly destroy value nobody was tracking.
Rebuild. Pages with a real audience or purpose but poor execution — slow, confusing, outdated information, weak structure. These get rebuilt from the ground up: new copy, new layout, new technical foundation.
Kill. Pages nobody visits, that duplicate other content, or that exist because someone thought the site needed more pages. Every unnecessary page is another thing that can be slow, outdated, or confusing. Cutting these is usually the single fastest improvement to a site's overall quality, and it costs nothing to do.
Why "just make it look modern" is the wrong brief
A visual refresh on top of a slow, poorly structured site just repaints the same underlying problems. We wrote about this directly in why your site is slow — the Kanata Mews rebuild went from a 6.2-second load time to 1.4 seconds, and that speed improvement drove a 40 percent increase in inbound calls attributed to the website. None of that came from the site looking prettier. It came from fixing what was actually broken underneath the design.
If a redesign brief is purely aesthetic — new colors, new fonts, more modern layout — and does not touch speed, structure, and the schema markup that helps both Google and AI answer engines understand the business, it is solving the smallest part of the problem. We covered why that structural layer matters more than ever in the schema markup most local business sites get wrong.
What actually needs a full rebuild versus a refresh
A full rebuild makes sense when the foundation itself is the problem: a CMS that is slow or hard for your team to use, a site structure that has grown into an unusable maze of pages, or a codebase old enough that basic performance fixes are not realistically possible on top of it.
A refresh is the right call far more often than owners expect: the bones are fine, the content mostly works, but the visual layer feels dated and a handful of specific pages need real work. Confusing these two paths is the most common way redesign projects balloon in cost and timeline without a corresponding increase in what the client actually gets.
Content that should never get rewritten just because it is old
Owners often assume old content needs replacing simply because it has been there a while. If a service page or blog post is still bringing in inquiries, its age is not the problem — check the actual numbers before touching the words. We see redesigns quietly kill converting content because nobody checked performance data before the rewrite. That is an unforced error a fifteen-minute analytics review prevents.
How to scope your own redesign this week
Pull up your site's page list and your analytics. Sort every page into keep, rebuild, or kill based on actual traffic and conversion data, not gut feeling about how it looks. That sorted list is the real brief — it tells you where the design and development effort should actually go, and it usually cuts the scope of "start from scratch" down to a fraction of the site.
If you want a second set of eyes on that sort before committing to a full rebuild, see our web development work or book a 30-minute call and we will tell you honestly which bucket your site actually needs.