If the foundation is sound — the site structure makes sense, it loads reasonably fast, it works on mobile, and the CMS isn’t fighting you — and the problems are a handful of specific, fixable things, repair it. If the foundation itself is the problem, a redesign will almost always cost less time and money in the long run than continuing to patch around it.
The trouble is that “loads reasonably fast” is a feeling, and feelings are a bad basis for spending a few thousand dollars. So before the judgement calls, here is how to get an actual number.
Ten minutes with real numbers
Google publishes the exact thresholds it uses to decide whether a page delivers a good experience. They are called Core Web Vitals, and there are three:
| What it measures | Metric | Good | Poor |
|---|---|---|---|
| How fast the main content appears | LCP | ≤ 2.5 s | > 4.0 s |
| How quickly the page responds when tapped | INP | ≤ 200 ms | > 500 ms |
| How much the layout jumps around while loading | CLS | ≤ 0.1 | > 0.25 |
One detail that matters: Google assesses these at the 75th percentile of real visits. So it is not asking whether your site is fast for you, on your laptop, on your office wifi. It is asking whether it is fast for three out of four of your actual visitors — including the ones on a phone with two bars in a parking lot.
How to check yours: run your homepage and your busiest service page through PageSpeed Insights. If your site has enough traffic, the top section shows real visitor data. If it doesn’t, you will only get the lab test, which is still useful as a rough guide.
Then read it like this:
- Green across the board, and the problems are cosmetic → repair.
- Failing on mobile, and it is one specific thing (an enormous hero image, one heavy plugin) → repair.
- Failing on mobile, and the report blames the theme or page builder itself → that is a foundation problem, and no amount of plugin tweaking fixes it. This is the single most common reason a repair turns into a rebuild halfway through.
Our free site check runs this for you and puts it in plain language if the raw report is hard going.
Signs a repair is enough
A slow contact form, an old headshot, a couple of broken links, service pages that need rewriting, or a site that just looks dated — these are all repair jobs. None of them need the underlying structure touched.
Rebuilding to fix things like these is usually overkill. It costs more, takes longer, and puts your existing search rankings at risk for no good reason.
Signs you actually need a redesign
A few signals point the other way, and they tend to compound each other rather than showing up alone:
- The site is built on a page-builder or theme that’s genuinely slow, and no amount of plugin tweaking meaningfully improves it
- The navigation and page structure don’t reflect what the business actually offers anymore
- Mobile visitors — often the majority of traffic — get a cramped or broken version of the desktop site
- Making even small content changes requires a developer because the CMS setup is fragile or overly custom
- The site was built years ago around SEO practices that no longer apply, and restructuring it page-by-page would take longer than starting clean
The middle case: a modernization instead of either extreme
Not every case is “fix it” or “start over.” Often the right move is keeping the content, the page addresses, and the search rankings you’ve already built up, while replacing the technical foundation underneath — new theme, faster setup, same pages in the same places. When the content is fine but the platform is holding it back, this is usually the quickest and safest route.
It is also the cheapest of the three, because the expensive part of a website is rarely the design. It is the writing, the structure and the accumulated search history — all of which you keep.
What a redesign should never cost you
A careless redesign can quietly wipe out years of search visibility: page addresses change without redirects, page titles and descriptions go missing, links between pages break.
None of that has to happen. Google publishes guidance on moving a site, and the important parts are short enough to hold anyone to:
Use permanent redirects. Google recommends server-side permanent redirects — a 301 or 308 — from every old address to its new one. Not a meta refresh, not a JavaScript redirect, and not a link to the homepage. If ten pages disappear and all ten point at your homepage, you have not migrated those pages; you have deleted them.
Keep the redirects for at least a year. Google’s own guidance is “as long as possible, generally at least 1 year.” In practice, leave them permanently — they cost nothing to keep and old links live longer than anyone expects.
Expect it to take weeks, not days. Google says a small to medium site “can take a few weeks for most pages to move.” A dip in the first fortnight is normal and is not, by itself, evidence that something went wrong. Panicking and changing things again mid-move is how a normal dip becomes a real one.
Get the inventory before anything comes down. Export your existing page addresses from Google Search Console while the old site is still live. This is the step people skip, and it cannot be done afterwards — once the old site is gone, the list of what used to exist is gone with it.
A quick self-check list
- Have you run PageSpeed Insights on your homepage and a service page, on mobile?
- Are your problems specific and fixable, or systemic to how the site is built?
- Can you make basic content updates yourself, or does everything require a developer?
- If you redesign, does the plan include a
301for every existing URL, not just the main ones? - Do you have an exported list of your current pages, saved somewhere off the old site?
If you’re not sure which side of that line your site falls on, that’s a fair thing to ask us directly — we’ll say plainly if a repair is genuinely enough, because talking someone into a rebuild they don’t need is a bad way to start. Otherwise, our website modifications & redesigns service covers both paths, and the pricing page explains what moves the number either way.