Signs Your Website Needs a Redesign (and When It Doesn't)
"Every three to five years" is a rule of thumb, not a diagnosis. Here is a ten-point self-audit using free tools, a decision tree for fix versus restructure versus rebuild, and the honest cases where the right answer is to leave the site alone and repair three specific things.
Most advice on when to redesign a website is a schedule: every three years, every five years, “when it looks dated.” Schedules are how businesses end up rebuilding sites that were working and patching sites that never will. A redesign is a decision with evidence behind it, and the evidence is easy to collect.
Here is a ten-point audit you can run in an hour with free tools, a decision tree for what the score means, and — because it is the more common case — a clear description of when the right answer is not to redesign.
The ten-point self-audit
Score one point for each “yes.”
1. Does it fail Core Web Vitals on mobile after the easy fixes? Run PageSpeed Insights on your homepage and top service page, mobile, and read the real-user section. If you have already compressed images and removed unused scripts and it still fails, the weight is the foundation. (Why Is My Website Slow? walks through the easy fixes first.)
2. Does it look broken or cramped on your own phone? Not “responsive” in theory — actually usable. Tap targets you can hit, text you can read without zooming, forms that work with a thumb.
3. Can you not change a price, a photo, or a paragraph without calling someone? If routine edits require a developer, the site is a liability every time the business changes.
4. Has it broken after an update, been hacked, or shown security warnings in the last year? Recurring incidents on the same stack are the stack talking.
5. Is there no page for a service you now sell, and no sensible place to add one? Architecture problems look like content problems. If adding a service means stuffing another paragraph on a “Services” page, the structure is out of room.
6. Are impressions or clicks in Search Console flat or declining for a year while you’ve kept publishing? Open Performance, compare the last twelve months with the previous twelve. Sustained decline on a site you maintain points to structure or speed, not effort.
7. Does the form fail, or does nobody know if it’s working? Submit it from a phone. If it fails, or if you’re not sure where submissions go, the site’s most important job is unverified.
8. Is the imagery stock, dated, or obviously not your business? Visitors decide credibility in seconds, and stock photography of models in hard hats tells them you had nothing real to show.
9. Does it make the business look smaller than it is? Read it as a stranger. A competitor with half your experience and a better site is winning the comparison. (Why Your Website Makes Your Business Look Smaller Than It Is lists the specific signals.)
10. Is there no conversion tracking, so you can’t tell what the site produces? No analytics events, no call tracking, no idea which pages or campaigns generate enquiries. You cannot improve what you can’t see.
What the score means
Fix, restructure, or rebuild?
- Yes Rebuild — the platform is the ceiling.
-
No
Does it rank and convert, but look dated or thin?
- Yes Restructure — keep URLs and rankings, upgrade design, depth, and conversion paths.
- No, it performs well Fix the specific leaks — forms, speed, tracking, weak pages. Don’t touch what works.
0 – 1 points: leave it alone. Fix the one thing you found. A site that passes nine of these checks is a good site, whatever its age.
2 – 3 points: fix, don’t rebuild. These are usually specific, bounded problems — the form, the images, the tracking, one missing page — that a developer can address in days on the existing site. Rebuilding for these is buying a new car because the tyres are worn.
4 – 6 points, mostly cosmetic and content (items 5, 8, 9, 10): restructure. Keep the platform and the URLs. Rework the architecture, add the missing pages, replace the imagery, build the conversion paths and tracking. This preserves what search already trusts while fixing what visitors see.
4 or more with structural failures (items 1, 3, 4): rebuild. When the site is slow after fixes, can’t be edited, or keeps breaking, the platform is the ceiling and no amount of work inside it changes that. Plan a rebuild — on a foundation that doesn’t have those problems — with a proper migration.
The three questions that decide fix vs rebuild
If the score leaves you between categories, three questions settle it:
Is the problem in the pages or in the foundation? Weak copy, missing proof, bad photos, no service pages: pages. Slow after optimization, breaks on update, can’t edit, no room to grow: foundation. Page problems are fixed on the current site; foundation problems are not.
Does the site currently earn anything you’d lose? Rankings, links, a page that brings enquiries. If yes, any rebuild must preserve URLs and redirect the rest — and the migration plan must exist before the design does. The redesign SEO checklist is that plan.
What would you do with the site in twelve months? New services, new locations, paid campaigns, content for search. If the current structure can’t hold that plan, you’re rebuilding either way; better to do it now, deliberately.
When a redesign is the wrong answer
This is the section most agencies won’t write, so here it is plainly.
- The site looks dated but converts, ranks, and loads fast. That is a refresh: new typography, colour, imagery, and layout polish on the same structure. A rebuild would put working rankings at risk to fix a cosmetic problem.
- Three specific pages underperform. Fix three pages.
- A competitor launched a new site. Their site is not evidence about yours.
- A new platform is fashionable. Platforms are tools. A site that works on a mature platform does not need to move because a newer one exists. (We build on Astro and still say this — see WordPress vs Astro for when the move is and isn’t worth it.)
- The real problem is traffic, not the site. If nobody arrives, the fix is search or ads, and a redesign changes nothing. Why Isn’t My Website Ranking on Google? diagnoses that separately.
Plenty of our own engagements end with “your site is fine; here are the three things costing you.” That answer is cheaper for you and it is the honest one more often than the industry admits.
If you do rebuild
Three things decide whether it goes well:
- The migration plan comes first. Crawl the old site, map every URL, keep slugs where you can, redirect the rest one-to-one, keep staging out of the index, and remove the block at launch.
- The foundation should remove the reasons you’re rebuilding. If speed, security, and editability were the problems, the new stack has to solve them by design, not by promise. Static-first sites served from the edge, with content you can edit, are how we do it.
- Scope, timeline, and price in writing. A rebuild with open scope is how a two-week project becomes a four-month one. The cost guide shows what moves the number; the launch plan fixes it before work starts.
If you’d like the audit done for you — with a straight answer about fix, restructure, or rebuild — the free website roast reviews all ten categories and ranks the fixes by what will actually change.
Questions people ask
How often should a business website be redesigned?
There is no correct interval, despite the "every three to five years" figure that appears everywhere. Sites age at different rates depending on how they were built and how the business changed. Audit rather than schedule — the checks below take an hour and tell you more than a calendar does.
What's the difference between a refresh and a redesign?
A refresh changes what the site looks like — typography, colour, imagery, layout polish — on the same structure and platform. A redesign changes the structure, the content, often the platform, and sometimes the URLs. A refresh is cheaper and quicker and fixes cosmetic problems; it cannot fix a slow foundation or a missing architecture.
Will a redesign hurt my search rankings?
Only if the URLs change without a redirect plan, or a staging setting ships to production. Done properly, a redesign usually improves visibility because the new site is faster and better structured. The migration checklist matters more than the design.
My site looks dated but works fine. Should I rebuild?
If it converts, ranks, loads fast, and you can edit it, "looks dated" is a refresh problem — a design pass on the existing structure. Rebuilding a site that works is expensive, and the risk to rankings is real. Fix what people see; keep what search and customers already trust.
How much does a redesign cost?
Roughly the same as building the equivalent site new, plus the migration work — URL mapping, redirects, and testing — that protects your rankings. The cost guide breaks down what moves the number; the migration line is the one to make sure is in the quote.
This article is part of the website strategy & cost library. If it describes a problem you have, the service behind it is here: see packages & pricing.
Keep reading
The Website Redesign SEO Checklist (Including WordPress to Astro Migrations)
A redesign is the most common way a business loses rankings it spent years earning, and the loss is almost always avoidable. This checklist follows Google's own site-move guidance step by step, then adds the WordPress-specific traps — the image URLs, archives, and plugin pages that vanish when you change platforms — that generic checklists leave out.
Why Is My Website Slow? What You Can Fix Today and What Needs a Rebuild
A slow website is usually slow for three or four specific reasons, and they are visible in a free test in under five minutes. Here is how to read the results, which fixes you can make from a dashboard this afternoon, and which ones are the architecture of the site itself.
How Much Does a Small Business Website Cost in 2026?
Published prices for a business website run from a few hundred dollars to well past fifty thousand, and the number alone tells you almost nothing. Here is what actually moves the price, what each tier includes, and how to compare quotes that look nothing alike.