Signs you need a new website: the short answer
You need a new website when the problem sits in the foundations. The platform no longer gets security fixes, the structure can't hold what your business now sells, or nobody can edit the site. Most other signs, including fewer enquiries, slow pages and a dated look, have smaller fixes on the site you already have. Check your own data before you decide, because a redesign carries risks of its own.
This guide is for founders, managing directors and operations leads at UK and EU companies of 10 to 200 people. Perhaps the site is a few years old, enquiries feel flat, or an agency has told you it's time.
We build websites for companies in the UK, France and Albania, so a redesign is work we'd be paid for. This guide still sorts the signs into the few that call for a new site and the many that call for something smaller. A supplier who recommends the smaller job is one whose advice you can check.
Fix, refresh or rebuild: three different jobs
Every website problem has three possible answers, and they differ in size and in risk.
- A fix changes one thing on the site you have: a form that doesn't send, a slow page, an out-of-date service description.
- A refresh changes how the site looks, such as the fonts, colours, photos and the layout of the main pages. The platform and the page addresses stay the same.
- A rebuild replaces the platform, the structure or both. This is what most people mean by "a new website".
Only a rebuild changes the page addresses Google knows, and usually the routes your regular customers use. That's why it needs the strongest reason. Google's guide to site moves warns that with any significant change to a site, rankings can move while it recrawls and reindexes the pages.
One of the lists ranking for this question in September 2026 counts three or more years since the last redesign as a sign by itself. Age doesn't tell you which of the three jobs you need. Your data does.
Start with your data, not with how the site looks
An hour with five checks tells you more than a year of looking at your homepage. Run them before you ask anyone for a quote.
| Check | Where to look | What it tells you |
|---|---|---|
| Clicks from Google over the last 16 months, this year against last | Search Console, Performance report | Whether search traffic really fell, and on which pages and searches |
| Enquiries per month, and the pages they start from | Google Analytics key events, or a count from your inbox or CRM | Whether the site is failing at its main job |
| Speed for real visitors on phones | PageSpeed Insights, field data | Whether slowness is real, and which measure fails |
| The route to an enquiry | Your own phone: send the form, tap the number | Whether enquiries can reach you at all |
| What customers ask on the phone | The person who answers it | What the site fails to say |
Google's guide to debugging drops in search traffic suggests the 16-month view, so you can set a month beside the same month last year. It lists seasonality and ranking updates among the usual causes, and points to Google Trends to see whether interest fell across your whole market.
Key events are the actions you tell Google Analytics matter to your business, such as a form sent or a phone number tapped. If nobody set them up, count the enquiries in your inbox for the last six months. Our websites page promises analytics wired to a question you actually have. For most companies that question is simple: how many enquiries came from the site, and from which pages?
Say you run a 30-person electrical contractor in Bristol. Web enquiries have halved since spring, and the site is five years old, so a redesign looks obvious. Search Console shows clicks from Google flat against last year.
You send a test enquiry, the site shows its thank-you message, and nothing arrives. The form still sends to the inbox of someone who left in March.
That's a fix of a few minutes. A new site would have found it too, months later, as a side effect of a much larger project. A thank-you message proves only that the page loaded; the test that counts is an enquiry arriving in the right inbox.
The signs that do call for a new website
These five point at the foundations. One of them justifies at least an audit. Two or more usually mean a rebuild.
- The platform no longer gets security fixes. PHP, the language WordPress and many other sites run on, stops security support for version 8.2 on 31 December 2026. Moving to a newer version is normally routine. It turns into a rebuild when the theme, plugins or custom code won't run on any supported version.
- The structure can't hold the business you are now. A five-page site for a company that now offers seven services in two countries needs a page per service, and perhaps a second language. That changes the menus, the templates and the page addresses together.
- Nobody can edit it. Changing a sentence needs the original developer, or the editing screen is so fragile your team avoids it. Our comparison of WordPress and custom-built websites covers what good editing looks like on each.
- It fails on phones, and the code can't be adapted. Since July 2024, Google has crawled sites for Search with its smartphone crawler. It says content that isn't accessible on a mobile device can no longer be indexed. A site built before phones mattered usually needs rebuilding rather than patching.
- You can't take it with you. If the site lives on an agency's own system with no export, every change depends on that agency. Run the checks in our guide to whether you own your website before you decide anything else.
Nielsen Norman Group, the usability research firm, advises choosing incremental change over a major overhaul in most cases. It names three exceptions: technology so outdated that critical changes are impossible, an architecture that has become a tangled mess, and small changes that no longer gain anything. They match the list above.
Signs that usually need a fix, not a new website
Most signs on the usual lists belong here. Each has a check that takes minutes, and a fix that leaves the platform and the page addresses alone.
| What you notice | Check first | The usual fix | When it does take a rebuild |
|---|---|---|---|
| Fewer enquiries, visits steady | Send a test enquiry from your phone, and tap the phone number | Repair the form or where it sends; put the number and the form on every page | Every template buries the route to contact |
| Fewer clicks from Google | Search Console by page and by search, against last year | Rewrite the pages that lost clicks; fix the errors Search Console reports | Only if the platform can't set page titles or redirects |
| Slow on phones | Speed data from real visitors to your top pages | Smaller images, fewer scripts and plugins, better hosting | The theme or page builder is itself the weight |
| Accessibility failures | The W3C's Easy Checks | Image descriptions, form labels, link text, contrast | Seldom |
| Old services, old photos, old words | Read the site as a new customer would | Rewrite pages, add service pages, take new photos | The structure has no room for new pages |
| It looks dated | Ask five customers, not your colleagues | A refresh on the same platform | Only alongside one of the five foundation signs |
Speed means what real visitors experience: the field data from the speed check in the first table covers real Chrome users over the previous 28 days. Google counts a page as good when the main content loads within 2.5 seconds and the page responds to a tap within 200 milliseconds. Layout shift should stay at 0.1 or below. The report names any measure that fails, and that points to the fix.
Accessibility is the clearest case for fixing in place. WebAIM's 2026 scan of the top million home pages found detectable WCAG failures on 95.9% of them. The six most common were low-contrast text, missing image descriptions, missing form labels, empty links, empty buttons and a missing page language. Each of the six can be fixed on the site you have.
When not to redesign: the false alarms
Some reasons feel strong and tell you nothing about whether the site is working.
- "It's three years old." A site that someone updates, edits and measures can keep earning enquiries for years. Age tells you when it launched.
- A competitor launched a new site. That tells you about their budget and their agency. Your enquiry numbers tell you about your customers.
- You're tired of looking at it. You see the homepage every day. A customer sees it for a few minutes, once, before they call.
- Traffic dipped for a month. Compare it with the same month last year, then check Google's Search Status Dashboard, which lists ranking updates with their dates. When we wrote this guide, it showed a spam update that began on 24 September 2026.
- You have a new logo. Change the logo and the colours on the site you have. That's a refresh.
- An agency offered you one. Ask what data it looked at. If the answer is your homepage, you have an opinion, not a diagnosis.
You may not need us, or any agency. If the checks point to a broken form or thin service pages, your current developer or a good copywriter can put it right. Keep the rebuild for when a foundation sign appears.
What a redesign puts at risk
A rebuild resets three things that took years to build up: your place in Google, the routes regular customers know, and the jobs the old site did quietly.
Rankings come first. Google says that for a medium-sized site, it can take a few weeks or more before new page addresses replace the old ones in its results. It recommends permanent redirects from every old address to its new one, kept for at least a year. Our guide to moving or rebuilding a website without losing your Google rankings covers the steps.
Familiar routes matter as much. Marks & Spencer relaunched its website in February 2014. In the 13 weeks to 28 June, its online sales fell 8.1%. Its chief executive said the "settling in" of the new site had had an impact on sales.
One retail analyst told Marketing Week the site presented outfits well but made it hard for customers to see how to buy. Your version is smaller and works the same way. A customer who reorders every quarter, or a partner who sends clients to your quote page, arrives to find the route has moved. Nielsen Norman Group sums up the general point in four words: "People don't like change."
The quiet jobs are the easiest to lose. An old site may copy enquiries into a CRM, carry the tracking your adverts depend on, or hold redirects someone set up years ago. None of it shows on the homepage, so nobody lists it unless someone goes looking.
If you do redesign, keep what already works
When a foundation sign is real, plan the rebuild around what the old site does well.
- Before anyone touches the site, write down what works: the 20 pages with the most clicks, the searches behind them, and monthly enquiries for the last year.
- Keep page addresses wherever you can, and map every address you can't keep to its new one.
- List everything the site sends data to: inboxes, a CRM, analytics and advertising tags.
- Put those lists in your brief. Section 8 of our website brief template has a place for them.
- Release in parts where you can. The UK government's design principles say that "iteration reduces risk". New service pages can often go live on the old site before the new design does.
- After launch, compare the same numbers with the same weeks last year, and send yourself a test enquiry on launch day.
When we're asked to take over a site, we start with an audit and a written recommendation. It includes the case for rebuilding when rebuilding is genuinely the better path, and it says so when it isn't.
This week, open Search Console, set the date range to the last 16 months, and compare this quarter with the same quarter last year. Then send yourself an enquiry through your own site, from your phone. Those two checks settle most redesign questions before anyone draws a new homepage.
Questions
How often should a business website be redesigned?
There's no fixed interval, and a calendar is a poor reason to rebuild. Once a year, run the checks in this guide and fix what they turn up. Refresh the look when it no longer matches your brand. Rebuild when a foundation fails: the platform stops getting security fixes, the structure can't hold your services, or nobody can edit the site.
Should we rebrand before we redesign the website?
If a rebrand is due within the year, finish it first or run the two together, so the site isn't reworked twice. A new logo and new colours are a refresh. A new company name on a new domain is a site move: every page address changes, so plan redirects exactly as you would for a rebuild.
Should the agency that built our site decide whether it needs replacing?
Ask them, because they know the site best. Remember that they have a stake either way: a rebuild is paid work, and a fault they built is awkward to admit. Ask for the data behind the recommendation and which checks failed. If you're still unsure, get a second opinion from someone who isn't quoting for the rebuild.
How we wrote this: drafted with AI, then checked and edited by the people who build our projects. Figures come from our own work or from the linked sources.