Most website redesigns are proposed for the wrong reason. Someone senior has grown tired of looking at the site, or a competitor launched something with more animation, and a budget line appears. Eighteen months later the organization has a site that looks current and works about as well as the one it replaced, because nothing that was actually broken got diagnosed.
A redesign is a large, disruptive, expensive instrument. It is the right instrument more often than people think, and it is the wrong one more often than agencies admit. Here is how to tell which situation you are in.
Eight signs you need one
1. Your team cannot publish without a developer
This is the strongest signal there is, and the one organizations mention last because they have normalized it. If adding a page means filing a ticket, your site is not a communications tool, it is a bottleneck. The cost is not the developer’s time, it is every update that never happened because it was too much trouble to request.
Measure it before you argue about it. How long does it currently take to publish a routine update, from the moment someone writes it to the moment it is live? If the answer is measured in days, you have your business case.
2. It fails on mobile, measurably
Mobile is the majority of traffic for most organizations now, and “it has a responsive theme” is not the same as working. The test is Core Web Vitals on a mid-range phone on a real connection, not a desktop audit score. Largest Contentful Paint should land under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1.
Check your own numbers in Search Console rather than trusting a one-off lab test. If real users are failing those thresholds on your main templates, that is a structural problem and it usually does not respond to tuning.
3. You now have an accessibility requirement it cannot meet
This has moved from a nice-to-have to a contractual and regulatory condition for a growing share of Canadian organizations. If your colour palette cannot reach the contrast requirement, your components were built without keyboard support, and your document library was never tagged, you are not looking at a remediation project. Retrofitting accessibility into a design that ignored it routinely costs more than building it correctly from the start.
4. The structure fights the content
You can see this in the symptoms: pages that exist only because there was nowhere else to put something, navigation that mirrors your org chart instead of what people came for, three different page types doing the same job because each was built by a different person in a different year.
This is a content model problem, and no amount of visual design fixes it. It is also the failure that returns fastest if a redesign skips the information architecture and goes straight to templates.
5. Plugins are holding it together
Count them. If the number is past thirty, and several are doing jobs that overlap, and at least one has not been updated by its author in two years, the site is being held up by other people’s abandoned code. Every one of those is a security exposure and a reason an update cannot be applied.
6. Nobody dares update it
A site that cannot be updated safely is already failing, it just has not failed publicly yet. When core and plugin updates get deferred because the last one broke the homepage, you are accumulating risk on a schedule you do not control. The eventual incident will not arrive at a convenient time.
7. The organization changed and the site did not
You serve different clients than you did when it was built, or you sell something you did not sell then, or you merged, or you now operate in two languages. A site describing an organization that no longer exists is a real problem, as distinct from a site that is merely unfashionable.
8. You cannot tell whether it works
No analytics, or analytics nobody reads, or no definition of what a successful visit is. If you cannot say which pages produce enquiries, you cannot say whether the current site is underperforming, and you will not be able to prove the new one is better either. Fix the measurement before you spend the budget, not after.
Three signs you do not
You are bored of it
You look at your site far more than any customer does. Familiarity reads as staleness to the people who work there and as consistency to everyone else. This is a real feeling and a poor reason to spend a year’s budget.
A competitor launched something
You cannot see their conversion rate, their maintenance burden, or whether the project ran over. Copying a competitor’s redesign is copying a decision you have no evidence about.
One page is underperforming
If your service page converts badly, rewrite the service page. Rebuilding forty templates to fix one is the most common form of overspending in this industry. Targeted work on the pages that carry your traffic is cheaper, faster, and gives you evidence about what actually moves the number.
How to tell the difference
Before committing, spend two weeks gathering five things:
- Your top twenty entry pages and what happens after someone lands on them.
- Core Web Vitals from real users, per template.
- Total page and document counts from an actual crawl.
- The publishing turnaround time, measured rather than estimated.
- A keyboard and screen reader pass on your three most important templates.
If those five turn up problems concentrated in a handful of pages, do targeted work. If they turn up problems that are structural, present on every template and traceable to how the site was built, you need a redesign, and now you can describe it in terms a budget holder will recognize.
What a redesign should not cost you
If you do proceed, four things are frequently lost and should not be:
- Search visibility. Every existing URL maps to exactly one new URL, in a single redirect hop. This is not a launch-week task.
- Content. Migrated by script, not retyped. Retyping loses things quietly.
- Integrations. Inventory what connects to the site before anyone designs anything.
- Your team’s capacity. A redesign that makes publishing harder has failed, whatever it looks like.
And where the platform allows it, rebuild in place rather than building a parallel site for a year and switching over in one night. Section by section is less dramatic and much harder to get badly wrong.
If the diagnosis points to a rebuild, our website redesign and modernization page sets out how we run one, including the URL mapping and the measured baseline.