Your Site Isn’t Broken. That’s the Problem.
The websites hardest to replace are the ones that still (mostly) work
If your organization’s site were actually broken — returning errors, failing logins, locking out half your users — the decision to rebuild would be easy. Leadership would approve the budget, and the work would start right away.
But that’s not how most websites fail. They fail slowly, in ways that are easy to ignore. A workaround here. A manual process there. Change requests that used to take a day now take a week, and nobody can quite explain why. The site is still running — it just works against you at every turn.
We’ve had a version of the same conversation with our public-sector clients for years: “We know we need to do something, but we’re not sure if we’re at the point where we need to start over.”
That question is harder than it sounds, and getting it wrong in either direction is expensive. Rebuild before you need to, and you’ve wasted money. Patch past the point of no return, and you’ve only deferred a larger, messier project.
Not every problem calls for a rebuild. Many are just the result of neglected maintenance. The updates aren’t exciting, but a well-maintained site on a solid platform will serve your organization for years.
If your site’s core architecture still fits what you need it to do — you can publish content, manage users, integrate the tools you rely on, and meet your accessibility baseline — then patching and ongoing maintenance is probably the right track.
Maintenance also makes sense when the problems are specific: a module that needs updating, a content type that needs restructuring, a design component that’s aging out. These are solvable issues that don’t require tearing down the foundation.
The bigger question is whether the problems stay specific, or start pointing at the foundation of the site itself.
Sometimes you just know the site is unsalvageable. But if you’re unsure, here are three signs your website has crossed from “needs maintenance” into “time to start over.”
1. The platform can’t do what you need it to do
Sometimes the feature you want simply isn’t possible with your current build. Personalization, AI-powered search, enrollment or CRM integrations, multilingual content — they aren’t supported, or they require extremely complex customizations.
This gap isn’t operational. It’s the distance between what your organization needs from its content management system and what the platform was designed to support.
The platform you run today was built for a different organization than the one you’ve become. Keeping up with updates won’t close that gap.
2. Governance has become a people problem instead of a system problem
On a site with dozens or hundreds of contributors, consistent content management requires the platform to enforce standards.
When it doesn’t, you’re relying on individuals to self-regulate. Even well-intentioned editors drift, and no one person can wrangle them all. Pages get longer and messier, and users get lost.
The end result is content sprawl that gets worse with every new contributor, and can’t be cleaned up without first having a system that helps prevent the next round of it.
3. The site only works because one person knows how it works
How many people in your team know how to update your website? It often comes down to a single person.
Someone on the team (often in IT, but sometimes in communications) has spent years learning and skirting the platform’s quirks. They know which overrides conflict, which content types break if you touch them, and which workarounds are holding things together.
When they’re out sick, the site gets harder to manage. When they leave, you find out what you’ve actually been running. A platform that requires a single keeper isn’t a platform you can maintain at scale. It’s a liability.
Any one of these is a signal worth taking seriously. All three together is a clear answer that it’s time to rebuild.
For government and higher ed in particular, ADA Title II has made the rebuild-or-patch question more concrete.
An older platform built before modern accessibility standards doesn’t become compliant by bolting on a third-party accessibility widget. Those tools apply surface-level fixes without touching the underlying code — which is exactly what enforcement looks at. If compliance requires structural changes the platform can’t support, patching isn’t a path to compliance. It’s a delay tactic.
Accessibility requirements have teeth, and can’t be ignored. Legacy platforms often make structural compliance difficult, and the quick fixes tend to pile up into bigger problems under the next audit.
If you’re realizing you need to act, your first instinct might be to call a vendor or issue an RFP for a redesign.
Before you do, start with some internal work. The organizations with the best rebuilds do two things before they talk to a vendor.
1. Talk about what you need (and write it down)
Deciding what you need in a rebuild and deciding who should build it are different decisions, made at different times with different information. Organizations that try to answer both at once are setting themselves up to fail.
Figure out what you need internally first. Hold the meetings, talk to your stakeholders, do a little research. You’ll still want help before any final decisions — but the vendor conversations go a lot further when you arrive with clarity about the problems you need to solve, rather than asking an agency to define them for you.
2. Decide who owns the project (internally)
A rebuild of a large website can take anywhere from six months to over a year. A project that long, with that many moving pieces and people, needs a project lead — someone with the authority to make decisions, gather input across departments, and keep things moving when other priorities compete.
The lead doesn’t need to be technical. But they do need to be empowered, and they need to be named before the engagement starts, not after it’s already in motion.
Try our free CMS assessment.
That uncertainty is normal, and it’s worth resolving before you start a vendor conversation or an internal budget process.
The assessment takes less than five minutes, and surfaces the patterns above in the context of your specific site.
When done, you'll have a clearer read on whether you’re in maintenance territory or approaching the line where rebuilding becomes the more practical path.
About the author
Dan Moriarty is the co-founder, CEO and chief strategist with Electric Citizen.