Every owner of an ageing system faces the same question: keep fixing it, or start over? Vendors who only build new systems will always recommend rebuilding. Here's how to decide on evidence instead.
Start with what the system is worth
List what the current system does well: the workflows staff rely on, the data it holds, the integrations that work. Replacement cost isn't just the build — it's migration, retraining, parallel running and the features you'll forget until they're gone.
Then price the real problems
Get the faults diagnosed properly. A R30,000 fix list on a system worth R400,000 of embedded process knowledge is obviously worth doing. A R300,000 repair on an architecture that can't support next year's plans usually isn't.
The four questions that decide it
1. Is the foundation sound — database, architecture, security model? 2. Can the system support the next 3 years of plans? 3. Is the cost of fixing less than half the cost of replacing, including migration? 4. Do you have (or can you get) developers who understand the stack?
Three or four yeses: fix and modernise in stages. Two or fewer: price a rebuild seriously — and run the old system properly maintained while you build.
The hybrid most businesses miss
Strangler-pattern modernisation — rebuilding module by module while the old system runs — is often the best answer. New rostering module this quarter, new reporting next quarter, old system retired piece by piece. Lower risk, earlier value, no big-bang cutover.
Get the decision in writing
Whatever you choose, get a written assessment with options, costs and risks. It protects the budget, the timeline and the person making the call.