Modernising legacy systems without the 'rip and replace' risk
A full rewrite of a legacy system sounds decisive. In practice, full rewrites fail 70 to 88% of the time — and the reasons are almost always organisational, not technical.
A "big bang" replacement means running the old system right up until a single cutover moment, then switching everything to the new one at once. It's disruptive by design, puts enormous strain on people and resources in a short window, and gives you no way to course-correct once you've committed — which is exactly why it fails so often.
A phased alternative: modernising without stopping the business
The strangler pattern — named for the way a strangler fig gradually grows around a host tree without it ever falling — replaces a legacy system one capability at a time. A routing layer sits between users and both systems, sending each request to whichever one currently owns that piece of functionality. New components are built and proven incrementally while the legacy system keeps running everything that hasn't moved yet, until eventually there's nothing left for it to do.
Why this actually reduces risk
Breaking replacement into small increments means problems are confined to whatever piece just moved, not the entire system. It also means the project can change direction if business needs shift, without losing all the work done so far — something a big-bang rewrite can't offer once it's underway. And because value gets delivered every quarter rather than only at a distant final cutover, the project stays easier to defend to the people funding it.
The two risks worth watching
Keeping data synchronised between the old and new systems while they run in parallel takes real planning — it's the most common technical snag. The bigger risk is organisational: stalling once the highest-value pieces are done, leaving both systems running indefinitely and quietly doubling ongoing maintenance costs. The pattern only works if the team holds the discipline to keep moving toward full retirement of the old system, not just partial relief from it.
What a realistic timeline looks like
Phased, business-case-led modernisation typically reaches positive ROI in 12 to 14 months, against 36 to 48 months for a full rewrite — while the production system keeps running throughout, rather than betting the business on a single cutover weekend.