Modernising legacy systems one piece at a time

The State of Digital Government Review, published in January 2025, estimated that legacy systems made up around 28% of central government systems in 2024, up from 26% the year before. In police forces the proportion ranged from 10% to 70%. Legacy systems are costly to maintain, harder to secure and increasingly difficult to staff, so the case for modernising them is rarely in doubt. The question is how to do it safely.

The risk of a single switchover

Replacing a core system in one go concentrates risk at a single moment. Programmes of this kind tend to run for years, deliver value only at the end, and chase a moving target, because the old system and the organisation around it keep changing while the replacement is built.

Replacing it piece by piece

The GOV.UK Service Manual describes a well-established alternative, sometimes called the strangler pattern: "initially introduce the API we want as a wrapper around the legacy system", then replace the legacy system behind it independently of the new services. In practice:

• establish what the system actually does, including behaviour that was never documented, by examining its data and working with the people who use it
• put a stable interface in front of it, so that new services depend on the interface rather than on the old system
• move one capability at a time behind that interface, starting where the value or the risk is greatest
• run old and new side by side, comparing their results before switching over
• migrate data in stages, with reconciliation checks and a way to roll back
• retire each part of the old system as soon as nothing depends on it

Avoiding the next legacy system

Government guidance on managing legacy technology recommends "an iterative or phased migration" and "designing your system with components that you can swap and move easily". The same discipline that makes a modernisation safe also stops the new system from becoming the next one to replace: clear interfaces, open standards, documentation kept current, and regular updates rather than occasional overhauls.

brytloop's approach is to integrate first and replace second: connect the systems an organisation already has, deliver value from them early, and then retire old components at a pace the organisation can manage.