Replace a legacy system without stopping the business

Map what it actually does, including the undocumented behaviour people rely on. Then replace it in slices with both running side by side, so each slice can be reversed on its own instead of betting the business on one cutover.

The problem

Something critical runs on software that is a decade old, unsupported, or written by someone who left. It cannot be switched off for a weekend and nobody fully knows what it does.

How we approach it

Map what it actually does, including the undocumented behaviour people rely on. Then replace it in slices with both running side by side, so each slice can be reversed on its own instead of betting the business on one cutover.

What changes

The old system retired, the data migrated and reconciled, and no day where the business could not operate.

The service this falls under

Software development

Web apps, SaaS platforms, APIs and the systems underneath them.

1 to 7 days for most builds. A platform with many moving parts runs weeks to months, and the largest have gone to eight. It depends on the scope.

Cloud and infrastructure

Edge, databases, single sign-on, CI, observability and cost control.

1 to 5 days for most work. A migration is scheduled around your business, and platform work is often ongoing rather than a project with an end.

Last reviewed · Site changelog