Six Rs Decision

The value of the framework is that it forces a per-application decision rather than a programme- wide strategy, and the honest answer for a large estate is a mix.

Rehost moves it as-is: fastest, lowest risk, and it carries every existing inefficiency along with an infrastructure bill that is often higher than before. Replatform makes targeted changes — a managed database, a container runtime — for meaningful benefit at modest risk, and is the sweet spot for a large share of applications. Refactor re-architects for the target environment: highest cost, highest benefit, justified only where the application's future warrants it. Repurchase replaces it with a SaaS product, which is frequently the right answer for non-differentiating capability and is under-selected because it is a procurement decision rather than an engineering one. Retire applies to more of the estate than anyone expects. Retain leaves it where it is, which is a legitimate decision when a mainframe application works and the business case does not close.

The trap that produces disappointing programmes: choosing rehost for everything to hit a deadline, then discovering the promised savings do not materialise because nothing was optimised, and that the modernisation is still ahead of you with the political capital spent.

The framing that improves the decisions: decide by the application's expected remaining life and rate of change. Something being replaced in eighteen months should be rehosted or retained. Something central for a decade justifies refactoring.