An architect must justify continued investment in a modernisation programme after eighteen months with the end state not yet reached. What should the argument be?

The argument

Evidence of value already delivered, plus the cost of stopping now.

If the programme was structured incrementally, the first part is available: which slices are complete, what each one changed, and what it is worth. Change lead time in the migrated areas, incidents eliminated, cost removed, capabilities enabled that were previously refused.

If none of that is available, the programme was not structured incrementally, and that is the finding to report honestly rather than the argument to make.

The cost of stopping

This is the stronger half of the argument and the one usually omitted.

Stopping mid-migration is the worst outcome available. The organisation operates both systems, pays for both, maintains a synchronisation layer, and holds all the complexity of the transition with none of the end state's benefits.

Perpetual coexistence is not a neutral pause; it is a permanent tax. Quantifying it — the cost of operating both, the reconciliation effort, the change cost of every feature that must be built twice — is frequently what makes continuation obviously correct.

What else the argument needs

1. A revised, credible end state with a date, informed by what the first eighteen months actually taught. If the original plan was wrong, saying so and replanning is far more credible than restating it.

2. Explicit re-scoping. Eighteen months in, the programme has almost certainly accumulated work that is not on the critical path to the end state. Cutting it is evidence of discipline and improves the date.

3. The remaining slices with sponsors and benefits, so continuation is a sequence of fundable steps rather than an open-ended commitment.

4. A stopping point that is not the end state. If funding is genuinely constrained, which subset of slices reaches a stable configuration — one system for these capabilities, one for those — where coexistence is bounded rather than permanent? Having that answer is what makes the conversation a negotiation rather than a binary.

The framing

The decision is not "continue or stop" but "reach a stable end state or pay coexistence costs indefinitely". Presented that way, the question becomes which stable state to reach and how quickly — which is a far better conversation than defending the original plan.