For the CFO
The software estimate may be based on an old labor model.
Most implementation estimates are defensible arithmetic performed on inherited assumptions about team size, duration, and coordination. The arithmetic is rarely the problem.
The move
Reduce the cost of being wrong before increasing the size of the commitment.
- Challenge the model, not the price
- Negotiating a number leaves the underlying delivery model intact. Examining the model is where the material difference sits.
- Make the first decision reversible
- A smaller first stage can produce evidence while the commitment remains comparatively easy to reverse. Stop remains a legitimate outcome.
- Price the cost of not building
- Deferred systems have a running cost: manual work absorbed by staff, decisions made without information, revenue never captured.
- Stop can be a successful financial outcome
- Evidence that prevents an unjustified programme protects capital. A decision to stop early, on real information, is a return — not a write-off.
- Better evidence improves capital allocation
- A Second Opinion is relevant before a material commitment is approved: it examines the delivery assumptions behind the number rather than negotiating the number itself.
Questions to ask now
Useful questions to put to your own organization — and to any provider.
- 01
How much capital is being committed before useful evidence appears?
- 02
How much of the estimate reflects labor-model assumptions?
- 03
What is the cost of delay?
- 04
What is the cost of being wrong?
- 05
Can the first commitment be smaller and more reversible?