The most expensive automation is the one that ships, costs to maintain, and never pays back. The cheapest one to kill is the one you score out before you build it.
How should you score an automation candidate?
Score it on payback, not novelty, the time and margin it returns, not how it looks in a demo. We rank every candidate initiative on the value it returns, and the unglamorous ones usually win.
Why does build effort belong in the ROI model?
Effort is a real axis because a slow high-payback initiative often loses to several fast medium ones. A high-payback initiative that takes a year is often beaten by three medium ones that ship in weeks. We plot payback against effort and sequence accordingly.
Is the ROI model a one-time gate?
No, the ROI model is a living scoreboard you run the program against, not a one-time gate. It gets updated as the system goes live and the real numbers come in.