A multi-year program rarely fails because the original plan was wrong. It usually fails because the program kept executing the original plan faithfully, long after the conditions that made that plan correct had quietly changed underneath it.
This happens for an understandable reason. Changing direction on a large, funded, multi-year initiative is organizationally expensive: it requires admitting the original assumptions need revisiting, it disrupts a delivery rhythm that took real effort to establish, and it's much easier to defer that conversation than to have it. So programs tend to keep moving in the direction they were pointed, measuring success against the original plan rather than against whether that plan still serves the outcome it was built to achieve.
The programs that avoid this build in a specific habit: a recurring, scheduled checkpoint whose only job is to ask whether the original outcome is still the right outcome, separate from the much more common checkpoint that asks whether the program is on schedule and on budget against its existing plan. Those are different questions, and a program can pass the second one for years while quietly failing the first.
Acquisition and contracting structures can either support this habit or actively work against it. A contract vehicle that locks in deliverables far in advance, with no real mechanism for revisiting them as conditions change, all but guarantees the program will optimize for compliance with the original scope rather than relevance to the current outcome. Building flexibility into governance and acquisition strategy from the start isn't a nice-to-have. It's usually the difference between a program that's still worth running in year three and one that's just still running.