the old linear models of thinking are no longer sustainable
changing a constraint or constraints and seeing what happens.
Which is why evaluation should really start with identification and mapping of constraints coupled with risk assessment which includes the percentage of current behaviour attributable to dark constraints (we can see the effect but not the cause or modulating factors). This is the real danger in political policy making and rarely if ever articulated.
One of the key changes is to shift from major initiatives to minor initiative that become major as they start to create a consistent change in direction for the better. That way we mitigate the impact of unintended consequences and the excessive investment which reduces the possibly of early admission of failure.