Which is the CORRECT approach for managing a large improvement initiative as smaller iterations?

Each iteration should be designed before starting the initiative and implemented without feedback

Feedback should only be taken into account when one iteration fails to meet its objective

Feedback should be reduced for large improvements as it is unlikely that circumstances will change

Each iteration should be continually re-evaluated based on feedback