Why is the data-science process commonly described as iterative?

A. Every stage must be completed exactly once B. Findings at later stages often require revisiting earlier assumptions, data, or features C. Models cannot be evaluated until deployment D. Iteration means repeating the same error unchanged