Rebuild or refactor? A decision framework that survives the board meeting

Every legacy system eventually prompts the same question, and the answer is almost always argued on instinct.

Ask four things instead. Can you deploy safely today? Do you understand the data model? Is the cost of change rising month over month? And is the business logic still correct?

If the answer to the last question is yes, refactor. Business logic is the expensive part, and rebuilds throw away years of accumulated correctness in exchange for a cleaner stack nobody is paying you for.