An analysis nobody can audit is an opinion

A business case is not a document produced at the end of an analysis. It is a structure imposed at the beginning: what decision is being made, by whom, by when, and what threshold would change the answer.

Without that structure, studies drift toward thoroughness. They answer adjacent questions exhaustively and the actual question obliquely, and they arrive after the decision has already been made informally.

The discipline that makes an analysis hold up is unglamorous. Every figure traces to a source. Every assumption is named where a reviewer can find it rather than buried in a model. Sensitivity says plainly which assumptions move the recommendation and by how much.

And when two alternatives are indistinguishable within the uncertainty, say so. A false precision that collapses under questioning costs a program more credibility than an honest non-answer.

The test we use internally: could a reviewer who dislikes the conclusion reproduce it? If not, it is not finished.