3 ideas
A decision is rarely a choice between right and wrong; at best it is a choice between 'almost right' and 'probably wrong', and more often between two courses neither provably better. Because of this, starting from 'the facts' is an illusion: people begin with opinions, and the effective executive treats them as untested hypotheses to be checked against reality rather than suppressing them in favor of a fact-gathering search.
Drucker holds that no decision should be made unless there is disagreement, citing Alfred Sloan's practice of adjourning a GM meeting where everyone agreed so the group could develop disagreement and understand the issue. Dissent protects the decider from being trapped by one view, supplies ready alternatives when the chosen course fails, and stimulates imagination. The executive should assume each opponent is reasonable and ask what that person must be seeing to hold their position.
The first question is 'What is this about?', and the most common error is solving the wrong problem or treating a generic situation as a string of unique events. The executive classifies each problem as generic, needing a rule or policy, or as truly exceptional, then specifies the boundary conditions the decision must satisfy. Many decisions fail because they address symptoms plausibly while missing the actual underlying problem.