The spreadsheet looks more certain than the inputs
A ranking can feel objective because it ends in a number. But reach may be guessed, impact may reflect one stakeholder’s enthusiasm, and effort may change after technical investigation. The arithmetic can be correct while the decision remains uncertain. Good prioritization keeps that uncertainty visible.
Agree on what the criteria mean
Before scoring, define the goal, time horizon, and scale. What does high impact mean for this decision? Does reach refer to accounts, people, or events? Are effort estimates comparable? Inconsistent definitions can create apparent differences that disappear when the team is discussing the same thing.
Make evidence part of the discussion
For each estimate, record its source and whether it is observed, modeled, or assumed. A well-supported opportunity should be distinguishable from an exciting possibility. Missing evidence is not always a reason to reject an idea, but it may be a reason to investigate before making a large commitment.
Change the uncertain inputs
Try plausible lower and upper values. If a small change reverses the order, the result is fragile. The next useful step may be gathering evidence rather than debating decimal places. Consider dependencies, risk, and strategic fit explicitly so they do not reappear as hidden reasons to ignore the ranking.
Let AI inspect the logic
Give your approved assistant the criteria and supporting evidence. Ask it to find inconsistent scales, unsupported assumptions, and tradeoffs the table misses. Do not ask it to invent customer reach or confidently select a winner from missing information. A useful critique makes the human discussion more informed.
Document why you chose
End with a short rationale in plain language. State what you selected, what you deferred, and what evidence would change the choice. If the explanation only says “it had the highest score,” the team may not have finished the decision work.