Day 99: Build a decision matrix you can challenge.

Hey, Ivy here. A decision matrix can help when several options look reasonable for different reasons. The catch is that a polished total can feel more objective than it really is. Today's exercise keeps your judgment visible and gives you a way to challenge the result. Paste this prompt into your AI assistant: "I am choosing among [OPTIONS]. First, ask me for the criteria and a weight for each criterion, with weights totaling 100%; then ask me to score every option from 1 to 10 on each criterion; build a weighted decision matrix; show the arithmetic for every weighted score and total; identify the criterion driving the result most; rerun a sensitivity test with that criterion's weight reduced by 50%, redistributing the difference proportionally across the other criteria; compare the rankings, state what changed, and flag any assumptions or missing information; do not make the decision for me." Start by replacing the bracketed options with the real choices. Answer the questions about criteria, weights, and scores, and make sure the weights total 100%. When the matrix appears, check one or two weighted scores yourself by multiplying the score by the weight. Then read the sensitivity rerun closely. If cutting the result-driving weight by 50% changes the ranking, you have learned that the choice depends on that preference. If the ranking stays put, the comparison is more stable under that specific test, though it is still not proof that the top option is right. The outcome is a decision record you can inspect, explain, and revise instead of a recommendation you have to trust. Tomorrow, I will show you how to find the story hiding in a small spreadsheet.