A Practical Guide to Better Decision-Making

1. Frame the decision before comparing options

Write one sentence that names the choice, the person responsible, and the time by which it must be made. If the sentence contains several choices, split them. “Should we launch, hire, and change pricing?” is three decisions with different evidence and reversibility.

List the status quo as an option when doing nothing has real consequences. Also distinguish a permanent commitment from a small experiment that could produce new information.

• What changes if no decision is made by the deadline?
• Which constraints are fixed, and which are preferences?
• Can any option be tested at lower cost or smaller scale?
• Who has relevant information that is currently missing?

2. Turn assumptions into questions you can test

Most uncertain decisions rest on a few load-bearing beliefs: customers will behave a certain way, a new role will provide growth, a project will take a certain amount of time. Name those beliefs explicitly.

For each assumption, ask what evidence would raise or lower your confidence. The goal is not perfect certainty. It is to spend research effort where a new fact could actually change the choice.

3. Compare tradeoffs and reversibility

Matrices can help organize criteria, but do not let an arbitrary total score conceal uncertainty. Keep a note beside every rating explaining the evidence, the confidence level, and the cost of being wrong.

Reversibility affects how much analysis is justified. A low-cost, reversible choice often deserves a fast experiment. A difficult-to-reverse choice deserves a wider search for failure modes and qualified advice where appropriate.

4. Keep a decision journal and review it

Before acting, record what you expect, why, and how confident you are. Set a review date tied to when useful evidence should exist. Without this snapshot, memory tends to make the original outcome feel more predictable than it was.

At review time, evaluate the reasoning separately from the result. Identify what was knowable, which assumption failed, what was luck, and what you would repeat next time.

Where AI fits—and where it does not

AI can suggest a missing option, articulate a counterargument, or turn a vague concern into questions. Those are prompts for further thinking, not verified evidence. Check factual claims and remember that unstated context is invisible to the model.

For medical, legal, financial, safety, or other high-stakes choices, involve qualified professionals. A journal or AI response cannot replace context-specific professional advice.