How to Keep a Decision Journal That Sharpens Judgment

A decision journal works because memory lies. You remember your good calls as skill and your bad calls as bad luck. Write the decision down before you know the outcome, and that story falls apart. You get a clean record of what you actually believed, why, and what you expected. Read it back six months later and you learn more about your judgment than any personality test will ever tell you.

I run about twenty companies alone. I make a lot of decisions and I forget most of them. The journal is how I catch my own patterns before they cost me a quarter.

What goes in a decision journal

Keep it short or you will stop. Four fields per entry.

The decision. One sentence. "Killing the paid tier on product X." Not a paragraph of context. The context is in your head today; the point is to reconstruct it later.

The reasoning. Why this, why now, what you expect to happen. This is the part that stings when you reread it. Write the real reason, not the reason you would give a board. "I am tired of supporting it" is a valid entry. Hide it and the journal is useless.

Your confidence. A number. 60 percent, 80 percent. Vague confidence is how people avoid ever being wrong. A number pins you down.

The expected outcome and a check date. "Revenue flat or up within 90 days. Check April 1." This is the whole game. Without a check date you never close the loop, and an open loop teaches you nothing.

That is it. Two minutes an entry. If it takes longer you will quit, and a journal you quit is worth nothing.

How to review it so it changes your judgment

Writing is half. The review is where judgment actually improves. On the check date, open the entry and mark it: right, wrong, or too early to tell. Then answer one question. Was I right for the reason I thought, or right by accident?

Being right by accident is the dangerous case. It feels like skill and it trains you to repeat a bad process that happened to work once. The journal is the only place that distinction survives, because your gut will quietly rewrite the accident as foresight.

Do this for a few months and patterns show up. You are overconfident on hiring and underconfident on pricing. You rush decisions on Fridays. You talk yourself into things when someone smart pushes back. These are not character flaws you read about in a book. They are your own logged behavior. That is the difference between self-knowledge that sticks and self-knowledge that stays an excuse.

Why a journal beats an assessment for judgment

An assessment gives you a snapshot: here is how you tend to operate. Useful, and tools like AstraTalk are built to give you that read fast. But a snapshot cannot show you your error rate on real decisions. Only a running log does that. The two work together. The assessment tells you where to look, the journal tells you whether the pattern is actually costing you.

If you already keep a journal, feed it back into the assessment. When a tool tells you that you decide fast under pressure, your entries will show you whether fast decisions were the right ones or the ones you flagged as wrong. That is how you turn a self-assessment into an operating system instead of a horoscope you read once and forget.

Common mistakes that kill a decision journal

People overbuild it. A spreadsheet with twelve columns and a rating rubric. You will abandon it by week two. Four fields, plain text, done.

People only log big decisions. The small repeated ones teach you more, because you get many samples. How you respond to a customer complaint, whether you say yes to a meeting. Patterns need volume.

People skip the confidence number because pinning it down feels uncomfortable. That discomfort is the signal you are doing it right.

And people never reread. A journal you write but never review is a diary. The review is the product. Block thirty minutes a month, reread the entries that came due, and mark them honestly. That single habit does more for your judgment than any course. If you want the wider practice, pair it with a personal quarterly review so the small entries roll up into something you can act on.

Start today. One decision, four lines, a check date. In three months you will know things about how you think that no test could have told you.

I run about twenty companies alone. I make a lot of decisions and I forget most of them. The journal is how I catch my own patterns before they cost me a quarter.

What goes in a decision journal

Keep it short or you will stop. Four fields per entry.

The decision. One sentence. "Killing the paid tier on product X." Not a paragraph of context. The context is in your head today; the point is to reconstruct it later.

Your confidence. A number. 60 percent, 80 percent. Vague confidence is how people avoid ever being wrong. A number pins you down.

The expected outcome and a check date. "Revenue flat or up within 90 days. Check April 1." This is the whole game. Without a check date you never close the loop, and an open loop teaches you nothing.

That is it. Two minutes an entry. If it takes longer you will quit, and a journal you quit is worth nothing.

How to review it so it changes your judgment

Why a journal beats an assessment for judgment

Common mistakes that kill a decision journal

People overbuild it. A spreadsheet with twelve columns and a rating rubric. You will abandon it by week two. Four fields, plain text, done.

People only log big decisions. The small repeated ones teach you more, because you get many samples. How you respond to a customer complaint, whether you say yes to a meeting. Patterns need volume.

People skip the confidence number because pinning it down feels uncomfortable. That discomfort is the signal you are doing it right.

Start today. One decision, four lines, a check date. In three months you will know things about how you think that no test could have told you.