Which decisions are worth writing down in a decision journal?
Short answer: a decision is worth writing down when you are genuinely unsure how it will go, something real rides on being wrong, and you will actually find out the answer later. Most decisions in a normal week fail at least one of those three, and trying to log all of them anyway is the single fastest way to abandon a decision journal — not because the format is hard, but because a system that demands too much input quietly turns into a chore you stop opening.
The instinct to log everything, and why it backfires
Once the idea clicks — write down what you expect before you know the outcome, so you can check your reasoning later instead of your memory of it — the obvious next move feels like more coverage. If a few entries are useful, surely logging every decision of the day is more useful still. In practice this is where almost every decision journal dies. A dozen daily entries about what to eat, when to leave, which task to do first are each true and none of them reward rereading, because nothing about them was actually in question. The volume buries the handful of entries that would have taught you something under a pile of ones that could not have taught you anything, and a system that takes real effort for near-zero return gets quietly dropped within a couple of weeks.
The three-part test
Before writing an entry, check the decision against three conditions. All three need to hold, not just one.
1. You are genuinely unsure how it goes
If you already know the answer and are just going through the motions of deciding, there is nothing to record. The test is not "does this feel important," it is "could I be wrong, and would I be surprised to find out I was." Whether to have coffee this morning fails this test even though you make the choice every day. Whether a difficult conversation will land the way you hope does not.
2. Something real is riding on being wrong
Not every uncertain decision needs a paper trail — some genuinely do not matter enough to be worth the two minutes. A useful rough filter: would getting this wrong cost you something you would still notice in a month — money, a relationship, time you cannot get back, a plan you built other things on top of. If the honest answer is "I'd shrug and move on," skip it.
3. You will actually see the outcome
This is the condition people miss most often. A decision journal only works because you eventually compare the prediction to what happened, and some decisions never resolve into anything checkable — the outcome is too diffuse, too slow, or too entangled with everything else going on in your life to ever cleanly confirm or deny what you expected. If you cannot picture the moment, weeks or months from now, where you would look back and say "that's the answer," the entry has nowhere to land. Write about decisions that will hand you a legible result, even a partial one.
What passes the test, in practice
A job offer with a real alternative attached, not just "should I take this" but "this instead of what." A hard conversation you keep rehearsing instead of having. A financial commitment large enough that being wrong about it changes your next few months. A decision to end something — a project, a role, an arrangement — where you are not sure whether you are being appropriately decisive or just impatient. What links these is not size. It is that in each case, you could describe right now what you expect to happen, and you genuinely do not know if you are right.
What does not, even when it feels significant in the moment
A decision made from clear preference, however emotionally loaded, does not need an entry — if you already know exactly why you are doing it and have no real uncertainty about the reasoning, there is nothing later to check it against. Neither does a decision you have made the same way a dozen times before with a track record you already trust; writing it down again does not add information, it just adds an entry. And a decision that is really an emotional reaction wearing the shape of a decision — the response written at 11pm to a message that upset you — is usually better served by waiting until morning than by being logged in the state it was made in, since the entry would record the mood more than the reasoning.
The recurring case: when it is a principle's job, not the journal's
The trickiest category is the decision that keeps coming back in roughly the same shape: whether to take on one more commitment this month, whether to respond to criticism right away or sleep on it, whether to say yes to a request you resent in advance. The first time one of these shows up, it can genuinely be decision-journal material — you are testing your own judgment on something new. But once you have made the same call three or four times and can already predict which way you will go, the journal has done its job and handed you something more useful than another entry: a pattern worth turning into a standing rule. At that point, writing a fresh prediction every time is not testing anything anymore, because you already know the answer before you write it. That is the moment to convert it into a principle with a behavior standard attached, and let the daily check-in track whether you kept it, instead of asking the decision journal to relitigate a question you have already answered for yourself.
The rule of thumb, if you do not want to run the checklist every time
Ask one question: am I actually about to find out something about my own judgment, or am I just documenting a Tuesday? If the honest answer is the first one, write the three lines before you decide. If it is the second, let the moment pass — a decision journal that only records the choices that were genuinely in question is one you will keep using; one that tries to record everything is one you will quietly stop opening, usually right around the point it would have started paying off.
CreedOS's daily check-in leaves room for a short note next to whatever you are already tracking, which is enough space for a decision-journal entry when a decision actually warrants one — and when a recurring one has stopped being a question and started being a pattern, that is exactly what a creed's behavior standards are for, with a review that shows whether you are keeping it. It is free, with no in-app purchases.