What a useful entry contains
An entry takes fifteen minutes at the point of decision, and the fields below are the working core. The discipline is writing them before the outcome arrives; every one of them can be gamed by a later hand.
| Field |
What to write |
Why it matters later |
| The decision and date |
What was decided, by whom, against which alternatives |
Pins the decision to a moment, before events resequence it |
| What was known |
The facts, constraints and unknowns as they stood, including what you chose not to find out |
The core defence against "we knew all along" |
| Expected outcome |
What you expect to happen, by when, with a rough probability or range |
Vague expectations cannot be scored; a number can |
| The reasoning |
Why this option, and what evidence or event would prove it wrong |
Lets you separate a bad process from bad luck |
| State of the decider |
Tired, rushed, confident, under pressure to agree |
Mood shapes judgement, and it is the first thing memory launders |
| Review date |
When the entry will be scored against reality, and by whom |
An entry nobody revisits teaches nothing |
Why contemporaneous records beat memory
The journal defeats hindsight bias by removing the material it works on: with expectations written down, the retrospective claim to have seen the failure coming has to survive contact with a dated page saying otherwise. Two further effects follow over time. The record separates decision quality from outcome quality, so a sound decision that failed can be distinguished from a reckless one that happened to pay off, and the organisation stops promoting luck. And a run of scored entries reveals personal base rates: the timelines that were always optimistic, the competitor responses that were always underweighted, the confidence that ran fifteen points ahead of the hit rate. That calibration is unobtainable any other way, because memory will not supply it.
The decisions worth journaling
- Consequential and infrequent commitments (senior hires, vendor selections, pricing moves, market entries) where feedback arrives too slowly for instinct to learn from.
- Committee decisions where what was agreed, and why, becomes contested within a quarter. The entry is cheaper than the argument.
- Any class of decision the organisation makes often enough to build a track record: the journal turns repetition into calibration.
A six-month illustration
A distribution business decides to switch its secondary supplier network. The journal entry, written the day the board approves it, records: margin improvement expected within two quarters, transition friction limited to a handful of key accounts, confidence around 70 per cent, and a flag that the incumbent's response is the least-evidenced assumption. Six months later, margin is flat and the friction is wider than forecast. Without the entry, the narrative would already have healed over ("we always knew the transition would be rough"). With it, the miss is specific: the incumbent matched terms to defend the accounts everyone assumed were sticky, exactly the assumption flagged as weakest. That finding, written into the next channel decision as a mandatory question, is the journal working as designed.
How the practice gets corrupted
- Entries are written after the outcome is visible. A backfilled journal is a press release addressed to the future.
- Expectations are kept vague ("we expect meaningful improvement") so that any outcome can be read as confirmation. If nothing could count as a miss, nothing will be learned.
- Only the comfortable decisions get recorded; the risky ones stay out of the book, which removes exactly the entries with the most to teach.
- Review dates pass unobserved, so predictions accumulate unscored and the journal becomes an archive rather than a feedback loop.
- Entries are drafted for a future audience, written to look prudent to whoever might one day read them. A journal that has turned into reputation management has stopped being a journal.
From private habit to organisational discipline
The journal scales into two organisational artefacts. A decision log, a shared register of significant decisions with owners and rationale (the decision log template sets out a working format), gives governance the same protection the journal gives the individual. A post-decision review, a scheduled scoring of expectations against results (the post-decision review template structures it), makes the review date a calendar event rather than a good intention. Independent perspectives enter at the moment of writing: the two hardest fields, the expected outcome and what would prove us wrong, are exactly what selected senior operators can pressure-test from experience of comparable commitments. A client brief to the Global Board before sign-off gives the entry a reasoning section that has already survived challenge, which is a better inheritance for the review meeting than hope.
