There is a recommendation in the decision-science literature so widely endorsed that it has the air of received wisdom: keep a decision journal. Write down what you decided, why, what you expected, what would change your mind. Daniel Kahneman has advocated for the practice. Annie Duke devotes a chapter of Thinking in Bets to it. We have lost count of the leadership podcasts that close with the host gravely advising the audience to start one.
And in roughly a decade of working with leadership teams, we have seen exactly two that actually kept one in any sustained way. We have seen many more start them. We have seen the Notion pages, the shared Google Docs, the bespoke Airtable bases, the consultant-built templates with twelve fields and a probability slider. They were started in January, abandoned in March, mentioned ruefully in November, and quietly archived before the next planning cycle.
The conventional explanation is discipline. Leadership teams are busy; journals require discipline; therefore journals don't get kept. We do not find this explanation persuasive. The same teams keep their CRMs current, their board books current, their forecast models current, and their compensation spreadsheets disturbingly current. The discipline is there when the artifact is institutionally useful. The decision journal is the artifact that, as commonly designed, is not.
Why individual journals don't survive the leadership team
Most decision journal templates are built for individuals. They reflect the original use case — a single decision-maker, often an investor, recording their own reasoning for later self-review. The fields make sense at that scale: what I decided, what I felt, what I expected, what I'd update on. The fields presume one author, one reader, one continuous mind.
A leadership team is not a continuous mind. It is a fluctuating committee of five to fifteen people, several of whom will rotate within a three-year window, all of whom are accountable to a board that rotates faster, and any of whom may have to defend a decision that someone else in the same room actually made. The individual journal template, transposed onto this institution, fails in three predictable ways.
It fails on audience. An individual journal is written for the future version of the same person. A leadership-team journal is written for the next CFO, the new board member, the regional GM who joins in eighteen months. The reflective tone of an individual journal — I felt uncertain about this; my gut said go; in retrospect I should have asked more questions — is exactly the tone an institutional reader cannot use. They need the data, not the feelings.
It fails on commitment timing. Individual journals are often written retrospectively, in a quiet hour, after the day's events. Leadership decisions need to be journaled at commitment time, when the framing is fresh and the alternatives are still in living memory. A decision recorded a week later is a decision narrated, not a decision captured.
It fails on retrieval. The individual journal is read by the person who wrote it, in chronological order, when they feel like reflecting. The institutional journal needs to be read by people who didn't write it, in topic-ordered batches, on a schedule. It needs to be a searchable, durable, retrievable record. The format that supports this looks less like a journal and more like a one-page memo, written to a template, filed in a known place.
The leadership-team decision journal that actually gets used is not the individual journal scaled up. It is a different object.
The five-field one-pager
The version that, in our experience, survives contact with a real leadership team has five required fields. We have tried longer templates. They get filled in for three weeks and then they get skipped. We have tried shorter ones. They produce entries that are too thin to be useful in the readback. Five fields, on one page, written at commitment, is the dose that takes.
One: what was decided. Two or three sentences. The actual choice, in concrete terms. Not "we discussed expansion options" but "we approved the German market entry at the budgeted twelve million, with a six-month soft-launch in Berlin first and a go/no-go gate at the end of that period." The test for this field is whether someone who was not in the room can tell what the room agreed to.
Two: what alternatives were rejected, and why. This is the field that does the heaviest lifting and is the most often skipped. List the genuine alternatives that were on the table — meaning, things that could have been done instead — and the reason each was rejected. If the answer to "what alternatives were rejected" is "none," the entry is a flag, not an entry. A decision with no rejected alternatives is, by definition, not a choice. It is a recommendation that the room ratified.
Three: what would change the call. The specific evidence, between commitment and the next review, that would cause the team to revisit the decision. Not "if things go poorly." Specific signals: "If the Berlin soft-launch closes fewer than four anchor accounts by month five, we pause the Munich and Hamburg rollouts and reassess." This is the field that converts a vow into a bet.
Four: what we expect to see in six months that tells us we were right or wrong. Related to field three but different. Field three is the kill criterion. Field four is the forward indicator — the leading data the team intends to watch as the decision unfolds, both positive and negative. The discipline of writing the positive case explicitly matters: it prevents the post-hoc redefinition of "what success looks like" once the numbers are in.
Five: named owner. A person, not a function. The owner is responsible for the readback at the next monthly review, for surfacing the leading indicators, and for raising the flag when field three is approaching. "The product org" is not an owner. "Sarah Chen" is an owner. The field that says "we'll all keep an eye on it" is the field that, in twelve months, surprises everyone.
That is the template. One page. Five fields. Filled in at commitment time, before the room leaves. Filed in a known location, accessible to the leadership team and the board, indexed by date and by topic.
The rituals that keep it alive
The template is the easier half. The harder half is the rituals — the small, scheduled, institutionally enforced touchpoints that keep the journal from becoming the artifact that everyone respected and no one read.
Three rituals do almost all of the work.
The first is the monthly five-minute readback. At the start of the monthly leadership meeting, the team reads the entries from the prior month. Aloud. Five minutes is enough — not for discussion, just for the room to be reminded of what it decided and what it said would change its mind. The point of the readback is not to relitigate. It is to keep the recent past in the room's working memory, where it can interact with the decisions on the current month's agenda. Without the readback, last month's commitments fade. With it, the team begins to notice — without prompting — when this month's framing conflicts with last month's reasoning.
The second is the quarterly audit. Once a quarter, the team pulls the entries from the prior three months and re-reads them with the question: which of these decisions has reached its first review checkpoint, and what does the leading data say? This is not yet the annual decision audit (we have written about that elsewhere). The quarterly is shorter, lighter, and forward-looking: are we tracking against the things we said would tell us we were right or wrong? Have any kill criteria been hit? If yes, why haven't we acted on them?
The third is the annual cross-decision review. Once a year, the leadership team reads every entry from the prior twelve months, in one sitting, and looks for patterns across decisions. Are the same load-bearing assumptions showing up in different bets? Are the same people being named as owners and missing the same indicators? Is there a class of decision that the team consistently under-frames? The annual review is the moment the journal stops being a tool for individual decisions and becomes a diagnostic for the team's reasoning.
The journal that no one revisits is a journal that no one writes. The template is the easier half. The rituals are the part that has to be defended on the calendar against everything else that wants the time.
We belabor the point because we have seen it fail this way more often than any other. A team that adopts the template without the rituals will keep the journal for about a quarter. A team that adopts both will, in our experience, still be keeping it at the three-year mark, by which point the institutional learning compounds in ways that are hard to describe to a team that hasn't experienced it.
What changes once the journal is real
Three things change, and they change in roughly this order.
First, decisions get slower at commitment and faster everywhere else. The room takes longer to write the entry — because the entry forces the room to be specific about alternatives and kill criteria that the room would otherwise have left implicit — but the resulting decision moves through execution faster, because the things that would normally cause meetings six months later have already been settled in the entry.
Second, the conversation in monthly meetings begins to refer to entries. "We said in February that the metric we'd watch is X. X is not where we expected. What do we do?" This is the sound of a leadership team using the journal as designed. It is, in our experience, the single biggest indicator that the journal has taken root. Until that sentence becomes routine, the journal is still a project. Once it does, the journal is institutional.
Third, the annual decision audit becomes dramatically easier. Most of the reconstruction work — what did we know, what did we consider, what did we assume — is already on the page, written at commitment time, free of hindsight bias. The audit moves from forensic archaeology to a comparatively quick scoring exercise. Teams that journal for two years find their third annual audit takes a third of the time their first one did.
Kahneman's original case for the journal was about individuals. We have come to think the case is stronger at the institutional level, where the memory is more easily lost, the turnover is higher, and the cost of un-recorded reasoning compounds across more people. The mechanics are different. The discipline is different. The artifact is different. But the underlying logic is the same: a decision that is not recorded at the moment it is made is a decision that, when revisited, will be silently re-narrated to fit what happened next. The journal is the only known defense.
If your team has tried this before and abandoned it, that is a more common outcome than not. The template you used was probably a journal designed for an individual. The rituals were probably never installed. We help leadership teams build the version that survives — both the template and the calendar around it. It is one of the highest-leverage organizational investments we know of, and one of the smallest in terms of effort. The cost is one page a month. The return compounds for as long as the team holds together.
Writes about decision quality at Bayeseon. Reach the team at hello@bayeseon.com.