In short: The operative reason for many decisions is political, personal, embarrassing or legally sensitive. A system that only records the sayable reason produces clean, coherent and quietly false records. No design fully fixes this, but several make it much better.
Key points
- Unsayable reasons fall into four kinds: political, relational, embarrassing and legally sensitive.
- Legally sensitive reasons need counsel's design, set with legal and compliance from the start, not more capture.
- A completeness flag lets people mark "stated reason, not the whole story" without writing the whole story.
- An optional cynic's question gives a deniable way to record the real driver.
- Outcome write-back exposes decisions whose stated reasons repeatedly fail to predict results.
Ask anyone who has run decision reviews: the stated reason and the operative reason often differ. The deal was approved partly because the sponsor is leaving and wanted a win. The vendor was chosen because the alternative's team burned someone two years ago. The launch slipped because nobody had the capacity, not because of the market conditions in the memo. None of that gets typed into a shared system.
This matters more for judgment systems than for any other kind of record. If the stated reasons are systematically incomplete, precedent misleads. The next person sees a coherent rationale and copies a decision whose real driver no longer applies.
Four kinds of unsayable reason
It helps to separate them, because each needs a different response.
Political: sponsorship, internal alliances, timing around people.
Relational: history with a counterparty, personal trust or distrust.
Embarrassing: lack of time, lack of skill, a mistake being quietly corrected.
Legally sensitive: reasons that should not be written casually at all, because records are discoverable.
The last category deserves a direct statement. Some reasons genuinely should not be in a broadly accessible record, and the right answer there is counsel's design, not more capture. A judgment system in a regulated firm should be designed with legal and compliance from the start, including what is written, where, and who can see it.
What helps with the rest
A completeness flag. Let people mark "stated reason, not the whole story" without requiring the whole story. It is honest, cheap, and tells later readers to weight the precedent accordingly.
Scoped visibility. Some reasoning can be visible only to the decision owner and the review gate. Narrower audiences produce franker records.
The cynic's question. Ask, as an optional field: "What would a sceptical colleague say was the real reason?" It gives people a deniable way to record the operative driver, and it is often answered with surprising candour.
Outcomes as a truth serum. Over time, outcome write-back exposes the gap. Decisions whose stated reasons repeatedly fail to predict their results are telling you that the stated reasons are not the real ones. You do not need anyone to confess; the pattern shows itself.
The builders go first
The first test of any judgment system is whether its own builders record their unflattering reasons. If the team running the programme writes only polished rationales, everyone else will too. Model the behaviour publicly, in your own records, before asking it of anyone else.
In practice
- Add a completeness flag and the optional cynic's question to the record.
- Agree visibility scopes with legal and compliance before launch, not after the first awkward record.
- Quarterly, list decisions whose stated reasons failed to predict their outcomes, and review them privately with the owners.
- Write three of your own honest, unflattering records in the first month.
What would prove this wrong. Evidence that stated reasons in a well-designed system predict outcomes as well as private, candid accounts of the same decisions. Then the gap between stated and operative reasons is smaller than practitioners believe.
Common questions
Why are decision records often misleading?
The stated reason and the operative reason often differ, because the real driver is political, personal, embarrassing or legally sensitive. If stated reasons are systematically incomplete, precedent misleads.
How can a decision log capture the real reasons behind decisions?
Add a completeness flag, scope visibility so some reasoning is seen only by the owner and review gate, ask an optional cynic's question, and use outcome write-back to expose the gap over time.
Should sensitive reasons be written into a shared decision system?
Some should not be in a broadly accessible record because records are discoverable. A judgment system in a regulated firm should be designed with legal and compliance from the start.