Decision record template (decision log)
A decision with no written trace disappears within months — and with it the reasoning, the discarded alternatives and the signals that drove it. The decision record fixes that: one page per important decision, filled in ten minutes at the moment the decision is made.
The format comes from engineering (ADRs, Architecture Decision Records), but it applies to every structural decision: pricing, positioning, hiring, features, acquisition channels.
The template
Copy the block below (the duplicable Notion / Google Docs version is available at the bottom of this page).
Header
| Field | Content |
|---|---|
| Decision | One declarative sentence: "We are switching to annual pricing by default" |
| Date | When the decision is made |
| Decider | Who makes the call (a person, not a committee) |
| Status | Proposed · Accepted · Superseded by [link] · Dropped |
| Reversibility | Reversible (cheap to undo) · Hard to reverse |
1. Context
Three or four sentences: the situation at the moment of the decision. What forces a call now? This is the section your future self will thank you for — context is exactly what evaporates first.
2. Signals and evidence
What motivates this decision? Customer feedback, numbers, constraints, assumed intuitions. Be specific: "12 prospects cited price as a blocker in May" beats "customers find it expensive." A decision linked to its evidence can be honestly reassessed later; a decision without evidence can only be re-debated.
3. Options considered
| Option | Pros | Cons |
|---|---|---|
| A (chosen) | ... | ... |
| B | ... | ... |
| C | ... | ... |
The discarded alternatives are half the value of the document: they save the team from re-proposing, six months from now, an option that was already studied.
4. The decision and its why
The chosen option, and the reasoning in two or three sentences. Which criterion tipped the scale?
5. What would change our mind
The most powerful and rarest section: under what conditions should this decision be revisited? "If monthly churn goes above X", "if fewer than Y signups by September." You turn the decision into a testable hypothesis, and the future re-debate into the simple reading of a criterion.
How to keep it alive
- One log, not scattered files: all decisions in the same place, sorted by date. The value comes from consultation, not from writing.
- Link decisions together: when a decision supersedes another, link them. The chain of decisions tells the company's strategic story.
- Reread before re-deciding: before any debate, check whether the topic was already settled. If so, the discussion restarts from section 5 ("what would change our mind"), not from zero.
Keeping this log by hand takes discipline. That is exactly the work a business memory like Verbasil automates: decisions get captured with their evidence, linked to the signals that confirm or contradict them, and resurface when the topic comes back.
FAQ
What is the difference between a decision record and meeting notes?
Meeting notes capture a discussion; a decision record captures a decision: the conclusion, the reasoning, the discarded options, the evidence and the conditions for revisiting. Nobody rereads meeting notes; a decision log gets consulted precisely when a topic comes back.
Does every decision need a record?
No, only the structural ones: decisions that commit the company beyond a few weeks or that are hard to reverse. A good heuristic: if the team might re-debate the topic within six months, the decision deserves a record.
What goes into "what would change our mind"?
Observable, dated conditions: a metric threshold, a market signal, a deadline. This section turns the decision into a testable hypothesis and gives an objective criterion for reassessing it, instead of reopening the debate through sheer fatigue.