For non-trivial product decisions: kill / continue / pivot, build vs. buy, refactor vs. rebuild, pricing changes, scope expansion.
Designed to be short. A two-page decision memo beats a 15-slide deck for most decisions.
[Short, specific. The decision in one line.]
One sentence, owned. I recommend we [decision], for [primary reason].
Take a position. Hedging is not a recommendation.
What changed that makes this decision necessary now? (2–4 sentences.)
What are we deciding between? List the real options. Not "do nothing" as a default — describe what "do nothing" actually means.
- Option A: [...]
- Option B: [...]
- Option C: [...]
Is this a one-way door or two-way door?
- If reversible: what does the unwind cost?
- If irreversible: what's the rigor we've applied to match the irreversibility?
For one-way doors, the rest of this memo deserves more weight than for two-way doors.
Two or three most important tradeoffs across the options. Honest comparison.
For each tradeoff:
- What's gained
- What's lost
- Which option does best on this dimension
What would have to be true for the recommendation to be right?
- Assumption 1 — evidence / hypothesis / guess?
- Assumption 2 — evidence / hypothesis / guess?
- Assumption 3 — evidence / hypothesis / guess?
If two or more are guesses, the memo is not yet ready to commit.
Three risks max. The ones that matter most.
For each:
- The risk
- The signal that would tell us it's materializing
- The mitigation
Pre-committed.
- Success: [metric, value, timeline]
- Failure: [metric, threshold, timeline]
- Kill criterion: what would force us to stop?
If the recommendation requires support from leadership, name it specifically:
- Headcount: [n people, roles]
- Air cover: [specific narrative or commitment]
- Other: [specific resource or constraint relief]
Vague support is not actionable.
What's still unresolved. For each: owner, deadline.
Write the recommendation first. Then write the rest as the defense of the recommendation.
If, while writing the rest, the recommendation visibly breaks — change the recommendation. The memo is a forcing function for clarity, not a sales document for a decision you've already made.