Track — <name>

Up: 70_Tracks-MOC · Board: Portfolio-MOC

A track is a bet in the portfolio. Short bets de-risk delivery; one deep bet goes for depth. Every track must be able to end in a business decision, including "stop here."

The bet

• Question / hypothesis: what would be true if this works?
• Horizon: short (de-risk, weeks) or deep (months)
• Why it's worth a slot: business relevance + what it teaches me

Definition of done (decide this now, before starting)

• Shippable-or-killed criterion: the concrete result that ends the track either way
• The decision it will inform: (predict vs cause? keep vs drop a feature/window/segment?)

Plan (minimum to iterate + reproduce — Don't over-engineer the research infra)

• Data + ground truth:
• Method(s) to compare:
• Eval (split by user & time): Evaluation theory

Experiments

• (link experiment notes as they're created — 80_Experiment-Template)

Traps to pre-empt

• Leakage checklist · Offline→online checklist · Predict vs cause checklist · For-profit tension checklist

Outcome

• Result:
• Business decision produced:
• What I learned about being wrong: