Engineering, production, governance, cost, communication · insight-1300
Gate project ranking on decision owner and label readiness before scoring value
- Evidence type
- practitioner heuristic
- Difficulty
- medium
- Roles
- Data scientist, Data analyst, ML engineer
- Industries
- cross-industry
Summary: situation, action, reason, boundary
Situation
A backlog of fourteen ideas is ranked by sponsor seniority and novelty (hypothetical).
Action
For each idea check three gates first: a named owner of the decision, features available at decision time, and a label source with known delay. Only then score decisions per month, value per decision, run cost and probability of success, and rank.
Reason
Expected value is volume times value per decision times the chance it works, and that chance is set mostly by data readiness and adoption, which are cheap to check up front.
Boundary
Quarterly planning and intake for data science and analytics teams.
Symptom or risk
Effort goes to exciting items with no labels, no person who changes behaviour, or too little volume to repay the build, while dull high-volume items wait.
Exceptions
Platform or enabling work that creates the labels can pass a gate on a plan, not on present state.
Trade-offs
Gates feel bureaucratic to sponsors; keep them to three yes-or-no questions.
Pitfalls
Letting an optimistic probability of success stand in for a failed gate.