Spark Prioritization Delta
Purpose: Evidence and ambition-preservation rules for Spark proposals. Impact-effort and RICE mechanics are model-known.
RICE Contract
RICE = Reach × Impact × Confidence / Effort- Reach uses the affected segment and one consistent time period.
- Impact uses an agreed metric and anchored scale; do not make every proposal “high.”
- Confidence is backed by analytics, experiments, research, or comparable launches. Unvalidated ideas remain explicitly low-confidence.
- Effort includes design, implementation, testing, release, documentation, and likely maintenance; report a range rather than a universal buffer.
- Compare like with like. Infrastructure, compliance, and strategic options need separate lanes when direct user reach understates their value.
Ambition Guard
Tag proposals H1 incremental, H2 adjacent, or H3 transformative. Rank within a horizon before comparing across horizons. Do not let RICE's confidence term silently remove all novel options; a slate with no H2/H3 alternative must state that conservatism explicitly.
Required Output
For each proposal provide inputs, units, evidence, calculation, sensitivity range, horizon, limitations, and the decision that the score informs. Scores support judgment; they do not replace it. Route strategic portfolio tradeoffs to Magi/Magi and tech-debt architecture to Atlas.