SCALE and REPORT
SCALE
Package: Reuses tools-impact-engine-measure
SCALE is not a separate component — it is the orchestrator calling MEASURE a second time on the subset of initiatives that passed ALLOCATE, with larger sample sizes.
Process
Take selected initiatives from ALLOCATE
Run MEASURE on each (with larger sample size / longer time window)
Compare scaled effect_estimate to pilot predictions
Generate outcome report
Inputs (from ALLOCATE)
Selected initiatives
Predicted returns (from pilot MEASURE → EVALUATE)
Budget allocated
Produces
Actual returns at scale (from scaled MEASURE)
Prediction error (actual - predicted)
Outcome report
Outcome Report
The final deliverable documents predicted vs actual outcomes:
┌──────────────────────────────────────────────────────────────────────┐
│ OUTCOME REPORT │
├──────────────────────────────────────────────────────────────────────┤
│ Initiative: Product Description Enhancement │
│ ──────────────────────────────────────────────────────────────────── │
│ Pilot Results: │
│ Effect Estimate: +12.5% conversion │
│ 95% CI: [8.2%, 16.8%] │
│ Confidence Score: 0.78 (quasi-experiment) │
│ Sample Size: 500 products │
│ │
│ Scale Results: │
│ Sample Size: 5,000 products │
│ Actual Effect: +11.8% conversion │
│ │
│ Comparison: │
│ Predicted (median): +12.5% │
│ Actual: +11.8% │
│ Prediction Error: -0.7% (within CI) │
│ │
│ Budget: $50,000 allocated │
│ ROI: Estimated $425,000 incremental revenue │
└──────────────────────────────────────────────────────────────────────┘