Propensity & ScoringADA Asia2022 – 2024Production
Opportunity Size Prediction
Predicts booking dollar value for leads and early-stage opportunities, enabling per-record expected value computation.
The Problem
Many opportunities have missing or unreliable CRM amounts. Without predicted sizes, pipeline projection expected value calculations were inaccurate for early-stage records.
What Was Built
XGBoost Regressor trained on account attributes, historical booking patterns, and product prediction inputs (from P14). Heuristic segment-average fallback for low-data scenarios. Feature selection via Information Value.
Business Impact
Enabled reliable per-record expected value computation (propensity × size) in the Pipeline Projection Engine for all leads and opportunities, including those with missing amounts.
Tech Stack
PythonXGBoostscikit-learnBigQuery
Domain Tags
Deal Size PredictionRevenue EstimationPipeline Intelligence
Details
- Role
- Primary Owner
- Status
- Production
- Tier
- Tier 1
- Period
- 2022 – 2024
- Employment
- ADA Asia