Methodology Disclaimer
Last updated: February 22, 2026
This page summarizes key methodological boundaries for interpreting attribution outputs.
1. Purpose of the Model
The model estimates probable relationships between DRTV spots and digital donations under configured windows and data quality constraints. It is designed to support, not replace, expert judgment.
2. Causality Limits
Attribution metrics are not proof of absolute causation. External drivers, seasonality, media interactions, and reporting latency can influence outcomes.
3. Decision KPI vs Diagnostic KPI
Decision KPIs are governed by explicit policy thresholds and signed governance records. Diagnostic indicators are directional and should not be used alone for irreversible spend decisions.
4. Data Quality Dependencies
Results depend on schema compliance, timestamp quality, campaign code consistency, and valid payment status filtering. Invalid or stale inputs reduce reliability.
5. Out-of-Scope Uses
The platform is not a substitute for legal, tax, accounting, or regulated financial advice, and is not certified for clinical or safety-critical decisions.
6. Validation and Review
Production use requires periodic revalidation, drift monitoring, and cross-functional signoff across product, data science, and governance owners.
7. Change Management
Formula, threshold, and schema changes must be versioned, documented, and tested before release to avoid untracked KPI drift.