Enterprise AI Administration
Governance & Lineage — Revenue at risk (churn)
How this KPI is sourced, calculated, governed, predicted and acted on — one traceable chain.
Current value
USD 6.8M
No stored target
Mission area
Churn & Retention
Decision journey
Churn & Retention
Source systems
2
Billing · CRM
Lineage path
From source system to value realization — counts of governed records linked to this KPI
- 2 Sources
- 3 Mappings
- 7 Variables
- 1 Formula
- 2 Rules
- 1 Models
- 5 Scenarios
- 3 Actions
- 1 Value records
Formula
Governed calculation for this KPI
Revenue at Risk (churn) APPROVED
Monthly revenue of subscribers whose churn propensity is at or above the governed high-risk band.
Σ average_monthly_revenue WHERE churn_propensity ≥ 0.65
Version 1.2 · Grain Subscriber-month · Owner Head of Retention · Sources: Billing service revenue; Churn Propensity Model scores
Source-to-target lineage
Source system, object and field feeding this KPI, and how each is transformed
| Role | Source | Transformation | Target | Grain / refresh |
|---|---|---|---|---|
| Subscriber revenue | Billingbilling_service_revenue.amount | Sum by revenue_type and billing_period; excludes equipment and VAT | revenue_kpi_mart.service_revenue | Account-month · Daily |
| Churn outcome | CRMcrm_subscribers.termination_date | Churn event when termination_date falls in period; voluntary/involuntary from reason code | subscriber_dimension.termination_date | Subscriber-month · Daily |
| Segment | CRMcustomer.segment | Map to governed segment codes | subscriber_dimension.segment_code | Subscriber-month · Daily |
Business rules
Thresholds that escalate this KPI and the alert each raises
| Rule | Severity | Alert raised |
|---|---|---|
| Revenue at Risk Escalation Revenue at risk ≥ $2.5M / month | HIGH | High-value urban prepaid churn up 0.5pp in 30 days |
| High-Value Churn Escalation Propensity 0.65; revenue at risk ZAR 5M / month | HIGH | High-value urban prepaid churn up 0.5pp in 30 days |
Models
Predictive and simulation models used for this KPI
| AI model | Validation | Simulation model |
|---|---|---|
| Churn Propensity Model XGBoost · v3.2 | VALIDATION REQUIRED | Churn Sensitivity Model |
Governed variables
Model features derived from the source fields above
Scenarios
Stored scenarios for the Churn & Retention journey
- Mass price matchalternative
- Targeted retention bundleRECOMMENDED
- Bundle + network fixalternative
- Do nothingno action
- Base casebase case
Actions & value
Governed actions and value records for the same journey
- Build target list from propensity modelCRM Analytics · Done
- Configure offer and USSD/SMS campaignHead of Retention · In progress
- Track save rate weeklyRetention Ops · Not started
- High-value churn reductionUSD 2.1M expected · USD 0.78M recorded actual — unverified · AT RISK