Signals_v4 Framework Overview
Modular framework for detecting early churn warning signals across Aeris direct customer accounts. Below are the latest static reports generated for all monitored signals.
| Signal | Name | What It Detects | Severity / Baseline | Report |
|---|---|---|---|---|
| S1 | Device Activity Divergence | Device count or subscriber trend declining while the other is stable | Severe / Notable H6 vs current month |
View Report |
| S2 | Subscriber Dormancy | Active SIMs (IMSI) far below billed subscriber count | Severe / Moderate / OK Current month snapshot |
View Report |
| S3 | Data-per-Device Compression | Each device sending less data than its own historical norm | Severe / Moderate / OK MA(3) vs MA(9) |
View Report |
| S4 | Protocol Pattern Changes | Application protocols appearing or disappearing from account traffic | Noted H6 vs current month |
View Report |
| S5 | AR Payment Deterioration | Newly delinquent or escalating overdue invoices | Sales / Collections / Resolved 2-month comparison |
View Report |
Detailed Signal Definitions
Signal 1 — Device Activity Divergence
Compares H6 baseline device count and subscriber count against current month. Flags divergence: devices declining while subs stable, or vice versa.
Signal 2 — Subscriber Dormancy
Utilization = COUNT(DISTINCT imsi) / SUM(subs_count) × 100. Severe: < 30% utilization | Moderate: 30–60%.
Signal 3 — Data-per-Device Compression
Overlapping MA(3) vs MA(9): recent 3-month average vs 9-month long-term. 25% flag threshold (compensates for overlap dampening).
Signal 4 — Protocol Pattern Changes
Compares H6 protocol set (all protocols seen in 6 months) against current month. Informational only. Platform-wide filter applied.
Signal 5 — AR Payment Deterioration
Tracks overdue invoice trends (NEW delinquency, ESCALATING balances). 2-month comparison window with billing-float filter.
Additional Reports
RAT / Device Profile Analysis
Interactive Radio Access Technology (RAT) and device profile analytics for IoT devices.