Price risk accurately and Cut loss ratio
Here are 10 numeric, per-policy/per-location (or per-claim) variables that are high-value in Insurance (e.g., Guidewire, Duck Creek, Sapiens, Majesco, Verisk/ISO, RMS/Moody’s) and well suited to visualize using Immersion Analytics:
| # | Variable | What it is (numeric) | Why it matters | Good IA mapping (suggestion) |
|---|---|---|---|---|
| 1 | Underwriting Risk Score (0–100) | Composite peril/behavioral score | Primary pricing/acceptance signal | X-axis (higher → right) |
| 2 | Expected/Actual Loss Ratio (%) | Incurred losses ÷ earned premium | Core profitability metric | Y-axis (higher → up) |
| 3 | Claim Frequency (per 1,000 exposures) | Normalized claims count | Identifies hotspots & trend shifts | Color (hotter = higher) |
| 4 | Avg Claim Severity ($) | Mean paid + reserved per claim | Captures tail risk exposure | Z-depth (closer = higher) |
| 5 | Premium / Exposure ($) | Earned premium or TIV by risk | Scales impact on portfolio | Size (bigger = higher $) |
| 6 | Cat/Peril Exposure Score (0–100) | Wind/flood/quake/hail model score | Guides re/insurance & limits | Glow (brighter = higher) |
| 7 | Documentation Completeness (%) | Required docs/inspections complete | Confidence in decision quality | Transparency (hollow = incomplete) |
| 8 | Prior Claims Count (#) | Historical claims on risk/entity | Predicts future frequency | Satellites (more satellites = more claims) |
| 9 | Open-Claim Aging (days) | Days since FNOL to close (open) | Drives reserve accuracy & cost | Pulsation (faster = older/urgent) |
| 10 | Fraud Probability (0–1) | Model score for fraud likelihood | Prioritizes SIU review | Shimmer (stronger = higher risk) |
What combined-ratio improvement could you unlock by seeing all ten—simultaneously—across every policy, location, and claim in your book?
