KASPR Datahaus · use case
Build parametric outage and slowdown insurance products
Parametric insurance pays a fixed amount when a measured index crosses an agreed threshold, with no loss adjustment. A third-party connectivity index — the measurement behind this dashboard — can act as the trigger for internet-outage and slowdown cover. This example prices one district end to end.
Scroll to read ↓
The instrument
A payout defined by measurement, not adjustment
Indemnity insurance pays an assessed loss, established after the event by a claims adjuster. Parametric insurance pays a predefined amount when an independent, observable index crosses an agreed threshold, irrespective of the actual loss. Payout is fast and needs no documentation, because a third party — not the insurer or the insured — verifies the trigger.
The category already exists for digital downtime. Parametrix underwrites parametric cloud-outage cover on its own third-party monitoring; in 2026 AIG launched a parametric cloud-outage endorsement with a two-hour waiting period and no monetary retention. What these products require is an objective index of record.
A parametric trigger has four parts: an index, a threshold, a minimum duration (the waiting period), and a payout schedule. Connectivity measurement supplies all four. A district connectivity index below a set fraction of its own baseline, held there past the waiting period, defines an outage; latency and jitter thresholds define a slowdown.
The record
Six years of connectivity for Malvar, Batangas
KASPR measures a monthly connectivity index for every district against its own baseline of 100. For Malvar, Batangas the index falls below the 90 trigger line in 29 of 73 months. Each breach is a period a parametric policy would have paid.
Index against a baseline of 100. The dashed line is the trigger threshold; shaded months breach it.
A six-year monthly connectivity index with a trigger line at the threshold; months below it are shaded.The loss distribution
Frequency times severity
The district recorded 60 negative connectivity anomalies. Most are short — a day or two below baseline — but the tail is long: the worst ran 32 days, and the deepest cut the index by almost its whole value. Outage frequency multiplied by this severity distribution is the expected loss a premium prices.
Outages by downtime days below 90% of baseline. A short body and a long tail.
A histogram of outage severity: most outages last one or two days, a few last far longer.The price
From distribution to premium
Ranking the outages by severity gives an exceedance-probability curve: the chance a single outage exceeds a given number of downtime days. An underwriter reads a payout cap and a retention off this curve.
Premium is expected loss plus loadings for expenses, capital and profit. The finer and more objective the index, the smaller the basis risk — the gap between what the policy pays and what the insured actually lost.
Probability that a single outage exceeds a given number of downtime days.
A downward exceedance curve: high probability of short outages, low probability of long ones.Building the product
The two layers behind the trigger
Regional Intelligence
The connectivity index, its baseline, and the classified anomaly record — the loss history a premium is priced from — are the Regional Intelligence layer, by district and month.
Open Regional Intelligence →Alerts
The live trigger runs on the Alerts layer: connectivity and latency events detected in near real time, with start and end timestamps to measure duration against the waiting period. This is what fires a payout.
Open Alerts →District: Malvar, Batangas (PHL). The index is the KASPR monthly connectivity index against each district’s own trailing baseline; outages are the classified negative anomalies. Figures are descriptive; a live product would set the threshold, waiting period and payout in the policy wording.