Alerts & KPI Monitoring
Automated threshold checks on key performance indicators
These alerts are generated by running threshold checks against the current dataset. In a production BI system, these would trigger on new data ingestion. Here they run client-side against the pre-computed aggregates to demonstrate event-driven monitoring.
Critical
3
Warnings
4
Informational
4
US geographic concentration exceeds 40%
critical65.6% of companies are US-based. The dataset has severe geographic bias: insights for non-US markets are less reliable.
Major Seed → Series A attrition
criticalOnly 8,172 of 11,102 seed-funded companies reached Series A (73.6%). This is the primary bottleneck in the funding pipeline.
US captures 75% of total funding
criticalThe US accounts for $442.3B of $591.5B total tracked funding. Non-US funding analysis has limited depth.
Extreme funding distribution skew
warningAverage funding ($17.1M) is 7× the median ($2.4M). A small number of mega-rounds dominate aggregate totals: medians are more reliable than averages.
Market concentration: Trading
warningTrading represents 0.0% of all companies (20). Heavy sector concentration may bias aggregate success rates.
2,146 companies have closed
warningThese companies received funding but ultimately failed. At an average dataset funding of $17.1M, this represents significant capital destruction.
High attrition from Series A to Series C
warningOnly 33% of Series A companies reach Series C (2,705 of 8,172). The middle stages are a significant filter.
86% of companies are still "operating"
infoMany may have closed without updating their Crunchbase status. Survivorship bias is likely; the real closure rate is probably higher than 5.3%.
3 countries have < 50 companies
infoSuccess rates and funding averages for these countries (MYS, TWN, BMU...) are statistically unreliable due to small sample sizes.
Dataset covers companies founded through ~2014
infoPost-2014 dynamics (crypto boom, AI explosion, COVID, ZIRP-era mega-rounds) are not captured. Trends and predictions may not reflect current market conditions.
Success rate is 61.0% ; but "acquired" ≠ success
infoThe model treats acquisition as success and closure as failure. But acqui-hires (fire-sale acquisitions) are counted as "success," and many "operating" companies are effectively dead. The real picture is more nuanced.
How Alerts Work
In a production Business Intelligence system, alerts are event-driven: they trigger automatically when new data enters the warehouse and a KPI crosses a predefined threshold.
This demo generates alerts by checking the current dataset against rules: geographic concentration limits, funnel conversion benchmarks, distribution skew ratios and data quality indicators. Each alert includes the metric name, the expected threshold and the actual observed value, following the standard BI alerting pattern of metric → threshold → action.