Verisk cuts processing time and storage costs with Amazon Redshift and lakehouse

This article describes how Verisk, a catastrophe modeling SaaS provider, implemented a lakehouse architecture using Amazon Redshift and Apache Iceberg to dramatically improve analytics performance and reduce costs.

• Processing time reduced from hours to minutes for billion-record aggregations
• Storage costs decreased through columnar Parquet compression
• Lakehouse architecture separates compute (Redshift) from storage (S3)
• Apache Iceberg enables schema evolution and historical data access without downtime
• Multi-tenant security uses schema-level isolation and metadata restrictions
• Three-stage pipeline: pre-processing, modeling, and post-processing aggregation
• Redshift Serverless handles concurrent background analyses across multiple clusters
• ACID-compliant operations maintain consistency during concurrent updates

Verisk's implementation demonstrates how separating compute from storage addresses enterprise analytics at billion-record scale while maintaining security isolation for multiple insurance clients.