Data in Action with Scout24: How Scout24 Reduced Costs and Enabled Conversational Analytics with MCP

Scout24’s data platform has evolved significantly over the past five years — from Presto to Trino, and ultimately to Starburst — in pursuit of better performance, cost control, and operational visibility.

After migrating their data lake to Delta format to support GDPR-driven anonymization requirements, Scout24 began encountering performance and cost challenges running Trino at scale in a self-service environment. Limited query observability and inefficient SQL workloads drove infrastructure costs to nearly $15,000 per month.

In this session, Kiran Gourish shares Scout24’s journey to Starburst, where autoscaling and enhanced query logging capabilities helped reduce monthly costs to approximately $8,000 while improving workload transparency and governance.

Kiran will also demonstrate how Scout24 built an internal MCP server integrated with ChatGPT, enabling users to query Starburst through conversational AI — eliminating the need for direct UI access and expanding analytics accessibility across the organization.

Attendees will learn:

• Why Scout24 migrated from Presto to Trino to Starburst
• How Delta format supports GDPR anonymization in their data lake
• How Starburst’s autoscaling and query history improved cost efficiency
• How Scout24 built and deployed an MCP server for conversational analytics
• What this architecture enables for the business moving forward

The session will conclude with a live MCP demo and audience Q&A.