Shifting Left: Getting Your Data Earlier, Smarter and Cheaper

This 42-minute talk from the Data Science Festival explores the concept of "shifting left" in data processing, presented by Tom Scott from Streambased. Learn how to streamline your data architecture by bringing offline workloads closer to online systems, eliminating the inefficiencies of traditional multi-hop architectures that often create a Data MESS rather than a Data Mesh. Discover the comprehensive 5Ws and 1H framework for implementing this approach: Who needs, governs, and owns the data; What data should be shifted left; Where bottlenecks exist and truth should reside; When processing occurs; Why this approach improves efficiency and consistency; and How to implement it today. Designed for technical practitioners, this session from the Data Science Festival Game On event 2025 provides practical insights for creating a leaner, smarter, and more adaptable data stack that delivers faster insights with fewer complications and a more unified data experience.