Thomas in't Veld

Beyond the Warehouse Wars: Building Marketing Data Stacks That Actually Deliver ROI

Many marketing teams claim to be "data-driven," only to end up overwhelmed by a £100k data platform bill—without gaining any real clarity on attribution. This talk aims to cut through the vendor hype and showcase what a truly effective, modern marketing data infrastructure looks like in practice. You'll see real-world performance benchmarks comparing Snowflake, Databricks, and BigQuery for marketing workloads, and learn about the critical—but often overlooked—tools beyond the data warehouse, such as dbt, Fivetran, Reverse ETL, and CDP integrations. Through case studies, we’ll explore how one large mobile app client reduced data costs by 75% while tripling query performance, and how Paired leveraged real-time multitouch attribution to unlock new revenue streams. The session also offers a decision-making framework to align platform capabilities with your actual marketing analytics needs, particularly around attribution. You'll learn why most marketing data platforms fail to deliver value within the first year—and how to avoid those common pitfalls. Attendees will leave with clear criteria for choosing the right data platform, a complete blueprint for a modern marketing data stack, a cost modeling framework to estimate total cost of ownership, and practical migration strategies for moving away from legacy systems.

From Dashboards to Decisions: Building Agentic Marketing Analytics with GenAI

While much of the buzz around AI in marketing focuses on using ChatGPT for copywriting, a quieter but more transformative shift is happening in marketing analytics—AI agents that can analyze data, not just generate content. This session explores how to build autonomous analytics systems that convert natural language questions into actionable insights, eliminating the need for manual SQL writing. You'll learn about the architecture behind agentic analysis—how large language models (LLMs) combined with structured data can automate insight generation—and see a live demo of a marketing mix modeling agent that updates itself daily. We’ll discuss why having clean, centralized data is essential for effective AI analytics, and walk through the process of moving from a proof-of-concept to production-ready systems. A case study will show how agentic workflows enabled faster, more cost-effective campaign analysis. Attendees will gain a technical blueprint for implementing agentic analytics on their current data stack, a clear understanding of where AI agents outperform traditional BI tools, practical code examples to get started, strategies for managing risk in AI-supported decision-making, and a roadmap to move from reactive reporting to truly proactive analysis.

Thomas in't Veld is co-founder and CEO at Tasman Analytics, where he leads data transformation projects for high-growth companies across Europe. With over a decade of experience building data capabilities from the ground up, Thomas has helped 70+ organisations — from Series A startups to FTSE250 companies — turn their data into competitive advantage. His work spans the complete data value chain: architecting modern data stacks, building attribution models that actually work, and helping marketing teams prove ROI with confidence. Recent client successes include supporting Paired subscriber growth through improved marketing analytics, Yubo’s internal data stack capabilities, and guiding PensionBee's data strategy through to IPO. Thomas holds deep expertise in both the technical foundations (Snowflake, BigQuery, dbt, BI platforms) and commercial applications of marketing analytics. He's particularly passionate about making advanced analytics accessible to non-technical stakeholders — because the best insights are worthless if they don't drive decisions. Based between Amsterdam and London, Thomas regularly speaks on practical data strategy, marketing attribution, and the real-world application of AI in analytics.