Thomas in't Veld | Data Track | Compass AI & Data Summit 2025
summary
Practical blueprint for marketing stacks that actually deliver ROI: performance benchmarks across Snowflake, Databricks and BigQuery, concrete tooling (dbt, Fivetran, RudderStack, Snowplow, AVO, Omni, Reverse ETL, CDP) and a migration path off legacy systems. Case studies show one client reduced data costs by 75% while tripling query performance, heavy emphasis on CI/CD and testing (70–75% of effort), and building a domain model plus a single semantic layer for consistent metrics.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:09Data projects failure reasons
- 8:29Lean data stack for scaleups
- 13:08Power of data modeling
- 21:06Query-driven development pitfalls
- 27:03Static targets in data models
- 32:24Semantic layers as truth
- 39:14Q&A on semantic layers
- 43:06Agile in AI/data environments
- 44:50Tools: in-house vs open-source vs SaaS
- 46:36Cybersecurity in data stacks
- 49:29Dashboards shouldn’t be endpoint
- 50:32Data retention and GDPR
- 51:37Domain models vs semantic layers
- 55:54Data model recommendations
- 57:22Choosing database and storage
- 58:24Mapping data relationships
- 59:45Closing remarks
related talks
50mMasterclass: Building a data stack in 2021
25mPerfecting Our Modern Data Stack Recipe: What We Learned On The Journey At Huel
22mSanjeev Mohan, SanjMo & Tony Baer, dbInsight | Databricks Data + AI Summit 2024
01h 17mData Engineering in the AI Era: Building a Robust Foundation for AI Applications
23mDelivering Retail Analytics & Data Projects With Data Created For Purpose
41mModern Data Stack: Optimising and Scaling Data in a Tech Company
23mData Team as an Optimisation Problem
01h 28mA Guide to the Modern Data Stack