Scaling Data Teams: Three Pitfalls and How To Avoid Them
summary
Three concrete failures decide whether to watch: teams reinvent the stack instead of centralizing a warehouse (Snowflake plus ingestion like Fivetran/Stitch or RudderStack, transformation in dbt, activation via reverse ETL) and waste months; teams let data model chaos creep in by modeling sources instead of the business static target; teams ship dashboards nobody uses because they skip prioritised revenue reporting, governance for RevenueCat/AppsFlyer/event tracking, and active onboarding. Claim: 85% of projects fail; build quick revenue metrics, domain models, and measured activation.
AI-written summary · about
chapters (19)
- 0:00Scaling data teams overview
- 2:07Talk purpose
- 3:33Three data team challenges
- 6:41Mistake: Unnecessary stack reinvention
- 10:11Prioritize business value
- 13:42Data source assumptions
- 16:53Tracking plans and governance
- 20:27Mistake: Neglecting data modeling
- 23:57Domain modeling benefits
- 29:16Mistake: Ignoring stakeholder adoption
- 33:32Roadmapping data projects
- 36:45Actionable dashboards
- 39:32Key takeaways
- 40:15Q&A: Data-driven readiness
- 44:00Q&A: Clean code standards
- 47:08Q&A: Tracking plans alignment
- 48:33Q&A: Cross-department definitions
- 49:57Q&A: Docs vs development
- 51:20Closing remarks
related talks
31mWhy Most Data Projects Fail and How to Avoid It
31mWhen the data levee breaks: Managing cognitive load in data-intensive projects
56mData Unleashed: A Developer's Perspective on Navigating the Architecture Maze
22mBridging Business and Tech for Data Science Success
57mWhy BI Projects Fail (And What You Can Do About It)
50mMasterclass: Building a data stack in 2021
28mPractical steps to establish and scale your data & ML team
40mThe Hidden Cost of Data Debt