Why Most Data Projects Fail and How to Avoid It
summary
Claims that 85% of data projects fail and that success requires three teams — data engineering, data science, operations — in a specific ratio of 5–10 data engineers per data scientist. Advises focusing on one to three projects, delivering in tranches, and calculating a target 10x ROI. Warns cloud vendors and tool pitches (Spark, Databricks, Snowflake, S3) are biased, data-warehouse teams often stall AI work, and organizational buy-in is required for hiring and execution.
AI-written summary · about
chapters (13)
- 0:00Intro
- 1:0185% of data projects fail
- 2:19The right questions
- 2:51The right people at the right ratios
- 5:01Data teams
- 6:05What is the business value?
- 8:46When will you generate value?
- 11:02Where will data be processed?
- 14:27How will the plan be executed?
- 19:30Why is data valuable?
- 23:30Not all gaps are technology
- 26:53Should you get help?
- 31:36Outro
related talks
31mWhy Most Data Projects Fail and How to Avoid It
53mThe State of Data Engineering
25mBuilding A Successful Data Science Team
29mData Science Best Practices : From Proof of Concepts to Production
26mThe Hidden Foundation: How Effective Data Management Determines AI System Success
32mWhy ML Projects Fail and How to Ensure Their Success
25mWhy Software Engineering Best Practices Fail in Data Engineering
48mBecoming a Data Head, a reflection on the trade offs and decision