Going all-in on data products
summary
Auto Trader builds a rich event stream alongside the operational product, emitting an event whenever state changes, pushing it down an event bus and ingesting into the data platform so product teams own a data product with human‑readable naming. That yields near‑instant answers to analytic questions, a production “black box” for debugging, and ready hooks for billing and monetization, at the cost of costly naming debates, a large surface area that breeds fear of change, and initial data‑engineering bottlenecks without distributed skills.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:26Auto Trader's data landscape
- 3:53Digital retailing project
- 5:00Data lake to data products
- 6:24Guaranteed part exchange case study
- 8:52Analytics as first-class citizens
- 10:36Data modeling challenges
- 13:46Naming pain in data products
- 15:12Large data product surfaces
- 17:22Balancing change and stability
- 18:24Skills gaps in data engineering
- 20:09Benefits uncovered
- 22:38Evolving with data insights
- 24:03When to go all-in
related talks
22mData as a Product: Transforming Analytics for Business Success
19mDevDataOps - Data-Driven Product Development for Software Engineers
07mTransforming Big Data into Real-Time Product Insights
31mData Unleashed: A Developer's Perspective on Navigating the Architecture Maze
39mData Mesh: Data Analytics for Engineers
51mPrototypes to products: Build & deploy at-scale DS driven products
36mReal time insights for better products, CX & resilient platform
14mBuilding a Data-Driven Culture: An Anthology of Strategies and Insights