Silja Märdla | Data Track | Compass AI & Data Summit 2025
summary
Bolt standardises business logic in the dbt semantic layer, implements roughly 5,000 metrics (~1,000 business terms), and generates 18 pre-aggregated tables (city/country/global × hourly–yearly) to avoid querying huge raw fact tables. The team maps additive, semi‑additive and non‑additive behaviors, enforces unique names, attaches timestamps to every metric, and automates Looker syncing via Uno so definitions flow into BI. Remaining pain points are metric-level backfills, convenient timestamp/partition tooling and richer integrations beyond dbt semantic layer.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:10Why Bolt semantic layer
- 2:56Defining metrics and entities
- 8:37Shift tables to metrics
- 13:13Analytics requests to metrics
- 16:47Time alignment challenges
- 24:13Materializing metrics strategies
- 25:17Entry barriers for teams
- 27:44Ownership and definitions
- 31:22Sustainable metrics platform
- 37:00Advice for first data people
- 43:04Self-service analytics impact
- 48:48Q&A
related talks
49mdbt: A Software Engineering Approach To The Analytics Workflow
28mHow Open Source Enables Innovation A Case Study with dbt and Lightdash
23mDelivering Modern Data Stacks With dbt to Operationalise Analytics
26mMetrics-driven Data Architectures
27mMetric Store
32mDataEngBytes 2023 - PER-T1-11 - Mark Kitson
01h 07mHow to Count Things with dbt - Data Build Tool
40mThe Tools Empowering 5000 Employees