Metric Store
summary
Company standardizes analytics into a four-layer pipeline (L1 ingest → L2 validated → L3 curated → L4 metrics) and builds a metric store as wide denormalized wide table containing ~150 metrics across five metric-tiers, fed by dbt metadata and Snowflake CDC via Snowflake streams to avoid full joins. Adoption metrics show 80% of recurring questions answered by the store, operationalized with monitoring, score-based SLAs and data portals, and it reduces duplication while still requiring new on‑the‑fly definitions and governance for analyst‑authored metrics.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 7:19Standardization architecture
- 11:36Implementing metric store
- 17:39Customer access
- 22:02Impact and effects
- 25:38Future and AI vision
related talks
40mMetrics and Semantics: An evolution in data consumption
26mMetrics-driven Data Architectures
35mThe Internals of Metric Store
35mData Pipelines for Operational Metrics
28mStop querying tables: give AI governed metrics instead
35mBuilding a Real Time Metrics Database
39mBREAKING YOUR DATA TEAM OUT OF THE SERVICE TRAP
50mMasterclass: Building a data stack in 2021