RBC · Senior Product Manager
Data Foundation
3
Data Layers
Months → weeks
New Source Onboarding
Shared hub
Product Events
Overview
Established the quality gates, lineage and onboarding standards for the data foundation underneath a six-product-line lending portfolio, so product teams could use consistent, validated information.
The Challenge
Data sources were onboarded through bespoke pipelines with inconsistent validation. Low-quality data could reach production, while product and reporting teams had no common event model or dependable service levels.
The Approach
Defined quality gates between staging, canonical and production data, added lineage and named owners for service levels, and created a reusable method for converting unstructured sources into structured data. Standardised product events in the shared hub.
Key Outcomes
- Introduced staging, canonical and production layers with a quality gate between each.
- Reduced new data-source onboarding from months to weeks with reusable patterns.
- Added lineage for compliance and service levels with named owners.
- Streamed product events into one hub so funnels could be measured consistently.
The Result
New source onboarding moved from months to weeks, poor-quality data was stopped before production, and the data foundation became the reusable template for subsequent lending sources.