Skip to content
Yash Singh Ramgadiya

Case study

Organisation
Burda Luxury, Asia Pacific

APAC Unified Commerce Analytics Platform

A governed analytics platform unifying Shopify, Amazon Seller and Vendor, subscription, customer, marketing and marketplace data across 8 brands and 5 markets, with 22+ transformation layers and approximately 45 semantic models.

~45governed semantic models

What changed

Commerce, subscription, marketplace, customer and marketing reporting moved onto one governed analytical foundation across eight brands and five markets.

Analytics EngineeringSemantic LayerSelf-service
Unified commerce analytics platform, conceptuallyShopify, subscription, Amazon, marketing and customer data flow into 22+ source-specific transformation layers, then into approximately 45 governed semantic models, which power executive reporting, self-service analytics, drilldowns and AI-enabled analytics.ShopifySubscriptionsAmazonMarketingCustomerdata22+ source-specific transformation layers~45 governed semantic modelsExecutivereportingSelf-serviceanalyticsDrilldownsAI-enabledanalytics
Unified commerce analytics platform, conceptually. Definitions live once, in the semantic models, so every output inherits the same numbers.

Context

Eight brands operating across five APAC markets, each with commerce, subscription, marketplace and marketing data arriving from different systems, and leadership needing one reconciled view.

Challenge

Reporting had to work brand by brand and market by market while still rolling up to a regional view. Metric definitions needed to be consistent everywhere, and the platform needed to support self-service without losing governance.

My Role

I own the platform as part of leading the regional analytics organisation: its architecture, its semantic-layer and KPI governance, and the analytics engineering team that builds and operates it.

Approach

  • Model each source in its own transformation layer so that source-specific logic stays isolated and testable.
  • Consolidate into governed semantic models that define measures and dimensions once for every brand and market.
  • Expose the semantic layer to BI and self-service tools so that reports inherit definitions rather than redefining them.
  • Automate recurring reporting on top of the governed models.

Architecture / Analytics Design

The platform is built on 22+ source-specific transformation layers feeding approximately 45 semantic models. Sources include Shopify, Amazon Seller and Vendor, subscription systems, customer data and marketing platforms. Governed self-service reporting and marketplace analytics sit on top of the semantic layer.

Business Impact

  • One governed view across 8 brands and 5 markets.
  • Governed self-service reporting that inherits definitions from the semantic layer.
  • A governed semantic foundation that AI-enabled analytics can build on.

Technology

  • dbt transformation layers
  • Cube.js semantic layer
  • Amazon Redshift
  • Business intelligence tooling

Key Lessons

  • Source-specific layers make multi-brand, multi-market platforms maintainable.
  • Governance is easiest to enforce in the semantic layer, not in dashboards.
  • Self-service only reduces analyst load when definitions are trusted.

Related work