Gaming Analytics Platform
Medallion architecture processing 1B+ records across 12 sources with 99.4% financial accuracy
1B+ records
34 pipelines
The Problem
Fragmented data across 12 sources, manual reporting, no unified view of player monetization or campaign ROI
Fragmented data across 12 sources, manual reporting, no unified view of player monetization or campaign ROI. A mobile gaming company with multiple active titles had revenue and user acquisition data scattered across 12 different sources - major ad networks, app stores, payment processors, attribution platforms, webstore APIs, and internal telemetry. Reporting was manual, time-consuming, and frequently inaccurate. Finance could not reconcile revenue numbers. Attribution data expired after the provider's retention window, erasing long-tail monetization visibility. Executives lacked a unified view of business performance across titles, campaigns, and cohorts.
We designed and built a medallion architecture (Bronze / Silver / Gold) on Google Cloud Platform. Raw data from all 12 sources lands in the Bronze layer unmodified. The Silver layer handles cleaning, deduplication, currency normalization, and schema enforcement. The Gold layer produces business-ready aggregations - revenue trends, campaign performance, and cohort ROAS from day-zero through day-365 - optimized for dashboard performance.
34 automated pipelines run twice daily on Cloud Run, orchestrated with scheduling, distributed locking, and dependency management. 19 executive dashboards with 280 data visualizations in Metabase give the leadership team real-time visibility into revenue, user acquisition, retention, cohort analysis, and profit and loss across all titles. Financial reconciliation accuracy reached 99.4% - replacing manual spreadsheet work that previously took days.
We also built a permanent player-level attribution system that decouples the company from its attribution provider's retention window. By ingesting raw install and in-app event data independently and cross-referencing player IDs across mobile and webstore transactions, the platform preserves lifetime value visibility that the third-party provider would otherwise expire after 24 months.
Records processed
Automated pipelines
Executive dashboards
Why it mattered
Operating need
Executives, finance, and live ops needed one trusted view of revenue, acquisition, retention, cohort ROAS, and P&L.
What changed
The platform turned 12 fragmented data sources into governed Bronze/Silver/Gold pipelines, dashboards, and permanent player-level attribution.
An analytics layer built for executives, finance, and live ops at once
The system mattered because finance, product, and live ops could use the same numbers. The permanent attribution layer also preserved lifetime value visibility after third-party retention expired.
Decision surface
Revenue, acquisition, retention, cohort ROAS, and P&L were brought into one surface with 280 visualizations instead of parallel spreadsheet reports.
Data path
Bronze, Silver, and Gold layers made the progression from raw events to trusted metrics visible, with player-level attribution preserved independently.
Operating use
Dashboards supported pacing and capital allocation from daily revenue through D365 cohort performance.
Context
A mobile gaming company with multiple active titles had revenue and user data scattered across 12 sources - ad networks, app stores, payment processors, webstore APIs, and attribution platforms. Attribution data expired after the provider's retention window.
Constraint
Manual reporting was inaccurate and slow. Finance could not reconcile revenue. No unified view of business performance existed. Long-tail monetization was invisible after attribution expiry.
Intervention
Built a medallion architecture (Bronze/Silver/Gold) on GCP with 34 automated pipelines, 19 executive dashboards, permanent player-level attribution, and D0–D365 cohort analysis.
Outcome
1B+ records processed, 99.4% financial reconciliation accuracy, 19 dashboards with 280 visualizations, and permanent attribution across mobile and webstore.
Architecture
Medallion architecture from raw ingestion to executive intelligence
Bronze Layer
Silver Layer
Gold Layer
Analytics Layer
Tech Stack
Data Warehouse
Google BigQuery
Compute
GCP Cloud Run (34 scheduled jobs)
Visualization
Metabase (19 dashboards, 280 cards)
Language
Python
Architecture
Medallion (Bronze / Silver / Gold)
Data Sources
12 integrated sources
Results
Planning a Similar Mandate?
A direct discovery call about the problem, the constraints, and the fastest credible path forward.
We respond within 4 hours during business hours