The Challenge
The platform's CFO and Head of FP&A made a complete data rebuild their top organizational priority. The company was processing 200,000 to 600,000 subscription events monthly, but no one trusted the numbers. Finance couldn't confidently report ARR because results varied depending on which table you queried. A senior finance leader admitted the team had no strong understanding of basic metrics like new subscriptions and cancellations. With 99% of the company dependent on accurate membership data, the executive team could not wait for incremental fixes.
The Constraint
The legacy membership table was riddled with systematic quality failures: duplicates throughout subscription tables, records that disappeared or changed values over time, and critical edge cases that broke financial reporting. Failed payments kept memberships appearing active with no revenue. Refunds caused entire records to vanish. Same-day cancellations created conflicting customer statuses. Analysts were manually adding DISTINCT clauses to every query just to get usable results, spending more time double-checking data than analyzing it.
The Approach
Mammoth Growth rebuilt the entire subscription data infrastructure in BigQuery, designing a modern data stack with clear separation between raw sources, business logic, and analyst-facing models. We started with five standardized reference tables from raw Stripe data, then built an event-driven subscription model that captured every lifecycle state change. Every model was validated against both Stripe's dashboard and the platform's existing reports to build organizational confidence.
The Outcome
The platform went from weeks-long analysis cycles to hours. Finance, Product, and Executive teams now operate from unified subscription models with consistent definitions, eliminating the metric debates that had paralyzed decision-making. The infrastructure accurately processes 8 years of historical data and scales to handle continued growth.
What This Unlocked
Previously blocked initiatives, now unlocked. Revenue recognition modeling, product metrics development, and user table buildout are on the roadmap with a solid foundation beneath them.
Dashboards tied directly to Q4 OKRs. Active memberships, cancellations, billing patterns, and customer join dates tracked in real time for goal-setting.
Architecture that scales with the business. The event-driven infrastructure handles hundreds of thousands to millions of subscription events without becoming the next bottleneck.
Services
Tech Stack
Results
Analysis cycle time
Reduced from weeks to hours
Data models delivered
18 production-ready in 5 weeks
Historical coverage
8 years of subscription data (2017–2025)
Cross-team alignment
Finance, Product, and Exec on single source of truth