Automating Pitch's Reporting

Analytics Engineering

Context

Pitch's reporting ran on a weekly cycle. Someone pulled numbers from seven different tools, combined them by hand, and sent a summary. By the time it landed, the data was already a week old. Anomalies were spotted late, decisions came slow.

Scope

Automate the whole process. One dashboard, all sources, updated daily.

Approach

We connected seven data sources into Snowflake through Fivetran and Airbyte, then built a transformation layer in dbt — source, staging, and reporting models — feeding a Metabase dashboard. The dashboard covers the key metrics across marketing, sales, product, and revenue, with filters for date, acquisition channel, company size, and revenue tier.

Automated reporting data pipelines

Outcome

Reporting went daily. KPI anomalies now surface in hours instead of a week. The manual work — 8 hours per week, 416 per year — is gone. After delivery, Pitch migrated fully to Metabase and later used the data stack as a foundation for broader self-service analytics.