Rocket Money exceeds data accuracy KPIs through consistent proactive data testing with Datafold

Rocket Money is a financial wellness company. The personal finance app utilizes AI to enable users to manage their personal finances and improve their financial health. A Series D startup, Rocket Money has raised $83.9 million with a user base that has doubled from one million to two million users since November 2020. The company analyzes $40 billion in monthly transaction volume.

The money app that works for you
Data team size
Total Employees
Implementation Partner
Data Stack
Google BigQuery
Datafold encourages good behavior. We value moving fast over doing things absolutely perfectly the first time. So we would just not do the right thing. Datafold helps us do the right thing.
Gilbert Watson
Senior Data Scientist
Data accuracy and quality KPI achievement
Faster testing and code review
Of code changes tested before merging to production
Ready to learn more?

The Problem

Despite having a code review culture, the Rocket Money Data Team could only evaluate proposed code changes to a certain degree. Particularly as the team grew, and code reviews were done by data practitioners who weren’t deeply and intimately involved in the task at hand, it became more complicated to understand the context of the problem and potential unintended consequences.

While the team had a baseline understanding if data was missing or not unique, it was hard to get visibility if a subtle change in the value of something had shifted the distribution in a given way. Following data issues, it became clear that Rocket Money needed to improve data observability, to understand how changes were impacting downstream users, reports, and analytics. When data practitioners were making changes, they needed high assurance that those changes were correct, accurate, and meaningful.

The Solution

Avid dbt users, the Rocket Money Data Team was making use of some data observability elements of that platform. After asking around in the dbt community, they quickly discovered and adopted Datafold with the hope of streamlining the code review process and improving data observability.

Using an on-prem deployment, the Data Team could immediately use the powerful features of Datafold’s Data Diff. This provided stats and summaries about the percentage of columns or values that were different as well as in-depth views of which columns will change.

The Results

  • Met and exceeded data accuracy and quality KPIs. Qualitatively measured, these KPIs ensured there was an acceptable level of data quality with minimal data incidents. Thanks to proactively testing all code changes, the Rocket Money Data Team met and exceeded these KPIs.
  • 90% faster testing and code review. What would take an analyst hours to do by hand Datafold does in a fraction of the time. Datafold’s Data Diff ensures that the team does comprehensive testing, taking only one hour instead of ten.
  • Consistent testing on every code change. Because the company prioritizes moving fast, code changes weren’t always tested across the Data Team. Now with Datafold, every change is thoroughly tested in a fraction of the time, with all data practitioners adopting a standardized way of ensuring data quality.
  • Streamlined onboarding for new hires. As the Rocket Money Data Team rapidly expands to meet the needs of the business, new hires get up to speed on existing datasets and distributions with Datafold’s Data Catalog. Furthermore, they can start contributing code faster, knowing that they won’t break anything because all code is tested with Data Diff for improved testing and code review.
Datafold helps you find the hidden changes you didn't know you made to your data, helping you if they’re unintended or understanding what's causing them.