Consulting Webflow Template - Galway - Designed by Azwedo.com and Wedoflow.com
Chime
Transforming Banking: Cloud-Native Solutions for Web Analytics
Transforming raw data into actionable results through AWS-driven services.
Date
January 20, 2025
Topic
Chime

Background

Chime Financial, Inc. is a San Francisco–based financial technology company that provides fee-free mobile banking services through two national banks, Stride Bank and The Bancorp Bank. 

Architecture

Chime utilizes an enterprise data warehouse powered by Snowflake to build powerful insights into various arms of their business including partner integrations, member details, marketing analytics, and so much more.  

To transform large raw data streams into usable intelligence, Chime utilizes Amazon Web Services to power these workloads. AWS Glue jobs take raw data into Spark, perform transformations, and refresh data sets across the platform. 

Implementation

Eric Jones was brought onboard to transform raw website data to power multiple dashboards in Looker in April 2024. 

The website data is created using Google Analytics 4 (GA4) which captures website traffic for digital properties on the internet. The digital property has many properties that can be adjusted in GA4 to discern discrete attributes about user engagement with the website. Customer data is shaped by attributes such as geographic location, engagement time, links clicked, and enrollment funnel steps completed. This customer engagement information is created in its raw form in Google Clould’s BigQuery. Bigquery is an enterprise data warehouse managed by the Google Cloud team at Chime. 

An AWS Glue pipeline was created to bring this raw data into Chime’s enterprise data warehouse. From there, Eric created over 15 Glue pipelines to transform the website data into a queryable source of customer activity on Chime’s global website. The largest table within the staging area for the data set scaled to over 26 billion records!

In the process of building this solution, Eric took the initiative to create additional artifacts that would be beneficial to other engineers. These artifacts include an entity relationship diagram, troubleshooting notes, and scaling attributes. 

The result of all of this work was immense. Chime’s marketing department spends over $150 million dollars per year to get customers to become new members. This project opened the door for the internal marketers to do deep dives on customer engagement metrics that were previously unavailable. The marketing team is now equipped to answer questions such as acquisition cost, engagement time, popular stories, enrollment events, and top 5 geographical locations for members.

Supporting Images

Chime's Entity Relationship Diagram
Chime's Airflow Dependency Table
Chime's Key Insight Tiles