Fair

Data Ingestion with a Cloud Data Platform

Fair is a financial technology company that offers a new way to shop for a car, get approved for a loan, and pay for the car. Its unique smartphone app gives customers the freedom to drive the car they want for as long as they want and gives them the flexibility to turn in the vehicle at any time. Data is essential to this business model: The company ingests and analyzes billions of data points from more than 500 data sources.

Previously, Fair's legacy data warehouse could not keep pace with the company's rapidly expanding appetite for data. This led to frequent contention for scarce server resources, resulting in the creation of data marts - siloed copies of some of the data stored in a data warehouse to offload analytic activity and preserve performance. In addition, dealer inventory data was imported only once per day to avoid system overload, making real-time analytics impossible. And Fair's analytics team spent hours troubleshooting cluster failures and waiting for ETL jobs to run.

Seeking to enable real-time analytics for data science, marketing, and operational reporting, Fair subscribed to a cloud data platformas-a-service that could easily ingest, analyze, and query all its data, including semi-structured JavaScript Object Notation (JSON) data from tables that contain billions of rows. The cloud data platform enables new data pipelines that ingest dealer inventory data continuously. ETL jobs, which previously required several hours to run, now execute in five minutes or less. The platform can ingest the JSON data in its native format. And having a cloud architecture that separates storage and compute resources eliminates resource contention for data engineering workloads, enabling marketers and data scientists to extract minute-by-minute insights from the data.