Every day, this multibillion-dollar company ingests data about millions of transactions into a data warehouse environment to support partner and co-branding programs, merchant services, reward points programs, and more. Thousands of analysts require immediate access to this data throughout the day, with workloads highly mixed.
Too many legacy platforms, not enough speed
Previously, the company used Netezza, DB2, and Apache Hive to access and analyze data in a MapR-based data lake. Instead, the company envisioned the use of a single modern data warehouse platform to meet speed, scalability, and price/performance requirements for all its use cases.
- Previous platforms couldn’t adequately support thousands of users running mixed workloads on large data sets, with users becoming frustrated.
- Some data loads took multiple hours, while some queries completed in minutes (or if at all).
- Disparate systems added to management complexity and overhead.
Faster analysis on more data with Yellowbrick
After evaluating Yellowbrick, Greenplum, Vertica, MemSQL, Databricks, IBM, and Teradata on four criteria (performance, concurrency, ease of use, and ease of migration), the company selected Yellowbrick after a successful POC in which a single Yellowbrick Data Warehouse instance beat a 20-rack Teradata 6800 in 238 out of 240 queries. Yellowbrick is now the company’s analytics platform of record for Accounts Receivable/Compliance, Merchant Reporting, Credit Card Rewards Points Calculation, and Consumer Loan Analysis, with additional applications under consideration.
- Faster decision making across the business, with some queries running up to 100X faster across multiple departments at a fraction of the cost-per-query.
- More accurate results, with five years of transactions (totaling nearly 2PB) available for queries.
- 30X faster access to new data, with some data loads that used to take 10 hours now completing in less than 20 minutes and available for queries immediately –enabling much fresher data for analysis.
- Rapid migration (lasting approximately one month), with most queries and stored procedures moving to Yellowbrick with minimal changes required.
- Significant data center savings, achieved through sharply reduced rack space, power, and cooling requirements