Revolut banks on Exasol for enterprise-wide analytics solution

Exasol provides fast-growing Fintech business 100x faster data queries, enabling self-serve data analytics for all employees across business functions.

Wednesday, 14th November 2018 Posted 6 years ago in by Phil Alsop

Digital banking alternative, Revolut, has selected Exasol’s in-memory database as the central data repository for its organisation-wide data analytics requirements. One of the fastest growing FinTech businesses in the world, Revolut has adopted data-driven decision making, and has chosen Exasol to provide comprehensive and up to date data to every employee from its data science team to business heads to frontline staff.

Revolut previously used Postgres, but it was not feasible to use the transactional database when computing complex aggregations with multiple joins across hundreds of millions of rows of data. Revolut’s data science team sought a high performance Online Analytical Processing (OLAP) database and trialled a free Docker image of Exasol. It was impressed that even the limited version of the database, running on a single node, outperformed its former database. Furthermore, since the trial began in May, Exasol has never failed to perform or required maintenance.

An Exasol cluster is now fully implemented in production on the Google Cloud Platform, providing the central repository for all data required for analytics throughout the business. Since implementation, the database has improved decision-making processes with query times that the data scientists estimate are 100 times faster than its previous solution. With more than 2 million people using the Revolut app, the ability to use and analyse large datasets spanning several sources helps Revolut with fraud detection, improving customer satisfaction and financial reporting.

Abhi Thanendran, lead data scientist at Revolut, said: “Revolut is extremely data-driven, and we needed a universal database that could keep up. When we tested Exasol in the Docker trial, to say we were extremely impressed would be an understatement.”

Thanendran continued: “We’ve saved countless hours across all our departments. Reports that previously took hours to generate can now be loaded in a second, and we have a much better grasp of our company KPIs and industry trends. It’s also proved incredibly stable since the start and we’re looking forward to seeing continued results as we increase the demands we place upon the database.”

Unusually, Revolut provides every employee with an open source BI (Business Intelligence) tool, and self-service access to the central Exasol repository. This data underpins the key performance indicators (KPIs) for every team. The data science team also works from this central database as a “single point of truth”, and benefits from being able to quickly query and download extracts from the large live database at any time. Exasol has saved Revolut’s data scientists a significant number of hours work during ETL (Extract, Transform and Load) processing, which enables them to run their further analysis in proprietary tools and use advanced functionality such as machine learning algorithms for analysis.

Aaron Auld, CEO, Exasol, commented: “Revolut is a high-growth, cutting-edge business that has put data at the heart of its culture by overcoming the limitations of conventional data analytics solutions. Its data science team has very ambitious plans for the future and we are excited about working with them to make full use of Exasol’s capabilities to apply machine learning tools to its processes and empower employees with the data analysis capabilities they need.”