3x more queries processed, with 90% faster performance than before
4,000 users undertake more than four million Looker queries per week
<3 seconds for Looker dashboards to load versus several minutes
Improved data governance and sharing with an open, decentralized data model and automation
Increased operational efficiency, resilience, scalability, and cost savings
Revolut adopted an open, scalable architecture on Google Cloud, creating a flexible data platform for innovation and freedom of choice, improving performance, and supporting scalable systems for global expansion.
Revolut adopted an open, scalable architecture on Google Cloud, creating a flexible data platform for innovation and freedom of choice, improving performance, and supporting scalable systems for global expansion.
Scaling data management for global expansion
Over the past 10 years, Revolut has grown rapidly from its origins as an energetic UK startup to a leading global fintech operating on five continents. During this time, its innovative services, including banking, payments, and currency exchange have attracted millions of customers worldwide.
As this audience expands, so have Revolut's computing and storage resources, as well as the tools used by employees to extract insights that guide specific projects and steer wider business strategies. These include state-of-the-art business intelligence applications such as Looker, which enables employees to uncover valuable data-driven insights that drive decisions, improve operations, and manage risk.
Looker is also incredibly versatile, and soon became an essential tool for Revolut employees. In one month alone, demand grew from 300,000 weekly requests to 1.5 million.
To support this growth, the team saw an opportunity to modernize its architecture by adding a more elastic, unified platform to enhance data access, sharing, and governance at scale.
This also presented an opportunity to free up teams from time-intensive data and infrastructure management. By adopting a more flexible, high-performance system with a broader set of tools, Revolut could empower teams to focus on innovation rather than maintenance.
Optimizing data infrastructure with Google Cloud
If we have large data processing or run machine learning tasks, then we can run Spark on GKE. It removes the heavy maintenance work that would otherwise be required with hardware and infrastructure.
Denis Agiev
Lead Software Engineer, Revolut
To overcome potential scaling issues that could constrain employee access to critical intelligence, Revolut implemented an open, modular architecture based on Google Cloud and open-source software. "We wanted reusable, loosely coupled components that were easy to deploy in new regions and interacted through open standards and interfaces," explains Denis Agiev, Lead Software Engineer at Revolut. "This would ensure that we were able to support innovation and new tools as they became available. Google Cloud embraces this open approach and delivers the scalable, resilient, and secure services we require to support our new self-service data infrastructure-as-a-platform model."
Revolut deployed its new solution in less than a year. This includes an open data lakehouse with 17,000 Apache Iceberg tables stored in 400 Cloud Storage buckets. Revolut teams use a mix of regional and dual-regional buckets, depending on resilience requirements. For further data protection, Revolut uses the built-in Soft Delete feature in Cloud Storage to restore deleted data. "Anywhere Cache in Cloud Storage is also great because it provides scalable, zonal SSDs, so we can co-locate compute and storage in the same zone to minimize latency and avoid egress costs," Agiev says.
To optimize storage management, Revolut implemented an event-driven automation framework using Google Cloud Pub/Sub and Spark running on Google Kubernetes Engine. This intelligent system handles resource-intensive tasks such as data compaction and automated expiration of obsolete data — delivering optimized and cost-efficient storage.
At the compute layer, Revolut deployed the open-source Trino query engine on GKE, providing on-demand scalability and powering fast query processing across 13,000+ Looker dashboards. For large-scale batch data processing and machine learning workloads, Revolut teams orchestrate more than 10,000 Spark jobs on GKE, primarily leveraging Spot VMs to achieve significant cost efficiencies.
Improved BI performance and cost savings
Implementing open data architecture has enhanced Revolut's data analytics capabilities, supporting increasing demand for Looker. With Trino, for example, Revolut can process three times more Looker queries at speeds 90% faster than the previous system. This company-wide BI platform is designed to deliver fast SQL analytics and scale to over one million queries per day.
With Looker's powerful semantic modeling layer, Revolut teams gain consistent metrics and self-service analytics at scale. This architecture also decouples compute from storage, enabling multiple specialized Trino clusters to serve different teams efficiently, driving high throughput and substantial cost savings through a shared, low-cost storage layer.
Google Cloud offers great flexibility and scalability. We can run any kind of open source solution and it works just as well without the need to scale the underlying infrastructure.
Denis Agiev
Lead Software Engineer, Revolut
Revolut also benefits from the dynamic scaling capabilities of Google Cloud. This helps eliminate unnecessary overprovisioning, delivering cost savings, and maintaining consistent performance even during periods of peak demand. As query volumes change over time, Revolut's elastic architecture can scale accordingly to deliver rapid response times without unnecessary infrastructure investment.
Empowering teams through automation and decentralized data ownership
Revolut's data platform offers data and product teams greater autonomy while maintaining strict governance standards. Product managers, analysts, and developers can access data directly supported by well-defined, automated processes that validate the requests against Revolut's governance models and then prompt the data owner to review and approve the request.
Automation also comes into play when engineers create new compute instances, storage buckets, or analytics dashboards. Here, the system checks these requests for operational efficiency, security compliance, and cost optimization. Automated processes also monitor and audit data usage patterns, ensuring compliance while freeing up valuable engineering resources for more valuable activities.
With Google Cloud, we're able to optimize resource usage through automation, enforce best practices, and balance innovation with cost efficiency.
Denis Agiev
Lead Software Engineer, Revolut
Above all, Revolut benefits from agile automated workflows while retaining control over critical governance processes. This compliance strategy helps drive innovation, readying the company for further growth while maintaining stringent security and regulatory standards essential for a leading global financial institution.
Revolut is a global fintech, helping people get more from their money. In 2015, Revolut launched in the UK offering money transfer and exchange. Today, more than 65 million customers around the world use dozens of Revolut's innovative products to make more than a billion transactions a month.
Across personal and business accounts, Revolut gives customers more control over their finances and connects people seamlessly across the world.
Products: Cloud Storage, Dataproc, Kubernetes, Pub/Sub, Spark, Looker
Industry: Financial Services
Location: Global
