Global Fintech Ends Analytical Silos with Kyvos

About The Customer

Fortune 500 fintech leader supporting millions of merchants

A US-based fintech powerhouse that operates in over 100+ countries and enables digital payments for over 3.5 million merchants and financial institutions. Processing more than 73 billion transactions annually, the organization plays a critical role in global payment services.

Challenges

Siloed reporting and infrastructure limits hindered decision-making

The company struggled to achieve holistic visibility of merchants worldwide due to factors like scalability constraints and data limits. They maxed out the capacity of their GCP environment. Despite running on a large server, there was no room for additional data.

Lack of a consolidated view: Analysts couldn’t get a single consolidated view of their data as they were using 20 different Strategy cubes for core reporting.

Delayed insights: Publishing a Strategy cube, even for a single data source, took more than 6 hours.

High downtime: Every data refresh or update triggered a 24-hour reprocessing window during which the system was inaccessible to users.

High costs: Cloud computing costs were extremely high.

Business Goals

Break down data silos and enable fast, unified financial insights

Bring all data into one consistent view for instant, scalable and accurate analytics. Strategic goals included

Reduce time-to-insight for revenue and transaction analytics, year-over-year analysis, etc.

Scale analytics to handle high volumes of transactions and thousands of concurrent users efficiently.

Control cloud computing costs while improving analytics performance and scalability.

How Kyvos helped

Consolidated data, eliminated silos and accelerated analytics at scale

Kyvos semantic layer on their existing cloud data warehouse, Google BigQuery, extended Strategy’s scope significantly. Users could plug directly into the Kyvos semantic models and instantly access insights on Strategy.

Improved performance: Allowed users to get their queries answered from Kyvos directly, delivering <10-second responses on 1.6B records.

Unified data model: Consolidated 20 Strategy cubes into a single semantic data model, covering all data sources, locations, merchants and customer information. Data modeling time also reduced drastically.

Cloud-native architecture: Delivered a scalable, elastic architecture that efficiently handled expanding data volumes and user concurrency.

Extended time-series analysis: Supported year-over-year and multi-period analytics across any amount of historical data without scaling constraints.

Cloud cost optimization: Eliminated in-memory limitations and ensured price-performant querying while keeping the existing BigQuery pipelines intact.

Seamless Strategy integration: Allowed users to plug in dashboards directly into Kyvos models with no manual settings or model recreation