1433x Faster Retail Analytics at Scale

About the Customer

A retail services powerhouse with presence in 40+ countries

One of North America’s largest retail services providers supports 4,000+ brands and 85,000 retail locations. The company manages massive retail data volumes across regions to drive sales, marketing and insights for leading consumer brands and retailers.

Challenges

Analytic bottlenecks impeded decision making

Despite investing in a modern data platform, the company struggled to turn its massive retail data into timely insights. They faced challenges like:

Slow Dashboard Performance: Power BI dashboards querying Databricks took ~45 seconds to return for a single region, and even longer for global views. They were dealing with multi-minute queries at the monthly level.

Slow Complex Calculations: Time-intelligence functions (e.g., YoY and YTD) stalled on modest datasets (10 million rows, 5 dimensions).

Sub-Query Overload: Power BI generated 300+ sub-queries, overwhelming the platform and degrading performance.

Manual Workarounds: Analysts had to make multiple hops, exporting SSRS reports before they could work in Excel, introducing delays and inconsistencies.

Data Summarization Trade-offs: They explored “massive summarization” of the data, reducing the dataset size by rolling up large portions of information. This still resulted in losing valuable detail-level insights while still not getting the desired performance. 

Buisness Goals

Making data work for everyone, everywhere

They wanted a solution that could deliver speed and scale while leveraging the technology in which they had already invested.

Ensure dashboards and queries return results fast enough for timely decisions.

Support growing data volume across multiple business units.

Improve performance on their current platform.

Analyze data without exporting SSRS reports to Excel.

Enable ease of use for design and management of data models

Make time-intelligence queries like YoY and YTD fast and accurate

How Kyvos Helped

Delivered fast, scalable analytics across the enterprise

Kyvos’ modern architecture removed bottlenecks and enabled the organization to analyze massive data volumes with consistently fast performance. This transformed how insights were delivered across the enterprise. Kyvos + PBI was 1433x faster than Databricks + PBI.  We delivered sub-second queries at the daily level with 30x more data.

Enterprise-scale performance on billions of rows: Kyvos was tested on a dataset of 6 billion rows, delivering speeds previously thought possible only with in-memory systems. Queries that had run for 45+ seconds returned in under a second for summary views, while detailed queries exceeded performance of expensive in-memory alternatives—all without data summarization.

Reduced BI infrastructure costs, no over-commissioning: The organization had to import massive datasets into Power BI, which significantly increased its premium-capacity footprint on MS Fabric, as they were using 2xF128 and 1xF256 instances. By eliminating the need for running multiple high-tier compute instances to maintain dashboard performance, Kyvos helped them achieve 20%+ reduction in overall analytics costs, even after factoring in Kyvos infrastructure and licensing.

Simplified complex period-over-period calculations : The platform supported advanced time intelligence functions such as YoY and YTD. What had previously forced workarounds in Power BI could now be defined in just a few clicks, empowering analysts to run complex calculations instantly.

No-code semantic modeling for rapid deployment: Designers with strong SSAS experience built functional data models with Kyvos in just one day. The no-code interface allowed the team to quickly create and maintain semantic models, accelerating rollout as additional business units were added.

BI tool connectivity with reliable performance: Excel users connect directly via Analysis Services, eliminating the need to export SSRS reports. The semantic model provided a consistent layer of business logic across Power BI and Excel, reducing discrepancies and ensuring reliable reporting.

By delivering high performance on disk, Kyvos enabled the organization to lower infrastructure requirements while still providing analysts with sub-second access to large-scale data.

A foundation for self-serve analytics: The combination of speed, scale and usability enabled business units across the organization to explore data independently.