How Kargo identified a $500K/year savings opportunity and delivered 5X ROI with SELECT

Eight years of Snowflake adoption created cost visibility challenges

Kargo has been using Snowflake since its inception, making them one of the platform's earliest adopters. Their architecture centers on processing massive volumes of raw JSON logs from microservices that handle their supply-side platform operations.

The platform processes auction data and outcome data (impressions, clicks, etc.) across web and CTV advertising channels. All raw data flows into Snowflake for JSON parsing, transformations, and aggregation before serving analytics through Looker.

As a mature Snowflake user, Kargo initially managed costs through warehouse separation - different microservice logs processed in dedicated warehouses to enable cost breakdowns. However, this approach proved inefficient for resource utilization.

Leadership pressure eliminated cost visibility through warehouse consolidation

When Diana joined three years ago, leadership was concerned about Snowflake costs and worked with Snowflake RSAs to optimize their setup. The recommendation was warehouse consolidation to improve resource utilization.

While consolidation improved efficiency by maximizing usage of already-running warehouses, it eliminated the ability to see cost breakdowns. Teams moved from microservice-level visibility to a single lump sum with no granular insights.

The Snowflake RSAs provided basic dashboards, but these were inadequate for the level of analysis needed to optimize their complex, high-volume processing pipelines.

Teams couldn't answer fundamental cost questions

The lack of granular cost visibility created constant frustration when stakeholders asked basic questions about processing costs. Everyone kept asking. We had no freaking idea. Said Diana while describing the situation before SELECT.

Leadership frequently asked questions like "How much does it cost to process targeting?" or "What's the cost per event?" Without proper attribution, these questions required manual analysis that was time-consuming and often inconclusive.

The team recognized they needed a solution that could provide the granular insights they'd lost through warehouse consolidation while maintaining the efficiency gains from their optimized infrastructure.

SELECT delivered immediate 5X ROI through automated savings

SELECT's Automated Savings feature provided instant value, reducing Kargo's Looker warehousing costs by 30% with zero effort required from the team, and giving Kargo an immediate 5X ROI on SELECT.

This immediate cost reduction not only paid for SELECT but demonstrated the platform's value in identifying optimization opportunities that weren't visible through traditional monitoring approaches.

SELECT granular cost visibility revealed $500K/year single event opportunity

The Kargo team implemented comprehensive query tagging across their pipelines, tagging down to individual Airflow tasks to enable precise cost attribution across their consolidated warehouse infrastructure.

This granular tagging immediately revealed their most costly processing pipeline: bid request events, which were costing $500,000 annually - far more than anticipated.

This insight enabled a strategic architecture decision to move their most expensive workload from Snowflake to Spark processing, fundamentally changing how they approached their most costly operations.

Intuitive interface eliminated friction from cost analysis

SELECT's user experience proved transformative for how the team approached cost optimization, eliminating the barriers that had made cost analysis difficult in the past.

The platform's intuitive design enabled natural exploration and analysis, allowing team members to drill down from warehouse-level costs to specific workloads without complex queries or manual data manipulation.

Usage groups enabled aggregate cost monitoring for leadership

SELECT's usage groups feature solved a critical reporting challenge by enabling cost aggregation across multiple pipelines and warehouses for high-level business reporting.

Usage groups allowed the team to group related workloads and provide accurate aggregate costs to leadership, even when those workloads spanned multiple warehouses and processing pipelines.

Before-and-after visibility supported systematic refactoring

As the team undertook systematic refactoring of their data pipelines, SELECT provided crucial before-and-after visibility that enabled them to measure optimization success.

This capability was essential for validating that optimization efforts were successful and quantifying the impact of specific changes.

Team empowerment through self-service cost insights

SELECT enabled democratization of cost insights across the data engineering team, allowing individual engineers to track the impact of their optimization efforts independently.

This self-service capability transformed how the team approached optimization work, making cost impact visible and measurable for individual contributors rather than requiring centralized analysis.

Looking ahead: Expanding beyond Snowflake

As Kargo continues moving expensive workloads to Spark and other technologies, they're eager for SELECT to expand beyond Snowflake to provide similar visibility across their entire data infrastructure.

The success they've achieved with Snowflake cost optimization has created demand for similar capabilities across their AWS, Kubernetes, and Spark environments as their architecture continues evolving to optimize for both cost and performance.