E-waste Recycler Leverages Kyvos Semantic Layer on AWS

About The Customer

A global innovator in e-waste recycling

For over a decade, this e-waste recycling leader has transformed how consumers trade in electronics responsibly. With 5,000+ kiosks across major U.S. and European retailers, the company enables instant cash-for-device exchanges. To date, it has collected over 40 million devices, preventing millions of pounds of electronic waste from reaching landfills.

Challenges

Rising data volumes and complexity slowed insights and increased costs

The company struggled to scale analytics on Snowflake, AWS and Tableau while handling years of complex kiosk transaction data.

Sluggish analytics: Query performance degraded significantly, when analyzing transaction data in Snowflake.

Lack of semantic consistency: Without a unified semantic layer, it was difficult to define consistent KPIs and data models across dashboards and reports.

High and unpredictable query costs: Ad hoc queries triggered heavy compute loads on Snowflake, resulting in unexpectedly high costs.

Complex data modeling: Multi-level hierarchies and one-to-many relationships between fact tables made it difficult to model data efficiently.

SSAS scalability constraints: SSAS couldn’t scale with growing data volumes, impacting query performance and manageability.

Business Goals

Get faster high-grain analytics performance at scale

The company aimed to leverage its kiosk analytics to deliver timely, actionable insights for business and operational decisions. They wanted to:

Scale analytics across four years of kiosk transaction data efficiently.

Optimize kiosk placement using historical transaction data to identify high-impact locations.

Reduce customer wait times by analyzing operational patterns to streamline the kiosk experience.

Tailor pricing strategies by leveraging user acceptance trends for more accurate, personalized offers.

Drive operational efficiency by transforming analytics into actionable insights that support growth and sustainability.

How Kyvos helped

Accelerating e-waste analytics with speed, precision and unified insights

Kyvos semantic layer on AWS with Snowflake as the data source enabled instant query responses across four years of kiosk transactions and provided a unified source of business logic for all BI tools.

Optimized semantic models for advanced analytics: Pulled data from Snowflake to build semantic models with complex hierarchies, calculations and schemas, supporting YoY analysis, multi-level hierarchies, 1-to-many joins and custom calculations.

Instant query responses: As all the aggregates were calculated in advance, ad hoc queries on all their data returned in less than 5 seconds.

Cost-efficient analytics & monitoring: Price-performant querying, incremental data refresh, partitioning and real-time model monitoring helped control Snowflake costs.

Improved forecasting logic: Enabled in-depth analysis of payout collections, transaction volumes, traffic quality and machine idle times to optimize kiosk placements.

Granular analytics: Provided deep transaction-level insights that helped improve price acceptance, minimize drop-outs and optimize kiosk performance.