Shawn Crenshaw

Director of Data , Yello

Yello’s talent acquisition platform enables the world’s leading brands to deliver personalized candidate experiences to every job seeker, resulting in quality hires and faster fills. The centralized platform is easy to use, enabling recruiters to collaborate with one another to attract and engage top talent. Key recruiting metrics provide meaningful insights that lead to more informed, data-driven decisions, while staying on budget.

Supernova Award Category: 
Data to Decisions
The Problem: 

Yello faced significant challenges with their outdated legacy data warehouse, which was built on a Postgres database within AWS. This setup was complex, prone to failures, and required constant maintenance, leading to outdated insights, lack of standardization, and an inability to provide real-time data access to their clients.

The Solution: 

Yello identified the inefficiencies in their data infrastructure, including the lack of real-time insights and standardization, by analyzing the frequent maintenance issues and slow reporting times.

To modernize their data platform, Yello moved from a legacy data warehouse to an open data lakehouse architecture. They chose Starburst Galaxy and Apache Iceberg for their scalability, cost-efficiency, and support for open standards.

The Results: 

Before transitioning to a data lakehouse, Yello’s data infrastructure was inefficient and outdated. The legacy Postgres-based data warehouse required constant maintenance, suffered frequent failures, and resulted in slow reporting times and stale data. This hindered Yello’s ability to provide accurate and timely insights to both internal teams and clients. After implementing Starburst Galaxy and Apache Iceberg, Yello achieved significant improvements: reporting times were reduced by 98%, compute costs were reduced by 12X, and data access became near real-time. This modernization enabled self-service analytics, enhanced data storytelling, and positioned Yello for future growth initiatives, transforming their data services into a revenue generator.

Metrics: 

Before the project, reporting times could take up to eight hours. After implementing the new data platform, reporting times were reduced to minutes, achieving a 98% improvement. Compute costs were significantly reduced, with the new system operating at 12 times lower costs than the previous infrastructure. Additionally, the total combined compute run time decreased from 28 hours to just four hours. Data refresh rates, which were previously limited to four times a month, now occur in near real-time.

The Technology: 

Starburst Galaxy served as the end-to-end analytics platform, providing a powerful computational layer. Apache Iceberg was used for its open table format, enabling a more flexible and scalable data architecture. The data was stored and processed in AWS, ensuring a robust cloud environment. For data transformation, dbt was utilized, while Dagster managed the orchestration of data workflows. Looker was employed as the BI tool for creating interactive dashboards and visualizing data insights.

Disruptive Factor: 

Implementing the new data lakehouse architecture posed challenges, including migrating from a legacy system and ensuring minimal disruption during the transition. However, the impact has been transformative. The project has drastically improved data accessibility, reduced reporting times, and significantly cut compute costs. This transformation challenged the status quo by moving away from a cumbersome legacy warehouse to a modern, efficient platform. The ability to provide near real-time insights and enable self-service analytics has set Yello apart from competitors, enhancing both internal operations and client services.

Shining Moment: 

“We are most proud of the TalentVault initiative, which leverages our new architecture to deliver embedded dashboards and advanced analytics to clients. This capability not only improves client satisfaction but also demonstrates the transformative power of our data platform. Seeing our team overcome the challenges of transitioning to a modern data lakehouse and witnessing the tangible improvements in efficiency and client satisfaction has been incredibly rewarding.” -Shawn Crenshaw

About Yello

Yello’s talent acquisition platform allows world leading brands to deliver personalized candidate experiences, resulting in quality hires and faster fills. The centralized platform enables recruiters to collaborate with candidates to attract and engage top talent. Key hiring statistics provide meaningful insights that lead to more accurate, data-driven decisions while staying on budget.