This year, I had the privilege of collaborating with HealthEZ, a third-party administrator specializing in custom, self-funded benefit plans that reduce employer costs while enhancing member experiences. Our project focused on developing a data roadmap to support a new operational system, improve reporting, and establish a robust data infrastructure to fuel HealthEZ’s ambitious future plans.
Executive Support: The Foundation of Success
From the start, the executive leadership team’s full commitment was instrumental. Their belief in leveraging data strategically drove the early success of the data roadmap, ensuring alignment with business goals and fostering a culture of data-driven decision-making.
Phase One: Building a Modern Data Architecture
The first phase centered on creating a modernized data and integration architecture. In just six months, our team of four full-time employees delivered a comprehensive data hub/data fabric, mastering:
- 5 business domains
- 22 subject areas
- 242 business glossary entries
- 4 database layers (ingestion, staging, product, and access)
- 81 tables
- 2,110 columns
This architecture laid a scalable foundation for HealthEZ’s data ecosystem, enabling seamless integration and efficient data management.
The "Shortcut": Accelerating Value Delivery
A key innovation was the "shortcut," a point-to-point integration between the data hub’s integration and access layers. This approach delivered multiple benefits:- Rapid Data Ingestion: Enabled quick data profiling by ingesting source data without transformations.
- Project Agility: Kept the data hub off other projects’ critical paths by providing fast integration.
- Early User Feedback: Provisioned data to business users quickly, initiating a feedback loop for data quality and transformation needs.
- Time for Robust Development: Allowed the team to build a comprehensive data hub solution with error handling, change data capture, automation, and cataloging.
- Seamless Transition: Replaced the shortcut with the full pipeline without no impact to the data sourcing or provisioning data to the target system.
This pragmatic approach ensured early wins while maintaining long-term scalability.
Data Catalog: The Backbone of Governance
A robust data catalog was another critical investment. It integrated:
- Business glossary and data definitions
- Data profiling
- Data models
- Source-to-target transformations and mappings
Our rule: no development without prior definition, profiling, and mapping in the catalog. We meticulously mapped every data movement across the data hub’s layers, ensuring complete end-to-end lineage. This lineage, accessible via a web portal, empowers business users and IT developers to perform impact analyses and understand data flows comprehensively.
AI-Powered Efficiency
With over 240 business terms to define and limited availability from business stakeholders, we leveraged generative AI to create initial business definitions. Most were directly integrated into the catalog, with minimal tweaks for business-specific nuances. AI-generated definitions were flagged for future refinement, ensuring flexibility as business needs evolve.
Overcoming Challenges: The Matching Engine
An unexpected challenge was the need for a matching engine to align third-party data with internal systems. Two factors were critical to our success:
- Team Expertise: The internal team’s deep, albeit undocumented, knowledge of data and systems.
- Iterative Feedback via Shortcut: Frequent user feedback on mismatched records enabled rapid iteration of matching rules.
This approach achieved a 98%+ match rate quickly, addressing an unscoped issue. The validated rules were then integrated into the full data hub solution, ensuring reliability and scalability.
Adapting to Change: A Resilient Data Hub
Mid-project, the business decided to pivot away from the planned transactional system due to its limited capabilities. This shift, while challenging, highlighted the data hub’s flexibility. By modifying only the access layer, the data hub seamlessly supported the business' new direction for their needs and expanded to accommodate additional data domains, aligning with the revised scope and supporting other projects.
Looking Ahead: Powering Enterprise Analytics
With the data management foundation in place and initial data assets in production, the next phase focuses on leveraging these high-quality, reusable assets for enterprise analytics. This positions HealthEZ to unlock deeper insights and drive strategic growth.
Doing Data Right
Few organizations get a clean slate for data management. HealthEZ’s journey demonstrates that with strategic vision, executive support, and innovative approaches like the shortcut and AI-driven definitions, it’s possible to transform inherited systems and data into a powerful asset for the future.
