Each year, Zeenea organizes exclusive client events, fostering an environment for collaborative discussions and sharing experiences and best practices. The Zeenea Exchange France took place on June 18th, 2024, and the International Edition on June 27th, 2024. These online sessions gathered clients from multiple industries.
In this article, we will review and provide insights into what was discussed during these client round tables, the topic of which was “Challenges and Best Practices for Implementing Data Mesh & Decentralized Data Management.”
International Edition Panelists
> Special Advisor at a Norwegian National Grid Operator
> Head of Data Governance at a European Electric Vehicle Cell Producer
> Head of Data Governance at a German Music Rights Association
France Edition Panelists
> Data Tools Project Manager at a French Banking Group
> Data Lead at Global Humanitarian Organization
> Data Governance Program Manager at a European Electric Vehicle Cell Producer
> Digital & Business Development at a French Retail Cooperative
Round Tables Highlights & Key Takeaways
Previous Data Governance Initiatives and Lessons Learned
Panelists shared insights on their data governance experiences, emphasizing the need for ownership and responsibility in maintaining data quality. The panelist at the cell producer highlighted the initial enthusiasm for data catalogs but noted challenges in ensuring data ownership. Another panelist discussed transitioning to a decentralized model to balance central coordination with business-specific needs. In the financial sector, a data leader stressed the importance of a coordinated data office, especially in a highly regulated environment, to balance central governance and local autonomy.
Takeaway: Effective data governance requires a balance between centralized coordination and decentralized ownership, with clear responsibilities to maintain data quality and compliance.
Maturity in Data Governance
Leaders at the round table were asked to share insights on their organizations’ data governance maturity. The panelist from the French banking group noted that data maturity varies, with a growing need for a unified data catalog to support governance and analytics driven by regulatory pressures. The data lead in the humanitarian organization highlighted significant progress in data collection and governance but mentioned challenges in prioritizing GDPR compliance in the humanitarian field. Another panelist in the retail sector indicated that their organization’s data governance is in its early stages, emphasizing the importance of precise data definitions and quality to support AI initiatives and prevent errors.
Takeaway: The level of data governance maturity varies widely among organizations, often influenced by regulatory pressures and the specific industry context. Continuous improvement in data practices is essential to support advanced analytics and compliance. Clear data definitions and quality are foundational to effective governance and AI initiatives.
Approaches to Data Management Decentralization
While most international panelists are moving towards or have already implemented data decentralization, French-speaking panelists favor a federated data governance approach. One panelist described their organization’s federated model, balancing central coordination with business-specific needs, allowing flexibility while maintaining coherence. Another supported this approach, emphasizing sharing knowledge and best practices across business units. A third panelist highlighted their exploration of a federated model to balance central governance with local autonomy, which is essential for their mutualist structure.
Takeaway: A federated approach to data management decentralization provides the flexibility needed to meet local business needs while maintaining overall coherence and compliance. Sharing knowledge and best practices across units enhances the effectiveness of data governance.
Challenges in Data Ownership and Responsibility
Panelists at both round tables discussed challenges in data ownership and responsibility. The panelist from a global humanitarian organization noted that assigning responsibilities to functions rather than individuals helps ensure continuity. Another emphasized defining data ownership within job roles to maintain data quality and accountability. The panelist in the financial sector stressed the importance of formalizing data governance roles within business units. Additional insights included the need for clear ownership boundaries and integrating data quality into KPIs, as highlighted by panelists from the electric vehicle cell producer and the music rights association. Strong executive-level support and bridging the gap between business and IT were also mentioned as crucial factors for success.
Takeaway: Clear and well-defined data ownership and responsibility are essential for effective data governance. Function-based ownership can mitigate staff turnover issues while assigning data ownership to specific roles and departments ensures continuity and accountability.
Data Mesh Literature and Theoretical Inspiration
Our clients mentioned being inspired by Zhamak Dehghani’s works, but most stressed adapting principles pragmatically to fit their specific needs. The panelist from the automotive industry emphasized using theoretical insights as a guide rather than a strict rulebook. The project manager at a banking group highlighted a highly pragmatic approach, tailoring implementation to regulatory and operational requirements. The representative from the retail sector stated they prefer incremental changes that align with their cooperative structure and business needs.
Takeaway: Theoretical frameworks provide valuable guidance but must be adapted pragmatically to fit each organization’s specific context and needs. Practical solutions and incremental changes often take precedence.
External Support for Data Initiatives
Most panelists agreed that external consultants have been crucial in shaping their data governance framework and ensuring best practices. The representative from the retail cooperative shared that they initially used an external consultant but have since transitioned to internal management. Another highlighted their reliance on a dedicated internal data team for flexibility and alignment. Others mentioned collaborating with external consultants for expertise and validation.
Takeaway: While external consultants provide valuable initial guidance and expertise, developing internal capabilities is essential for sustainable data governance.
Definition of a Data Product
While the term wasn’t widely used by most of our French-speaking panelists, and a clear definition of a data product wasn’t necessarily known, the maturity around data products was a bit more advanced for our international customers. For a data leader in a music rights association, a data product is a consumable data package described from technical, business, and usage perspectives. They emphasized the importance of data contracts to ensure quality. Similarly, the head of data governance in the automotive industry defines data products within a medallion architecture, with distinct domains responsible for curation. The third panelist discussed using a domain-driven design, managing information products through an information catalog with metadata acting as data contracts.
Takeaway: The concept of data products is central to effective data governance, ensuring data is consumable, high-quality, and aligned with business needs. Defining and managing data products through clear ownership and data contracts enhances usability and accountability.
Cultural Transformation
Our customers from both round tables expressed the need for acculturation in the organization regarding data. Changing an organization’s mindset and shifting data culture is sometimes challenging. The Norwegian national grid operator aims to foster a “sharing is caring” culture, encouraging employees to contribute to broader data quality and bridging the gap between business and IT through training. Another panelist in the automotive sector focuses on decentralization, shifting control to individual units and emphasizing local responsibility for data products. Decentralization was necessitated due to data complexity and volume at the music rights association.
Takeaway: Cultural transformation is essential for successful data decentralization. Encouraging a shared responsibility and ownership mindset supported by training and clear communication fosters a data-driven culture.
Measuring and Maintaining Data Product Value
Most of our customers agreed that measuring data product value was complex. The head of data governance from the cell producer explained that they categorize data products by legal requirements, time savings, and productivity gains, using business line collaboration to estimate value. They emphasize integrating data quality into business processes. Another panelist mentioned that they struggle with internal product valuation and need a framework to quantify the impact. The data leader in the energy industry uses iterative improvement and continuous feedback to refine data products and maintain quality.
Takeaway: Measuring the value of data products is complex but essential. Practical methods like collaborating with business lines to quantify time savings, productivity gains, and iterative improvement processes help maintain and enhance data product value.
Conclusion
The discussions made clear that while the shift towards decentralization offers greater flexibility and responsiveness to business needs, it also requires a robust data ownership and governance framework. Key themes included the importance of clear data ownership and accountability, the critical role of data catalogs in facilitating data discovery and accessibility, and the need for a tailored approach to data governance based on organizational context. The concept of data products was particularly significant, with organizations recognizing the value of treating data as a product to ensure quality, usability, and alignment with business needs.
