This article is the third in a multi-part series on data governance by Arun Krishnamurthy, Managing Director at Grandview Analytics. The series focuses on how asset and investment managers can establish a robust data governance practice and a culture of data.
Have you ever found yourself overwhelmed by the sheer volume of data your organization manages? You are not alone.
The more data your organization manages, the harder it is to maintain quality, manage access, and maintain security. Data governance is a necessary discipline to manage your data through policies, standards, and procedures to establish ownership and accountability, reduce risks, and ensure data can be trusted and leveraged for decision-making. Effective data governance ensures quality, reliability, controlled access, and transparency for your data.
That is why we kicked off our data governance series. In our first article we covered key concepts, terminology, and the critical importance of a robust data governance practice in asset management firms. In our second article, we took a closer look at the framework that supports a strong data culture, exploring the organizational structures, standards, and policies that make it all work.
We also discussed how to assess where your organization stands today—a crucial first step in planning for the future. Once you understand your organization’s current data governance maturity level, you are ready to begin your data governance journey and must next consider and decide which data governance operating model is best suited for your organization.
There is no single data governance operating model that works for all organizations. In our last article, we introduced the three most common data governance models: centralized, decentralized, and federated. Each model offers its own strengths and challenges. In the discussion that follows, we will walk you through the pros and cons of each one, so you can figure out which is the best fit for your organization. The figure below summarizes the three most common data governance operating models.
Centralized data governance the traditional approach
In a centralized data governance model, a single team takes the reins, handling all data governance and data management activities across the organization. In highly regulated industries the centralized consistency of this model may be beneficial. This approach has its benefits:
- Consistency: A uniform set of standards, policies, and procedures ensures the entire organization will be implementing, measuring, and reporting their governance capabilities using a standard approach.
- Enhanced compliance: Centralized oversight by a single governing body simplifies regulatory compliance and strengthens data security.
- Efficient decision making: When one team is in charge, there is less coordination required across your organization; and dedicated resources can be allocated efficiently.
- Unified data strategy: A centralized team can maintain data repositories that offer a clear, accurate picture of your organization’s data.
But, of course, there are some drawbacks:
- Slower decision making: Centralized control can sometimes slow things down. So even though less coordination and consensus building may be required, a central group cannot unilaterally make decisions. Coordination is paramount, especially if the group wants to be effective across the entire organization. This approach could be a roadblock to quick decision-making.
- Lack of flexibility: A one-size-fits-all approach might not suit the unique needs of local departments.
- High costs: Maintaining a centralized team that meets the needs of the entire organization can be expensive and resource intensive.
Decentralized data governance model: distributed throughout the organization
A decentralized data governance model distributes data ownership, decision-making, and accountability across multiple teams or business units, rather than centralizing control within a single data governance team or committee. This model is particularly appealing in large organizations or sectors where diverse data needs, rapid innovation, and cross-functional collaboration are essential.
Some of the benefits of a decentralized model are:
- Enhanced responsiveness: With the data governance decisions closer to the data’s operational users, issues can be addressed more rapidly, and policies can evolve more flexibly.
- Domain expertise utilization: Subject-matter experts in different departments understand the specific data they work with, leading to more relevant and practical data governance practices.
- Scalability and flexibility: The model scales well across large, complex organizations where a single, centralized team may struggle to govern effectively across all departments.
- Improved cross-functional collaboration: Teams responsible for data governance within their departments often collaborate more naturally, leading to a better understanding of data dependencies and usage.
But there are challenges to a decentralized model:
- Risk of fragmentation: Without a strong central framework, there is a risk of inconsistency in data standards, policies, and practices. This could lead to silos and potential data quality issues.
- Complexity in compliance: Ensuring regulatory compliance is more challenging when different units have their own governance practices. This could lead to gaps or inconsistencies in meeting regulatory requirements.
- Increased operational costs: Each department may need resources, tools, and staff for governance. This leads to potential redundancy and increased costs compared to a centralized model.
- Potential for misalignment: Alignment and consistency across the organization will be weakened when decentralized data governance practices diverge from a centralized standard, impacting data integrity across the organization.
Federated data governance: A collaborative model
A federated data governance model spreads the responsibilities across data steward teams that are closer to the action. The data steward teams across your organization work independently but in concert with policies and guidelines from a central data governance team. In larger complex organizations, the ability to draw in experts across the organization may be optimal. This model comes with its own set of benefits:
- Closer to the data: Data stewards who work with the data every day are better positioned to see where it can deliver the most value while staying consistent across the larger organization by following a central team’s policies and framework.
- Aligned with business needs: This model aligns well with treating data as a product, making it easier to move toward a data mesh paradigm.
- The data mesh approach is a way for large organizations to handle and grow their data effectively. Instead of relying on a single, central data system, it encourages each team or department to manage its own data. This method makes data management more flexible and scalable by using the knowledge within each area and sharing the responsibility across the organization.
- As new business use cases are identified, the data stewards and business teams that work closely with and “own” that specific data or data domain can be utilized to bring the expertise to do the work and implement the data governance practices required.
However, a federated approach can also present challenges:
- More management effort: It takes a significant amount of work and coordination to keep everyone on the same page in a federated model. Collaboration, communication, and transparency are key.
- Risk of data silos: Without careful coordination, federated governance can lead to fragmented data management and inconsistencies. Implementing an actively managed data catalog will drive collaboration and communication across the organization and mitigate the risk of data silos.
Which model is right for you?
So, which model is right for your organization? It depends on a few key factors:
- Size and complexity: Large, complex organizations might benefit from a federated approach, while smaller firms could find centralized governance more efficient, at least in the early stages of adopting a data culture.
- Regulatory requirements: In highly regulated industries, the consistency of centralized governance might be more optimal.
- Your business use cases: Understanding your specific needs will help guide the choice.
- Organizational culture: Your organization’s culture will play a significant role in determining which model works best.
One strategy to consider is starting with a small, centralized team and evolving into a federated or hybrid model as your data governance needs grow and the business users and leaders in your organization adopt a data mindset. This approach combines the strengths of both models while minimizing their weaknesses.
In the initial stages of establishing a data governance practice, a centralized team may be more effective in setting up the policies, procedures and practice. But as the data culture matures and different use cases crystalize, being able to leverage the expertise of data stewards and business teams closer to the data may allow your organization to cover more data domains and maintain alignment to the central data governance team’s standards and framework.
For example, we once helped a financial services client who was pursuing a centralized data governance operating model, which worked well for them in the initial stages of their data governance practice. But as the data culture of the organization matured and different use cases became evident, they felt like it was slowing down innovation in their more agile departments.
Grandview Analytics was engaged to help them transition from a centralized operating model to a hybrid federated operating model. This is a great example of how what works for one company or at a certain point of a company’s data governance journey might not be the best fit for another, and to honor this journey as one that can and should evolve and adapt to your business needs. The transition is not always easy, but the payoff can be worth it—more flexibility, better alignment with business needs, and greater value from their data.
Conclusion: Finding the balance
There is no perfect answer when it comes to choosing among centralized, decentralized, or federated data governance models. The best approach is the one that aligns with your organization’s needs, goals, and culture. Sometimes, a hybrid approach offers the best of both centralized and decentralized models. Where you start your journey may not be where it is in a few years and that is okay. Focus on the data – let the data and its use cases be your guiding light. Where you start your journey is just that, a starting point and evaluating and evolving the effectiveness of your data organization will be a continual process.
What is your experience with data governance models? Have you found a particular approach that works best for your organization?
Stay tuned for the final article in this series, where we explore specific activities assigned to each component of the data governance framework. If you have any comments or questions about your data governance journey, feel free to reach out to Arun Krishnamurthy.
At Grandview Analytics, we are passionate about helping our clients navigate the complexities of data management. Whether it is implementing innovative solutions, integrating top-tier asset management platforms, or designing resilient data architectures, we are here to share insights you can use right away. We have seen first-hand how strong data governance can be a game-changer, and we want to help you optimize your investment data, processes, and technology across every corner of your business.
Read Part 4 of this series, “Data governance in action,” to see how the essential components of the data governance framework come together in practice.
