As businesses undergo digital transformation, effectively managing and leveraging their growing data is crucial for success. Data mesh and enterprise data products are gaining traction, focusing on decentralized data management, scalability, and governance. By treating data as a product, organizations can unlock greater agility and make more informed, data-driven decisions.
To dive into these concepts, we spoke with Dani Humphrys, Enterprise Data Architect at TDC Net! In this interview, we explore how data mesh reshapes enterprise data strategies. Dani shares her insights on how organizations can develop effective data products and drive long-term success. She is also a featured speaker at the 10th jubilee edition of the Data Innovation Summit 2025!
Read on to hear Dani’s thoughts on the challenges of implementing data mesh and how to build enterprise data products. She also speaks about the future of data governance in a decentralized landscape!
Hyperight: Dani, can you share a bit about yourself? What is your professional background and what are you currently focused on in your work?
Dani Humphrys: I have been passionate about technology and software since I was a little girl. Over the years, this passion has evolved into a fascination with architecture. I’ve had the opportunity to explore complex landscapes filled with multiple systems, applications, and integrations and how they all fit together. The challenges and complexities of this field never cease to captivate me.
Moreover, I worked in different teams, from Delivery to Support to Manage Services. I also worked in the roles of both Service/Software provider and Customer. Moving from the Operational side to Design gave me a unique perspective. This experience helps me find the most practical and fit-for-purpose solutions.
Currently, my focus is on enterprise end-to-end architecture, focusing on data and developing comprehensive data management strategies for our organization. I am a big fan of the “good enough” practices and approaches. These approaches emphasize practical solutions and adaptability in the face of real-world constraints.
Hyperight: At the Data Innovation Summit 2025, you will present on Data Mesh in Action: Creating Enterprise Data Products for Enhanced Business Agility. What can the delegates expect from your presentation?
Dani Humphrys: In this session, we will delve into the principles of data mesh. This includes domain-oriented data ownership, treating data as a product, and establishing a self-serve data infrastructure. I will share our organization’s journey with data mesh so far, highlighting successes, challenges, and the lessons we’ve learned. We will discuss strategies we used to address these challenges and how our approach evolved to introduce the concept of enterprise data products. I’ll provide an update on where we currently stand in this journey and our future aims. The goal is to equip delegates with the knowledge and actionable insights to enhance their data strategy and foster agile, data-driven decision-making within their enterprises.
Hyperight: Can you explain the concept of data mesh and how it differs from traditional data architecture approaches?
Dani Humphrys: Data mesh is all about decentralization. It shifts the responsibility for data management to domain-specific teams, treating data as a product. Unlike traditional centralized models where a central data team handles data ingestion, processing, and governance, data mesh empowers individual domains (e.g., departments or business units) to own their data end-to-end. This includes data governance.
Key differences from traditional data architectures:
- Decentralization: Ownership and management of data are distributed across teams.
- Data as a Product: Each domain treats their data as a product with clear ownership, quality standards, and service-level agreements.
- Self-Serve Infrastructure: Centralized infrastructure provides tools and platforms to allow domains to easily manage, process, and share their data.
- Federated Data Governance: Data Governance policies are standardized across the organization but applied in a flexible manner by the domains.
This approach enhances scalability, agility, and data quality by leveraging the domain expertise closest to the data. We have adopted those guiding principles. However, we have put a twist on it and that is where the Enterprise data products come into play.
Hyperight: What are some of the key challenges businesses face when creating enterprise data products, and how can they overcome them?
Dani Humphrys: One of the key challenges businesses encounter, whether dealing with data products or other initiatives, is ensuring consistent terminology and understanding across the organization. Often, teams may interpret terms differently, leading to miscommunication and inefficiencies. Our aim is to shift the focus from system specific thinking to data specific thinking, which makes speaking the same language crucial. To design effective enterprise data products, it is vital to address this issue. We are dealing with this by establishing a unified data model that standardizes definitions and terminology. Ensuring everyone speaks the same language within the organization facilitates clearer communication and better collaboration. By adopting a unified data model, businesses can streamline the process of developing and implementing data products, ultimately enhancing data consistency and reliability across the enterprise.
Hyperight: One of the highlights of your session is the concept of a self-service data environment. How do you ensure that business users can seamlessly “mix and match” data products without overwhelming them with complexity?
Dani Humphrys: Within data governance, we often say that the business users are the rock stars on the stage – they create, understand, and live with the data daily. They have intimate knowledge of business goals, drivers, and objectives. The enterprise data products approach is designed to aid these users by shifting their focus from underlying systems to the actual information they need for insightful analytics. To ensure business users can seamlessly “mix and match” data products without being overwhelmed, we prioritize the development of enterprise data products that abstract away complexities.
This means business users don’t need to worry about the data’s origin, quality, or whether they are using the best source. All of these aspects are meticulously managed in the design phase. By providing a self-service data environment, we enable business users to concentrate solely on leveraging the data to meet their specific needs and objectives, thus enhancing their ability to generate valuable insights.
Hyperight: In your experience, what are the best practices for implementing governance and data quality controls within a data mesh framework?
Dani Humphrys: I recently got familiar with the term “good enough practices,” which I am becoming very fond of. Data governance and quality are absolutely crucial for any trusted and valuable insights, no matter the underlying architecture. Our strategy is to incorporate data governance at the enterprise data product level. This ensures the products serve as solid foundations for creating consumer-aligned data product. This approach involves establishing clear ownership and accountability for each entity within the enterprise data product. It also includes defining standardized governance policies across domains and setting robust data quality metrics.
Embracing good enough practices means focusing on practical solutions that address key issues without over-engineering. Thus maintaining a balance between thorough governance and operational efficiency.
Hyperight: How do you see the future of data mesh evolving in the next few years? What technologies do you believe will have the biggest impact on enterprise data strategies?
Dani Humphrys: I would really like to think data mesh will evolve to become a core component of enterprise data strategies. While numerous tools and technologies are being developed to support this shift, it’s important to remember that technology alone will never be the complete solution to a problem. It might sound a bit old-school, but when implementing a new strategy or technology, we always need to consider people, processes, and technology. If one of these elements is missing, even the best tools and technologies won’t help the situation. Ensuring adoption and operational efficiency should be a high priority in the overall vision, as the success of data mesh depends on how well these three aspects are integrated and managed.
If you’re interested in learning more, don’t miss Dani’s presentation at the Data Innovation Summit 2025! She’ll explore the principles of data mesh and share how organizations can create enterprise data products to boost business agility. Dani will discuss how decentralizing data ownership and treating data as a product can transform data strategy. Thus enabling more agile, data-driven decisions.
Join Dani to hear about her organization’s journey with data mesh, the challenges they’ve faced, and how they’ve introduced enterprise data products to drive success. If you’re looking for practical insights on creating a self-serve data infrastructure, improving governance, and fostering better collaboration across teams, Dani’s session will offer valuable takeaways!
Add comment