In our previous article on Data Products, we discussed the definition, characteristics, and examples of data products as well as the necessity to switch to a product-thinking mindset to truly transform your datasets into viable data products. Amid this shift towards a Data Mesh architecture, it is important to highlight a very important part of data product management – data product ownership. Indeed, it is crucial to appoint the right people as stakeholders for your enterprise data products.
In this article, we go over the human side of data products – the role, responsibilities, and required skills of a Data Product Owner.
What are the role and skills of a Data Product Owner?
As the name suggests, Data Product Owners are the guarantors of the development and success of data products within an organization. They act as a bridge between data teams, stakeholders, and end-users, translating complex data concepts into actionable insights that drive value and innovation. To do so, Data Product Owners have unique sets of technical skills, including the ability to extract insights from data & identify patterns, understand programming languages such as Python or R, and have a strong foundation in data technologies such as data warehouses, databases, data lakes, etc.
In addition to technical skills, a Data Product Owner has great business acumen, with the ability to understand the business context, objectives, trends, and overall landscape and develop data strategies that are aligned with said context. They therefore use data for decision making by correctly collecting and analyzing data.
Lastly, Data Product Owners have great communication skills, with the ability to convey data insights to the different stakeholders in the company such as data scientists and developers but also non-technical roles such as business users and analysts. They usually also have experience in agile methodologies and problem-solving skills to deliver successful data products on time.
What are a Data Product Owner’s core responsibilities?
The multifaceted nature of a Data Product Owner as described above makes them have a variety of responsibilities. In Data Mesh in Action, by J. Majchrzak et al., they list Data Product Owners’ tasks as:
- Vision definition: They are responsible for determining the purpose of creating a data product, understanding its users, and capturing their expectations through the lens of product thinking.
- Strategic planning of product development: They are in charge of creating a comprehensive roadmap for the data product’s development journey, as well as defining key performance indicators (KPIs).
- Ensuring satisfaction requirements: Ensuring the data product meets all requirements is a critical responsibility. This includes providing a detailed metadata description and ensuring compliance with accepted standards and data governance rules.
- Backlog Management & Prioritization: The Data Product Owner makes tactical decisions regarding the management of the data product backlog. This involves prioritizing requirements, clarifying them, splitting stories, and approving implemented items.
- Stakeholder Management: They must gather information to understand expectations and clarify any inconsistencies or conflicting requirements to ensure alignment.
- Collaboration with Development Teams: Engaging with the data product development team is essential for clarifying requirements and making informed decisions on challenges affecting development and implementation.
- Participation in Data Governance: The Data Product Owner actively contributes to the data governance team, influencing the introduction of rules within the organization and providing valuable feedback on the practical implementation of data governance rules.
While the principle dictates one Data Product Owner for a specific data product, a single owner may oversee multiple products, especially if they are smaller or require less attention. The size and complexity of data products vary, leading to differences in the specific responsibilities shouldered by Data Product Owners.
What are the differences between a Data Product Owner and a Product Owner?
The relationship between a Product Owner and a Data Product Owner can vary based on specific characteristics and requirements. While in some instances, these roles overlap, in others, they distinctly diverge. In the book Data Mesh in Action, they distinguish between three different scenarios:
Case 1: The Dual Role
In this scenario, the Data Product Owner also serves as a Product Owner, and the development teams for both the data product and the overall product alignment. This configuration is most fitting when the data product extends from the source system, and the complexity is manageable, not requiring separate development efforts.
An example would be a subscription purchase module providing data on purchases seamlessly integrated into the source system.
Case 2: Dual Ownership, Separate Teams
Here, the Data Product Owner holds a dual role as a Product Owner, but the teams responsible for the data product and the overall product development are distinct. This setup is applied when analytical data derived from the application is extensive, requiring a distinct backlog and a specialized team for execution.
An example would be a subscription purchase module offering analytical data supported by a ML model, enabling predictions of purchase behavior.
Case 3: Independent Entities
In this scenario, the roles of the Data Product Owner and Product Owner are distinct, and the teams responsible for the data product and the overall product development operate independently. This configuration is chosen when the data product is a complex solution demanding independent development efforts.
An example would be building a data mart supported by an ML model for predicting purchase behavior.
In essence, the interplay between the roles of Product Owner and Data Product Owner is contingent upon the intricacies of the data product and its relationship with the overarching system. Whether they converge or diverge, the configuration chosen aligns with the specific demands posed by the complexity and integration requirements of the data product in question.
Conclusion
In conclusion, as organizations increasingly adopt Data Product Management within a Data Mesh architecture, the effectiveness of dedicated Data Product Owners becomes essential. Their capacity to connect technical intricacies with business goals, combined with a deep understanding of evolving data technologies, positions them as central figures in guiding the journey toward unleashing the full potential of enterprise Data Products.
