What are data products?

From a broad perspective...

Data products are anything that creates added value from raw data.  

‍

From a technical perspective...

Data products are specialized tools or applications designed to generate, process, or provide data as a service. They can range from simple dashboards and visualizations to advanced machine-learning models or analytics platforms. These tools transform raw data into actionable insights, valuable information, or services that drive decision-making.

‍

From the perspective of business end-users...

For end users, the technical details don’t matter—what matters is that their questions can be answered with data and no added discussions with people from different teams. A data product is a reusable data asset, designed to provide a reliable dataset for a specific purpose. It pulls in data from relevant sources, processes it, and makes it immediately available to authorized users.

‍

By decoupling the dataset from its underlying systems (data warehouse), data products simplify access for consumers, making data easy to discover and use while shielding them from the complexity of the data infrastructure.

‍

Data products principles

Zhamak Dehghani, the creator of the data mesh concept, defines it as "a sociotechnical approach to share, access, and manage analytical data in large-scale environments." At the heart of this approach are key principles for designing effective data products:

1. Discoverability: A data product must be easy to find and understand, with details like the owner, origin, location, and quality metrics clearly documented. This builds trust and enhances security.
2. Valuable Data: Only store data that adds value or serves a purpose—either for current use or future profitability. Unnecessary data increases storage and maintenance costs.
3. Trustworthy Data: Reliable decisions need trustworthy data. Regular checks on quality and accuracy ensure the data remains reliable over time.
4. Understandability: Data should be well-structured and intuitive, so users can easily access what they need without extra help.
5. Accessibility: Data formats must suit different needs, such as machine learning models or business dashboards. Ensuring appropriate access levels is key.
6. Governed Access: Not all users should have access to all data. Governance ensures the right people have access while maintaining security.
7. Addressability: Each data product must have a unique identifier, making it easy to locate and use.

‍

Components of a Data Product

Every data product is built on these key components:

• Metadata and Documentation: Descriptions, business context, and operational details about the data.

‍

• Data: Raw and processed datasets used as the foundation.

‍

• User Access: Dashboards, APIs, and search tools for usability.  
• Lineage and transformations: Tracking the data's origin and transformation journey.

• Categorization and Security: Access controls and privacy measures.

• Maintenance: Monitoring, error logs, and regular updates.
• Contracts: SLAs (such as time frames when the data is ready to be used for other data products to stay up to date), policies, and usage agreements.

‍

These principles ensure data products are scalable, secure, and user-friendly, forming the backbone of a decentralized data architecture—a concept central to the data mesh philosophy.

‍

Why Data Products Matter

Data products offer significant advantages to both data consumers and organizations by simplifying data use. For data consumers, pre-built products save time by providing verified, trustworthy data, enabling quicker insights and fostering real-time situational awareness for better decisions. Moreover, upfront guarantees of data quality and compliance ensure governance is seamlessly integrated into their use.  

‍

For organizations, data products drive efficiency and profitability by fostering reuse and reducing overhead, ensuring data architectures remain adaptable and future-proof. They also bridge the gap between business and IT, creating a shared understanding and reducing uncertainty about data integrity. As McKinsey reports, implementing data products can accelerate new business use cases by 90%, cut total costs by 30%, and minimize governance-related risks and expenses (McKinsey).

‍

Data products act as a unifying framework, connecting physical systems, data models, and business processes. They eliminate fragmented approaches to data management, decentralize operations, and enable data to be applied flexibly across diverse scenarios with minimal preprocessing.

‍

Achieving these benefits requires adopting an agile strategy—start small, release quickly, iterate, and gradually expand capabilities.  

How data products fit into the data mesh framework?

Data mesh Principles  

Let’s go one step back to the umbrella term “data mesh”. Each principle is interconnected, addressing specific challenges and dependencies to create a cohesive and scalable approach to managing data.

1. Domain-oriented ownership: This principle establishes responsibility for data within specific business domains. By giving teams ownership of their data, it prevents siloing and encourages greater engagement while reducing isolation between domains.
2. Data as a product: Treating data as a product ensures that it is discoverable, reliable, and valuable to end users. This principle empowers domain teams to produce high-quality, reusable data assets while reducing the cost of ownership through streamlined processes and infrastructure.
3. Self-serve data platform: This provides the necessary tools and infrastructure for domains to manage their data independently. It lowers the technical barriers for domain teams and allows them to focus on generating value, rather than struggling with operational complexities. More about it in this article: Decentralized Data Architecture and Its Role in Data Mesh
4. Federated computational governance: This ensures consistency and compliance across the mesh. Enforcing policies at a global level, aligns decentralized data practices with organizational standards, providing reliable and scalable governance.

‍

‍

These principles are connected through several dependencies:

• Preventing data siloing depends on domain-oriented ownership and empowering teams to manage their data as products.
• Interconnecting data products adds higher-order value, enabling domains to collaborate and share insights effectively.
• Consistent and reliable policy enforcement at the mesh level ensures trust and compliance without compromising the flexibility of individual domains.

This model demonstrates how implementing these principles in harmony addresses the challenges of managing data in complex, large-scale environments, ensuring both autonomy and alignment across the organization.

‍

Tools for Supporting Data Mesh

Dawiso and data products  

Dawiso is: Data Governance and Catalog Platform

Features:

• Metadata management, business glossary, and lineage tracking.
• Tools for collaboration, transparency, and democratization of data.
• Focus on usability, affordability, and accessibility for all domains.

Use in Data Mesh: Dawiso can serve as the connective tissue of a self-serve data platform, helping domains document, discover, and govern their data products in alignment with data mesh principles.

‍

Collaboration with other platforms

Our platform allows users to define data products, document them, and make them available in a marketplace-like format. For instance, when integrated with tools like Keboola or Confluent Kafka, Dawiso can automatically generate physical data flows based on user-defined products.

‍

This flexibility supports both centralized and decentralized models, enabling data teams to collaborate while empowering individual business units like marketing or sales to take ownership of their data. Governance remains a critical layer here—without it, decentralized data management can quickly become chaotic.

‍