At the core is quality.

What is a Data Product?

Ask the next three people who mention “Data Product” to explain what it is. Most will struggle. The term is widely used but still poorly understood. Even reading through sources such as the Data Mesh book or community articles often leaves gaps — the definitions are broad, the language is repetitive, and the practical “how” remains unclear.

What is clear, however, is the core problem: anyone who has worked with analytical data knows that far too much time is wasted on loading, cleaning, and reconciling data. Chapter 3 of the Data Mesh book highlights this precisely — and argues that quality (“usability”) must be designed where data are produced, not where it is consumed.

Across published sources and best-practice frameworks (such as Collibra’s Data Mesh 101 and Microsoft’s definition of Data Products) one principle stands firm:

A data resource qualifies as a Data Product only if it fulfils essential quality and usability characteristics. 

Everything else is secondary.

01. Find Data

It is impossible to work with a data resource that cannot be found (“discovered”) in the first place. A Data Product must be discoverable.

With other words, data users must be able to:

• Explore available data resources
• Search and locate the ones they need
• Access minimum information to assess suitability, such as:
  • Source
  • Owner(s)
  • Timeliness / runtime freshness
  • Quality metrics
  • Sample datasets

How a team ensures that Data Products can be found(catalogues, portals, registries, etc.) is not as important as ensuring it exists and works reliably.

02. Understand Data

You need to know what you are dealing with – it is pointless to work with data without clear understanding of their meaning, how they are structured and how they are encoded.
A Data Product must provide clarity on:

• The classes/entities it delivers
• How these relate to one another (within and across products)
• Their elements/attributes and business meaning

Call it a model, call it a contract, call it a schema — but just “as a rose by any name is still a rose” a model by any name is still a model and short of such model understanding of Data Products won’t be feasible.

No hand-holding must be required. A Data Product must stand on its own through the resources it provides: documentation, catalogues, schemas, metadata or contracts.

03. Combine (Join) Data

Most valuable insights do not live in a single dataset. They come from combining information across domains, processes, and systems.

To support reliable integration, a Data Product must adhere to:

• Enterprise-wide standards
• Stable and consistent identifiers
• Cross-domain alignment of reference data

This is what transforms isolated datasets into enterprise-level value streams.

04. Uniquely Identify and Reference Data Products

A Data Product must have a permanent, unique address — one that supports semantic and structural evolution over time. This address accommodates semantic and syntax changes (schema evolution). The same applies to the records within a Data Product.

These identifiers must follow organisation-wide conventions to ensure interoperability and long-term stability.

05. Trust Data

Trust is the foundation of any data-driven decision. Therefore, Data Products must guarantee quality through automated validation, cleansing, and integrity checks at creation and during each update cycle.

Service Level Objectives (SLOs) should be published, covering:

• Update frequency
• Timeliness (event → availability)
• Completeness of key attributes
• Descriptive statistics (record volumes, ranges, distributions)
• Provenance and lineage
• Precision and accuracy over time
• Operational metrics (freshness, uptime, performance)

Trust is not a slogan — it is engineered.

Achieving Quality: What Does It Take?

High-quality Data Products do not happen by accident. They require a focused combination of:

• Methods (governance, standards, SLOs)
• Metadata (context, meaning, structures, lineage, ownership)
• Focus (domain stewardship, monitoring, care)

These ingredients — and how organisations can systematically apply them — will be the focus of the next Q Insights entries.

Figure : High-quality Data Products require three pillars: method, metadata quality, and disciplined focus. Together, these form the foundation for reliable, trusted data.

Closing Thought

Organisations succeed with Data Products not by adopting buzzwords, but by delivering quality data with clear, measurable, user-centred standards.
When teams can find, understand, combine, reference, and trust their data, the Data Product model stops being an abstract concept and becomes a practical accelerant of business value.

Your data has potential. We help you unlock it.

The next competitive edge won’t come from more data — but from better data.
Quaylogic enables organisations to create value-driven Data Products, combining governance, engineering, and product thinking to deliver insight-ready data across the enterprise.

Let’s talk and build the data foundation your business needs to scale. 🚀

About the Authors

Dr. Hans Lux, Data Architect – Quaylogic

Dr. Hans Lux is a pioneer in enterprise data architecture and the creator of highly profitable data solutions and strategies. With decades of experience driving large-scale initiatives across Data Mesh, metadata and model repositories, federated governance, and automated analytics, he combines deep technical expertise with strategic vision. His work has enabled organisations to productise data, automate complex systems, and embed data governance at scale.

Hans is a systems thinker applying the same precision that shapes enterprise data frameworks to his passions in design, engineering, and teaching, making him a true architect of both data and ideas.

Professor Geoff Smith, Partner – Quaylogic

For more than 30 years, Geoff Smith, has led commercially focused data privacy and empowerment functions, developing, operating, and transforming data systems and culture. He is a Visiting Professor at Loughborough University focusing on Digital Ethics and Trust Economies, Innovation, Leadership and Empowerment Technology.

Geoff is an advocate for sustainable development he is an ethical advisor on the UK Government National Digital Twin Program and a member of the working group for digital decarbonization.