What Is a Microsoft Purview Data Product? — Data SkyLab Studio Definition

Sasha Korniak of Data SkyLab Studio defines a Microsoft Purview Data Product as a business-facing concept that brings together related Microsoft Purview Data Assets around a defined business purpose or use case. The definition below was created as a practical data governance definition informed by Microsoft Purview, Data Mesh, DAMA and data-democratisation principles.

Definition created by Sasha Korniak, Data SkyLab Studio.

Microsoft Purview Data Product definition — Sasha Korniak / Data SkyLab Studio

A Microsoft Purview Data Product is a business-facing concept that groups related, valuable, relevant and trusted Microsoft Purview Data Assets around a defined business purpose or use case, and adds the business and governance context needed to discover, understand, assess and appropriately use the data.

These Microsoft Purview Data Assets can include structured, semi-structured and unstructured information, such as databases, schemas, tables, views, columns, Excel spreadsheets, CSV files, JSON files, Parquet files, PDFs, Word documents, data lake files, Databricks tables, Microsoft Fabric assets, Power BI reports and semantic models, together with other supported data sources and asset types registered and governed within Microsoft Purview.

Why create a practitioner definition?

Microsoft provides the platform foundation. In Microsoft Purview Unified Catalog, a Data Product is a business concept that logically groups related Data Assets around a use case. That gives organisations a way to present technical data through a business-facing structure rather than expecting people to navigate source systems, databases, schemas and platform terminology first.

In practice, however, successful governance needs to answer more than which assets have been grouped together? It also needs to answer why the product exists, whether the assets genuinely contribute to that purpose, what value the information provides, whether the information can be trusted, and what somebody needs to know before using it.

The Data SkyLab Studio definition therefore builds on Microsoft's product model and brings in principles from Data Mesh, DAMA data management and governed data democratisation.

Why “business-facing concept”?

A Microsoft Purview Data Product should not simply reproduce the technical architecture of the organisation.

The underlying information may physically exist across Azure SQL, Azure Data Lake Storage, Databricks, Microsoft Fabric, SharePoint, Power BI and other platforms. Those technical objects are represented as Data Assets. The Data Product provides a logical, business-facing layer that brings the relevant assets together around something the organisation recognises and values.

This separation matters. A database, Lakehouse, Databricks catalogue or Power BI semantic model does not automatically become a Data Product simply because it exists. The Data Product should explain why a meaningful group of assets exists from a business perspective.

Why “related, valuable, relevant and trusted”?

Each word is deliberate.

• Related means the assets have a meaningful relationship within the product rather than being grouped together simply because they share a platform, schema or storage location.
• Valuable means there is an identifiable reason for the organisation to invest effort in governing, maintaining and exposing the information.
• Relevant means each asset contributes to the stated business purpose or use case of the Data Product.
• Trusted establishes a governance expectation: people should be able to judge confidence in the data through evidence such as ownership, stewardship, metadata, lineage, classifications, known limitations and Data Quality information.

Trust should not be assumed merely because an asset has been registered in Microsoft Purview or associated with a Data Product. A product can exist while governance gaps remain, but those gaps should be visible and progressively addressed as part of the product's lifecycle.

Why a “defined business purpose or use case”?

A well-designed Data Product should be able to answer a straightforward question:

Why does this Data Product exist?

This prevents organisations from turning their business catalogue into a copy of their technical estate. A Data Product should represent a business purpose, decision, process, outcome or use case, and bring together the Data Assets needed to support it.

One Data Product may therefore contain assets from several technical platforms. Equally, an individual Data Asset may legitimately support more than one Data Product where it contributes to several business purposes.

Why “discover, understand, assess and appropriately use”?

This final part of the definition describes the outcome that the governance model should enable.

• Discover — people should be able to find that the information exists.
• Understand — they should be able to understand what it means, why it exists, who is accountable for it and how it relates to other information.
• Assess — they should be able to evaluate relevance, ownership, stewardship, lineage, quality, classification, limitations and fitness for their intended purpose.
• Appropriately use — they should be able to use or request access to the data in a way that respects organisational policy, security, privacy and other governance controls.

The word appropriately is important. Data democratisation does not mean unrestricted access to every dataset. It means making governed information easier to find, understand and use while retaining the controls and accountability required by the organisation.

How Data Mesh influenced the definition

Data Mesh contributes the principle of data as a product. Product thinking focuses attention on usability, discoverability, understandability, trustworthiness and fitness for purpose rather than treating data as a technical by-product.

A Microsoft Purview Data Product is not, by itself, the complete implementation of a Data Mesh data product. Data Mesh can encompass operational ownership, interfaces, code, service expectations and other capabilities beyond catalogue governance. However, Microsoft Purview provides a strong business and governance layer through which many of those product principles can be represented and communicated.

How DAMA influenced the definition

DAMA's data-management principles reinforce the idea that data is an organisational asset whose value needs to be managed, protected and enhanced. Governance, metadata, Data Quality, ownership and stewardship are therefore not secondary documentation activities; they are part of what makes information understandable, dependable and useful.

This is why the definition deliberately includes valuable, trusted and business and governance context.

How data democratisation influenced the definition

Data democratisation is sometimes reduced to the idea of giving more people access to more data. That is incomplete.

Effective democratisation requires people to be able to discover suitable information, understand its meaning, determine whether it is appropriate for their needs and follow the correct route to access and use it. This must happen without removing security, privacy, policy or accountability.

The Data Product therefore acts as a bridge between the technical Data Assets and the people who need to make informed decisions about using them.

From cataloguing data to governing data

A catalogue can tell somebody that data exists. A governed Data Product should help answer:

• What is this information?
• Why does it exist?
• What business purpose or use case does it support?
• Why is it valuable?
• Which Data Assets contribute to it?
• Who owns and stewards it?
• Can it be trusted for the intended purpose?
• What quality, lineage, classification and access information is available?
• How can it be appropriately used?

That is the difference between simply cataloguing technical metadata and using Microsoft Purview to create meaningful, business-facing Data Governance.

Sources and influences

This practitioner definition was developed by Sasha Korniak, Data SkyLab Studio. It is informed by Microsoft Purview's Unified Catalog and Data Product model, Data Mesh principles, DAMA data-management principles and approaches to governed data democratisation.

• Microsoft Learn — Data Products in Microsoft Purview Unified Catalog
• Data Mesh Principles — Zhamak Dehghani / Martin Fowler
• DAMA — What is Data Management?

This is a practitioner definition created by Sasha Korniak / Data SkyLab Studio. It is not presented as Microsoft's official wording.

Related definition

For the distinction between a general Data Asset, an Information Asset, a Dataset and a Microsoft Purview Data Asset, see Data Asset, Information Asset, Dataset and Microsoft Purview Data Asset.