Skip to main content


Data Architecture should define the common business vocabulary around data, develop the critical relationships between data and processes, and create multi-layer models to represent them.

Implementing data architecture must be a partnership between all functions including, of course, IT. For this to be successful a data architecture and data governance function must be developed collaboratively.

Examples of data architecture activities are shown below:

  • Developing, maintaining and advocating a published data strategy.
  • Creation of enterprise wide data models representing data at conceptual and logical models.
  • Developing and maintaining principles to guide and support operation of data throughout the organisation.
  • Setting and maintaining the design rationale around key areas of data management including master data management.
  • Developing and integrating the business rules with business and process analysts.
  • Developing and maintaining an integration architecture for all internal and external data flows.
  • Acting as the conduit between ‘data groups.’
  • Leading impact analysis on data and associated models on all programme of change and large operational activity shifts.
  • Leading or supporting the education and training around the value and power of data.
  • Supporting data owners in identifying data services and products.
  • Working with wider technology teams to ensure data architectural integrity across all IT change and operations.
  • Working with business architects on consolidated models and value chains.
  • Participating/leading data forums such as data governance board and technology architectural reviews.

These activities will look, on first review, complex, expensive to implement and requiring many new skills and capabilities for many organisations. However, the key is to pick the most important aspects for the organisation and build this data capability first. Data architecture provides the link between a corporate strategy and a way of enacting it from a data perspective.