Direkt zum Inhalt springen

Why is data modeling a fundamental activity in supporting your data-driven business and how to approach it?

Data modeling is a shared responsibility of business and IT. A common misconception is that it is primarily a necessary technical step in developing applications and a task reserved for IT professionals. The truth is, only a tight cooperation with the respective experts from the business yields data-driven applications supporting the business goals. Data models serve as a communication tool in achieving that aim.  

Domain-driven data modeling is a modern approach, with overarching elements such as the Enterprise Conceptual Data Model and data model governance. The approach is structured and yet pragmatic, guided by the areas with the highest business value. 

Why is data modeling relevant today?

Data models are a communication tool between people, most notably from the business and IT departments, as well as a blueprint for building data-centered applications. These broad expectations on data models point out that data modeling is a fundamental activity. However, in reality it is often undervalued and underrepresented. In projects people tend to focus more on the syntax and technical aspects of data modeling, instead of being driven by the business value of it. Data exploitation derived from the business targets leads to a competitive advantage in the marketplace.  

The benefits of data modeling are diverse and are summarized in the illustration below. 

Data modeling supports strategy. Good data modeling practices lead to more adaptable data structures and higher efficiency when implementing strategic changes. Data transparency facilitates merger and acquisitions strategies.

The collaboration between different functions and units, especially between business and IT is promoted by communication data models in a technology-neutral way, but still with enough detail to enable a creation of physical data structures.

Increased performance and reduced costs. In the modern data storage solutions, the most expensive component is not storage but rather computation. In general, a good data model will reduce the cost of the queries, as well as lead to less complex code and thus increase application speed and reduce implementation and maintenance efforts.

Understanding and improvement of business processes. Data modeling forces the company to articulate its business and its processes. It is not possible to define the structure and dynamics of data if it is not known how business is operating

Faster time to value and increased efficiency. With data modeling, business users can have a direct hand in defining core business rules which means fewer revisions are needed at implementation time. This increases efficiency and reduces costs.

Improved data quality. Data modeling can help to identify errors and inconsistencies in the data, which improves the overall quality of the data.

Data modeling became a much more complex endeavor in recent years. The increased complexity and dynamics of the data architecture landscape, exponentially higher data volumes, as well as new technology trends and advancements in the area of AI/ML are the most prominent factors. Combined with the continuously changing business requirements, data modeling must be executed in a structured way in order to be effective.

The good news is that reaching ideas how to build data models per use case became significantly easier. There are many resources about data modeling patterns and industry-specific standards. Furthermore, there are tools suggesting physical data model structures based on provided data, and GenAI is able to generate parts of data models based on the provided requirements. However, all of this does not replace the need for data modelers. Need for data modeling as activity remains. Creating data models has not only a pure requirements and technical perspective, but is also a product of an agreement and common understanding between the relevant stakeholders from Business and IT.

How to implement and manage data models within a holistic approach?

Enterprise data is spread across various data stores, managed centrally or by individual decentral units. Enterprise-wide analytics requires that the underlying data from different sources can be integrated and combined in a desired way. In order to achieve that, a common data model is needed, representing a collective understanding of business entities, relationships, and rules, thus achieving overall acceptance.

Enterprise Data Model (EDM) is such overarching common data model, created as a collaborative effort of different levels and units of the business. It is a holistic layer-based model, whose layers represent the enterprise, subject area, and application levels, and in addition correspond to a conceptual, logical, or physical data model type. The EDM provides a common vocabulary and reconciles differences between different understandings of the same business entity by providing corresponding mappings. For example, the entity Customer in a central department could be a Client in some decentral unit.

The top Enterprise Conceptual Data Model is the only centrally managed component. The subject area conceptual data models (CDM), as well as the corresponding logical (LDM) and physical data models (PDM) below are managed by the individual functional areas of an organization, such as Finance or Marketing.  

The subject area models enable the autonomy of the corresponding business domains over its data while adhering to the broader organizational standards. They are not standalone artifacts but are rather interlinked between each other, enabling the enterprise-wide integration.


This domain-driven design is also incorporated in the data mesh approach, a decentralized sociotechnical data architecture that reinforces the concept of data as a product of individual domains.

The policies, procedures, standards, and metrics that ensure the effective and efficient use of data modeling within an organization are managed by the data model governance. It focuses on the oversight and regulation of the data modeling efforts and is a part of the overall data governance function. One of the areas of data model governance is a setup of internal data modeling guidelines, covering the aspects such as the data modeling notations, naming conventions, design standards, and documentation requirements. A regular quality control is another key aspect, where metrics like the Data Model Scorecard are applied when reviewing data models.

Guiding principles

From the general point of view the following guiding principles are recommended in a data modeling approach:

Make sure that the relevant subject matter experts (SMEs) from the business, as well as the IT experts are part of a collaborative effort to create or adjust the conceptual and logical data models. 

Use generic data modeling templates from the industry or the ones created with AI as an inspiration and a checklist and not as a first draft of your data model.

The EDM itself does not include all the possible data models that exist in a company and focuses on the core enterprise data.

Keep the components of the Enterprise Data Model consistent with each other after every agile iteration.

Creating the Enterprise Data Model is not an academic exercise, but rather a common-sense and practical endeavor. Be pragmatic and keep the benefits and efforts in balance. Start with quick wins first – areas with the highest business value.

The data modeling capability of an organization is a journey often characterized by levels in a corresponding capability maturity level model. The maturity assessments enable the evaluation of the current state and show a roadmap on the way forward. From the initial ad-hoc efforts, the organization typically develops by establishing standards and subsequently governance policies, by introducing more advanced tools and continuous improvement practices.

AdEx Partners offers a comprehensive approach for assessing the data modeling capabilities of your organization and steering them forward with you to meet your specific business needs. We start with the analysis of the current state and drive the transformation initiatives accompanied by the best agile and project management practices.