The Importance of Data Governance

How do I ensure high quality of master data in my organization? The answer to this question is key for any data-driven organization. A successful MDM strategy also requires solid execution. In our previous blog, we discussed setting up a master data strategy. In this blog, we focus on data governance, the management and oversight of high-quality master data.

Finally, master data is a moving target because customers continuously raise the bar and market players such as GS1 and Amazon are constantly evolving. Therefore, the third blog will address lean master data management (lean MDM).

What is Data Governance?  

Data governance refers to the entire framework of processes and methods used to ensure data quality. Good data governance encompasses several aspects:

• It defines responsibilities for officials and departments.
• It specifies how data quality is measured and what goals apply.
• It includes mechanisms to detect and correct deviations.
• It includes a management model for changes in the data model and business rules within the MDM system.
• It ensures compliance with agreements at all levels, from management to operations.

Successful data governance is not limited to data alone. It involves a combination of people, agreements, procedures, and technology. Two crucial questions are: how do I start with data governance and what organizational model should I choose?

How Do I Start with Data Governance?  

Commitment from the management and all involved departments is essential for a successful MDM strategy. The next step is for the organization to recognize that high data quality requires a good governance structure. For this, one should formulate answers to the following questions (per data domain):

1. What is the importance of good data for our customers? What are the potential consequences if customers receive incorrect or incomplete data?
2. What is the importance of good data for our own business processes? What are the potential consequences and extra costs if we work with incorrect or incomplete data?
3. What is the current level of data quality in our organization? Where are the bottlenecks or inefficiencies due to the lack of direct and reliable access to necessary data?

Successful implementations often start small, secure buy-in, and expand step by step. Stakeholders from management and execution are involved from start to finish in establishing data governance. Useful tools for this can include:

• Define ownership and responsibilities by domain and field, for example in a RACI matrix.
• Establish clear rules and guidelines for master data fields and share them with the entire organization.
• Set up clear processes and systems for storing and managing master data.
• Use technology to your advantage by setting up automated data quality checks in the MDM platform.
• Provide executors with direct insights into data quality by setting up dashboards.

Remember that data governance is not a one-time project. It is an ongoing, business-driven process. The organization and the environment are constantly evolving, so successful data governance will continually adapt to changes. Step by step, this is integrated into the corporate DNA and data governance matures. The collaboration, as depicted in the diagram below, is increasingly refined.

Which Organizational Model Should I Choose?  

There are three commonly used models for organizing data governance:

1. Centralized

   Operational management is handled by a single specialized team, such as the Data Management department or the PIM Team. This model is often used in retail and smaller organizations.

2. Hybrid

   The hybrid model combines elements of both the centralized and decentralized models. Data is centrally managed where possible, and delegated to other departments where necessary. For example, a manufacturing company exporting to many different countries might centrally manage technical product data, while local commercial content is added at country-specific locations.

3. Decentralized

   Departments are responsible for the control and enrichment of data within their own domain, such as logistics for logistics data, e-commerce for online data, and procurement for supplier data. This model is often used when many different departments are involved in the product creation process, particularly in the manufacturing industry.

The diagram below provides a comparison of the different models.

Conclusions  

High data quality is ensured by effective data governance. Key factors for successful data governance include:

• Start the journey toward a professional data governance organization with small steps. Begin with a process where a lack of (quality) master data leads to bottlenecks in the organization or dissatisfied customers. Analyze the process, address the bottlenecks, and implement improvements. Share the successes to build support for the next steps.
• Ensure that the business takes ownership of each MDM domain, with support from IT.
• Choose a data governance model (centralized, decentralized, or hybrid) that fits the organization’s needs.