Data management modernization — pivoting from PPT to 3EDC framework

The Traditional People Process Technology approach for data management operationalization is not aligned with the advancement of the data technology landscape. This is my thinking on how this gap can be resolved.

The traditional data management approach is based on “People, Process, Technology” (PPT) concepts. In general, PPT based approach is focused on 1) establishing data organization with data forums, roles, and responsibilities, 2) defining and implementing operating models for data processes, and 3) deploying technology tools to support data processes. The evolution of the data realm (volume, variety, velocity), data technologies (big data, cloud), and data architecture (e.g., cloud data warehouse, data lake, data lakehouse, data mesh, data fabric) imposes significant challenges to the traditional data management approach. For example, we can establish data forums, roles, and responsibilities, but if we don’t establish data culture, the operating model will fail in execution. Another example: we can define and implement an operating model with data processes, but without advanced automation capabilities, the data program cannot scale with data complexity as we have today. In my opinion, the traditional data management approach (PPT focused) and “by the book” implementation is not sufficient to ensure the success of the data management program. Hence, the approach has to change to catch up with the (r)evolutive changes in the data realm.

People, Processes, and Technology (PPT) are and should remain key pillars of the data management approach. However, considering the (r)evolutive changes in the data realm we need to take a more pragmatic approach to ensure that data management capabilities are aligned with the organizational data landscape. In my opinion, PPT should pivot to 3EDC to resolve this gap. 3EDC stands for Establish Data Culture (1st EDC), Enable Data Capabilities (2nd EDC), and Enhance Data Content (3rd EDC). Depicted below is an overview of the transformation process:

• Establish Data Culture (->People): Defining the data policy and standards and establishing data forums, roles and responsibilities remain steppingstone for data management operationalization. However, the secret sauce for success is to establish a data culture where every individual in the organization puts data-related activities on top of its agenda in daily operations — in other words, an individual becomes a data citizen with the right to access data, but also with the responsibility to contribute to data management. Data citizenship is a very important concept — a great example is that one of the leading companies in the niche, Collibra, has been appointed the Chief Data Citizen role (assigned to one of the co-founders Stijn (Stan) Christiaens). A pragmatic approach to make that happen is to define and operationalize metrics that will measure the impact of data on business outcomes.
• Enable Data Capabilities (->Process and Technology): Processes and technology are tightly coupled to define data management capabilities. The secret sauce here is the combination of a traditional operating model with a detailed definition of data processes, along with advanced technology capabilities such as automation and AI/ML. Automation and AI/ML plays a pivotal role to enable scalable data program. One example is Data Quality capability where the traditional manual definition of data quality rules has to be enhanced with intelligence-driven AM/ML capability able to “understand” the data, identify anomalies and automatically generate data quality rules associated with critical data (a good example is Collibra Data Quality). Another example is Metadata Management capability where traditional static and passive metadata discovery and intake into data catalog, can be enhanced with intelligence-driven(including inference to enhance baseline content) and actionable metadata that lives together with data and can span across the entire data landscape.
• Enhance Data Content (->): This aspect of the 3EDC framework has no direct mapping in PPT. However, the intention is to improve data consumption and leverage maximum data potential to drive positive business outcomes from data. That can be improved by introducing domain-based data management (a good example is the Data Product concept in data mesh architecture) as well as linking business semantics with data (a good example is introducing a semantic layer on top of data lakehouse with Stardog virtual knowledge graph on top of Databricks data lakehouse or another one is Cambridge Semantics Inc. Anzo platform).

In conclusion, the 3EDC framework is not brand new, but rather an evolutive enhancement of the traditional People, Process, Technology approach to ensure data management capabilities are aligned with “cutting edge” technology. That alignment is needed for two reasons: 1) to make the data management program scalable and 2) to drive positive business outcomes by leveraging maximum data potential.

What are your thoughts? Comments are welcome and appreciated.

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the opinions or positions of any entities author represents.