Musketeers of the Data World

There are two concepts that have been on the agenda of professionals working on data management for a while.

Data Governance and Data Mesh

When we look at it in very general terms, the agenda of today's data world, especially in our country, is focused on topics such as data warehouse, data lake and recently data lakehouse. Alongside these topics, analytics, reporting and Business Intelligence (BI) applications are also priority areas.

This is of course perfectly normal. Focusing on technologies that touch business outputs and produce tangible benefits and producing studies on these issues has become a necessity when we consider today's economic conditions.

The two concepts I will discuss in this article are a bit more difficult to explain and to make people accept. It can be challenging to explain the indirect benefits, especially to company management. Let's try to elaborate on these concepts a bit more.

Data Governance

In my previous articles, I have talked extensively about the importance of data governance and how we have to apply these principles if we want to create value from data. In this article, I will aim to provide a more concise and concrete overview of the topic.

You can think of Data Governance as the traffic rules of the data world. Just as there are rules that everyone must follow in traffic, similar principles apply in data management.

This approach includes:

• Data definitions and standards  (i.e. everyone in the organization should speak the same language, everyone should understand the same thing about any concept or business term)
• Data quality  (our data should be free from errors, corruption and omissions, and should provide accurate and reliable information) 
• Data security  (Data must be protected against unauthorized access and misuse) 
• Data ownership  (who can access which data, who is the decision maker about which data) 

The important paradigm shift here is that the people who should make decisions about our data should no longer be IT, but business units. IT should be responsible for managing the infrastructure here. For example, the relevant business unit should decide which criteria to use when calculating revenue. As it understands the relevant business processes, that unit is best placed to determine whether VAT or e-commerce sales should be included in the calculation. The relevant business unit should therefore take responsibility for this definition and assume ownership of the data. In this way, we will have created a standard for many of our business terms, and everyone in our organization will understand and calculate the same thing for a business term. This prevents conflicting values from being reported for the same business KPI. The great thing about data governance is that it keeps everyone on the same page, which means more consistent and accurate decisions, better data quality, and easier legal compliance.

The great thing about data governance is that it keeps everyone on the same page, which means more consistent and accurate decisions, better data quality, and easier legal compliance. 

Data Mesh

Data Mesh is actually a concept that is thought to be implemented by companies that have reached a certain level of maturity in data governance. It is important to have been implementing data governance for a certain period of time and to do this throughout the organization. The starting point of Data Mesh is to be able to produce a solution for issues where Data Governance is not sufficient.

It is impossible to discuss this concept without mentioning Zhamak Dehghani, who introduced the Data Mesh approach during her time at Thoughtworks. We can say that this concept, which was first mentioned by Martin Fowler, one of the most well-known people in the field of Agile, on his own website, gained interest and acceptance in a short time. 

Here, in a very simple way, we can say that everyone is responsible for their own data. Just as micro service architecture has created a change in the software world, it is possible to say that data mesh has created a similar effect in the data world.

The concept of Data Product is also one of the terms used by this new concept. Units are responsible for their own data and its management. Data is actually a product and can be served as desired within or outside the company. For example, unit A is responsible for the quality and security of its own data in every way, and if unit B wants to use any of this information, it can consume it like a product. Here, unit A, of course, undertakes all the quality etc. regarding the data it offers or serves, as I mentioned. Unit B trusts this information and uses it as it wishes. 

Data Mesh offers a distributed and domain-centric approach to data management. In traditional data platforms, all data is collected in a central data warehouse or data lake, and data processing and analytics are usually handled by a central team. Data Mesh aims to move beyond this centralized structure and distribute data ownership and management to teams operating within specific business domains. These teams become the owners and custodians of data specific to their domains. 

The basic principles of Data Mesh can be summarized as follows:

• Domain-Driven Ownership : Data responsibility is distributed across the various business areas of the company or organization. This allows data teams to better understand the needs of the business domains they are responsible for and develop solutions that meet those needs.
• Data as a Product:  Data is designed not just as raw information but as user-friendly, reliable and multi-purpose products. 
• Autonomous Data Platforms:  The technological infrastructure allows each area to produce and manage data independently. 
• Federative Data Management:  Standards such as data sharing, quality and compliance are carried out in a coordinated manner by cross-teams. 

It’s safe to say that Data Mesh is one of the most groundbreaking approaches to modern data management. The decentralization of data ownership and management provides organizations with greater flexibility, speed, and innovation. However, to implement this model successfully, organizations need to transform not only their technological infrastructure, but also their ways of working and organizational culture. Assessing the potential offered by Data Mesh is a critical step for the data-driven organizations of the future. 

In conclusion;

Our data is extremely valuable. As institutions, we have learned, accepted and internalized this over time. Now it is time to look at how we can manage it better and create more value from it.

It is essential for a data-driven perspective to become part of the organization’s strategy. It is equally important for organizations to have an approach to how they manage their data. In the coming period, the importance of data seems to increase and continue to be the focus of more institutions.