Within the field of data governance, there are different ways of approaching data and the definition of truth. In this post, we will look at different approaches to data and also how truth can be defined with a data governance framework.
Defense
A defense approach to data is focused on controlling data. This can involve security and stringent governance of data through a highly centralized setting. In addition, the defensive data approach is concerned with minimizing risk and ensuring compliance with standards and expectations. Preventing theft and tracking the flow of data through an organization is also important.
ad
When analytics are used they are used to detect fraud and unusual activity. How defensive an organization is depends on the field or industry. For example, banking and health care are highly defensive due to the type of data they gather.
Offense
An offensive approach to data is focused on developing insights with data. The goal is not to protect but to develop insights for decision-making. An offensive approach to data is characterized by flexibility and being focused on the customer. This style of approaching data is generally emphasizing a decentralized style of data governance.
Organizations that find themselves in highly competitive environments often are forced to become more offensive as they search for insights to maximize profits. How much offensive and defensive an organization needs does vary. However, in general, most if not all organizations start defensive and slowly become more offensive in nature.
Truth
Whether the approach to data is offensive or defensive it is important to determine what is the truth when it comes to data in an organization. Every organization needs a single source of truth (SSOT) for critical data. The SSOT is language used within data that is the same across an organization. For example, sometimes the same name can be entered in multiple different ways in an organization's data. Take the company AT&T as an example it could be entered in some of the following ways
ATT
att
Att
AT and T
AT&T
Each of the examples above can be considered different and can lead to chaos when it is time to analyze data for insights. This is because redundant names can lead to redundant costs. For example, if AT&T was a vendor for our fictitious company there might be several different contracts with AT&T with several different divisions who all spell AT&T differently. To prevent this the SSOT will define the one way to code AT&T into the system and determine what it represents.
However, keeping the offensive approach to data in mind. There are times for the purpose of analysis that the SSOT can be modified. Doing this leads to what is called multiple versions of truth (MVOT). An example of MVOT is a department that classifies our example of AT&T different way from the SSOT. Accounting might see AT&T as a vendor while marketing might see AT&T as their internet provider, etc. Since everyone knows what the SSOT is they are aware when they make a MVOT for their distinct purpose.
Conclusion
Each organization needs to decide for themselves what approach to data they want to take. There is no right or wrong way to approach data it really depends on the situation. In addition, every organization needs to determine for itself how they will define truth and there is no single way to do this either. What organizations need to do is address these two topics in a way that is satisfying for them.
