Data Governance in Action

Overview 

This is the 3rd article in the series “Digital Transformation in Action” and this article focuses on the implementation of Data Governance (Data Management). The overall idea of this article is to identify the practices and processes to introduce proper data management within the organization. 

The first article focused on an overview of digitally transforming the processes in a company and the second article focused on business analysis. If you have not already read the articles, I suggest you read that first before you proceed. 

Constant improvement is the key. 

One of the key points that I have mentioned in my previous post is that, with connected systems and quality data in place, you will be able to get good insights and an overview of how your company is performing now and how it is likely to perform in the future. It is extremely important to rely on this information so that you can strategically plan the activities in the business to succeed.  

One of the critical challenges you will face while you are in the digital transformation journey will be data management. With all these connected systems and connected data, there is now a need to maintain the quality of data to get reports that provide real value. Again, the overall aim is to win more business whilst improving the profit by utilizing these reports to make better decisions. 

Maintain Quality Data 

How do we maintain this quality data? The answer to that is data governance.  

What is Data Governance 

In simple terms, it is a set of principles and practices to have and maintain high-quality data in your organization, which can be successfully implemented with a collective effort from Business & I.T. 

Once implemented within the systems we need to educate the users about the importance of it and urge them to follow it. 

Why do we need Data Governance? 

Better Analysis and Faster Decisions: Data Governance comes with simplified and accurate data, which will help the business to make decisions faster and better. 

Reduce Redundancy and Improve Operational Activities: Data Cleansing is a part of the Data Governance activity and by having a clean dataset, the chances of users using a duplicate record are minimal. This will reduce the efforts of the users thereby avoiding confusion in record selection. 

Increase the Profit: By improving the quality of data, we will reduce the efforts of the users, which increases profit.  

Now that we know the need for data governance, let us see how we can successfully implement data governance within an organization.

The first step is to identify and gather the information related to master and reference data sets that are used within business applications.  

Identifying Datasets 

Understanding Reference and Master Data. 

Reference Data 

Reference data is information that is used to define other data. If you see any stable dataset used within your application, that is not modified or created too often, you can treat that as reference data. One of the main characteristics of a reference dataset is that it possesses only a few attributes like Name and Description. A few examples that you see within an application are Country, City, Currency, Product, Industry, Categories or Classifications (Product Category, Job Category, etc.) or Types (Product Type, User Type, Job Type, Position Type), etc. 

Governance of Reference Dataset 

The ideal approach is that these reference data should be created and self-governed by a central authority. Only a limited set of users will be authorized to add/modify any information related to this reference data set. However, we could implement a process for the other users to send a request, if they find any discrepancies or missing records related to an existing reference data set.  

Master Data 

If you see a dataset where new records are added or modified frequently, this can be treated as master data. There will be a large set of users adding, changing, or deactivating these records. Datasets like Customer or Supplier can fall under this category. 

Governance of Master Dataset 

We need to identify a set of processes for implementing the Governance of the Master Dataset. Identifying the right data owner from the business is the key. Even if the data owner is a single person, there might be multiple individuals involved in the approval cycle.  

Identify the Users 

In a typical scenario, a user creates a record, and it should be verified and approved by another user, who should be the owner of the record. All the datasets should have a data owner, who verifies and approves the data. Depending on the size of the company, we can have a data steward, who is responsible for verifying the information.

Identify the system 

Now that we have the Master and the Reference datasets identified, we need to identify which system is suitable to hold these datasets. This depends on the overall technical architecture and the users within the organization. The reference data management should be maintained in a single system, which can be connected to all the other applications. However, for the master data management, we should identify the set of users who can understand the facts related to the dataset and thus finalize a system that is mostly used by these users and the system should have the capability to implement approval workflows. 

Conclusion 

Implementing Data Governance is a collective effort. One should focus on simplifying the process rather than overcomplicating it. Understanding the purpose is the key to simplifying things. Start with a smaller set of data and then add more datasets.  

Note: I have added several pieces of information to this article which I have collected through a lot of web searches and by speaking to individuals whom I consider experts in this area; however, the core concepts are defined based on my experiences.