Master Data Management (MDM): an Essential Part of Data Strategy

First of all, what is Master Data Management (MDM)?  Master data refers to the critical data that are essential to an enterprise’s business and often used in multiple disciplines and departments. MDM is the establishment and maintenance of an enterprise level data service that provides accurate, consistent and complete master data across the enterprise and to all business partners.The concept of Master Data Management originated around 2008 when data warehousing and ERP applications became popular in many organizations. With the increase of data volume and the number of databases, and thus the increase in the number of applications for users to enter and read the data, it became more and more important to make sure that correct master data definitions are used so that there is asingle truthin the data without discrepancies, duplications or being out-of-date. The first example is related to customer information. In a big organization, there could be multiple customer databases populated and managed by multiple applications or department silos, and the same customer in the real world could receive multiple direct mails or notifications from the same company.  As the data grows, the master data consists of not only the customer information, but also other key data assets, such as the data of prospects, suppliers, panelists and products. MDM has been a challenge to implement, because all three aspects of processes, technologyand tools are required to ensure that the master data is coordinated and synchronized across the enterprise.  

 

 

 

With the recent explosion of big data and rapid progress of analytics and IoT, the consistency of referencing and applying high-quality master data has become unprecedentedly crucial. An enterprise should not only need to make sure it has its key data assets efficiently and accurately managed, but also embrace new data assets to fully realize the economic potential by joining or referencing the existing master data that the company already owns. As such, MDM needs to be an essential part of the data strategy for a company to grow and profit as well as one of the core missions for C-Suites and Executives such as Chief Data Officer (CDO).

 

There have been 2 main reasons for failures of MDM initiatives in the past 10 years: 1) Relying on only technology and tools without buy-in and support from business units; 2) Focusing on fixing and solving current data issues, without forward thinking.  For MDM to be successful, it needs to be first a business-driven process and embraced by business departments and executives. In many cases, fundamental changes to business processeswill be required to establish and maintain unified master data and some of the most difficult MDM issues are not technical at all. Next,a forward looking strategy is crucial in placing MDM as an essential part of data management in an organization, since it proactively lays the foundation for future success. Numerous experiences tell us that the implementation of MDM is easiest and smoothest when a dataset has just been introduced for ETL and into a business intelligence project. Trying to fix for existing data assets and processes often require high cost and large effort, which also likely leads to big impact on the current deliverables. Below is a comparison to illustrate the big differences between implementing MDM at the beginning versus fixing the issues of existing data and systems:

 

Delayed Implementation of MDM	Implement MDM from the Beginning
Duplicate processes in silos and costly developments	Efficient and faster development with lower cost
Data quality issues everywhere with no easy way to track down	Fewer data quality issues that are faster to fix
Low customer satisfaction	High customer satisfaction
Data asset potentials are not fully realized	Generate more revenue opportunity
Difficult to migrate to new data platform	Much easier to migrate to new data platform when needed

With the above comparison, it is clear that MDM should be an essential part of any company’s data strategy, and should be forward-looking with long-term commitment. In other words, MDM needs to be treated as an investment, which will pay off in the long run and establish a solid foundation for a company’s growth and profitability in the areas of big data, analytics and IoT.

 

 

Oncea MDM strategy is set, the next step is to implement the master data management within an enterprise. This is a big topic that can be covered in much depth in a book by itself. In this article, I would like to give a very high level introduction and point out 4 steps that are essential for a successful implementation of MDM. Each of the steps will warrant its own topic in the future with more elaborations and detailed examples.

 

 

This is the most critical and essential piece of MDM, and also the most difficult one. To enforce MDM requires the commitment of a data governance committee, which normally has the following structure:   

 

 

   The main missions of the committee include the following:

 

 

Below lists some of the key areas that the data governance committee should make decisions for:

 

 

 

Always keep in mind that the simplest and most efficient way to make MDM successful is to enforce data consistency when the data is created. MDM is a long-term project and requires the long-term commitment of a company. Any attempt to change the data in an ad hocway renders the effort both ineffective and costly. Data governance policies and definitions are implemented throughout 2 channels: 1) via any new projects and application development;  2) by using a data governance software. Many organizations’ MDM implementations stalled because of the high cost and effort they faced when trying to fix the existing systems and issues; they did not realize that the best way to start with MDM is to apply it for going forward for new projects, which will test it out first and enable the organization to build up expertise and experience.  

 

In addition, most of the data governance should be implemented directly as part of data related projects into applications and reports. For example, data governance should enforce and propagate its definitions, policies and principles into the following technical implementations:

 

 

 An ideal MDM software should have the following functionalities:

 

 

There are many tools on the market that can do 1) and 2), but it is not easy to do 3) with the same tool. This is the reason why a MDM software can be also a data integration tool at the same time, orvice versa. Recent rapid progress in artificial intelligence (AI) has made such software more powerful with enhanced data management, which has a bright future in the coming years.

 

Keep in mind that a tool is still a tool. Without the Governance committee and sponsorship from the C-Suite and executives, the software itself cannot play the magic and is not sufficient. In addition, constant communications and reinforcement by information stewards and data stewards in each department also play more essential roles than the software itself.

 

 

Many companies may not have the luxury to create new master datasets from scratch, which means they need to revamp the existing database and related applications. Applying MDM to the existing data assets often requires a large amount of effort, which could also fail or abort due to complexity and cost. To make the MDM effort successful, careful planning is required to establish a road-map with multiple phases. Sometimes, it may be a better strategy to apply MDM only partially, until the data or system is migrated to the new platform, while focusing on applying MDM to new master data that are being added or new applications and processes that are being built for the enhanced and new data sources to be joined with the master data.

 

 

MDM has become a necessity for an organization to fully realize its datasets’ revenue and profit potential, but it is not easy to implement. MDM first needs to become a permanent part of the data strategy in order for the company to have a long-term commitment. Next, it requires consistent governance and sponsorship from the top management, as well as persistent efforts from information stewards of IT/CDO departments and data stewards of business departments. The challenge of fixing existing data and system issues should not stop the adoption of MDM for an enterprise. Instead, applying MDM to the new data sources and new applications will lay the foundation to gradually apply it successfully to the existing data and systems.