Master Data Management (MDM) is an information management discipline focusing on data quality and data governance. Its goal is to mitigate problems caused when Master Data is inconsistent, fragmented, or out of date. In this article, we’re going to explain what MDM is and provide some tips on how to start an MDM journey.
What exactly is Master Data Management (MDM)?
The discipline of Master Data Management (MDM) is relatively new and so there are numerous definitions for it. A simple, concise definition is not always easy, so instead, we’ve distilled the essence of MDM into the following themes:
- It relates to a company’s Master data
- The over-arching goal is to create a “single version of the truth” for your Master Data (eg customers, products, employees, and physical assets)
- It is a set of disciplines, processes, and technologies
- It covers applications, systems, and databases
- It spans both business and IT realms
- There is a focus on continuous data improvement
- It requires a set of data quality dimensions to be defined (eg timeliness, accuracy, completeness, meaning) by the business
- It often involves unifying data from multiple disparate systems or sources
- It is hard to do
The last point is not intended to put anybody off starting a Master Data Management. It does, however, convey the message that for large organizations it can be a large, complex undertaking – more on this later.
MDM solves business problems
Another important consideration with MDM is that it is used to solve your own particular business problem. The definition or scope of MDM can therefore be tailored to your own situation and the problems you’re trying to solve.
Some examples of business problems that MDM can help solve:
- When two companies merge, they each have their own set of master data; this needs to be merged into a single set. (Incidentally, one of the traits of organizations that are good at acquiring other businesses is that they have highly effective MDM strategies.)
- There is overlap in where master data is stored so that updates are required in multiple places.
- When reference data in system x is updated, it takes too long for the update to synchronize to system y, causing delays or errors.
- There is no clear ownership or responsibility relating to master data
- There is no clear process relating to when or how master data should be modified
What are Master Data and Reference Data?
The terms “Master Data” and “Reference Data” are often used interchangeably, but there are important differences. While both provide context for business transactions, reference data is concerned with classification and categorization, while master data is concerned with business entities.
Master Data
Master Data refers to business entities such as:
- People/Organizations (e.g., customers, suppliers, employees)
- Products (e.g., SKUs, product descriptions, inventory, parts, suppliers)
- Assets (e.g., equipment)
- Locations (e.g., offices, stores, distribution centers)
- Financial (e.g., Chart of Accounts)
Reference Data
Reference Data is used to categorize and classify other data, or to specify the allowable values. It is generally static and standardized across industries. Some examples include:
- Units of measurement
- Country codes
- Conversion rates
- Organizational or location identifiers (eg company abbreviations, airport codes)
There are some other data definitions that we should consider to help us contextualize Master and Reference data, and these include:
- Transactional Data: the everyday, operational information in your CRM, ERP, and HR databases, which could include orders, sales invoices, customer updates, new hire details, etc.
- Meta Data: data that provides information about other data; examples being ‘updated_by’, ‘date_modified’ fields when a Master Data field is changed.
- Unstructured Data: normally text-based data that is not easy to categorize, such as emails, social media posts, articles
Why are MDM Initiatives hard to get started?
If you’ve read this far and have been nodding at some of the problems that MDM tries to solve, then you are probably also aware of some of the implementation challenges.
A lot of organizations know that they have business problems caused by Master Data issues and many will have tried to fix them. Some of the issues they come up against include:
- Lack of Executive support: Senior executives need to understand the business benefits of effective MDM and drive it from the top. Senior interventions are often needed to overcome the “silo mentality” that exists where departments are unwilling to give up control to an enterprise-wide initiative.
- Complexity: Organizations typically underestimate how complex their Master datasets are or overlook the extent of the problem. Legacy systems, increased global distribution and language factors all add to the complexity.
- Governance: Whilst there may be high levels of governance across many areas of an organization; ownership and processes relating to data are often sorely lacking. The lack of data governance is often related to the complexity of data quality issues.
- Duplication: It is not uncommon for large organizations to store the same data (or versions of it) across many systems. Sometimes this is caused by legacy issues, or by departments keeping their own versions of the data because they “trust their version more”.
- Lack of Data Model: Organizations will often embark on an MDM project without first clearly defining simple structures such as which are the primary and secondary master data sets and which are the slave data sets?
- Unclear Data Definitions: It may seem simple to come up with a definition for “Customer” in your Master Data, but it’s not always the case. Unless there is clear direction and ownership, even this seemingly simple task can be difficult and time-consuming.
How should we start an MDM initiative?
There are no hard and fast rules on how to implement an MDM initiative. There are, however, some key focus areas that need to be covered if you want to be successful. These are shown in the following diagram and explained in more detail below.
Business Goals
The first step is to realize that MDM is a business challenge, not a technology problem. Technology will ultimately play a key role in the initiative, but it is not the starting point. You need to make sure that business people are driving it from the outset. This involves defining your goals, strategy, and roadmap.
The strategy and roadmap will need to cover things like:
- Goals
- Priorities
- Definitions of success criteria
- Executive sponsorship
- Establishment of a Data Governance body
- A phased approach to implementation
Identification & Analysis
The next step is to identify what your existing master data and reference data is and where it resides. It’s useful to have an evaluation matrix to help objectively decide what is and isn’t master data. This will also help later when you define the data models.
The identification process also needs to determine how master data currently gets produced and consumed – what people, processes, and technology is involved?
It’s useful early in the process to stress the importance of good metadata. When you are looking at your existing master data, make sure you capture metadata (definitions, sources, consumers, rules, owner, etc).
Ownership & Governance
Without clear ownership, any initiative is likely to fail. This is where strong executive support is important, as discussion around data ownership can often lead to conflict or delays. Likewise, it is important that the Data Governance structure is established early on, even if only at a low level initially.
Infrastructure
Finally, you can start defining your architecture and data model. The choice of technology that you use will likely determine the specific details of the data model (eg specific terminology or object hierarchies).
Summary
In this article, we’ve introduced the discipline of Master Data Management (MDM) and explained why it’s important. We also highlighted some of the challenges in trying to implement an MDM initiative, as well as provided some guidance on how to get started.
In the next blog article, we’ll discuss Microsoft’s platform for MDM – Master Data Services. We’ll also explain how you can use a simple tool like SQL Spreads Excel Add-In to achieve many of your goals.
