To MDM or not to MDM, that is the question…or is it?

To MDM or not to MDM, that is the question. Well, actually it’s not the question. And really, it’s not a question at all. Rather, it’s a statement. And that statement is: “Pay me now, or pay me much more later.” If you have a business, you’re already involved with Master Data Management (MDM) whether you realize it or not. Whether we’re talking about critical data managed in two applications, or reference data (the more common scenario), the question remains: how is it managed? The answer is: poorly, most of the time, or not at all. In addition, how many times have we seen data being maintained in Access, in Excel, in a csv file, or in a lonely table shoved in a random database somewhere. This post will explain why it’s critical to identify your organization’s MDM need and approach—and that it be more than just a few line items buried in a project plan.

What is Master Data Management?

Wiki’s definition of Master Data Management is “a comprehensive method used to consistently define and manage the critical data of an organization to provide a single point of reference.” Define it once and reuse it everywhere else.

The Three Sisters.

To be effective and provide a solid return on your data investment, the following three processes must work in concert:

  • Data Governance
  • MDM
  • Data quality

Quality is always a clear need. Bad data will kill or diminish the value of an application. And Data Governance is the neglected child. But it’s important to always define all three processes—including Data Governance—which represents the final say in regards to data. Think of Data Governance as being analogous to a football referee.

Data Governance

Data Governance enables an organization to ensure high data quality exists throughout the complete lifecycle of the data, by managing and directing the following:

  • Metadata definitions
  • Master data goals
  • Data Quality goals

Key Data Governance process requirements:

  • Form a Data Governance Committee
  • Take stock
  • Identify Data Owners (Stewards)
  • Data Dictionaries, Quality standards,
  • Policies & Procedures
  • Identify critical Data Elements & Domains
  • Define Control & Success Measures

MDM – The choice

Back to MDM. The decision of how MDM should be implemented at your organization will be based on what your organization can best accommodate, not on what provides the most value to the business. Whether considering cost, effort, or overall impact to applications and organizations, organizations must balance what is available versus what can reasonably be implemented.

Steps to defining an MDM strategy:

  • Define the MDM requirements
  • Define your consumers
  • Define what data to incorporate in MDM
  • Define the Data Quality requirements
  • Select your Style
  • Define your infrastructure
  • Select your tools
  • Define your Data Governance

Some common goals associated with MDM:

  • Identify System of Entry
  • Identify System of Record
  • Maintain a single version of the reference data
  • Create the “Golden” record (e.g. the most comprehensive/accurate view of a customer).
  • Data consistency/Data Quality

Approaches/Architectural Styles to MDM

Once a choice has been made, the next step is to pick an Architectural Approach/Style to implement, of which there are many.

Here is the list of Styles, in order of least intrusive to highly intrusive to implement:

  • Registry
    • No data consolidation (minimal amount of data to identify uniqueness, pointers back to source for rest using federated queries)
    • Complex to set up
    • Not real time/latency in data
    • Read-only access/no loop back to sources
    • Least intrusive
  • Consolidation (also called Conformed)
    • Data consolidation/scrub and aggregate data
    • Not real time/latency in data
    • Feed downstream systems
    • No loop back to source systems
  • Coexistence
    • Data consolidation/scrub and aggregate data
    • Loop back to source systems as well as downstream
    • Not real time/latency in data
  • Hybrid
    • (combination of Registry and Coexistence)
  • Centralized (also called Repository, Enterprise or Transactional)
    • Data consolidation/scrub and aggregate data
    • Loop back to source systems as well as downstream
    • Real time access/integrated with applications
    • Most intrusive

For MDM to be effective, it must work across multiple business processes, functional areas, organizations, geographies and channels. Data Governance is key to managing the cross-business interactions that inherently take place.
The most common style is Consolidation. This is the de-facto choice as it falls in-line with individual project efforts and is commonly used for Business Intelligence/Data Warehouse efforts. We could go into greater detail on Architectural styles, but we’ll save that for a future blog.
Standardized data doesn’t go away with the advent of the Cloud, microservices, or SaaS providers. Marketers would have you believe that with new products in the cloud you can do things quicker, but the challenge will actually get worse. With control of the data (especially with SaaS managed data), data conforming is reduced, and the number of data sources dramatically increases. Long story short: you need a plan.
The goal here is not to define your MDM strategy, but to make you aware of your need for a strategy and start you down the process. So, does your organization have a strategy in place, or are you just going with the flow? If your organization needs help defining its MDM strategy but doesn’t know how or where to start, please don’t hesitate to reach out to Anexinet. We’d love to help you out.

Share on

linkedin sharing button twitter sharing button