What is Master Data Management?
Master Data Management, or MDM, (since every IT person loves TLA’s (three-letter-acronyms) is the process of cleansing all existing, managing the generation of, and governing the entry of information into a database. This process applies specifically to Conformed Dimensions, meaning that MDM doesn’t directly apply to transactional data; rather it is the process of managing the details surrounding your transactions. The goal of MDM is the creation of an easily repeatable protocol meant to facilitate the entry of cleaner dimensional data, which in turn facilitates the downstream Business Intelligence processes.
MDM offerings have exploded over the past few years, with larger vendors like SAP, IBM, Oracle, and Microsoft offering complimentary products to their existing Business Intelligence or Database platforms. Additionally, there are many other offerings from Informatica to Talend to SAS to Kalido to Tibco to Pitney Bowes…that’s right the people who make the large postage meter and mail sorter machines also offer an MDM software solution. However, this also points to the heart of MDM as a business solution, because without proper postal addresses or placement of the address in the appropriate quadrant of the envelope, postage charges and surcharges could be significantly inflated.
It’s a Business Process, not an IT Problem!
MDM, like so many other BI related topics and buzz words, has morphed into a space which includes both process related tasks and software product offerings. The reality is that MDM is a business process. Let me repeat that, MDM is a business process, not an IT function! While IT will invariably provide software offerings intended to facilitate the business process, true MDM is a business function. The sooner this is realized in an MDM project, the sooner you place yourself on a path to success.
In this way, just organizing along the lines of a classic Technical Project Plan ignores the subtle nuances of negotiation required. The way to work through these subtleties is to encourage the business to holistically own the content using an Organizational Change Management tools/frameworks such as the RACI Matrix (or Responsibility Matrix) . At this point, IT becomes a facilitator rather than an owner.
Garbage In = Garbage Out
If you’ve ever worked on a Data Warehouse project, then you’ve surely heard this mantra. It is an EDW Developer’s mechanism for saying “if we don’t get good inputs, then we can’t be held accountable for our systems’ outputs.” It begs the question, “Who’s in charge of getting better inputs?” This is exactly where MDM comes into play. With the implementation of the appropriate process, procedure, and validation/business rules, the inputs can be greatly improved thus improving the subsequent outputs.
The next logical question is how?
Truly Conformed Dimensions (and Measures)
In answering the “How?” posed above, there are a number of first steps, but the best first step is to gauge the business impact and calculate a potential Return on Investment (ROI) for the project. As an example, let’s assume your company uses a mail campaign to reconnect and market to your current & prospective client base and send out 10,000 mailers monthly. Again, assuming that your customer address data is incorrect just 10% of the time (which is low in some cases), you will be sending out 1,000 falsely addressed letters each month costing your company $10,320 annually ($0.43 postage cost x 1,000 erroneous mailings x 12 months), not to mention all the lost potential revenue due to an ineffective marketing campaign.
As a critical first step to an MDM project, the ROI is a way to both justify the cost and ensure an engaged business. As soon as the business has started to own the process, then the real work begins. This is where a little guidance can go a long way because we are about to embark on a difficult journey of agreeing upon some common (and often assumed) business terminology. For example, who is a Customer? What is an Account? How do we calculate Revenue? What is Profit? While benign at face value, get 15 of your key business leaders into a room and float one of these questions and you’ll surely get 8-10 unique responses. While a difficult procedure, this is the most critical to the success of the MDM implementation. All the software in the world won’t fix your data unless your business can agree to how it should be defined. However, with the appropriate amount of structure, the freedom to fully explore each topic, and the agreement of collaborative completion, this will be a process benefitting your company for decades.
Governance & On-Going Support
Now that you’ve completed all the hard work of defining your terms & calculations and implemented the processes & software required to correct the data in your systems, it’s time to decide who is going to manage the long list of updates that are waiting to be made or how are you going to support the process for when updates are requested? Governance and Support are critical to the longer term success of MDM. It’s best to identify a non-IT governing body, or committee, who will serve as the decision making entity for all clarification requests. The best tool to support them is a collaboration portal which is accessible by all in the organization and used to reference the details (Definition, Calculation, Derivation, Sources, etc.) of each artifact. The guiding principle of this MDM Committee is to ensure the adherence to the protocols set in place to provide clean data, clear procedures for entering & cleaning data, and support to handle exceptions and updates to the existing protocols.
Conclusion
MDM is a lot like laundry, in that every company has dirty data, just like every household has dirty clothes. The key is to make the cleansing and governance processes an integrated part of the daily tasks/chores, thus mitigating paralyzing piles. In the end, every company stands to benefit from some sort of MDM. The process needs to at least be put on the roadmap and discussed as a future endeavor. It is not something that should be done hastily or without proper priority and visibility, because half an MDM project can do more harm than good. The best-case scenario is an engaged and willing business community, including strong executive sponsorship, and the appropriate time and software to integrate these processes into your standard daily procedure.
What next?
Let Axian come to the rescue and help define your BI strategy, develop a roadmap, work with your business community to identify the next project, and provide clarity and direction to a daunting task. For more details about Axian, Inc. and the Business Intelligence practice, click here to view our portfolio or email us directly to setup a meeting.
[1] The RACI Matrix assigns project participants to one of four categories: Responsible (Owner/Decision-Maker); Accountable (Task Executor); Consult (Advisory Role); & Inform (Notified/Updated post decision). For more information on RACI, click here.