Michael C. Pelletier, MBA, MSCS
In the last article we introduced the concept of Master Data Management (MDM) and why it is relevant to a broad range of organizations. We ended the article with some of the indicators you might look for to determine whether you have an operational or analytical MDM challenge within your organization. If you're like many of the organizations we work with you probably asked yourself the following question when you finished the article, "So now what?" With that in mind, let's dive into what's next.
Perform A Viability Assessment
The very first thing to do is carry out a viability assessment. There are several questions that need to be answered to help an organization determine whether MDM is a viable solution:
Do all the systems to be integrated have a data import/export capability?
Without the ability to import and export data in a reliable, repeatable and schedulable way, an MDM solution can become very challenging to implement. Depending on the different MDM model you choose to implement, you may be writing data back to each of the ERP systems to keep their data in sync. A real-time interaction capability with a system is also sufficient and, in many cases, preferable. However, if you can't easily get data in and out it can pose a problem.
What are the short- and medium-term plans for some of the existing systems within the organization?
Consider an organization that has four different business units each running a different ERP system. Perhaps a near-term strategy is to implement an order management solution that sits in front of the ERP systems allowing them to take on more of a warehouse fulfillment role. MDM doesn't play as significant a role (if any) if what you were trying to manage is customers and product.
Is there common meaning within an organization as it relates to the various entities under evaluation (Customer, Product, Vendor, Territory, Region, etc.)?
Without a common definition for these it makes defining the master data attributes of these entities challenging. Further, it complicates the task of defining common processes for managing this data over time. Before venturing too far down the path of an MDM project, it's important to ensure that you have at least a solid baseline from which to work.
Is your organization ready? Does it have support from senior leadership?
While many individuals tend to think of MDM as a technology project/challenge, the technology is only one component. In order for an MDM initiative to be successful there will be process change and standardization. There are likely countless differences in the way the different organizations sell their products, represent customer relationships in their systems, manage commissions, etc. Without support from senior leadership to potentially make changes to how the business operates, an MDM project will likely not succeed.
Are you prepared to take on an enterprise application integration (EAI) initiative and/or a data warehouse/business intelligence (BI) project?
MDM projects executed in a vacuum provide almost no value to an organization. Specifically, if you install an MDM product and go through the process of importing a bunch of customer or product data into it and come up with a nice clean model with a golden record for your customers and products, you've done some good work but have little business value to show for it. If you're building this to address operational MDM challenges and aren't also planning to build the EAI framework through service-oriented architecture (SOA) or other approaches, then your data will never make it back to those systems and your goals for operational MDM will fall short. The same thing can be said for an analytical MDM project. If you don't also build the BI components using the MDM model to source the dimensions and the source systems to provide the facts, you'll, again, fall short.