What's Next for Master Data Management?

Master data management comes in two flavors: customer data integration (CDI) and product information management (PIM). But soon, you can expect to see these two products migrating toward each other, according to Karen Leightell, senior product manager for IBM Master Data Management Solution Group at IBM.

Leightell says one problem with many MDM solutions is that the vendors used a CRM or ERP design philosophy for MDM. IBM says it is different because its solution was designed from the ground up specifically for MDM:

Where they fail is in the underlying design principle, in that those databases do not have the functionality embedded within to manage the maintenance and quality and integrity of the customer data itself. Whereas CDI solutions and MDM solutions, as they're now referred to, were designed specifically to address that problem -- the accuracy, relevancy of the master data --and had no other functions other than that -- and they were essentially marketed as completely agnostic to the front- and backend systems. ... So, in order to get the best-of-breed solution for MDM, it has to be built for that purpose from the ground up -- not be a CRM tool or a data warehousing tool repurposed.

Read the entire interview:  What Master Data Management Can and Can't Do

The point of MDM is to address the problem of inconsistent and conflicting data across large organizations. For instance, customer service may have your old address, while sales has your new address, under a slightly different name. MDM, in effect, consolidates all the views of a customer or a product, or even a partner.

"Getting Started with Master Data Management," a white paper by Deloitte Consulting and Hyperion Solutions, talks about the importance of IT and business working together on MDM projects. It defines master data as the company's set of control data, used to categorize, aggregate and evaluate transaction data. Master data, it explains, allows IT and business users "to understand the meaning of each piece of data in a transaction."

But to be clear: Master data is not transactional data, as Leightell explained:

By definition, MDM becomes the system of record. I should also say that master data is not all of an organization's data. What these companies are doing is looking at a subset - the most critical data - and they may gradually add more types of data.... What master data management does not want to become is another huge data mart. That is not the idea. Again, they're transactionally based so they cannot act like a data mart or become another silo of information.

MDM's usage is growing. An ARC Advisory Group survey shows the master data management market grew 30 percent last year and projects it will reach $1.35 billion in 2011. Anureg Wadehra, vice president of marketing at Siperian, a vendor of centralized data hubs, told IT Business Edge that there are two factors driving MDM adoption

.

1. Despite further investments and upgrades to CRM systems, businesses still don't have a 360 degree view of their customers.
2. IT has stepped up efforts to integrate systems and dissolve information siloes, and that, in turn, has put a spotlight on the problems of inconsistent data.

But it's no longer just a solution that focuses on B-to-C; companies are now using it more for B-to-B. Leightell believes this is a side effect of regulatory compliance issues:

What was a real accelerant to getting an MDM solution put in was the regulatory requirements. On a dime, these organizations were not able to produce that information quickly. In the B-to-B, it might be a manufacturer that has a lot of organizations it deals with or a financial services company that deals with companies on the consumer side. It could be that a company needs to get a global view of their distributors.

Companies may be embracing MDM, but database consultant Rick Sherman, who writes the Data Integration Adviser column for DM Review, believes the current MDM market is about systems integration - not packaged solutions. He and others point out that MDM requires collaboration between business and IT as companies work out which data is correct, how data should be changed and which business policies govern the data.

In fact, he recently expressed doubts about whether MDM software would help or hinder the information silo problem:

The critical success factors for reference data management, the cornerstone of both MDM/CDI, are people and processes rather than products. A cynical view would be that new software solutions may result in new data silos, which is the exact opposite of their intent. A combination of people, processes and products (new or old) may lead to success, but product-only solutions (silver bullets) will only result in a shortfall in expectations and business ROI.