Questions to Ask Before Starting Master Data Management


When was the last time someone -- who wasn't a vendor -- told you a technology project would deliver a return on investment that exceeds your wildest expectations?


I've seen promises that you would get a ROI, but never that it would exceed expectations. And yet, that's what Jill Dyche, a partner and co-founder of Baseline Consulting, daringly promises in this excellent destinationCRM.com article about Master Data Management. The exact quote:

There's been a slower adoption curve than we expected, especially given the increased maturity of the software. But for those companies that have overcome these deployment woes, the ROI can exceed their wildest expectations.

As an outside observer -- which is to say, someone who doesn't actually work in IT -- I must say master data management looks like a no-brainer. I mean, of course you should have one place that reconciles data inconsistencies and acts as the central clearinghouse for customer data. As a customer, I have to wonder: Why in the world don't you have that already?


But you only have to read a bit more deeply to see why MDM is so intimidating, regardless how mature the solutions might be.


One factor complicating MDM is that, really, it's as much a business quandary -- who's responsible for cleansing data and determining whether this information is right -- as it is a technology implementation.


In fact, according to Dyche, failure to promote MDM as essentially a business issue, not an IT architecture upgrade, is impeding MDM at many organizations.


Apparently, business executives hear about MDM and think, "Well, isn't that what our data warehouse does?" I suspect that, just like me, they can't quit believe they've been operating without one clearinghouse for data quality.


Dyche suggests you overcome this misconception by focusing on the business benefits that will come with MDM's additional functionality. The business case will vary, of course, but she gives as an example applying MDM to health care: You create a master patient index so that everyone involved sees the same customer data. She doesn't mention accuracy, but that's a big selling point no matter what your industry.


(For a look at how the state of North Dakota applied MDM to its Medicaid system, read our interview with Walt Ellenberger, the vice president of Payer Practice for Initiate Systems.)


But if Ventana Research's statistics are right, I can't imagine it will be hard to build a business case for your specific industry. According to this Ventana white paper, 26 percent of companies have customer data in up to 20 types of systems, and only 11 percent have total confidence in the quality of their customer data.


But beyond the the business problems, there's a lot of work to be done before you even begin an MDM project, as this article explains. You'd do well to read the full piece, which goes into greater detail.


Here's my synopsis of questions you should ask yourself before beginning MDM:


1. Where are you going to start? The article recommends you focus on either customer-data integration (CDI) or product-information management (PIM), not both, but before you buy a solution, consider how far out you want to build.


Karen Leightell, senior product manager for IBM Master Data Management Solution Group at IBM, told IT Business Edge in July that MDM offerings developed from either PIM or CDI and their offerings reflect their origins. We're only just starting to see signs of the two MDM solutions converging in what IBM calls multi-form MDM. So, keep this in mind when talking to vendors.


2. Does everybody understand his or her role when it comes to handling data? The destinationCRM.com piece explains the importance of separating data governance -- the rules that apply to the data, including how it's labeled and who gets access to it -- from managing the data. Guess where IT fits in? If you guessed managing the data, give yourself five points on the "I Get IT/Business Alignment" score sheet.


3. Do you have underlying data-quality and integration tool sets in place? In the article, Ray Wang, principal analyst of enterprise applications at Forrester Research, points out that the foundation for MDM are enterprise-information integration (EII) and operational data stores. How far along you are with those pre-steps will affect how long and expensive MDM will be for your company.


4. Have you built a service-oriented architecture (yeah, right) -- or, at least, do you have Web services up and running so you can tie your data together? The article notes that SOA is a precursor for MDM because it enables data synchronization across different systems. SOA is a foundation "technology" for MDM, but MDM may prove to be an enabler for your SOA, since it can help with the payoff.