RESEARCH & RESOURCES

Conquering Your Fears of Master Data Management

MDM is being incorporated into big data and big data sources. What pitfalls must you avoid when deploying or using MDM, and what's ahead for this technology?

[Editor's note: This article was originally published by Enterprise Management 360° and is reprinted by permission. It has been edited slightly from its original form. ]

Master data management (MDM) applies to the regulations, processes, and technology that organizations use to maintain order and consistency among the data most crucial to operations and business intelligence, joining all relevant data, in essence, into a single file or single view.

One of the most rapidly growing trends regarding MDM revolves around its incorporation of big data and big data sources, so MDM vendors should be expected to see their offerings concentrated towards big data this year. Will there be any other surprises in the marketplace? EM360° caught up with Aaron Zornes, founder and chief research officer of analyst firm MDM Institute.

EM360˚: On the principle concept of MDM, how are we seeing the industry evolve to where we are today?

Aaron Zornes: The industry has gone from early adopter to mainstream. We are now picking up the "technology laggers" -- certain government agencies and certain geographies where they do not possess the capital or the competitive pressures to do leading edge IT. However, what I'm rather referring to is "mission-critical IT," with concentration in facilities such as data warehouses, it's commerce and other things that we would consider leading edge just years ago.

The thing with data management is that it is embraced by government agencies, home level, companies of all different sizes, and all different scopes of industries. It is very much scaling widely in terms of global, industry, and company acceptance.

If you're dealing with mass and critical elements of data, it is important to have a strategy that is succinct and in keeping with some of the issues that are encountering in organizations today. As mentioned, MDM product interpretations currently amalgamate some of the current trends we see today such as big data, social media, and cloud computing.

Absolutely. There is no silver bullet when it comes to technology. Everything bleeds into each other with concepts such as big data, MDM, and vice versa. MDM can benefit from big data, although big data gives you an indicative source that says who is talking about what and what they are actually saying. This is all master data related. The analytical insights that we get out of big data is why we do big data. Organizations don't just process the data -- we want insights and action.

The economics (Capex and Opex models) of cloud computing is something that affects MDM in terms of the applications that are moving in to the cloud. MDM has to help repatriate that data that is now locked up in the cloud. This could be on-premises data, back office systems on accounting, returns and customer service, and all the things that fit in with automation on the types of data that the company lives on. Sometimes that needs to be repatriated and reinvigorated with the rest of the company's data.

Then with the issue of social -- conducting social networking and analyzing the graphs and relationships of who is related to whom, who has issues with whom -- it is no longer what is your net value to the company but rather the net value of your network of friends and people you influence that companies want to know about and other related organizations. You are putting a lot of your trust and personal data into Amazon, LinkedIn, and Facebook which somewhat becomes your private or personal MDM hub, but wouldn't it be nice to have a way of linking all that together instead of having to/or not have to replicate it across several networking systems and you want to be able to keep it separate?

All of this blends in together at some point. In-memory computing has an impact on MDM through conducting customer matching to be able to possibly identify somebody for transaction purposes or for validations.

From an enterprise perspective, what would you say are the pitfalls when it comes to deployment and utilizing an MDM product that IT departments can find themselves in a tangle with?

Companies of any size are under very high-pressure situations when choosing a perfect MDM system. It is like buying a house for the first time. You have so many details you have to consider before selecting the MDM platform because the investment level is tremendous. This decision, rather unfortunately, gets brought to the IT person or business decision maker, where you have these very well-oiled individual vendors [such as] SAP, Oracle, IBM, and Informatica that are trying to sell you MDM components, so it is a little bit of a sticky situation to be in. It's very much about cutting through the vendor marketing to really understand what works and what doesn't work and asking yourself why you are doing this.

Another pitfall to worry about is how to do this because there are multiple ways to do master data. Before the realization that companies bite more off more than they can chew, MDM is almost ubiquitous. When you buy an SAP or an Oracle application, underneath is an MDM platform that is managing your master data. An MDM platform might not necessarily come from one vendor: these organizations have a mixture of MDM components from different service providers, but each of those application systems has its center of gravity in their MDM hub.

It used to be that you would only use one MDM hub to reintegrate all that stuff. Now we need an MDM hub to reintegrate the MDM hubs to process through all these applications that we buy, so that's a bit of a challenge because you can't expect a provider such as SAP to necessarily do a great job in integrating multiple applications into the SAP world, which is a disadvantage. That opens up the door for some third parties and best-of-breed type vendors to integrate the hubs with each other.

Another big pitfall for most of us is the lack of data governance software. These data hubs -- product, customer, and supplier hubs -- all require good governance practices in terms of determining what data you are mastering and where you get the trusted sources, etc. -- defining those rules and managing those rules, defining who can access what, who pays for what, requires governance. However, today's governance is primarily a manual process because the software industry simply hasn't come up with the requirement to design these hubs the way we design our databases, the way we design our applications and our data warehouses. We don't have the ingenuity to design the tools and the management tools that you would expect for something that is mission critical. We could be doing governance a lot better when it comes to MDM data.

About the Speaker

Aaron Zornes is the founder and chief research officer for the MDM Institute. Aaron is also conference chairman for the "MDM & Data Governance Summit" conference series that is a widely attended professional conference focused exclusively on MDM and data governance. Aaron is a noted speaker and author on Global 5000 enterprise IT issues, as well as a frequently quoted industry analyst on master data management (MDM), customer data integration (CDI), reference data management (RDM), and data governance.

Prior to founding the MDM Institute, Aaron was founder and executive VP of META Group's largest research advisory practice. He has also held line and strategic management roles in leading vendor and user organizations, including executive and managerial positions at Ingres Corp., Wang, Software AG, and Cincom Systems.

TDWI Membership

Get immediate access to training discounts, video library, BI Teams, Skills, Budget Report, and more

Individual, Student, & Team memberships available.