Stage Set for Master Data Management Adoption in the Enterprise
Done right, MDM delivers both strong ROI and improves the reliability of one's BI or data warehouse assets.
- By Stephen Swoyer
- December 10, 2009
With the economy showing signs of stability, some believe the stage is set for the too-long-prolonged adoption of master data management (MDM) into the enterprise.
MDM brings the identification, reconciliation, rationalization, and standardization of data across the enterprise, but it's a lot sexier than it sounds, advocates claim. Done right, MDM delivers both ROI bang-for-the-buck and improves the reliability of business intelligence or data warehouse assets. One problem, backers concede, is that MDM doesn't seem sexy. Some even point to what might be called a "demo-ability" gap from other technologies, such as BI, that have more currency with C-level executives.
"Master data management is really hard to sell. For example, it's a lot harder to sell MDM internally than some of the BI projects, because executives are more BI-savvy now and they get it faster. But when you start to talk about operationally reconciled data through a common master hub, it's a different conversation," observes Jill Dyché, a partner and co-founder with business intelligence (BI) and data warehousing (DW) analyst firm Baseline Consulting.
"In BI's defense," Dyché continues, "it's more demo-able. You can actually show somebody a slick scorecard tool. MDM, in a lot of ways, is infrastructure; it's a lot harder to paint the picture in real life for people."
Demo-ability is one thing, but MDM's biggest problem is one of timing. The adoption and spread of MDM was supposed to be in full-swing by now. It is in a sense -- most large shops have master data efforts of some kind (see http://www.tdwi.org/News/display.aspx?ID=9686).
Most of these are kludgey, however. They were developed and implemented at a time when discrete MDM tools were either primitive or non-existent. Home-built MDM projects -- which are usually yoked to that jack-of-all-data-management-trades, the operational data store (ODS) -- are likewise primitive. They're duct-tape-and-super-glue efforts that were basically designed to hold things together until something better -- preferably, something shrink-wrapped -- came along.
"Six years ago, there was no such thing as MDM, and companies had to build it. One of our clients, Royal Bank of Canada, needed to reconcile customers across [their different lines of business] -- so not just the banking, but the insurance, the capital markets, the international [operations]. There was no way to do that, so they had to build their own," Dyché explains.
RBC, of course, is an obvious exception, she insists. Instead of treating its MDM issue as something to be solved with duct tape and super glue, it cast it as a strategic initiative, building a powerful, flexible, and manageable master data solution from scratch. Most folks didn't take that route.
In fact, comparatively few could afford to do as much, given the costs both in human resources and IT budget dollars.
"Most of the homegrown MDM solutions we see are very high maintenance and require specialized people and specialized code, so most people are saying, 'We'd love to get off this and on to something that's easier to manage,'" Dyché notes.
MDM-ification is Coming. Really.
Advocates and analysts alike have been saying that MDM is on the horizon for some time. A couple of years ago, Dyché and other observers started talking up a coming wave of MDM-ification. They had it partially right, accurately forecasting the rapid maturation of MDM tools and the surprising diversity of MDM offerings. What they didn't foresee was the recession of late 2007, to say nothing of the financial crises of 2008 or the slow (and to this point jobless) recovery of 2009. That's one reason why MDM-ification has been a long time coming. Dyché, for her part, says things are clearly ramping back up.
"Our take is that MDM is still at an early adopter stage," she acknowledges, adding that some industries -- such as Big Pharma -- have historically been ahead of the MDM curve. "In most cases, companies have maybe narrowed down their lists of vendors, they're doing the program planning, they're getting the budgeting, they realize that the development is different, and they're gearing up development teams and [designing a] new MDM development methodology. We think we'll start to see a lot more adoption next year."
This time around, Dyché suggests, the timing is propitious, and although MDM probably won't ever be as demo-able as a highly interactive BI dashboard, its business case is altogether more obvious. Given the events of the last 24 months -- and viewed in the context of a future that may include increased regulatory oversight (which means compliance) and market consolidation (which means merger and/or acquisition activity) -- MDM is a technology proposition that sells itself.
"Making the ROI case for MDM in a lot of ways is more straightforward because if you do your business case right, the clarity for what MDM is going to buy you is a lot more obvious. With BI, it's very difficult to estimate [before the fact] the value of the [BI] investment. With MDM, [when you can show that] speeding up an acquisition for a company with strategic M&A goals by two months can save [that company] millions of dollars -- it's obvious. It can pay for the MDM project exponentially," she explains.
Many companies are already struggling under the burden of compliance requirements that rely -- to varying degrees -- on clean, consistent, and accurate data, Dyché points out. With policy changes on tap in the health-care, energy, and financial arenas, regulatory reporting is only going to get more onerous.
"Think about Pharma companies, [for which] a big issue is state reporting. They all have to report on how much money they spend marketing to every individual physician, and they have to do that on a state-by-state basis. If you multiply that by 50, and you take into account [the fact that] physicians can be licensed in different states, this becomes a big identity resolution problem," she suggests.
It can also become a big drag on the bottom line for non-compliant companies. "Just the compliance alone, the fines that these companies have to pay to federal regulators … are huge. In a lot of ways, situations like this make the ROI [case] for MDM much more straightforward," says Dyché.
So there's plenty of incentive to encourage MDM adoption at the C-level. What's interesting is that MDM is also making inroads at the data management (DM) level. DM pros, as a rule, tend to be suspicious of new technologies, especially when they impinge upon, disrupt, or threaten the orderly operation of their data assets.
This was especially true with MDM, which some proponents first positioned as a zero-sum replacement for (or refinement of) the enterprise data warehouse. It wasn't so much that MDM was going to replace data warehousing; it was rather that MDM threatened to swallow up (or subsume) the data warehouse, which is why DM pros felt threatened.
"The data warehouse wasn't designed to clean the data, to apply business rules to the data, or to link back to the operational systems," she explains. "All of the stuff that we define in our MDM stack is sort of outside the core data warehouse, so people saw [MDM as] this new platform that's going to swallow up the data warehouse."
The good news, Dyché asserts, is that this argument is no longer cast (or perceived) in zero-sum terms. MDM isn't going to replace the venerable enterprise data warehouse. Increasingly, in fact, it isn't so much perceived as a complementary but as an essential data integration service.
That's helped give it credibility among otherwise-hostile DM pros.
This is thanks, in part, to the diversity of MDM tools. In addition to best-of-breed players, for example, all of the big enterprise applications vendors (Oracle Corp. and SAP AG, along with midmarket applications specialist Microsoft Corp.) market or (in Redmond's case) plan to market MDM technologies. More than this, however, most of the big data integration players (IBM Corp., Informatica Corp., and SAP AG), along with many data quality vendors (such as DataFlux, a subsidiary of SAS Institute Inc.) have introduced MDM offerings.
Dyché sees this as both a good and a bad development. "The good news [for MDM] is that it's really starting to be institutionalized as part of the formal information management architecture. MDM has a permanent seat at the table," she argues.
"This also blurs the lines again: just as everybody was starting to distinguish the distinct value of MDM -- versus something like an operational data store, versus the data warehouse -- the lines will blur again as everything gets poured into that one data integration bucket."