Solve the Master Data Management Equation - CRM Magazine
John McLane stashed this in Big Data
The MDM Landscape
According to Gartner Research, global master data management software revenues will total $1.9 billion in 2012, a 21 percent increase from 2011. That number is expected to jump to roughly $3.2 billion by 2015. Much of this growth will likely come from three or four standout leaders.
In 2010, Informatica acquired Siperian for $130 million, which has continued to mold the company into "one of the leaders in the space," Wang says.
In May, Informatica MDM 9.5 was introduced to the market. The company says the solution will "open the door for organizations to engage in effective social e-commerce, take MDM to the cloud, enable the mobile workforce, and scale MDM for today's big data realities." The solution is available on-demand or on-premises, and offers an Informatica MDM iPad App, which allows users to grab customer transaction, social, or master data on the fly.
Naturally, Informatica is not alone. IBM has completed its fair share of MDM acquisitions (such as DWL in 2005 and Initiate Systems in 2010), positioning it as a strong competitor. DataFlux (a SAS company) is another master data contender. Plus, industry-specific solutions from companies such as Talend are sitting at the master data management table too. And, of course, one mustn't ignore industry newcomers, such as Pitney Bowes Software, which released MDM solutions in April.
Master data management solutions will have a pivotal year in 2012, according to The MDM Institute's "Master Data Management & Data Governance Strategic Planning Assumptions for 2012–2013." Vendors will need to increasingly support multidomain environments with reference management and better integrations with data governance and business process management (BPM) solutions. According to IT Business Edge, while BPM automates a company's business processes, by melding it with MDM for a complete view of its data, a company could, for instance, trigger an order or response when inventory falls below a set level.
"The BPM people link together workflow applications, but realize: 'What's the point of linking applications if you don't have a consistent view of the data?'" Zornes notes. "So you've seen BPM vendors like TIBCO and Progress Software add master data capabilities."
By the same token, Zornes says MDM solution providers such as IBM, Oracle, SAP, and Informatica are growing their BPM capabilities.
Although cloud-enabled MDM is in its early stages, there is room for growth in 2012 if the SMB market enters the picture; cloud-based MDM solutions offer a viable entry point with no long-term or expensive commitment.
In its "Next Generation Master Data Management" report, issued in April, TDWI says that next-generation MDM means nixing the so-called "Roach Motel" mentality—a scenario where data checks in, but never checks out. Single-domain MDM is inherently unidirectional. In other words, customer data flows from ERP or CRM systems into the database. If it comes out, it will usually just flow into a data warehouse or analytics mart. Unidirectional data works well for research and business intelligence purposes. However, it is not conducive to improving reference data, the report indicates.
By fostering a bidirectional, multidomain MDM system, the Roach Motel can be avoided. The goal of next-generation MDM is to create a real customer hub that allows data to flow freely and to be improved or fixed, and synchronized back to its original source, be it a BPM, CRM, or financial applications system. These efforts will undoubtedly further MDM's purpose—to create a single version of the truth.
Master Data Management is a tough industry for we lay-people to understand.
It's somewhere in between the databases we're familiar with and the big data science techniques that have been emerging in the past decade.
That's a really interesting point.
It seems we've struggled for decades with the disparate systems problem. There have been lots of point solutions, from 1:1 connections to federation and integration models. The traditional data warehouse has weaknesses as a 'roach motel', but it's a heck of a lot better than nothing. Think of the time and cost involved in building a DW, and the organizational change needed in most environments to keep it running clean and relevant.
My impression is that MDM is basically one more layer of abstraction over the spaghetti such that, finally, this time around, users really can get an integrated real-time view. Is this a correct description? If so, then isn't the technical challenge really one of systems integration, such that the MDM suite really is plugged into everything? If so I'm wondering where the big data science piece comes in. Maybe when an enterprise wants to go beyond the (previous holy grail!) of integrating structured data, to bring social/unstructured/big data into the picture?
Where I struggle is with the notion that any of the new approaches is going to solve the disparate systems problem once and for all. In most organizations, data is a mess and largely unmanaged. That problem is a result of history (M&A etc) and of underinvestment. The technical investments needed to get integrated are very high and have trouble competing for resources with other projects. The even bigger investment is usually a people change, fighting inertia and perhaps human nature. So the ROI of any such initiative may need to be pretty transformational in order to get the required leadership buy-in.
As long as there are legacy systems, there will be spaghetti abstractions.
Tibco has been peddling real-time integrated views for two decades.
Perhaps everything old is new again?