You could be forgiven for thinking data management software vendors are bored with master data management (MDM)....
MDM isn’t the sexiest concept for their marketing departments, and following a flurry of activity related to MDM tools in the latter half of the last decade, most of the product hype these days seems to have moved on to topics such as “big data” and data virtualisation.
In addition, MDM remains a fuzzy notion within many organisations, and that’s likely contributing to its reduced public profile.
“There’s a lot of misunderstanding as to what MDM is all about,” said Clive Longbottom, founder of UK-based analyst firm Quocirca. “Most people look at MDM and think it’s just about cleaning up their customer records so they have one set of referential data. It’s not. It’s about finding out what matters to the business.”
More on MDM tools and technology
MDM technologies thrown into relief by Informatica and IBM acquisitions
Learn how to develop an MDM strategy in seven steps
Take a look at Gartner’s view of MDM and cloud computing
For example, Longbottom added, customer data might not be the key master record for a company. “It might be the things you’re selling, your suppliers or most likely a combination of things -- in which case you need two or more sets of referential data,” he said.
But Aaron Zornes, chief research officer at the MDM Institute consultancy in the US, said, “MDM is very good at helping businesses do things that have traditionally been very difficult to achieve.” That includes cross-selling among different product lines and business units, gaining a single view of customers, patients and other entities, and dealing with regulatory compliance issues, Zornes said.
He also believes it is a more helpful concept than big data. “I rise in anger whenever I hear the term. It’s not like someone pushed a button and suddenly we’re in the era of big data. Data has been getting bigger every year,” he said.
MDM’s tentacles are today reaching out to embrace other key concepts like business process management, data integration, data quality, data stewardship and -- of course -- big data. And while it may have slipped under the radar somewhat, MDM is still an area in which there’s plenty of technological innovation under way.
Onward, upward for MDM tools
Zornes’ view is that there have been all sorts of advances in the technologies underlying MDM, such as identity resolution, which uses algorithms to help users understand things such as whether individuals are who they claim to be -- an essential capability for risk management and fraud avoidance. “A lot of the newer algorithms are very good at sussing out who’s who through indirect association -- for example, through a common phone number or by knowing who’s married to whom,” he said.
Through a series of acquisitions, IBM has gathered capabilities in identity resolution, and Informatica and Tibco have made it a priority as well, according to Zornes.
Data matching for identifying and consolidating differing versions of product data is another MDM category that has seen considerable investments by vendors, as highlighted by Oracle’s January 2010 acquisition of product data quality software vendor Silver Creek Systems.
But one of the biggest areas of MDM innovation is around understanding the mass of unstructured data being generated on the Web. Zornes, again: “Companies think that because they have bought, say, customer data from Dun & Bradstreet, that they have a 360-degree view of the customer. In fact, it’s more like 75 degrees. If you really want a 360-degree view, you need to sort through all that information out there on the Web -- Facebook, Twitter, LinkedIn and so on.”
MDM tools can help integrate text-based data from such sources with traditional structured data through a unified set of master data. And that can pay business dividends. “Being able to pull in information such as a customer’s hobbies, beliefs [and] what football team they support is extremely valuable for salespeople who want to get closer to their customers,” Zornes said.
In addition, MDM systems are now going in-memory to offer dynamic data analysis capabilities in real time, according to Zornes. SAP is putting its MDM technology on top of its high-performance HANA in-memory processing engine, for example. “We’ll soon see real-time MDM becoming not an expensive solution but the norm,” he said.
Another big shift is Microsoft’s planned integration of free MDM tools in the next version of Office, codenamed Office 15, which Zornes thinks will open up the MDM market to more small and medium-sized enterprises.
Volvo: Driving for MDM innovation
Having been working on in-house MDM since 2008, automaker Volvo is currently looking at products that could meet its future MDM needs. Belgium-based programme manager Christoph Balduck said that while the company sees a lot of potential value in the megavendors in terms of ensuring data quality, he thinks some of the more interesting innovations are happening among the niche players.
“For example, we’re looking at Talend, which has an open source offering that’s particularly good at data migration,” Balduck said. He also is impressed with what he has seen from MDM vendor Orchestra Networks: “They’re looking at semantic modeling, which I believe is the future for MDM. It’s also one of the few vendors with a cloud MDM solution.”
Orchestra’s DataSpace version-control feature, which lets users simulate the effect that changes to master data will have on live systems, is particularly interesting to Balduck. “You can see not only how it will affect your [data] model, contents and attributes but also how it affects your workflow, business rules and so on -- in multiple areas,” he said. “If the simulation doesn’t throw up any problems, you can then say the change is OK to go live. That ability to be as flexible as possible could bring a lot of value.”
But more MDM innovation is required: Zornes said one area in which vendors have been slow to respond to market need is in introducing MDM technologies that can help with data governance and data stewardship efforts. A number of smaller vendors are playing up their data governance credentials, but he said none of them is really there yet.
And there remain problems when it comes to integrating all the different sets of master data that have built up within organizations. “There’s currently no answer other than to buy yet another hub,” Zornes said. “Vendors are now talking about ‘uber hubs,’ but at the moment no one has the capabilities to integrate all the application packages at the level that the business wants.”