Business is confusing. Operating in a global market, where acquisitions and a high mix of suppliers are the status quo, can create an ocean of information that can throw even the best-managed company into a tailspin.
Take Airgas Inc., a Pennsylvania-based distributor of industrial, medical, and specialty gases that has completed over 300 acquisitions and operates through 13 different subsidiaries around the world. As the company built out its business, it needed a way to manage the influx of new information coursing through its IT infrastructure. Each business unit, for instance, had its own way of formatting and describing products and part numbers. That meant that when customers — which range from local gas stations to Fortune 100 pharmaceutical companies — tried to order the same product in a different part of the world, they encountered inconsistent descriptions and varying part numbers.
"We have all sorts of data sets out there, and we needed a single repository that would allow us to appear as if we are one company to the customer," says Steve Max, director of e-business marketing at Airgas.
More Than Just Location
The kind of repository Max alludes to often starts with a database. But companies have learned that simply throwing data into a single system doesn't solve the more pressing problem of reconciling that information so that there is a single version of the truth. In light of that, companies are now striving to create logical links among the disparate data that populates not just a database, but a variety of applications across the enterprise, and sometimes even down to the plant level. The concept has evolved into what industry observers call master data management (MDM). But what is MDM? Is it a database? Middleware? Integration technology? The answer is that it's all of that, and much more.
According to the Data Warehousing Institute (TDWI), MDM is the practice of acquiring, improving, and sharing master data — which is the reference data that defines a business entity. In an MDM survey that TDWI conducted via the Internet last year, 83% of the 741 respondents revealed that their organization has suffered as a result of having poor master data. The problems range from inaccurate data to internal arguments over which data is relevant to bad decisions based on incorrect data. Only about 20% of these companies have invested in a stand-alone MDM solution, but those that have started down the path are starting to see tangible results.
"MDM is more than just technology. It is about the entire process around how you manage your corporate data," says Sunil Gupta, director of solution marketing at SAP. "The data is really the crown jewels of your business, and it covers areas such as data strategy — how you manage it and how you define it conceptually."
To achieve a harmonized view of data, MDM requires that a governance strategy be built into the information flow so that a product SKU in one system is represented the same way in a separate system that might otherwise have listed it differently. Or, that a customer can be identified between the CRM system and the accounting system regardless of whether he is entered as "Joe Smith" or "Joseph Smith."
For help executing on an MDM initiative, companies can turn to their ERP companies, like SAP or Oracle Corp. There are also integration and data management specialists that manufacturers can team up with, including DataFlux, NetManage Inc., and Informatica Corp., to name a few. And emerging data management services from companies like ByteManagers Inc., which Airgas is using to help make sense of its existing data, can offload the burden of fixing flawed data.
"We knew that the problems were with the data, but we didn't know how to fix it and we didn't have the manpower to do it," Max says.
Indeed, reconciling customer data, product data, and supplier data can help manufacturers operate more efficiently. But there's an ocean of information out there in a variety of systems. So where do you start and how do you determine the ROI?
Data As an Asset
MDM is based on the concept of data governance. Viewed from that angle, it can be an overwhelming idea. "But you need to get started. Just don't boil the ocean," says Rob Stowell, vice president of operations at ByteManagers.
Of course, figuring out where to start can be difficult. But it might just be a matter of looking to your ERP partner. Oracle, for example, created the Adaptive Business Solutions Group to build an integration framework common to all of the applications it has acquired over the years. That framework underlies the company's Fusion middleware, and on top of that middleware layer, the company has designed an MDM application that consolidates data, cleans it, and distributes it as a single point of truth to all Oracle applications, officials say.
Oracle built the MDM application because middleware on its own can't solve the problem of data inconsistency. "The problem with middleware is that it can hook multiple applications together but it doesn't answer the question: 'Where do I go to find the trusted source for customer data?'" says Pascal Laik, vice president of Oracle's master data management strategy.
Oracle's MDM solution includes policy hubs for managing the information as well as tools such as Oracle Warehouse Builder and Oracle Data Integrator.
Similarly, SAP embeds master data within its NetWeaver platform, but there is an optional program a customer can buy called SAP NetWeaver Master Data Management (MDM), which provides a way to pull data from various structures, including an XML feed, a database, or a flat file. The SAP approach starts with a definition of data structure. It then cleans the data (a de-duplication process) and populates the correct data via Web services to the applications that need it. This is what SAP calls a syndication capability.
ByteManagers' approach to MDM is to build a process around data organization. The vendor analyzes, structures, and builds out schemas and a taxonomy. Ultimately, the process can change the way a business operates. "We are not out to reengineer your business, but the data and information can help reshape a culture organically once people realize the benefit to them," Stowell says.
ByteManagers is working with Airgas to improve the content of its data so that it is more manageable. The larger goal, though, is to create data that is identical when viewed from any access point in the organization. That way, when a customer deals with any of the Airgas operations, there is a cohesive look and feel.
Cleansed and Ready for Action
Ensuring quality data is an important aspect of the MDM strategy. Data can be pulled into a repository, compared, and distributed again, but if it's not the right information, it is building more waste into the system. To address that part of the challenge, SAP and Oracle call in reinforcements in the form of ByteManagers, DataFlux, or Informatica.
"We are riding the coattails of SAP, Oracle, and IBM... These guys don't want to dig into the dirty parts of the data," says David Harrelson, ByteManagers' vice president of sales and marketing.
In a perfect world, quality would be built in. Unfortunately, there is a lot of mismatched information floating around, and that can make it hard to conduct business. DataFlux, for example, applies business rules inside an MDM hub to reconcile information among different applications.
Informatica does something similar, storing and applying rules to metadata. "There is so much information coming into [an organization] every day and you need a way to cleanse it," says Karen Hsu, Informatica's principal product marketing manager. "Quality is tied to MDM... We supply the rules around the tools to enable the business user to manage the data." Having the rules is much different from just having a single central repository, she notes.
The biggest benefit of MDM is operational efficiency, industry observers say. But it's sometimes hard to measure the true ROI of an MDM implementation. Some reports begin to paint a picture of returns, including an independent study by GE's Consumer and Industrial business that estimated a $3.9 million savings within its organization as a result of applying MDM processes and a Six Sigma methodology.
Hard numbers aside, a growing number of companies, Airgas included, are realizing that the benefits of MDM are measured more by the sense of security that comes from knowing they are building a business around good clean data.