Manage Learn to apply best practices and optimize your operations.

Data integrity trumps MDM in midmarket

Improving data integrity for better sales leads and business intelligence prep takes priority over master data management for now.

Midmarket CIOs may secretly yearn for the holy grail of the "golden ticket," but at most midsized companies, data quality -- not master data management (MDM) -- is a more realistic goal. The motivation: the need for clean and consistent data.

Master data management is a cross-company approach that links all critical data

to one master file that serves as a repository for the "golden record" -- i.e., a set of consistent, accurate, redundancy-free data that is sometimes dubbed "one version of the truth." On the enterprise level, an MDM project can take years, cost millions of dollars and involve a battalion of consultants to build master data hubs and supporting infrastructures.

Perhaps because of this complexity, MDM projects are far from ubiquitous, even at enterprise-level companies. A 2008 report from IT consultancy Aberdeen Group Inc. found that 26% of 1,075 companies with more than $5 billion in annual revenues are using or plan to use MDM, and 10% have plans to adopt MDM "in the future."

In the survey, 20% of companies with revenues between $50 million and $500 million are using or plan to use MDM, while 21% said they had no plans to adopt MDM.

But even the nonadopters are probably taking steps toward improving the quality of their data, according to Michael Dortch, senior analyst at Boston-based Aberdeen Group. Adopting tiered storage and building structured document repositories are two methods that companies are using to improve data quality, he added.

"At the midmarket level, MDM is typically a combination of high aspirations and base motivations," Dortch said. "Customers tell us they want all of their data integrated, and they want all of their customer data to be accurate and consistent -- that elusive one single truth. That's fine, but in reality their problems are more basic, like: 'We've got eight different copies of information about customer John Doe. Which one is right?'"

However, just because midmarket companies aren't undertaking full-blown MDM projects, doesn't mean they don't understand the underlying value of their data in relation to customer knowledge and increasing revenues, Dortch emphasized. Sometimes a data quality issue can be as simple as fixing and standardizing the way customer addresses are stored.

For Rajeev Kumar, CTO at Boston-based database marketer Intellidyn Corp., the data quality issue came down to control. In 2000, the company decided to take its data operations in-house. Intellidyn, a privately held company with yearly revenues of $5 million to $10 million, wanted more control over its data and a higher accuracy rate. At the time, the company had no IT infrastructure at all, so Kumar set out to build one.

Unlike other smaller companies, though, Intellidyn started out with the goal of using MDM. The company buys consumer data from credit bureaus, compiles and cleans it, and then sells it to banks, retailers, insurance companies and mortgage firms. "Data integrity is a business strategy for us," Kumar said. "In a sense, our business is master data management."

Each day, Intellidyn generates a billion match codes from 250 million records. To handle the data load, the company uses a combination of tools from Cary, N.C.-based data quality vendor DataFlux Corp. and its parent, SAS Institute Inc. The tools reconcile address and name differences and discard duplicate records. The clean data is then loaded into the SAS Business Intelligence software for in-depth analytics, Kumar said.

DataFlux CEO Tony Fisher acknowledged that MDM is far from ubiquitous at companies of any size. "We're not seeing any large-scale adoption of MDM in the enterprise right now. It'll probably be three to five years before that happens," Fisher said. And on the midmarket side, insufficient IT staff and a lack of business analysis talent pose significant stumbling blocks. "MDM is not a shrink-wrapped solution by any means, and it probably never will be."

Size doesn't matter: Complexity is a constant

National Instruments Corp. had both the IT talent and the resources to undertake MDM. For this Austin, Texas-based maker of measurement and test equipment, the issue of improving data quality and implementing MDM came up in 2000. The company does business in more than 40 countries and has yearly revenues of more than $740 million.

Problems with customer data integrity were the prime motivation, and duplication was the main culprit. Analysis of customer data was shaky, and sales leads were proving unreliable, according to Deepa Srinivasan, applications manager at National Instruments. Getting "buy-in" from the business groups in the company was fairly easy, she added. Keeping expectations reasonable in terms of "time to use" was the main problem.

At the midmarket level, MDM is typically a combination of high aspirations and base motivations.

Michael Dortch, senior analyst, Aberdeen Group Inc.

In 2006, the company deployed Initiate Systems Inc.'s customer data integration hub in tandem with its central Oracle database. Since beginning the project, the company has seen a 20% improvement in the accuracy of customer data, and the rate keeps improving, noted Christine McClary, customer data manager.

The need to make customer data more reliable and, hence, more useful for sales is also a motivation at Monster Cable Products Inc. in Brisbane, Calif. CIO Oded Haner is in the early planning stages of a proposed MDM project that will start with an analysis of data quality issues. The goal is to use the data for business intelligence purposes.

"I am preparing a plan to discover and correct data quality issues as well as monitor when they occur, and I hope to go through it before the end of the year," Haner wrote in an email. The project will require some skill building within the existing 30-person IT team, particularly in the area of customer management. A cross-functional team will then work with the sales analysis team to understand the reporting requirements, he added.

The decision whether to build or buy an MDM solution will be based on this groundwork, Haner said. His initial inclination is to build the MDM application in-house, he said, but that may change in the months to come.

Those same months may bring additional changes in the MDM space that could make both data quality improvements and entry-level MDM more accessible, according to Dortch.

By Christmas, he said, SaaS versions of some MDM products could be in the announcement stage -- which could make a nice holiday gift for some midmarket CIOs.

Let us know what you think about the story; email

Dig Deeper on Small-business infrastructure and operations

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.