Data warehouses deliver: cook up some cool business intelligence by slicing and dicing your data.

WAYNE ECKERSON ESTIMATES THAT DATA WAREHOUSING represents approximately half of the activity in the computer industry. Because Eckerson is the director of research for The Data Warehousing Institute, Seattle, it's not surprising that he holds that opinion. However, because he applies the term data warehousing to all the tools that make it possible to analyze data in a way that empowers better business decisions, he probably has a point. Half of the tools capture information and the other half analyze it, creating the business intelligence that organizations need to identify business opportunities.

[ILLUSTRATION OMITTED]

"Typically," says Eckerson, "organizations buy [computer] systems and applications that each have their own databases, creating a basically fragmented view of the company from a data perspective. It's like looking through a kaleidoscope trying to figure out what your company is doing; all the pieces are there, but they don't fit together." Data warehousing--synonymous with business intelligence, business analysis, and other terms--is a common way to put the data together so that you can analyze it to identify trends, understand customers, and identify opportunities. In his report "Smart Companies in the 21st Century: The Secrets of Creating Successful Business Intelligence Solutions" (2003, 101communications, LLC), one in The Data Warehousing Institute Report Series, Eckerson asserts that in today's business climate it is the smart company that will gain a competitive advantage. "The secret weapon that these smart companies wield is information--more specifically, highly integrated information that empowers workers with new insights about what drives the business and how to optimize business processes to better meet strategic goals and objectives."

Put simply, says Ted Udelson, president, Integrity Computing, Washington, D.C., "Integrated data reporting reveals patterns that guide an association to ask the right questions. Once you ask the right questions, you can begin to make some key decisions."

In consulting with associations, Alan Browning, vice president of the association consulting group Management Advisors, Inc., Greenwood Village, Colorado, finds that the concept of data warehousing isn't broadly embraced or understood yet. And that may be in part due to the nomenclature. "I do find, however," says Browning, "that data warehousing is quite effective for organizations that are wrestling with multiple systems that are not integrated and find that some important data does not naturally flow through their association management systems." After all, he says, AMS databases are really optimized for transaction processing, not data analysis. "Several AMS vendors are starting to address this issue," says Browning, "and some, such as iMIS and Gomembers, are developing modules based on the data warehousing approach." Until association management systems become better tools for data analysis, association executives may want to consider data warehousing as a means to gather business intelligence. Several organizations are already using the concept.

Evolving intelligence for understanding members

With 19,000 members representing 11,000 health care organizations, the Medical Group Management Association, Englewood, Colorado, has had its data warehouse up and running for five years. Both a trade group and a professional organization, MGMA represents medical groups and the medical practice administrators who run the business side of the practice. Gary C. Fox, MGMA's senior vice president and chief information officer, explains, "The initial intent [of the warehouse] was to improve understanding of customer data--purchase trends and their profiles--as well as to help us understand the profiles of the group practices we represent." For years, says Fox, "we got a variety of analysis-type questions from staff about our members: 'What is the value of a customer? Who are our most profitable customers and members? What are the profiles of our new customers versus those that are leaving?'"

Typical of many associations, MGMA's association management system was not designed to answer those kinds of questions. The idea, says Fox, "was to extract all the data from our AMS into the warehouse and organize it in a way that would start to answer some of these analysis-type questions." These days most of MGMA's marketing efforts, campaign tracking, and so forth occur through the warehouse. "It's used in lots of ways that we didn't envision initially," says Fox. "The bottom line for me is that the warehouse is equally important as our AMS. I can't envision our organization running its operations without it."

Getting started. At the time that MGMA embarked on its business analysis quest, no AMS products were in the data warehousing space. "There were some data mining alternatives," recalls Fox, "but these were more of a one-time snapshot that did not help with our ongoing analysis." So MGMA turned to data warehouse consultants who helped the association's technical staff construct a data warehouse that included "a fundamental design that was tailored to meet our needs," he says.

According to Fox, the data warehouse team of six staff was careful to design a database to support efficient data queries. "You want your results to come back in a few seconds or less," he explains. "So you need to reorganize your data in the warehouse around the types of questions you want to answer." At the start, MGMA's goal was to analyze its information to better understand its customers and to help with its marketing and product development efforts.

Getting your money's worth. True to Eckerson's smart companies report, in which survey results found that a large portion of executives did not insist on rigorous cost-justification for their business intelligence projects because the projects were considered so strategic, MGMA did not do a formal analysis of its return on investment. "It was pretty apparent to everyone that we needed this type of tool," says Fox. Regardless, the investment was significant. "Aside from paying the design consultants and purchasing the hardware, the direct costs were minimal. However, allocating staff time for three full-time staff members for the better part of a year was a significant investment." Nonetheless, says Fox, "the investment has more than paid for itself."

Getting results. MGMA has reaped several rewards from its initiative. For example, by analyzing member data, it became clear that people are more likely to join the association after attending one of its meetings. "It used to be that if a nonmember would register for a meeting we sent out membership material right away," says Fox. Staff now knows that nonmembers are more likely to react to materials once they've had some experience with the association. As a result, MGMA now sends material a few weeks after the meeting--and has gotten much better results.

This experience led the organization to establish other triggers related to events. Says Fox, "The warchouse can flag significant events, which automatically trigger certain actions--send a mailing, make a contact, or initiate some other response."

An interesting unintended benefit has been the reduction in the side databases kept by staff who were reluctant to incorporate certain data into the association management system. Fox explains, "We found that because the information from the side databases didn't flow into the warehouse--and these folks saw some of the benefits of having the data in there--they had more of an incentive to use the main database."

And if at some point MGMA decides to move to a different AMS, the data warehouse will remain so that all of its reporting and business processes stay in tact. Having the data warehouse in place, Fox says, "takes some of the pressure off of moving to a new system."

Accomplishing the mission one step at a time

The overall mission for the State of Iowa Department of Revenue (IDR), Des Moines, says Rhonda Kirkpatrick, executive officer, "is all about the promotion of voluntary tax compliance." In a typical year, the department processes more than 4 million documents as part of its tax-collection function, and in 2003 the state collected in excess of $5.9 billion and issued $672 million in refunds.

In 1998, the various areas of the compliance division came together to figure out how better to identify individuals and businesses that were not compliant with Iowa tax law. "We identified a number of databases--not only the ones in our division but also some from external data sources--that were stand-alone systems," recalls Kirkpatrick. The business challenge: Take all of those data sources and put them together so that staff could match different line items, names, and addresses to identify the nonfilers. "We have IT resources," says Kirkpatrick, "but we really did not have the skills or technology to support an advanced effort of bringing multiple sources together into one database and then programming things from there."

Tracking down the tools. As a government agency, IDR was required to go through a competitive bidding process. The compliance group initially issued a request for information to about 30 vendors. "We basically said in the request, here's our business challenge; please provide us with a solution. We wanted to be open-minded, looking at all kinds of solutions to determine which would work best," Kirkpatrick says. While most vendors came back with proposals other than data warehousing, many did propose combining two, three, or four data sources.

"We quickly learned," she says, "that building a data warehouse is not simply plugging in a box that is ready to go. It takes a lot of internal resources--people who know your data and those who understand the external source data--and a lot of up-front design." Kirkpatrick was part of a six-person internal project resource team (five business process people and one person from IDR's IT shop) that assisted and supported the data warehousing consultants in identifying what data they needed out of which sources, how they would use that data, what design process was necessary, and what actual technical programming effort was required.

Taking a financial risk. Though the funding mechanism for IDR's data warehousing project is not one that the typical association can use, it does provide an interesting example of sharing risks with supplier partners. Basically, due to legislative restrictions, the only money that IDR could use to pay its data warehousing consultants was the additional revenue generated by using the data warehouse to identify nonfilers and get them to file their taxes.

Fortunately for IDR's vendor, the new taxes collected have more than paid the bill, which was $11.5 million. Kirkpatrick is quick to point out, however, that this price tag included much more than the data warehouse project. "We realized that the additional workloads for our staff to process a higher number of tax documents would require increased efficiency," she says. Consequently, the contract included related projects to achieve automation and other business efficiencies. Aside from these additional investments, Kirkpatrick notes, "when you talk about how much a data warehouse costs, there are many variables--size, disk space, and so forth. I'm hesitant to put an exact price on the project, but I also think that ours is not an unusual situation with the project including related initiatives to achieve other business efficiencies."

Tightening up compliance. According to Kirkpatrick, the project has been a great success in promoting tax compliance, with the state collecting, through January 2004, $35 million in revenue associated specifically with the data warehouse project. (The first dollar of that total was recognized in April 2000.) An added bonus: The data warehouse has identified additional areas of opportunity.

"We felt that our [initial] ROI was higher by identifying those who should be filing and were not. But we also know that many taxpayers do not file correctly," says Kirkpatrick. The department is now analyzing data to find out why people are not getting it right. "By providing more education and being more proactive in achieving compliance, we can get closer to our ultimate goal of consistent compliance," she says.

Outsourcing options

Don't have millions, or even thousands, of dollars to spend on a data warehousing project? Don't worry. Associations can, of course, outsource data warehousing just as they can outsource myriad other technology applications. (Go to The Data Warehousing Institute's Web site at www.dw-institute.com for a list of industry consultants.) Marsh Affinity Group Services, New York City, provides one example of a supplier reaching out to the association community with data warehousing services. A provider of affinity insurance programs to associations, the company uses a data warehouse to allow its clients to create targeted marketing programs for their members. Don Neal, chief marketing officer, Marsh Consumer Practice, explains: Our data warehousing capability allows clients to load their information--to the extent to which they are comfortable. Marsh then cleans up the data and creates a segmentation strategy that allows associations to more effectively market to their members.

"About five years ago," Neal recalls, "we recognized the need to develop our own data warehousing and marketing technology infrastructure to be able to support our association clients." This was due in large part, he says, to the fact that many of the company's association clients have numerous legacy database systems. Often information about members is stored in a different database than the one that tracks the products and services members buy. "We saw an opportunity to build a sort of institutional memory for associations to allow them to market to their members with more precision," says Neal.

As both the Medical Group Management Association and Iowa Department of Revenue did, Marsh evaluated the top-priority business requirements for a data warehouse that would analyze member data--in its case, specifically for a client's marketing purposes. Looking at all of its association clients, says Neal, "two priorities rose to the top: ways to increase member satisfaction through association-brand relationships and ways to sell more products and services to augment nondues revenue streams." Achieving those requirements meant coming up with a solution that served all of Marsh's constituents: insurance carriers, clients, internal partners, and consumers.

Matching those needs with technology required the company to evaluate its own technology platforms in light of external platforms, consider what other service providers were offering, and agree on a data warehouse technology platform that could achieve customized marketing. "What we developed is kind of a combination of [outside technologies and] homegrown and internally developed technology to meet the specific needs of associations," says Neal.

By taking advantage of Marsh's data warehousing system (in some cases for a fee and in others a complimentary service, depending on the nature of the relationship), association clients are now able to do the kind of data analysis that helps them provide more of a value proposition to members, says Neal. "They're not sending information on homeowners insurance to people who don't own homes; they're not trying to market bank cards to members who already have the association-sponsored card," he says. Rather, the association can develop individually relevant marketing communication that has a much better chance of attracting people to buy products or join the association. Neal, who has a background in customer relationship management, reports that associations using these techniques often experience higher response rates, particularly in marketing their affinity insurance programs, and see the program as a way of building their brand.

Would-be data warehousers beware 

IDR's Kirkpatrick warns: "There's a misconception that a data warehouse is delivered in a box and you just plug it in; that is not the case." Tackling a data warehouse project takes a huge organizational commitment from top to bottom, because it has an impact on lots of people in the organization. "For the project to be successful, you must have leadership support, total team development and teamwork, and good communication. Not only that, the vendor cannot come in and do it alone; it has to be a true partnership," she says.

MGMA's Fox offers a couple of additional caveats. "Design the big picture, and then take it slowly, populating the data warehouse one piece at a time," he says. "It can be an intensive undertaking to implement all the data extracts and downloads at once. I'd advise easing into it. You will see more immediate success that way."

To help with cost justification for a data warehouse project, Fox suggests going into the project with the idea that the data warehouse could eventually become the centralized reporting system for your organization. Reengineering some of your business processes around the data warehouse will also add value.

Finally, he recommends building solid training into the process. "While the end-user tools make use of the data warehouse fairly intuitive, understanding the data itself is another matter." Part of the learning curve is training staff to understand the nuances of your data. For instance, he says, "if you are selecting individuals who have purchased a specific product, you must make sure that you are also looking at the returns on those products. You can make mistakes if you really don't understand the data."

Fox believes that all the effort is worth it. "The data warehouse," he concludes, "is giving MGMA better information about the trends, patterns, and profiles of our members--and that really empowers staff."

Carole Schweitzer is executive editor of ASSOCIATION MANAGEMENT. E-mail: [e-mail].