Analytic Applications: Packaged Decision Support Solutions Take Decision-Making to a New Level

by Heather HayesDecision support systems, once viewed as extremely complex and generally confined to sophisticated users, are now finding their way into the hands of the masses, thanks to a new offering known as the analytic application. The analytic application offers a new spin on the decision-making process by combining traditional decision support systems ? including extraction, transformation and load tools, data warehouses and data modeling, query and reporting tools ? with function-specific business rules and logic, and making this "packaged solution" available via Web interfaces and enterprise portals.The result of all these ingredients working together is the ability to reach a decision rapidly and without much fanfare."Analytic applications enable you to drill down to the heart of the matter very quickly, to get to the level that you need to make a decision, and then share that information across the board," said Ben Plummer, vice president of operations for Cognos Corp., a business intelligence firm in Ottawa. In contrast to the traditional process, which required decision-makers to get a report, find the problem and get another report, "the information is right there, and you can really navigate through it," he said.This new commodity represents a genuine quantum leap in thinking, according to industry observers. "What we're seeing here is really a shift from build to buy within the decision support arena," said Madan Sheina, a senior analyst in decision support for the Aberdeen Group, a research firm in Palo Alto, Calif. "If you build something from a classical data warehouse model, the process can be very long and drawn out and not particularly effective in terms of matching the requirements of your organization quickly enough. Analytic applications provide more immediate benefits."Of course, the need for tools that can help government organizations get a handle on its reams of data and turn information into usable knowledge has never been greater, due in large part to increased expectations from constituents and the continuing need to do more with less employees and less resources.Analytical applications address these issues by relying on "pre-built knowledge," or the business logic and lessons learned that come with years of working in certain fields, such as financial, program and customer management. With a packaged solution, organizations no longer have to perform the same analysis over and over, while continually going back into the data warehouse to ask increasingly specific questions. The analytic application contains the inherent knowledge required to dig deeper into the data and ask the appropriate questions automatically. "It really makes sense to take all that knowledge and build an actual product out of it as opposed to offering customers a mere tool set," said Steve Jones, solutions specialist for the federal government at Oracle Corp. in Redwood Shores, Calif. "With this, we're really giving them a real jump-start on the decision-making process."XXXSPLITXXX-Not surprisingly, analytic applications are a hot commodity. Research company International Data Corp. of Cambridge, Mass., predicts that the business intelligence tools market, of which analytic applications are a part, will grow from $3.1 billion in 2000 to $12.7 billion in 2004, a 33.6 percent annual growth rate. Moreover, analytic applications probably are the strongest growth area in business intelligence right now, according to Keith Gile, senior industry analyst for the Giga Information Group, an e-business consulting firm in Cambridge, Mass."Everybody, if they don't have one already, is scrambling to either buy one or have one built," Gile said.The growing demand has spurred a number of players, both traditional firms providing database systems and decision support tools, as well as start-ups, to jump into the fray. These include Oracle, Informix Corp., Cognos Corp., MicroStrategy Inc., Informatica Corp. and Business Objects S.A.A number of factors are driving the whole trend, including the emergence of the Web, increased customer and constituent expectations and the growth of the packaged applications industry, including companies such as PeopleSoft Inc. and SAP AG."Because of the numbers of organizations using such products, it's enabling companies like us to build products that combine data from those different systems for analytic purposes," said David Lyle, vice president of product development within the applications business unit for Informatica in Palo Alto, Calif. Creating the business analytics required to run an organization is a very complicated problem, Lyle said."Organizations recognize that it is much less risky and more cost-effective to buy a solution that is supported by experts than it is to build it themselves," he said.That includes a number of federal agencies that are in various stages of implementation. Among these are the Naval Undersea Warfare Center, which has streamlined its financial management processes using an analytic application suite from Cognos, and the U.S. Postal Service, which is using MicroStrategy 6 to track stamp usage at different locations to better predict and remedy inventory shortages and surpluses.Meanwhile, the Department of Housing and Urban Development is working to implement an analytic application that can keep track of its various services, most significantly its Section 8 contracts that provide low-income rental housing to citizens. Like other agencies, HUD had to deal with a number of stovepipe systems and could not gain an integrated view of its entire environment.The disadvantages of such a system were palpable. Local housing authorities, for example, were hard-pressed to comply with an agency tasking to examine the gap between their needs and available resources, said Richard Burk, HUD associate deputy assistant secretary of operations."Trying to come up with what was already being provided is very difficult if you only have information on half the picture," Burk said.When in late August HUD unveils its Empowerment Information System, built on a MicroStrategy 6 platform, the agency expects that every employee, and eventually its business partners, will have access to the data that crosses over the various parochial interests. For example, a public housing authority with a HUD drug elimination grant could go to the Empowerment Information System data warehouse to find out what other drug elimination activities are going on in that particular neighborhood. "I would never know that just by looking at the public housing database," Burk said. "I will have that with the EIS data warehouse, which will finally bring all of HUD's information into one place."The demand among federal agencies right now is pretty intense, according to Tricia Reneau Iveson, industry solutions manager for government for MicroStrategy. Functional areas that are particularly strong include financial management, fraud detection and enforcement, logistics and service to the citizen. "The bottom line that everyone is really recognizing is that you can't do e-government without a data warehouse," she said. "It's the foundation you need to truly be able to meet the increased performance expectations of your constituents."XXXSPLITXXX-Analytic applications are expected to grow at the rate of 34 percent annually during the next three years, making it the fastest growing segment of the business intelligence market, according to Madan Sheina, senior analyst in decision support for the Aberdeen Group, a research firm in Palo Alto, Calif. He noted much of this industry growth comes not just from new players in the field, but also from cannibalization, whereby existing decision support vendors re-architect their products and strategies to provide solutions rather than tools.XXXSPLITXXX-Despite its benefits, analytic applications are not automatic, out-of-the-box cure-alls for organizations looking to improve their decision-making capability."There are a number of trade-offs that organizations face when they choose an analytic application over a more traditional build process," Gile said. "There are really no standards at present, so vendors in particular tend to define this field very loosely. "Just because someone says it is an analytic application doesn't necessarily mean that it is one," he said, "so you have to really look deep to find out if it is truly an end-to-end solution."Iveson said an analytic application capable of providing a solid e-government platform requires five components: ? a central data warehouse where the data resides;? an analytical engine that allows users to dig into the data and look for granular answers to their questions;? a personalization engine that can send the most relevant information to each individual user;? a broadcast engine able to deliver the information wherever it's needed;? and an interaction piece that allows employees or constituents to act on the information they've received.But not all products are alike, and Gile advised buyers to beware. Some products, for example, do not have all the sourcing defined in mapping libraries or a wide breadth of finished product analysis templates, and in such cases organizations could end up building half of the project themselves. Even with top-of-the-line products, though, there exists a strong need for customization and, as a result, integrators have a substantive role to play in tuning the implementation. "An analytic application pretty much fits the 80/20 rule, where it possesses 80 percent of the functionality that's required for a specific task, with the remaining 20 percent embedded and customized by a third party," Sheina said.Integrators not only must have a strong knowledge of relational databases and data warehouses, they need a specific domain knowledge, such as logistics or financial management, said Bill Mickler, director of the communications systems division for CACI International Inc., a systems integrator in Arlington, Va., that has been working with Cognos to implement its analytic application suite within agencies."The typical end user is too busy to do the development, so there is a little bit of setup where an integrator can really add value, including creating the metadata and adding the client's business perspective and business rules to the business intelligence software," he said.The availability of packaged products likely will cannibalize some of the integrator's previous design and build work load, Jones said. "It's a bit of a balancing act. There will be more projects going on, but fewer hours on each job," he said. "The flip side is that you'll get the projects up and running faster, so there's less risk of major complications, which of course will make customers happier and quick to send more projects your way."