Decision Support Systems

Putting Data to Work<@VM>Real-Time Knowledge

What Are Decision Support Systems?

Decision support systems are solutions comprised of extraction, transformation, load tools, data warehouses and data modeling, query and reporting tools. They allow managers and executives to pull together large volumes of data, pose questions to the data and use cognitive reasoning to turn that data into usable information. They are also known as business intelligence systems or analytics.

As federal agencies try to wrap their arms around the myriad data generated by government information systems, many are turning to decision support systems that gather data from across the enterprise to give a holistic view of an organization.

"For agencies today, it's all about getting that one version of the truth," said Joe Russell, business intelligence practice manager for Northrop Grumman Corp.'s Information Technology Sector in Herndon, Va.

Getting an enterprisewide view has become increasingly important to agencies seeking to understand and use information. In recent years, this has been driven by a desire to improve online customer service and satisfy congressional demands for better evaluations of government programs.

But to build the necessary information systems, agencies have to integrate back-end systems and connect heretofore disconnected pockets of data.

"By putting the focus on integration, what you're doing is eliminating these islands of automation or these multiple versions of the truth, which are not only inefficient, they're disruptive," Russell said.

Getting that one view of the truth requires integrating disparate databases residing on everything from legacy mainframes to modern servers, pulling that data into a single data warehouse environment and applying analytic tools.

Such an effort promises massive opportunities for systems integrators who possess a dual knowledge of business processes and legacy systems.

Unisys Corp., Blue Bell, Pa., has identified more than $500 million worth of opportunities over the next 18 months, including requests for proposals put out by the Agriculture Department and the Air Force. Both agencies are looking for integrators to build data warehouse interfaces with existing legacy databases as well as with Web portals, and to develop overriding decision support
systems.

Some agencies are already ahead of the curve. The U.S. Geological Survey's Water Resources Division, for example, recently unveiled the National Water-Quality Assessment Program. It uses Informatica Corp.'s PowerMart data integration tool to pull information from separate databases devoted to, among other things, water surface data, water quality data and groundwater; and Oracle's database ad hoc query and analysis tool to get an overall picture of water resource conditions.


Geoff Stilley


Previously, the organization had to pull the information and manually compile it in one location. Now, they can do the task in minutes and get accurate data.

"Despite all the investments in modern systems, precious few decision-makers have good data in their hands on which they can make sound decisions," said Geoff Stilley, director of sales and marketing for Informatica, Palo Alto, Calif. "That's because 95 percent of federal databases still can't talk to each other."

That is one of a number of factors driving the new trend. Government organizations also want to better understand their customers, who are demanding increasingly higher levels of service and information on government Web sites. And there is a continued push throughout government to operate more efficiently.

"Pulling to-gether all relevant data and analyzing it as enterprisewide data is the only way agencies are going to get a total picture of what they've got and where they stand," said Lee Cooper, vice president and general manager of sales for Unisys Federal Systems. "An added reward is they can also reduce their cost of operation."

The growing use of customer relationship management (CRM) systems in the federal arena is also pushing agencies to adopt a holistic approach to IT systems.

Keith Gile, senior industry analyst for the Giga Information Group, an e-business consulting firm in Cambridge, Mass., said that organizations initially were implementing customer tracking programs, call-center applications and other varieties of popular CRM software systems without integrating them with back-end databases and decision support tools.

"Now, both organizations and the vendors selling CRM solutions are saying in order for this thing to really work, there has to be integration, and there has to be data warehousing," Gile said. "And there absolutely must be data warehousing, because data warehousing normalizes, it dimensionalizes all of that operational data, and CRM is, to a great extent, operational.

"Every transaction about a customer is operational, so you have to analyze it, transform it, manipulate it, clean it, get it into a place where it can be analyzed," Gile said. "That calls for integration work."


Lee Cooper


Still, having the opportunity to develop these enterprisewide decision support systems and being successful at it are two different things. Systems integrators who will excel in this market require several characteristics that could be considered incongruous. Personnel must have an IT focus and a solid understanding of federal legacy systems and decision support systems, including newer add-ons such as analytic applications. But they have to understand business processes and operations and possess program management skills.

"You've got to be so well-rounded," Cooper said. "A lot of integrators try to drop in an [off-the-shelf] product
and, no matter how whiz-bang the product, they ultimately fail because they don't have any real expertise in legacy
systems."

Such a project boasts plenty of challenges, said Guy Creese, research director of Internet analytics for the Aberdeen Group, a research firm in Palo Alto, Calif. Challenges include locating relevant data in disparate databases and then defining it. On the analysis side, the toughest part is figuring out the most salient items.

"If you get kind of consolidation-happy, you'll end up with tons of data, which can be just as bad for making good predictions and decisions as not having enough data," Creese said. "So, knowing which data to pay attention to and which data to ignore takes quite a bit of business savvy."

Russell's firm is providing integration and building decision support systems for agencies such as the Defense Logistics Agency and the Air Force. He said integrators need to be aware that the trend of applying decision support tools to corporate data is part of a larger theme in today's world: the growing alignment of business and technology. As such, success with such an effort requires an incremental approach.

"The Big Bang theory doesn't work here," he said. "The nature of business intelligence is that you don't know all of your requirements upfront. So you've got to go slowly, because even as you get involved, you're learning about the business."The goal of information technology has always been to use silicon and software to try to replicate the human brain. Nowhere has that been more true than in the decision support realm, which is all about cognitive reasoning, or learning from the past to improve current and future processes.

Unfortunately, the effort has fallen short of its goal. The single largest flaw in the value of decision support up to now is its failure to act in real time.

When faced with dilemmas, people are able to think cognitively and intuitively and use decision-making skills, said Jim Martin, director of business development and alliances for public-sector business at i2 Technologies in Washington. By contrast, he said, "if a batter relied on traditional paradigms for computer-based decision support, he would not swing at a pitch until the floodlights went out, and the crowd had long gone home."

Such a slow response wouldn't work in baseball, nor in today's online world, said Guy Creese, research director of Internet analytics for the Aberdeen Group. Organizations, now forced to work at Internet speed, "need to react to customers and suppliers at the time of interaction," he said. "That calls for a completely different model. Instead of just reporting your findings in a report and delivering it the next day, they need to be able to extract the data and put it directly into operations."

More recently, the race to develop real-time decision support applications, as well as real-time data warehousing, has quickened significantly. Vendors such as i2 Technologies and Oracle are pioneering new models for real-time decision support applications, infrastructures and solutions.

And business intelligence vendors are seeking out partnerships with companies that provide message brokering systems. Business Objects SA and Crystal Decisions have recently partnered with IBM Corp., Armonk, N.Y., in hopes of eventually linking their decision data warehouses and analytic tools with the MQSeries product. Actuate Corp., San Francisco, has teamed with BEA Systems Inc., San Jose, Calif., and its WebLogic product.

In explaining the whys and hows of real-time application, Martin said: "We focus on building and populating decision support engines, or frameworks that apply a theory of constraints to drive best-value decisions that support discrete organizational priorities," he said. "These dynamic models provide the intelligence to drive automated, consistent decision analysis, presenting executives with decision options rather than data aggregation and reams of paper-based reports."

Such compensated best-value decision support is increasingly critical to the viability and value proposition of the complex supply chain and e-procurement marketplaces that are generating real traction in the federal arena.

The hurdles for achieving real-time in decision support systems, however, are steep, according to Keith Gile, senior industry analyst for the Giga Information Group, an e-business consulting firm in Cambridge, Mass.

The scope of the project can preclude real success, because as the data sets become more numerous and complex, cost accelerates rapidly. As a result, not all applications will be able to achieve real-time decision-making capability.

For those that can, the critical step will be "tighter integration with tools that supply real-time delivery of data, like message brokering systems," he said.

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above

What is your e-mail address?

My e-mail address is:

Do you have a password?

Forgot your password? Click here
close

Trending

  • POWER TRAINING: How to engage your customers

    Don't miss our Aug. 2 Washington Technology Power Training session on Mastering Stakeholder Engagement, where you'll learned the critical skills you need to more fully connect with your customers and win more business. Read More

  • PROJECT 38 PODCAST

    In our latest Project 38 Podcast, editor Nick Wakeman interviews Tom Romeo, the leader of Maximus Federal about how it has zoomed up the 2019 Top 100. Read More

contracts DB

Washington Technology Daily

Sign up for our newsletter.

Terms and Privacy Policy consent

I agree to this site's Privacy Policy.