Quick thinking

Clear goals plus more tightly focused systems hasten move to business intelligence software

Business intelligence software is among several technologies expected

to play an increasingly important role in the way federal agencies do business

and share information after the terrorist attacks of Sept. 11.

The software enables organizations to access, analyze and share information,

both internally and externally. Users can

query data to spot patterns and trends, make connections among vast amounts

of disparate information and then use that information or analysis to make

more informed decisions. The software is expected to be a vital cog in the

newly created Office of Homeland Security.

Although it can be applied to myriad tasks, some experts complain that

the software is tricky to use — and expensive and time-consuming to deploy

especially when used to build a data warehouse, which is the most common

type of business intelligence system used in the federal government.

More recently, however, users have been shifting to a newer, more focused

approach to business intelligence that reportedly brings a quicker payback,

and that trend is expected to gain momentum in the wake of the terrorist

attacks.

Technology must be put in place to make it "possible for the government

to understand what it knows in the sum of its partsÖand take action," said

R. James Woolsey, director of the CIA from 1993 to 1995, during an Oct.

17 forum on government integration and national security sponsored by webMethods

Inc. in Washington, D.C.

Data warehouses — repositories of information from multiple sources

structured for querying and reporting information — certainly are one way

to increase this type of understanding.

But even before Sept. 11, some federal agencies had begun gravitating

toward data marts, which are smaller and less expensive versions of data

warehouses, according to Darren MacLennan, manager of business intelligence

and data warehousing for Oracle Federal.

MacLennan said the company's philosophy about business intelligence

comes from its chief executive officer, Larry Ellison, who has been calling

for fewer databases containing better information. "It's the same approach

here with business intelligence," MacLen.nan said.

He said Oracle's work with the Justice Department's Office for Domestic

Preparedness began in 1999 and has blossomed into a model for business intelligence

projects seeking to bring together information from various government agencies.

The Justice office is responsible for enhancing the capacity of state

and local governments to respond to domestic terrorism, which requires an

integrated effort among the different levels of government.

"They are using databases, query and analysis solutions to evaluate

data, like populations by discipline, hazardous materials information, emergency

medical services teams, and fire and safety personnel competency levels

to be able to respond [to emergencies] by state, county or region," MacLennan

said.

The U.S. Geological Survey is using an Oracle product for its National

Water-Quality Assessment Program, which includes evaluating samples of $70

million worth of water-quality information from around the country.

Alan Shaeffer, area vice president for the public sector in Cognos Inc.'s

federal division, said he has seen a trend toward agencies asking for solutions

that can be deployed quickly, which means more requests for data marts.

"They want an immediate return on their [information technology] investment

dollars and to show results quickly," he said.

Shaeffer said a data mart can be built in a few months, but some customers

who want an even faster deployment have chosen data shelf solutions, which

can be up and running in one month. "It's all about responding quickly with

meaningful dataÖand government executives are held accountable for quick

results," he said.

Cognos is working with officials in an Air Force organization to help

them measure business processes, such as how long an interview process takes

and the life cycle of laboratory test results, and then analyze the data

to see if the cycles can be shortened.

Think First

Whether requirements call for a data warehouse, a data mart or a still

smaller data shelf, business intelligence vendors say the first step in

building a system is for agencies to come up with a clear idea of what they

want to accomplish.

As part of the homeland security initiative, government officials may

have to go through a "management decision process re-engineering," said

Bart Foster, senior vice president of worldwide marketing at Informatica

Corp., which has many federal customers.

This means finding out "what information in each agency is relevant

to the broader issue of catching bad guys," Foster said. He named the CIA,

the FBI and the Immigration and Naturalization Service as agencies that

could benefit from tying together some of their disparate databases in the

context of homeland security efforts.

But Foster emphasized that agencies must "know what they want to learn

before you start pulling schemas and data," because business intelligence

"projects are never successful if you attempt to answer all the questions."

Of course, this task is often easier said than done, especially if it

involves drawing on data that only resides within one agency's walls.

When building a business intelligence system, "the [obstacles] are invariably

organizational, not technological," said Dave Kellogg, senior vice president

of marketing at Business Objects SA. "The impediments to getting that sharing

have always been organizational."

Perhaps that will change as the stakes continue to rise. The events

of Sept. 11 illustrated an "absence of sufficient cross-agency collaboration,

and that means sharing information and [using] business intelligence tools

to share information," Kellogg said.

Jeff Babcock, vice president of government sector sales and marketing

at SAS Institute Inc., agreed. "It's not rocket science to build a solution.

The hard part is getting all the parties to agree to do it and [getting]

rid of the stovepipes," he said. "It shouldn't take six months to build

a data warehouse if the right work is done upfront."

Of course, the responsibility for making changes does not rest solely

on the shoulders of the federal agencies that use business intelligence

software. Former CIA director Woolsey said the private sector needs to rise

to the challenge as well.

"Much of our ability to deal with the war we're now fighting is the

speed and adaptability of American businesses, particularly in the high-tech

sector," he said.

At least some adaptation on the part of vendors seems in order, particularly

if a new business intelligence system is needed by agencies involved in

homeland security, according to Babcock.

Those agencies and the Office of Homeland Security need to "track and

analyze security-related information, and that's not what those [business

intelligence] tools are designed to do," he said. "Very few companies have

the capability to provide the kinds of things necessary for business intelligence

for homeland defense."

Tips for getting the right answer

Representatives from the Internal Revenue Service, the Labor Department

and the U.S. Transportation Command shared how data warehousing projects

aid their missions at the Data Warehousing Institute's spring conference.

The officials agreed that certain criteria must be met for such a project

to succeed. They include:

* Showing a return on investment early in the program.

* Getting support from executive leaders and communicating that support

to midlevel managers and users.

* Linking the business strategy to the technology.

* Ensuring timely and accurate data migration, cleansing and access.

* Properly training employees on how to use vast amounts of data.