(FORTUNE Magazine) – In summer 1998, Bill Braddy traded in the Pentagon for Schneider National, a big transportation and logistics outfit in Green Bay, Wis. The former Army colonel had just finished a stint as deputy director of a simulation facility where the Department of Defense trained high-ranking officials in crisis management. His mission at Schneider, while not exactly a matter of national security, would be a sensitive one: Figure out how to search through and harness mountains of valuable but disorganized corporate data quickly enough to make the business more efficient. "We were drowning in data but starving for information," he says of the crisis Schneider faced. "It was a significant emotional event for the company."

Data? Emotional? You heard right. Braddy isn't being cute or overdramatic when he talks about data management. Companies have been able to use technology to do some very cool stuff--to reach customers in new ways, to automate operations. But one thing many businesses haven't been able to do easily is use the data they've collected to find and stamp out waste across operations. Sifting through corporate data was supposed to make executives more efficient. Much of the time, though, it's just made them more confused.

To attack Schneider's problem, Braddy did what a lot of corporate IT folks are trying these days: He installed business intelligence software. BI, as it's often called, helps businesses view and catch and dissect and reassemble and write reports about all sorts of information lurking in the dark corners of their databases. If BI is linked to clean, carefully tended data--a really big "if"--it can work very well. At Schneider, Braddy is counting on BI to show managers where outsized costs (and benefits) exist and why. So far it's off to a good start.

To understand why a trucking company cares about gargantuan databases, it helps to understand what Schneider does. The family-owned company's main business used to be renting out its signature pumpkin-colored trucks. Today Schneider is the biggest transportation and logistics company in North America, sending goods by truck, rail, and sea all over the world; it also brokers freight to other carriers and provides some financial services. A big chunk of its $2.4 billion in annual revenues comes from Schneider Logistics, a division that makes sure its customers' wares get to the right place at the right time--taking care of freight payments for the likes of Wal-Mart and chemical giant BASF, coordinating shipment of aftermarket parts for General Motors and Ford.

Logistics and transportation are two of the most data-intensive pieces of Schneider's business. So when Braddy signed on, he turned to their information woes first. The company's computers were high powered, but they had to run complicated analyses on almost ten terabytes of data strewn across eight different databases. Schneider collected all kinds of documents: invoices, tracking paperwork, and late-payment notices, as well as delivery information zapped in via satellite from truckers, warehouse operators, and accounting offices all over the world, every minute of every day. According to Braddy, if you transferred all the bits and bytes Schneider was juggling onto 3.5-inch floppy disks, you'd have enough cargo to load ten of the company's 53-foot trailers.

The volume of data, in and of itself, wasn't the problem. The problem was that you practically had to be a rocket scientist (well, an SQL programmer) to find anything in there--and the process was very, very slow. Let's say a Schneider analyst wanted to figure out why it cost the company 20 cents per pound to make deliveries to a huge Ford dealership in Texas, but only 17 cents to most other locations. There could have been a number of reasons for the extra cost, but she could only pick one at a time to investigate, so she might have chosen to look into how long an average shipment to the dealership had been in transit over the past six months.

To get to the precise logs she wanted, she would have to write up a request and deliver it to one of her department's data engineers--who in turn would write a program to fish out the needed data, run it, dump the data he found into another report, and e-mail that report back to the analyst. The process could take as long as a week and would often prove fruitless (if, for example, the problem was related not to time in transit but rather to shipment weight). "If the information was good, you used it," remembers Schneider CIO Steve Matheys. "If it wasn't good, you probably used it anyway. That's the way things were run."

What really bugged Braddy was that this setup erected a wall between the data and the people who really understood them. "The guy who knew how to get to the data and the guy who knew how to use it were two different people," he says. "I had to find a tool that would let a business professional find what he was looking for without needing an IT professional."

With the advice of Stamford, Conn., technology-consulting firm Meta Group and a team of as many as 30 people from Schneider, Braddy spent five months evaluating some 23 BI solutions, eventually zeroing in on eight finalists. In February 1999, an executive group voted unanimously to use a suite of software from an Ottawa company called Cognos, mainly because one of the products in the suite, PowerPlay, seemed to offer the most advanced data-analysis capabilities.

PowerPlay excels at two things: presenting key data to laypeople and making it easy for them to analyze that data in depth. To do this the software takes records from a company's databases (specified by people in IT) and pours them into specialized information arrays it calls cubes--which are basically stacks of manipulable spreadsheets. A cube that the business analyst uses today, for instance, houses all kinds of data about load status--if a shipment is waiting to be assigned to a carrier, picked up, in transit, delivered, or paid for. Using the software she can look at the average time that loads spend in transit to the Ford dealership, click on a button to sort the data, see immediately that the problem lies somewhere else, and then--this is the beauty part--run a second analysis instantaneously. Once she finds where the problem is, she can drill down on elements in the cube until she gets to a single smoking-gun document. The entire process takes less than ten minutes.

Just five months after they chose the software, it was up and running--thanks in part to the fact that Schneider had already devoted two years and 60 people to a complete overhaul of its massive databases. Technicians, for example, had worked to smooth out the ten terabytes of information item by item (making sure that dates were written uniformly and so on). "That required real intestinal fortitude," says Bob Grawien, vice president of applications development, who worked with Braddy on the cleanup.

A few weeks after the PowerPlay launch, Braddy began seeing results. A business analyst in the trucking division was trying to track down reimbursements owed to Schneider for certain kinds of auto work--money that had been practically impossible to collect in the past. Using PowerPlay, the analyst was able to pinpoint, in a matter of two weeks, hundreds of cases in which payments had fallen through the cracks. From that one sweep, Schneider was able to recoup all the money it spent on Cognos software in the first place. (While Schneider won't reveal how much it spent, a similar large installation, says Cognos, might cost as much as $2 million.) Over the past two years, Braddy estimates, the software has helped Schneider grab a total of $2.5 million in "hard, quantifiable savings."

Now Braddy is extending the Cognos tools across the company; by the end of the year, four of Schneider's five business units, including transportation and brokerage, will be using the software. And more than 200 customers are tapping into the system through a tool called InterAcc that Schneider created using Cognos software. For a small subscription fee starting at a few cents per invoice, customers get access via the Web to cubes, analysis tools, and portions of Schneider's database so that they can track their own logistics. (While Schneider isn't planning to make it a major profit center, InterAcc is already breaking even.) Sven Meyer, a freight payment manager at BASF's Mount Olive, N.J., office, loves the program's speed: "When a carrier calls me with a complaint, I go immediately to InterAcc and look up their account. I don't have to put them on hold."

Ironically, Schneider techies say that the company's new databases deliver information so efficiently and flexibly that some users simply don't know what to do with it. Maybe that's why a few customers balk at studying their logistics, no matter how much money it might save. Says Mark Rourke, vice president of customer service for Schneider Logistics, with a grin: "They say, 'That's why we hired you.'"