(FORTUNE Magazine) – Now, what exactly is it you wanted to know?" Tom Payne asks. He's dressed in jeans and sipping coffee, and he has a friendly, questioning look on his face. Payne is president of VF Services, a division of the $5.5-billion-a-year apparelmaker VF Corp., and he's in charge of an elaborate corporate reengineering project involving some 100 people. It has taken four years so far and has run up more than $70 million in software bills alone. Already competitors, customers, financial analysts, and management consultants are admiring his work.

But a visit from FORTUNE has Payne a bit perplexed. He begins flipping through an inch-thick pile of transparencies, full of facts and figures and flow charts and time lines, and color-coded, pyramid-shaped diagrams that look a little like piled-up bricks. Payne isn't nervous about sharing what he and VF are up to. He knows his project is well conceived, and the implementation is too far along for any competitor to catch up. What's troubling him is being asked to wrap it all up in a neat little package.

Payne's frustration is one every good CIO has experienced: It's easy to explain what a company hopes to gain from reengineering and investing in new information systems, but telling the story of exactly how everything gets cemented in place is another matter. VF has a detailed plan to consolidate and centralize corporate functions, to make running those functions less costly, and to pump up marketing efforts. It aims to increase annual revenues to $7 billion by 2000 by homing in ever more closely on the needs of consumers. Like many companies today, it's hoping to economize by buying prewritten, "packaged" business applications.

That's where the simplicity ends. It will be two years before all the software is running, and Payne's employees are continually assessing and reassessing their progress, dumping old ideas and coming up with new ones. As for "packaged" software, well, that's kind of a misnomer. Like other big users of enterprise software, VF has cobbled together such an array of these packages that its system looks downright proprietary. It uses enterprise resource planning (ERP) from SAP, supply-chain functionality from i2, a demand-forecasting program from Logility, data mining from SAS, query tools from Brio, an apparel-design package from Gerber, demographic data from Spectra, and in-store display systems from a company called MarketMAX--to name a few.

VF Corp. is something of a stealth giant, a company that touches your life in dozens of different ways without your ever being aware of it. One hundred years old and headquartered in Greensboro, N.C., VF makes Lee, Wrangler, Britannia, and Rustler jeans. It sells about $2 billion of jeans a year--more than any other company. VF also makes Timber Creek khakis, Vanity Fair and Vassarette underwear, Healthtex clothes for kids, Jantzen bathing suits, JanSport backpacks, and Red Kap workwear (everything from mechanics' uniforms to bulky firefighter garb). VF's goods are sold in mass-market retail outfits such as Wal-Mart and Target as well as department stores like Macy's.

Payne's massive project grew out of a recognition that the apparel company needed shaking up. VF stock had long been a steady but not dazzling performer. But sales were almost flat, and costs were rising: As late as 1996, VF's 17 brands operated as independent entities, each with its own purchasing, production, marketing, and computer systems. Payne says that VF's information-systems costs alone rose more than 10% each year because "everyone had their own stuff and bought different systems. We were adding IS staff without any synergies." A lot of companies were running into such problems and turning to PC-based client-server systems to solve them. Maybe for VF the time had come too.

Computers weren't the only area in which there were redundancies. VF's divisions never bothered to consult one another before ordering raw materials. Nor did they coordinate manufacturing to maximize output. And the brands often competed for the same buyers. "We had all the disadvantages of a small company and none of the advantages of a big one," Payne says.

In 1996, VF got a new CEO, Mackey McDonald, who believed that a well-integrated VF computer system could do more than just cut unnecessary costs. McDonald thought it could actually become a revenue generator. If VF could find a way to aggregate and sift the right data--customer demographics, point-of-sale information such as common sizes and popular colors, and so on--it could, he figured, pinpoint what mix of products and sizes a particular retail location was likely to sell. In other words, a computer-savvy VF, once it identified consumer types at the store level, could tailor its production and distribution systems to serve them with maximum efficiency. Unlike its rivals, it would no longer have to depend on murky national and regional forecasts, which often created inventory gluts or shortages. "We had a tremendous opportunity to compete on knowledge," McDonald says.

With a micromarketing system, as it's called, McDonald believed that VF's computers would one day be able to predict that a particular Wal-Mart store would sell a given number of Wranglers with a 34-inch waist, in white, at the beginning of summer. VF would then make sure that those jeans were in stock: no chance of losing that sale to some other brand across the aisle. It was all part of a strategy McDonald called consumerization, and he thought it could spur growth, helping to produce revenues of $7 billion by 2000.

Processing all the data effectively would be an immense challenge, and Payne had to figure out how to make it--along with the cost cutting--happen. His first move, in January 1995, was to bring together 50 VF managers to develop a battle plan. Members of the group fanned out across the country seeking best practices to emulate at companies such as Procter & Gamble, 3M, FedEx, Caterpillar, and Rubbermaid (now a fallen angel). Armed with its research, the group locked itself in a conference room for about six months and came up with new business processes for VF. First, it created a central nervous system. The 17 brands would combine in five "coalitions": jeanswear (including workwear), intimates, playwear, knitwear, and international operations and marketing. Corporate headquarters, which had been in Wyomissing, Pa., since 1899, would move to Greensboro, which was closer to many of VF's operational sites. Payne, who had worked only for Wrangler, would head a new, cross-coalition information-technology and process-engineering organization called VF Services. (The division would also oversee some accounting, procurement, and human resources operations.)

The committee next worked out processes for product development, production planning, and distribution, as well as for routing information from department to department. Only then was it time to deal with the software. "We came up with the practices first," says John Schamberger, chairman of VF's jeanswear coalition. "The systems are just the enabler."

VF has already begun installing an SAP ERP system, which operates at the corporate level. SAP's software, R/3, will handle all of VF's financial operations as well as some order-management and materials-management functions. R/3 will eventually link with all the other software VF is installing, so that when a retailer places an order for backpacks, for example, planning and manufacturing will find out about it right away.

At the coalition level, the company is setting up the different packages one at a time to avoid making any single coalition into a sacrificial lamb. Intimates is using Gerber's WebPDM program and various design tools that link with it to cut costs in the product-design phase. Intimates and workwear are installing i2's Rhythm, which applies a sophisticated algorithm to the problem of optimizing materials utilization and assembly-line space. Jeanswear is using software from Logility, which forecasts consumer demand so that manufacturing facilities can produce what customers will buy. McDonald's pet project, the micromarketing engine, is known internally as retail floor space management (RFSM). It is made up of three packaged-software components as well as proprietary software VF developed for itself. Set to debut in the jeanswear coalition, it's currently a pilot project at a Midwestern mass-market retailer.

VF isn't using these software packages as is. Thanks to the company's size--and its relatively early entry into the reengineering game--it can make fierce demands on software suppliers, forcing them to write brand-new "apparel-industry standard" software to fit its needs. For example, VF and Reebok persuaded SAP, which leads the worldwide market in packaged ERP software, to write a program called Apparel Footwear Solution (AFS), an add-on for apparel makers. (Before AFS, R/3 was unwieldy for these businesses because it didn't speak their "language"--it had no way of naming different sizes or colors of a single product.) SAP has been working on AFS for two years, and VF is just now beginning to install it (the company rejected three previous releases). SAP has an incentive to cooperate: It wants to be the leading ERP provider in as many industries as possible. AFS has already attracted more than 30 apparel companies, including Warnaco, Bruno Magli, and Adidas. "Because of our relationship with VF, we think that we'll become the leader in ERP for apparel companies," says Jeremy Coote, president of SAP America.

Outside software providers aren't the only ones under pressure--VF's own programmers are too. "You start talking in acronyms, and you don't know how to say things any other way," says Bob Franks, who manages 30 or so VF employees devoted to getting just one "module" of SAP software up and running. Every member of Franks' staff has had at least 14 weeks of SAP training. At any given time two or three of them are stationed in Germany, where SAP is headquartered. Sometimes SAP programmers visit Greensboro. Blake Conklin heads the product-development reengineering process, which means he's struggling to fine-tune WebPDM. He, too, has loaner techies on-site. "You'll have a team of two or three of their people here for the rest of your life," he sighs.

The price of the software includes the services of the "loaner" programmers, but VF has also hired consultants from Deloitte & Touche and its information-technology division, ICS. In some companies, such consultants can cost as much as the software; Payne says he has kept costs down by limiting consultants to specific reengineering and implementation tasks. "Consultants are necessary," he says, "but you've got to plan how you use them."

So what will it look like when it's done? To give me some idea of how micromarketing will eventually work, Pat Garvey, who heads RFSM development, takes me to a Greensboro-area Wal-Mart. The store we visit hardly seems a hotbed of high-tech magic. It's late August, so mothers are snatching up notebooks, rulers, lunchboxes, and the other little accessories their kids need for school. Senior citizens stroll through the other side of the store, pausing as they pass gardening tools, magazines, and the occasional food item. In the men's apparel section, a grid of shelves--known in the retail business as "bins"--is stocked with VF jeans.

This is where it will all come together, Garvey tells me--where customers will find those white Wranglers on the very day they need them. It's kind of hard for me to imagine, looking at the disorganized piles of mixed-up products on the shelves. On the far left are spaces for the lower-priced Rustler jeans; to the right is where the Wranglers and Timber Creek khakis should be. Whoever is arranging goods at this Wal-Mart is not exactly on top of things. Garvey admits that retailers could be the weak link in VF's plan; it all depends on stores' willingness and ability to follow VF's directions for stocking the shelves. Payne concurs: "It's a burning issue. Getting product to the store is only half the battle." But Garvey thinks the retailers will want to cooperate, because high-tech "consumerized" displays can help them increase sales. Results from the RFSM pilot in the Midwest have been impressive. Because they now stock only what they are reasonably certain they'll need, VF and its retail partner have reduced on-floor inventory 11% and increased product turnover 15%. New product offerings have gone into the freed-up space, and revenues from the pilot have increased more than 20% in each of the past two years.

Payne says it will be a couple of years more before VF's $100-million-plus investment makes a significant impact. Just centralizing information-technology operations has bumped up operating margins almost two percentage points, but benchmarking from now on will be tricky. "The first year, we'll know by gut feeling how we've done," says Franks. "The second year, we'll build on that." In the meantime, hopes are high, if measured. CEO McDonald says, "Our people designed the products and the process. That's why I think we can have success." Leroy Allen, who leads reengineering for VF Services, goes further out on a limb. "If we pull this off, no one else will be close" in integrating data across operations, he says.

Allen talks inspirationally about how the new computer system will unite VF, making its coalitions all finally work as one. Tom Payne interrupts: "Knock on wood." Apparently he can't say yet that he sees a perfect end in sight. He raps his knuckles on the conference room table.