(FORTUNE Magazine) – Second-rate systems can be found even at Hewlett-Packard, one of the world's best managed corporations and currently No. 9 on FORTUNE's list of Most Admired companies. HP's North American distribution organization handles billions of dollars in products, from PCs to toner cartridges, from order to delivery. But there was a problem: On average, it took a languid 26 days for a product to reach the customer. Employees shuttled information through a tangle of 70 computer systems, some decades old--and this at a leader of the high-tech industry. The job of reengineering the process fell to two seasoned managers, Mei-Lin Cheng, 44, and Julie Anderson, 46.

They came up with an extraordinary strategy. Assigned to create entirely new processes, they bought freedom from their bosses at HP by agreeing to produce significant and measurable improvements in customer satisfaction--in less than nine months. Failure would mean a cutoff of funding. They assembled a team of 35 people from HP and two other companies, explained the ground rules--and then vigorously refused to tell anyone what to do. Result: They met the deadline and achieved their goal.

FORTUNE: What was wrong with HP's business as usual?

ANDERSON: We were just cranking the wheel, taking care of this problem, satisfying that customer, making the goals for this year and then next year doing the same thing all over again. Patching things together with bubble gum didn't feel good.

This complaint is not about HP. The whole economy is changing, society is changing, and organizations are stuck in a certain way of managing. We haven't got a system that matches where we need to go.

CHENG: We were both passionate about not doing things the same old way. I was striving for a much higher level of integration of my work and personal life. I'd come home after giving someone a negative performance evaluation and escape into jogging, so I could just listen to my heartbeat and my breathing.

We talked about how to structure this project, and pretty soon we said, "Maybe we shouldn't have a structure at all?" When people are free to define their own goals and roles at work, their commitment intensifies, and the job becomes more personally meaningful. You don't feel as if you're leaving your real self in the parking lot before you come in to work. Your contribution can be richer. You're energized at the end of the day, instead of feeling I'm dead, give me a drink.

FORTUNE: Don't remind me. How do you create this miracle of integration?

ANDERSON: Let me explain the situation first. We were creating a complicated system, using millions of dollars of off-the-shelf software from an outside vendor--SAP--to create a single, unified database covering everything from the customer order through credit check, manufacturing, shipping, and warehousing to invoicing. We had to reengineer all those systems, so it was incredibly complex. We wanted everybody--including team members from Andersen Consulting and our transportation and distribution partner, Menlo Logistics--looking at the whole system.

CHENG: Most of these complicated information-technology projects fail--the industry's failure rate is as high as 70%. You're taking too long, and you're over budget. And by the time you get it done, the business need has changed.

ANDERSON: We decided that if we couldn't get results in one year, we were probably taking the wrong approach. That's why, when we started last November, we agreed to put our heads in the noose by committing to real business results before the funding review scheduled for nine months later.

We ended up meeting that target in eight months. Our pilot program focused on one customer, an office-products distributor that accounted for a large share of the multibillion-dollar sales of this part of HP. We've consistently delivered products to them in eight days, vs. 26 days before. That enabled them to cut their inventories by nearly 20% while increasing service levels to their customers. And so we got our funding for the project's second year.

FORTUNE: Great--but what did you do to make the team so effective?

ANDERSON: We took things away: no supervisors, no hierarchy, no titles on our business cards, no job descriptions, no plans, no step-by-step milestones of progress. We located as much of the team as we could in one office, explained the time frame and the business measures, and left them to figure out what to do. The idea was to create a sense of personal ownership.

FORTUNE: And the result was complete chaos?

CHENG: Naturally. ANDERSON: We started with a two-week training and orientation program, to familiarize everyone with the existing processes and the needs of the business. For instance, we all physically followed an order through every step. When the order was sent to HP headquarters for batch processing on an outdated computer system, everybody took a bus to headquarters and talked to the people there about their jobs. The tour took three days, but everyone got a tacit understanding of the process. Practically everywhere we went we saw a building full of inventory. We met people whose entire jobs were devoted to reconciling information from one computer system to another.

We introduced a couple of key ideas, such as the value of considering diverse points of view, and MIT professor Peter Senge's systems thinking, which emphasizes learning skills and teaches a perspective encompassing the whole of a situation, not just the parts. We simulated a kindergarten classroom with little chairs and crayons, to remind people of how they felt when learning was fun.

CHENG: It set the tone: We're all in this together, guys. We're going to do something different.

FORTUNE: Let's hear about the chaos.

ANDERSON: When people came back from the orientation, they were anxious to start. Everyone felt the time pressure. They asked us how to organize the work, and we just said, Start. They spent the whole first month in one long meeting, trying to figure out what to do.

CHENG: Eventually they agreed that the first step should be conceptual design. So they tried out different ways of working together to do that. One approach was to have a group working on one module, let's say finance, and other teams working on other modules. But then we realized that if we worked on each piece separately, the traditional way, we'd run into the old problem of how to integrate them later.

One person told me she spent the first two months feeling we were wasting her talents. Then she realized we weren't actually preventing her from contributing, so why not just do it?

FORTUNE: They must have had smoke coming out of their ears.

ANDERSON: Oh, yes. Some people would come up to us and say, Why don't you just tell us what to do so that we can move on? This is a big waste of time. And we'd say, If we give you an answer, it might not be the right one. This was their introduction to the idea that no one of us can see the whole system perfectly. The more people focusing on it, the better the solution will be.

CHENG: If you think I am smarter than you, you are wrong. This is your chance to try something different.

FORTUNE: How much of that was sincere and how much was theater?

ANDERSON: Well, I really didn't know what the right answer was. I could have made some suggestions. I had some ideas that might have helped...

FORTUNE: Probably some very good ideas.

ANDERSON: Yes.

FORTUNE: And yet?

ANDERSON: I sat on my hands. Chewed on my fingernails.

CHENG: I had some internal struggles about not speaking out. I sometimes woke up in the middle of the night, wondering what the next step should be.

ANDERSON: I never slept better. I felt like a burden had been taken off my shoulders. I was out of the rat race. The difficulty was that we'd all been rewarded for being able to stand up with a chart and point to results. We couldn't do that anymore. As we struggled with our new roles, we tried to make our own soul-searching visible to the team, to serve as role models of living with that vulnerability.

FORTUNE: How often were you reporting in to your masters at HP?

ANDERSON: Not very often, if we could avoid it. It was very hard to honestly represent the situation in those early days in a way that would mean anything to people from outside the team. We would meet the project's sponsors in the hallways, and they would ask how things were going. We would say, Well, fine, you know, we're still working on it.

FORTUNE: How did the team finally get out of the rut?

ANDERSON: We'd made some progress, but within a couple of months we were in a tough spot. Some people suggested that, since we already had the SAP software, we could try to execute orders on it. We had some very lengthy discussions about that. Finally somebody said, This is like the movie Groundhog Day. Bill Murray plays a character who relives one day of his life over and over. He has to figure out how to get to the next day of his life, so he tries all sorts of things. He begins by learning simple things, like remembering not to step in a puddle on the street. Eventually he learns to stop living just in his mind, to listen to his heart and soul. Each time the day began again, his clock radio played Sonny & Cher's "I Got You Babe." That story became the metaphor for our team. Every day at 8:30 a.m. we'd play that song. And every day we tried to process an order through the system.

FORTUNE: So the real secret here is dealing with bite-size problems one by one?

ANDERSON: In a way. The idea was to try things out and learn as we go along--to learn by practicing, like an orchestra or a sports team. People in business too rarely get a chance to practice.

CHENG: At first we didn't have a complete system, so we represented the various pieces with toys. Mr. Potato Head was an HP business unit in Boise, a Lego Polynesian boat stood in for a Japanese supplier, a can of beans was Finance. The first time we put an order into the system, we couldn't even get it to the warehouse.

I remember one discussion then about whether to stop the practice. A team member said, Why do we have to be perfect right away? Imagine this as Bill Hewlett and Dave Packard starting the company. Do you think they waited until all their order-fulfillment processes were perfect before opening the door for business? The team kept at it, identifying problems and improving the system bit by bit.

The first time we successfully simulated getting an order all the way to the customer, the process took about ten days. Something would break down, and we would have to fix it. Finally we got to the point where we could do an order in a day. Then we started launching simultaneous orders, making it more and more complex.

FORTUNE: What have you learned?

ANDERSON: This idea of practicing is probably the most important idea we have introduced. Every morning there is a practice meeting where we discuss what happened yesterday, what did we learn, what are we struggling with, what are our priorities.

CHENG: Another lesson: You don't need to tackle complexity with complex solutions. Our approach is to jump in, bite off what we can chew, and learn as we go. You've got to start without knowing where the journey is going to take you.

ANDERSON: What we're trying to do is to build the capacity to learn, first in individuals and ultimately in the organization. That's the only way we know to get an organization that's nimble in responding to business needs. We're both still learning what our role as managers should be. One thing we know for sure, though, is that we don't make the decisions.

REPORTER ASSOCIATE Rajiv M. Rao