Gary Fulmer wasn't looking for a profit center when he introduced matrix management five years ago. Fulmer, executive vice-president of Crane Plastics Inc., in Columbus, Ohio, simply hoped it would quell the chaos that had erupted as the company struggled through a difficult transition in its manufacturing operation. Crane was beginning to make its own plastic compound for its diverse line of vinyl siding and custom extrusion products, and the changeover required intense collaboration among departments.

The solution that Fulmer landed on was matrix, a project approach to management that companies in a wide range of fields have adopted since it was developed some 30 years ago. But the technique has largely fallen out of favor, and even those companies that still use it complain that team members are forced to work for two bosses at the same time: their normal supervisors and the matrix-teamleader.

At Crane, however, matrix does work, largely because of Howard Bennett, vice-president of matrix. His sole job is running the 7 to 10 teams that are usually in operation.

Although he constitutes the entire matrix department, Bennett's rank is on a par with vice-presidents of other departments at the company, which expects sales of $80 million this year. His status underscores Crane's commitment to the matrix approach, and it gives Bennett the clout to fend off the department-versus-department power plays that have plagued other companies' attempts at matrix.

Under Bennett, the system has been so successful that the matrix department does operate as a profit center, typically recovering its implementation costs within a year at the longest. For example, one matrix team recently recommended spending $75,000 for extrusion hardware to improve productivity for one of Crane's popular siding accessories. The company reduced manufacturing costs by 25%, slashed the selling price by 10%, and projects an annual manufacturing cost reduction of $275,000.

But when Fulmer first introduced the matrix concept, he was just looking for a way to deal with his manufacturing problem. "Making the conversion in our large-volume custom products was driving me up a wall," recalls Fulmer. "Rarely had we done anything that cut across so many departments. It took compounding experts, toolmakers, extrusion people, quality-control people -- in all, I guess it took about five different disciplines to make this thing work. . . . But it wasn't working.

"People would have a meeting," Fulmer says, his voice rising in frustration at the memory, "and they'd have their assignments. They'd come back after two or three weeks, with all good intentions, but they just didn't get it done, because they had more important things to do in their own functional areas. And," he pauses, "there was no one to say, 'You didn't do what you were supposed to.' They were peers without a leader."

In the spring of 1979, Fulmer ran across a series of articles that described matrix management. "I had never heard of 'matrix' before, and neither had anyone else around here," he says. "The more I read, the more sense it made. I could see the negatives, but I thought, 'Lord, anything's better than what we're doing now."

Matrix, or project, management was developed in the early aerospace era. Companies in the 1950s and '60s found that assembling project teams from several functional areas was the most efficient way to produce intricate spacecraft components under deadline. But while a number of large companies still employ the matrix technique, it generally has fallen into disrepute, because of its inherently complex reporting structure. Conflicts arising from that structure often flare into disputes among matrix team members caught in the middle.

As Fulmer built his matrix program, he saw the usual interdepartmental competition develop. The matrix team leaders were drawn from various functional areas and "while they tried to be impartial," Fulmer says, "they all had a vested interest [to make their own departments look good.]"

The program didn't begin to click until Bennett took over the matrix operation. It was good timing, though, rather than strategic foresight that landed Bennett in his current job. The position hadn't even been conceived when matrix was introduced.

"The job was created for [Bennett] and by him," Fulmer explains. Bennett had just completed a stint as vice-president of operations at a Crane subsidiary and needed another assignment. The appointment meant that he would be the captain of every matrix team. Unattached to any functional department, he had no particular ax to grind.

Bennett guards against the eruption of interdepartmental friction by requiring each prospective team member's department head to agree formally to that person's involvement in the project, thus avoiding any scheduling conflicts. Fulmer gives the final go-ahead, but approves matrix projects only when recommendations from more than two functional departments are needed to get a high-volume, expensive product produced more efficiently. Routine problems are otherwise handled within functional departments.

Bennett's 35 years with Crane Plastics and his understated, nonthreatening style add to his effectiveness as matrix leader. The 56-year-old Bennett is familiar with the company's entire operation, having worked in all its departments except ac counting. A reserved man whose gray and black hair crests above his forehead, Bennett speaks in a quiet, slow baritone. He is unflappable, yet intolerant of excess.

In meetings, his contributions are minimal. He prefers the role of conduit to conductor. But Bennett is the focal point of every project, the person who handpicks the teams, sets the pace, controls the meetings, and is ultimately responsible for the outcome of every project undertaken.

Fulmer and other managers at Crane credit Bennett with the program's success. "It would be tough to do it without Howard," says Fulmer. "He's the most technically competent extrusion person in the country . . . and he's a father image for most of the people in this company. I don't know anyone who doesn't respect Howard."

But Bennett still needed to adapt to his new role. "When I assumed this position," Bennett says, "I was totally engrossed in manufacturing. I was focused on productivity, manpower utilization, manufacturing efficiency, and scrap control. I had no sympathy for marketing, accounting, or most of the other functions. This job gave me a broader outlook. I developed empathy for the individual within the project team.

"I realized early on," Bennett continues, "that a poor idea coming from someone who can implement it has a better chance of succeeding than a good idea coming from the person at the top, so I religiously try to keep my mouth shut. I began to see ideas coming from individuals that I didn't see before.

"Of course there's a psychology there," he says, noting the importance of the company's strong commitment to the matrix method. "They know the project is a corporate priority and that they're in the spotlight. Matrix assignments are not a chore, however. They know matrix projects will succeed, period."

For example, Bennett recently got the task of solving a packing problem on one of the Crane production lines. A matrix team had improved the production of one siding accessory by 25%, and the strips of warm vinyl are now spilling off the line faster than the line operator can pack them.

It is not a simple problem. Merely add ing another line operator would be only a stop-gap solution. So Bennett assembled a team of five managers from production, engineering, quality control, and marketing -- the operating departments that are involved in the production and sale of the product.

From this group, the solution began to emerge. The leading player at one early session was Byron Rose, a manufacturing engineer, who proposed adding a machine to the production line to streamline the packing chore.

"I want to take as much away from the line operator as possible," Rose says. "Now he is nesting, stacking, lifting, and packing the soffit [the vinyl panels applied beneath a roof's overhang], and not paying enough attention to quality control." Rose proposed building a $25,000 machine to assemble the packing cartons that would be incorporated into the conveyor system, so the operator doesn't have to handle the messy glue.

Bennett, however, wasn't convinced yet. He didn't like the price tag. He wondered about the mechanics of sealing the boxtops. Rose and the production supervisor produced sketches and samples of the boxes showing how the machine would work.

"It's not pressing, unless marketing starts pushing the soffit," Bennett says slowly. "We can solve things temporarily by adding another operator on the line." He scheduled another meeting two weeks away, with the understanding that Rose would explore ways to trim the cost of building the machine.

The group will meet several more times until it has reached a solution to the packing problem. Then Bennett will give each member an award -- a plaque decorated with the matrix symbol of a grid -- and disband the team.

Says Byron Rose, who has worked at Crane for seven years: "With other companies, matrix management is reduced to shouting matches. I've had no conflicts here, although I expected to. It's inevitable anytime you have two bosses. You know that what you're doing has a direct bearing [on your profit sharing.] Plus, you know matrix projects are a corporate priority. Howard's got talents you don't find that easily in a manager. He knows what's happening on the floor and can see problems from both ends. He's willing to listen."

Another employee agrees, but adds a caveat: "Yes, Howard gets things done, all right. Sometimes he can be like a dictator. But that's not all bad."