Father, Son & Co.

by Thomas J. Watson

  “Ever since your father put me on this board—”

  I said, “Gilbert, look. My father didn’t put you on the board; I put you on the board. Everything that happens around here gets attributed to my father. But I met you on the Mutual Life Insurance board and admired you greatly, and you were very kind to me even though I was a newcomer. When Dad asked me to suggest someone as a director from Chicago, I gave him your name.” My telling him this healed the rift between us, and Scribner remained a valued board member for many years.

  Fear of failure became the most powerful force in my life. I think anybody who gets a job like mine, unless he’s stupid, must be a little bit afraid. There is such a long way to fall. Yet the fear I felt when I went back to IBM took me totally by surprise. Before Dad died I was running the business and even chafing at not getting full credit because he was around. I didn’t realize how much I still needed him emotionally. I remember standing in the corridor outside my office soon after I came back from Alaska, looking dumbly up the stairs that led to his. Except for Dick’s operation at World Trade, the weight of IBM was now all on me; if Dad hadn’t died I could have been chief executive for years and never felt so burdened.

  I decided it would be foolish to act as though I could totally take his place. Rather than move into his big, wood-paneled suite, I kept my own office downstairs and ran IBM from there—later we turned his office into a library. I retired his title of chairman, keeping the one he had given me: president. There was the problem of what to do with his seat on the board; I solved that by asking Mother to join. She had been at Dad’s side for so long that many IBM people felt a personal loyalty toward her that I didn’t want to lose.

  The worst thing that can happen when a leader dies is for his followers to lose their inspiration and carry on like robots. I moved as fast as I could to prevent that. So before the year was out I called the top one hundred or so executives to a conference at Williamsburg, Virginia, and distributed power and responsibility even more broadly than before. In three days we transformed IBM so completely that almost nobody left that meeting with the same job he had when he arrived.

  I picked Williamsburg because it is a historic place and this meeting was meant to be a kind of constitutional convention for the new IBM. Almost everybody came knowing a few details of what we were about to do, and you could feel the anticipation and excitement in that rented conference room. After the tumultuous events that had already taken place that year—the settling of our antitrust case, my promotion to chief executive, and Dad’s death—everybody felt that this was the takeoff point. It was the first major meeting IBM had ever held without Dad present, and we knew how far we’d come from the days of his Hundred Percent Club meetings with their circus tents and banners and songs. Of the old guard, George Phillips was the only one at Williamsburg, and he was scheduled to retire the following month. There were a lot of young men and one young woman. My age and experience were typical of the people present—I was forty-two and I’d been in management barely a decade.

  What we created was not so much a reorganization as the first top-to-bottom organization IBM ever had. It was largely the work of Dick Bullen, the young MBA I’d named as organizational architect the previous year. Under his plan we took the product divisions that we’d already established, tightened them up so that each executive had clearly defined tasks, and then turned the units loose to operate with considerable flexibility. These were IBM’s arms and legs, so to speak. At the head of the corporation, to oversee plans and major decisions, we set up a six-man corporate management committee that consisted of me, Williams, LaMotte, my brother, Miller, and Learson. I gave each man responsibility for a major piece of IBM, while leaving myself free to roam across the whole company. Finally we superimposed a corporate staff that included experts in such areas as finance, manufacturing, personnel, and communications. Their task was to work as a kind of nervous system and keep our adolescent company from tripping all over itself, as had happened a few months earlier when we had two divisions unwittingly bidding against each other on the same tract of land for a factory.

  By the mid-’50s just about every big corporation had adopted this so-called staff-and-line structure. It was modeled on military organizations going back to the Prussian army in Napoleonic times. In this sort of arrangement, line managers are like field commanders—their duties are to hit production targets, beat sales quotas, and capture market share. Meanwhile the staff is the equivalent of generals’ aides—they give advice to their superiors, transmit policy from headquarters to the organization, handle the intricacies of planning and coordination, and check to make sure that the divisions attack the right objectives. Du Pont and General Motors began applying this system to business as early as the 1920s, but to the IBMers at Williamsburg in 1956 it was a revolution. Everybody in that conference room had been trained under my father in exactly the same way—we’d all started out as salesmen and we’d all been molded as line managers. The phrase you heard about a successful IBM executive was, “He knows how to get the donkey over the hill.” We all knew how to get the donkey over the hill. But when it came to thinking about which hill, or whether it might be wiser to go down and around a hill rather than up and over, we tended to be as stupid as donkeys ourselves. When the meeting convened I gave a little speech saying that times had changed: “We’ve been a company of doers. Now we must learn to call on staff and rely on their ability to think out answers to many of our complex problems.”

  We then created the staff right before their eyes. There were dozens of slots to be filled, and since IBM had very few specialists on the payroll, we “made” our experts simply by naming people to the jobs. Williams and I had rejected the idea of recruiting outsiders, except in highly specialized fields such as law and science. We’d spent years weeding out Dad’s yes-men and replacing them with fierce, strong-willed decision makers. If we’d hired a bunch of professors or consultants to come in and play staff against these people, the newcomers would have been eaten alive. Instead I put IBM’s best executives in the new jobs, starting with Al as chief of staff. It was a huge sacrifice to move him outside the chain of command—and for Al it meant swapping a position in which he had twenty-five thousand people working for him to one in which he was boss of only eleven hundred. But by shifting our stars into the staff we enabled it to command the respect of the divisions, and that was the key to making the whole thing work.

  The great strength of the Williamsburg plan was that it provided our executives with the clearest possible goals. Each operating man was judged strictly on his unit’s results, and each staff man on his effort toward making IBM the world leader in his specialty. So on every operating proposal we had financial men demanding to know how it would improve profits, public relations men fighting to make sure it enhanced IBM’s image, and manufacturing men insisting that we maintain the highest productivity in our plants and quality in our products. When the conference came to an end after a couple of days of workshops, there was no doubt that IBM had been totally transformed. To underline that fact we ran a special issue of the company newspaper featuring the first organization chart IBM ever had. Oddly, I had no feeling at all that we were going against Dad. I didn’t think he’d have been alarmed by anything we did. If I could have asked him, “How do you want this thing run?” his answer would probably have been “I don’t know, son. It was getting so big when I left that I could hardly understand it. Do whatever you think is right.” That was the mandate I thought I had. Not a day went by when I didn’t think about the old boy, but the only thing I really worried about was lousing the business up.

  My friend Al never worked harder in his life than he did whipping IBM’s new staff into shape. He had to take men of action and turn them into men of thought. There were thirteen people reporting directly to him, and he was working sixteen hours a day, six or seven days a week. The first thing he did was try to teach them to write. A good staff report is supposed to be intelligible and direct. It s
hould lay out a problem concisely and end with a clear recommendation, so that all top management has to do is read it and say, “Go ahead” or “Don’t.” Al told the staff, “You have to pay attention to this written stuff. Remember you’re going to hand it to a senior manager and he has to understand it. If he approves your recommendation he’s going to hand the report to somebody else to carry it out, and that person has to understand it too.” This advice didn’t take. After six months the reports were still so chaotic and obtuse that Al got desperate. He called the staff together and showed them a manila folder. “Every time one of you hands me a sloppy report,” he said, “I’m going to take a folder like this and put your report in the left-hand pocket. Then I’m going to rewrite it the way you should have done it in the first place, and put that in the right-hand pocket. Then I’m going to circulate the folder to each man in this room.” After that, if somebody did a lousy job, his twelve associates saw it. The quality of the written work went up dramatically.

  Al constantly had to arbitrate disputes between the staff and the line. A typical headache was that when any of the staff men spotted something wrong in a division, he’d revert to being a man of action and fire off a blistering memo dictating what ought to be done. This would cause the division executives to howl that the staff was trying to interfere. The staff also developed the bad habit of dragging their feet on projects they didn’t like. Al must have lectured his men a hundred times: “Quit acting like bureaucrats. You’ve got to facilitate things, not block things.”

  In the months leading up to Williamsburg, Al and Dick Bullen had spent many hours discussing how to transform the natural tension between the staff and the line into energy that would move IBM forward rather than friction that would slow us down. The organizational solution they finally came up with was as simple and brilliant as the transistor—a check-and-balance scheme that eventually became famous as the IBM system of contention management. Not only did it make staff versus line conflicts acceptable; it actually encouraged them. In this system, no operating plan at IBM was final without a staff man’s concurrence. And if he signed, his job was just as much at stake as the executive’s who made the plan. When an executive and a staff man couldn’t agree, the problem automatically would get kicked upstairs. They’d have to come in and air their differences before the corporate management committee, which did not suffer indecisiveness gladly. This was enough to make our executives hammer out all but the thorniest problems among themselves, while it forced major problems to surface quickly for top management to see.

  With so much contention built in, why did the new system work? For one thing, every IBM employee had job security, dating back to the days when Dad had refused to fire people during the Depression. If a man proved ineffectual at his new assignment, he wasn’t going to be put out on the street; instead we’d reassign him to a level where he could perform well. In doing this we would sometimes strip a man of a fair amount of dignity, but we would then make a great effort to rebuild his self-respect. We also abided by the IBM custom of promotion from within. As inexperienced as our executives might be in their new jobs, they had all come up from the bottom and knew what IBM stood for as well as they knew their own names.

  Money was another reason the contention system worked. On one occasion I was ranting in a management committee meeting about something—I hope something worthwhile—and a rather sarcastic and outspoken fellow from my office named Tom Buckley leaned over to Spike Beitzel, later a senior vice president, and asked, “Do you know why they all take this bunk from him?” Beitzel, who was anxious not to attract my attention just then, shrugged. Buckley said, “Because they’re all getting filthy rich!”

  I went out of my way to reward the people IBM counted on most. Not only were there lucrative stock options for those near the top, but before 1956 was over I met with the board and quietly set up a unique incentive plan. It involved my father’s pay, which now of course was my pay. He’d always taken a salary plus a percentage of the profits, and when he died, I’d made a great point of fighting to keep that arrangement intact for myself. But after Williamsburg, it no longer fit. If anyone had a right to a percentage of the profits, it wasn’t Watson but the Watson team. My cut of IBM’s profits that year was one fourth of one percent after dividends, or $298,000. I split it with Williams and LaMotte, letting them apportion the money in a way they decided was fair, and it worked out so that my own pay was only two thirds of what it would have been. In subsequent years we took the same money and distributed even more broadly, to the thirteen top men. This arrangement had a powerful symbolic effect, because it proved that the one-man show was over.

  From then on I managed IBM with a team of fifteen or twenty senior executives. Some of these men were my friends, but I never hesitated to promote people I didn’t like. The comfortable assistant, the nice guy you like to go on fishing trips with, is a great pitfall in management. Instead I was always looking for sharp, scratchy, harsh, almost unpleasant guys who could see and tell me about things as they really were. If you can get enough of those around you, and have patience enough to hear them out, there is no limit to where you can go. My most important contribution to the company was my ability to pick strong and intelligent men for these slots and then hold the team together by persuasion, by apologies, by speeches, by discipline, by chats with their wives, by thoughtfulness when they were sick or involved in accidents, and by using every tool at my command to make each man think I was a decent guy. I knew I couldn’t match all of them intellectually, but I thought that if I used fully every capability that I had, I could stay even with them.

  I was pretty harsh and scratchy myself. I wanted all the executives of IBM to feel the urgency I felt; whatever they did, it was never enough. I was a volatile leader, perhaps even more volatile than Dad, and I justified this by telling myself that I was never harder on any of my men than he had been on me. Only gradually did I learn the virtue of restraint. We had an executive in those days named Dave Moore, a ruddy-faced, two-fisted guy I’d known growing up. Our families had been friendly. Like his father before him, Dave was manager of our International Time Recording division, which mainly sold factory time clocks and went all the way back to the company’s origin. Up through the 1930s it had been one of IBM’s strongest units. But after the war, when the idea of looser, more flexible factory management came along, the time clock became a symbol of the sweatshop. Not only was this bad for IBM’s image, but a lot of companies simply quit using them. The division became stagnant and its profits fell because there were a lot of time clock makers and they were all competing for a smaller and smaller pie.

  Moore did his best to turn the situation around, but year after year the results were poor. Finally Williams and I decided Moore had to be replaced. We promised him an equivalent job at the same salary, but Dave was stunned. He sought out his immediate boss, Red LaMotte, and said, “Well, I’ve made a success of the time division, haven’t I?” Of course, if it had been a success we wouldn’t have taken him out of the job, but Red equivocated, being a gentle man. When I heard of their conversation I should have just let the thing ride. But I always lost my temper when an IBM man refused to face a problem squarely. I called Moore in and said, “I don’t know how the hell you can say things like that, Dave. Let’s be realistic. The time equipment business is going nowhere. How can you call that a success? You can say you did as well as you could under the circumstances. But did you make money? Did your division move ahead? No, you were not a success.” He left the meeting and my words festered in him. After a few months he quit, sold all his IBM stock, and took a civilian job in the Air Force. I’ve always regretted driving Moore out—especially since, in retrospect, it’s clear he had an impossible job. His successor couldn’t turn the business around either, and eventually the unit was sold.

  Gradually I learned to control myself better. But it would have been a fatal mistake to expect perfect harmony at IBM. You can’t run a business well simply by announcing,
“Tomorrow we’re going to do A and Friday we’re going to do B and next year we’re going to do C.” The best way to motivate people is to pit them against one another, and I was constantly looking for ways to stir up internal competition. This led me, a few years after Williamsburg, to one of my most controversial decisions. In spite of all our efforts at decentralization, the Data Processing Division, which was responsible for all our computers and punch-card machines, was just too big. It was in danger of becoming clumsy and bureaucratic. So I assigned a task force to figure out a way to break it into manageable parts. They came back to me after a month and said, “It’s impossible. It’s all one business. There’s no way to divide it.”

  “All right,” I said. “I’ll do it for you. All products that rent for over ten thousand dollars a month will belong to one division, and all products under ten thousand dollars will belong to another division.” I picked ten thousand dollars because that was our average rental price, and it made the two camps roughly equal, with about thirty thousand employees each. But apart from that, it was a very awkward split. It essentially broke our computer business into competing halves. The corporate staff had a terrible time figuring out how to divvy up our laboratories, factories, and sales force between the two new units.

  I didn’t think we were smart enough to manage IBM any other way. From then on we didn’t have to depend on the brilliance of top management to decide what needed to be changed; when you have separate units competing against one another, to a large extent they discipline themselves. The new arrangement put the hot breath of competition into every division executive’s ear and we could gauge IBM’s efficiency by comparing one division with the other. I could go to the head of the Data Systems Division and instead of saying, “Gee, are you sure your overhead cost is reasonable?” I could say, “How come the overhead in the General Products Division is lower than yours?” Or: “Why does it take you four years to develop a computer while it only takes General Products two years to develop a roughly comparable machine?” Much of the time it wasn’t even necessary to ask. Procter & Gamble was doing something like this with consumer products—they’d develop two or three brands of detergent and let them compete against each other in the grocery stores—and of course General Motors sold several different lines of cars. But it was radical thinking to apply internal conflict to the degree IBM did. A lot of people told me it would never work, but it was one of the secrets of how we got as big as we did.