Father, Son & Co.

by Thomas J. Watson

  I was beginning to learn that the majority, even the majority of top performers, are never the ones to ask when you need to make a move. You’ve got to feel what’s going on in the world and then make the move yourself. It’s purely visceral. I didn’t trust myself enough yet to insist, but I knew in my gut that we had to get into computers and magnetic tape. To my great surprise, this was echoed by one man at headquarters who kept prodding me to act: Kirk’s old crony Birkenstock. Talking him out of quitting after Kirk died was one of the best moves I ever made, because Birkenstock did more to put IBM into the computer business than any other man. After his demotion from general sales manager, he’d been sidelined in a little department called Future Demands. Its job was to help fine-tune the product line by keeping track of customer requests. It took Birkenstock only a few months to transform Future Demands into a watchdog of IBM’s future. He was no engineer himself, but he had a natural understanding of technical matters and an ability to articulate them.

  Birkenstock was constantly telling me that the punch card was doomed, and so were we unless we woke up. Customers wanted more speed, and we were reaching the limit of how fast our machines could go. When we made our punches punch faster, they wore out quicker; when we pushed our high-speed sorter up from six hundred cards a minute to eight hundred, the cards themselves started getting ripped to shreds. He was constantly goading me, calling my attention to all the activity in electronics, and asking if I really wanted to miss out. This was nerve-wracking to listen to all the time because I didn’t know how I’d convince my father. But I knew I’d be a fool to close my ears.

  Not long afterward I reached the conclusion that the smart way to protect our future would be to hire electronics engineers—large numbers of electronics engineers. Whether we ended up trying to commercialize computers and magnetic tape or not, IBM needed to understand what was going on; the field of electronics was advancing so fast, on so many fronts, that I thought a small group would never keep up. We needed a critical mass. But this was before Dad made me executive vice president, and nobody in his research and development operation would listen to me. IBM’s main laboratory, on North Street in Endicott, was a very peculiar place. Between three hundred and four hundred people worked there, but the whole thing was built around seven senior engineers whom Dad called his “inventors.” They were mostly self-taught and they’d been with T. J. for decades. Even though there was a laboratory manager and a vice president of engineering, these inventors reported directly to Dad. He was really the chief engineer. When he had an idea for a product, he’d call in one or two of these old birds and describe what he wanted it to do. Then the inventors would go back and try to “put it in metal” as they used to say. Each one had his own workshop and assistants, and Dad liked to make two or more of these teams compete when he had a technical problem that absolutely had to be solved. “Nobody is smart enough to determine such things in advance,” he used to tell me. It was an expensive but very effective way to develop products, and later on I used it myself.

  Unfortunately, none of Dad’s inventors understood electronics. The man who designed the SSEC, for example, didn’t know how to hook up vacuum tubes, and he literally had to go out and hire young engineers two and three years out of school to do the job. The inventors had been so successful for so long that they were set in their ways, and they pooh-poohed my concern about our lack of electronics expertise. The vice president in charge of engineering became a real thorn in my side. He was a Princeton graduate with a degree in electrical engineering who had been with us since 1930. He was extremely creative in his way, and had been a great success at promoting the use of IBM equipment in the world of scientific computation. It turned out that punch-card machines were useful on small scientific problems, and one of our punch-card machines was so popular among scientists that it was called the “poor man’s ENIAC.” The trouble with this vice president was that he was more engineer than executive: although he was interested in electronics, he never really understood that pursuing it adequately would require a fundamental change in direction for IBM.

  The only electronics experts we had on staff worked far outside the company mainstream. Their lab was in Poughkeepsie, in an old country mansion overlooking the Hudson, and they had to share it with engineers from our typewriter plant nearby. Most of the electronics men had wartime experience on things like radar. Their boss, Ralph Palmer, had worked on top-secret electronic circuit projects for what later became the National Security Agency. In 1947 and 1948 the main project at Poughkeepsie was improving my pet machine, the 603. The group had also started to experiment with UNIVAC-style computing and magnetic tape, but when Palmer asked for manpower and funding to expand this effort, the vice president of engineering told him no. He thought such work would merely soak up resources that the main lab in Endicott could use for improving our punch-card line. The boys at Endicott had high hopes for a number of projects. They had a new line of machines in the works that could use giant punch cards. These were twice the size of ordinary cards, could hold much more information, and were being counted on to help keep customers like Time Inc. loyal.

  IBM had so much built-in resistance to exploring electronic computing that we might have been better off simply buying out Eckert and Mauchly. Ironically, we were offered the chance to do just that. In 1949 their principal backer got killed in a plane crash and before long the two inventors ran out of money and came to Dad. I was in his office for their visit. I was curious about Mauchly, whom I’d never met. He turned out to be a lanky character who dressed sloppily and liked to flout convention. Eckert, by contrast, was very neat. When they came in, Mauchly slumped down on the couch and put his feet up on the coffee table—damned if he was going to show any respect for my father. Eckert started describing what they’d accomplished. But Dad had already guessed the reason for their visit, and our lawyers had told him that buying their company was out of the question. UNIVAC was one of the few competitors we had, and antitrust law said we couldn’t take them over. So Dad told Eckert, “I shouldn’t allow you to go on too far. We cannot make any kind of arrangement with you, and it would be unfair to let you think we could. Legally we’ve been told we can’t do it.” Eckert understood perfectly. He leaped to his feet and said, “All the same, thanks very much for your time.” Mauchly never said a word; he slouched out the door after an erect Eckert. A couple of months later came an announcement that they’d been bought out by Jim Rand. Remington Rand, our old rival in punch cards, was now getting ready to fill orders for six UNIVACs.

  It was the success of the 604 Electronic Calculator that convinced me that electronics was going to grow much faster than anyone had anticipated. Like its predecessor the 603, this machine was designed to fit into an ordinary punch card installation—right down to its clumsy, Victorian-looking black metal case. But there was nothing old-fashioned underneath that shell. Palmer and his men had produced an amazingly elegant design that made it easy to cope with vacuum tubes, which were constantly burning out or otherwise going haywire. They mounted each tube and its supporting circuits in a standard plug-in unit that could be mass-produced at very low cost. That way, every tube could be thoroughly tested before it ever went into a machine; if it went bad while the machine was in use, it could be easily replaced. The design also made it simple for us to increase production when customers started snapping up the 604.

  When we first brought it out in mid-1948, we expected to sell a few hundred over the machine’s lifetime. But by the end of 1949 we’d already installed almost three hundred and demand clearly was going into the thousands. The 604 rented for $550 a month, about the same as a sophisticated electromechanical tabulator, but it could handle division and other jobs that were almost prohibitively expensive to do mechanically.

  But Palmer’s operation was an anomaly. I did a complete review of all our development projects once I became executive vice president in 1949, and decided that IBM was still in the Dark Ages. I finally told Dad we needed
something different. By then I was terribly frustrated, so I criticized his organization in Endicott in the roughest possible way. I said to him, “All you’ve got up there is a bunch of monkey-wrench engineers. Don’t you see? The time for hacking machines out of metal is gone. Now you’re getting into a field where you have to use oscilloscopes and understand the theory of electron streams and scanning beams inside the tubes. You’ve got to do theoretical things, you’ve got to do them with able people, with different backgrounds from the people we now have. You’ve got to hire engineering graduates—a lot of them.”

  I was attacking a source of the old man’s pride. He was always telling people that IBM had the greatest engineering department in the world and that he put all his faith in engineering. But he didn’t answer me directly. Instead he buzzed his secretary and asked for the vice president of engineering to be sent down. He came about two minutes later and Dad said to him, “My son tells me that we don’t have any kind of research organization. Is that true?”

  The vice president thought awhile. He was slow in the way he gave his answer; he wasn’t being glib. He said, “We have the finest research organization in the world.” That was the end of him as far as I was concerned. All businessmen get asked a question like that sometime in their lives. They either answer it with courage and get fired or promoted, or they answer like a patsy. He’d just made a major mistake, because Dad wasn’t going to live very long and I was never going to want the vice president around me again. If I knew we lacked the organization we needed, he should have known it too. He knew research and he knew engineering and he ought to have been able to see enough of the future to understand we were in a bad spot. I think he was afraid to push Dad very hard; maybe he sensed that Dad didn’t want him to. But there are a lot of ways he could have hedged it. He could have said, “Mr. Watson, you’re both right. We have a superior engineering department for what we’ve built until now. But we’re going to have to move into vacuum tubes and electronic circuits and we’ve got almost nobody who knows about them.”

  I don’t know how long Dad and I would have stayed at an impasse on engineering if it hadn’t been for Al Williams. At that point he was still in finance, and he did a study comparing IBM’s research-and-development spending with that of RCA, General Electric, and other successful companies. It showed us falling behind. On average, the others were spending three percent of revenue—three dollars out of every hundred dollars in sales flowing back into R and D. We were spending about two dollars and a quarter.

  Williams took these figures to Dad. “Mr. Watson,” he said, “I don’t know if you’re aware of it or not, but we are slipping back on research.” Dad was totally noncommittal. But the next day he called a meeting of the executives and said, “Now, gentlemen, I’ve been thinking about our efforts in research, and we’re not putting enough into that. I want you to go out and build this up. Now Mr. Williams—Mr. Moneybags over there—may complain to you about the cost. But don’t let him stop you. I want you to build this research up.”

  Williams had acted without my knowledge, but I was delighted. Dad knew full well that this meant a major expansion in electronics. I looked across the room at the vice president of engineering and started thinking about who we had to replace him. Palmer wasn’t the right choice to manage the expansion—putting the leader of the Poughkeepsie mavericks in charge would be an unnecessary insult to Dad’s inventors in Endicott. Besides, Palmer was simply too important in the lab. It took me several days to settle on my choice: Wally McDowell, the head of the Endicott lab. McDowell had some vague connection to Dad because his father was a doctor in the town of Corning, about forty miles from where Dad grew up. But he was also an MIT graduate—one of the few we had at that stage—and while he had come to work at IBM in the same year as the vice president of engineering and had the respect of the old guard, he struck me as more farsighted. Even though he had no experience with hiring electronics engineers, I thought he could do it well.

  There was no formal procedure for getting McDowell promoted or the other man removed; I just waited for the right moment to ask Dad. It came in May 1950, when we were in Endicott for the Hundred Percent Club convention. There was a company field day, and Dad and I went down to the IBM country club to watch the sports. I spotted McDowell near the tennis courts and said to my father, “We’ve really got to move in expanding our research program, Dad. I think we ought to start with Wally. We don’t have anybody else. He’s got a degree from MIT, that’s something, and I don’t think the man we’ve got in the job has the energy and he doesn’t really see the need.”

  Dad had apparently thought it through and decided for himself that a change was necessary. He said, “That’s a good idea. Why don’t you go over and talk to him?”

  I walked around the tennis court, got hold of McDowell, and asked him if he’d move down to New York and hire engineers in quantity. “What do you mean by ‘quantity’?” he said. “A few dozen? I could do that from up here.”

  “No, I mean at least a few hundred and perhaps a few thousand.”

  Wally was surprised, but he agreed and we made him director of engineering. His predecessor stayed on as a vice president; ironically, having the burden of responsibility taken away seemed to free him up, and he became a strong and effective advocate of electronics research, particularly in the area of transistors. McDowell, meanwhile, got on with the hiring. It wasn’t easy to attract top electronics people at first because we didn’t have anything in place. We couldn’t say, “Come see what we have.” Our pitch was more like, “Come hear what we’re gonna do.” But we hired plenty because I told McDowell to take any man who had even a reasonable chance of being good. I didn’t give a damn where they came from. So he gathered all sorts in his net—Americans and Europeans and Egyptians and Indians. Some of the ones from very different cultures did badly—we had a higher percentage of successes with the British and Americans. But until we had a critical mass, that didn’t matter. The problems in electronic computing were so diverse and so vast that we couldn’t do anything without enough people on the job. Once we had them, we could sort them out. Thanks to McDowell’s efforts, we grew from five hundred engineers and technicians to well over four thousand in the space of six years.

  Our veteran sales executives and planners watched this wave of immigration with extreme skepticism. These were tough marketing men, the core of IBM, and with punch-card machines selling like hotcakes, it made no sense to them that we should be scrambling into electronics. They had nicknames for the new MIT types we were bringing in. They called them “double domes” and “long hairs.” But Dad let us do our thing, although I later found out he was watching very closely indeed. When Wally McDowell first moved down to New York, a messenger would show up at his desk each morning at 11:30 with two box lunches. At 11:35, Dad would come in, sit down, unwrap a sandwich, and start asking Wally questions. I never knew about it at the time, but that went on almost every day for months.

  We finally started producing computers after the outbreak of the Korean War. That was in June 1950, while Father was in Europe, busy setting up World Trade. He cabled President Truman, putting the resources of IBM at the government’s disposal and naming me as the man to contact. I assigned Birkenstock to go to Washington and find out what we could do. I knew he would volunteer IBM to build a computer for the war effort, and I thought that was fine. It seemed to me that if we could build a couple of one-of-a-kind machines under government contracts, we’d have a way of getting our feet wet. Birkenstock spent the fall of 1950 knocking on doors at the Pentagon and traveling around visiting government labs and defense contractors to ask about their computing needs. With him he took Cuthbert Hurd, a mathematician who had joined us in 1949 from the Atomic Energy Commission’s laboratory at Oak Ridge. Hurd knew all about scientific computing—in fact, he had helped spearhead the drive to get punch-card machines into engineering laboratories and workshops.

  They looked into a lot of defense-related field
s—atomic energy, guided missiles, cryptanalysis, weather forecasting, war games, and so on. What they found was that engineers and scientists were desperate for computing power. There was a terrible sense of urgency in America in those days because of Korea. Only five years had gone by since the end of World War II, but we had thoroughly demobilized—our soldiers were back in civilian life, our fleets were in mothballs, our tanks and bombers had been scrapped. When Korea tied up the only standing army we had, people worried that Russia would take advantage and attack in Europe or some other part of the world. So in the fall of 1950 there was a frantic rush to remobilize and rearm. As Birkenstock and Hurd traveled around, the situation in Korea kept getting worse—in November the Chinese launched a bloody attack that drove our army out of North Korea and down into the bottom one third of the peninsula.

  In the context of those grim days, Birkenstock and Hurd came up with a plan much bolder than I expected. They said we should build a general-purpose scientific computer to work in all the defense applications they’d studied. “Probably it won’t solve a hundred percent of anybody’s problems, but it will solve ninety percent of them,” Hurd told me. He thought we could find customers for as many as thirty machines.