8 SYSTEM 360: ONE OF THE GREATEST PRODUCTS IN HISTORY?
Today we are making the most significant product announcement in IBM history—the System/360.… The day so many of you have worked so hard for has now arrived.
—THOMAS J. WATSON JR., APRIL 7, 19641
IBM’S SYSTEM 360 was one of the most important products introduced by a U.S. corporation in the twentieth century, and it nearly broke IBM. A short list of the most transformative products of the past century would include it, competing with electricity and the light bulb and with Ford’s Model T car. IBM’s System 360 changed its industry and helped the world transform how it worked, making new tasks possible, and enhancing the productivity of companies and governments. But that was all in the future. In the early 1960s, System 360 was one of the scariest dramas in American business history. It took a nearly fanatical commitment—consensus—at all levels of IBM to bring forth this remarkable collection of machines and software.2 In hindsight, it seemed a sloppy and ill-advised endeavor, chaotic in execution yet brilliantly successful. Tom Watson’s father had made “bet your company” decisions, such as his commitment to tabulating equipment and to keep expanding through the Great Depression, and now it was the son’s turn. IBM could take on the System 360 because built into its culture was the knowledge that such gambles had to be made from time to time, and it was able to do so. Large bets on a company’s future were never welcome events. IBM’s CEOs preferred to ride waves of existing successful strategies, introducing incremental changes. The success of a bet depended on how quickly it was placed and how well it was implemented, in order for it to be cost-effective and timely to market. Executives understood their market realities and the strengths and weaknesses of IBM’s problems, strategies, and capabilities. Most got it right, or at least right enough.3
The story was about an old issue that kept coming up. In the 1920s and 1930s, there were internal disagreements about such bets, and the 1960s proved no different. While IBM CEOs solicited advice from all around them, had staff scurrying to collect market data, and sought consensus within their inner circle of advisers, ultimately everyone around a conference table looked to the CEO to say “let’s go with it.” Consequences of that decision were dealt with by thousands of employees committed to the decision or ordered to execute it, from dealing with the media to facing antitrust lawyers, from customers needing to be persuaded to embrace the System 360 to manufacturing the machines, writing software for them, and installing and maintaining them. Each of IBM’s CEOs faced the gaze of their senior executives waiting for the decision. Those were some of the loneliest moments in the life of any CEO, but they were also some of the most important ones.
It would be difficult to find any historian who would challenge the importance of the System 360.4 This one part of IBM’s long history has led to so much commentary that it is easy to lose sight of the forest because of the many trees. We can clarify our understanding about System 360 by limiting the discussion to this one chapter, because it is essential that we appreciate the transformative nature of this product on IBM and the computer industry at large. To keep from being mired in the details or having System 360 dominate the company’s larger history, several key messages need highlighting. First, this was an important product and chapter in IBM’s history. Second, in contrast to what historians of technology claim, while the technological innovations were important, how they were created and deployed bordered on disaster. We live in an age that celebrates innovation, so understanding how that is done remains a high priority for scholars in many disciplines. As William Lazonick puts it, “What makes a firm innovative? How have the characteristics of innovative firms changed over time?”5
Third, IBM’s experience presented here is of people clashing and collaborating in a rapidly growing company with unstable, and in some instances unknown, technologies. Keith Pavitt calls this behavior “tribal warfare,” and a growing number of scholars would agree with his notion that corporations can either innovate or become rigid. We understand the features of either behavior, thanks to Alfred D. Chandler Jr.’s typology and notions of “paths of learning.”6 His notion of the role of professional management, clearly evident in this chapter, extended to engineering and manufacturing.7 IBM illustrates that internal warfare was not just a behavior of one company versus another but can be better understood as occurring within any large enterprise. For example, we saw similar behavior between IBM’s engineering communities located in Endicott and in Poughkeepsie, and between sales and Poughkeepsie’s management in the 1950s and early 1960s. That behavior continued during the development and diffusion of the System 360 as uncertainty and ambiguity dogged the footsteps of all the protagonists in this drama. This chapter, however, departs from a recent trend in business history in that it does not discuss the S/360 through some broad interdisciplinary lens, with one possible exception: I present the story through the eyes of management, departing purposely, if partially, from the technological focus that engineers took when studying the S/360.8 Nevertheless, the chapter describes how IBM relied on its own employees to develop and service this new product line.9
It may seem that major actions were undertaken almost oblivious to external events, which is only true at some tactical level in the organization, but it is also misleading since IBM’s development of the S/360 was in response to customer demands and with an observant eye cast toward what its competitors were doing or might do.10 IBM’s experience demonstrated that its engineers operated within the confines of the firm, because IBM was big and diverse enough in talent, staffing, financing, and materiel while at the same time cognizant of external realities. IBM succeeded with the S/360 because in an almost entrepreneurial fashion it took advantage of emerging technologies about which it had knowledge, no matter where they were located within the enterprise. That experience contributes another case to a growing list that historian Geoffrey Jones argues function regardless of geographic roots, being based more on its capabilities.11 The one qualification to his thesis is that in IBM’s case all its leaders involved with System 360 worked in the United States.
We deal with six related issues. We need to understand the business and technical problems customers and IBM faced that motivated the company to develop the System 360. Second, once development had begun, the engineering and technological events unfolded as much as politics as they did as technical issues. Third, contrary to IBM mythology and many reports long after the fact, when this family of computers was introduced, customers took a cautionary “wait-and-see” attitude toward them. Fourth, I describe how IBM responded to myriad problems faced by the firm and its customers after initial deliveries, because those problems also, not just those that surfaced during the rush to develop new machines and software, nearly sank the company. I summarize the consequences of this product, which justifies the accolades it received from two generations of business managers and academics. As in earlier chapters, we end with an assessment of business results.
IBM’S BIG PROBLEM
By the end of the 1950s, computer users faced a problem that had never been seen before. Until it appeared in the data centers of large companies and government agencies, it was merely theoretical, hardly anticipated, not real. Nevertheless, it was such an intractable problem that if not solved it could have prevented computers from becoming widespread. It might have led IBM and its competitors to grow slowly, probably at 1–2 percent a year instead of by an order of magnitude or two over the next decade. It seemed that all other computer-related problems could be traced to this one. If it were not solved, thoughts of living in an Information Age would have been fiction. So, what was this key problem?
Customers acquired computers from vendors in great numbers in the late 1950s, turning the corner on automating many of the old punch card processes and doing more with data processing. The popularity of the IBM 1401 illustrates the rapid adoption of computing. Introduced in 1959 and sold until 1971, by the time it was over, IBM had installed over 12,000 of these systems. By 1961, one-fourth of all installed computers in the United States were 1400s (2,000 of them), and over 10,000 had been installed by the mid-1960s. Between 1960 and 1966, the company shipped four-fifths of the value of all computers. IBM’s profits tripled thanks to these systems, so any problems with the 1401 were serious. One of them was that 1401s, like their rivals, were too small.
Users found these machines so useful that they kept piling more work on them, reaching their system’s capacity. In some cases, data centers acquired an additional one or more, which they parked near their first one or in another data center. Thousands of customers in that situation had several options: move to a bigger system, such as an IBM 7000, install a competitor’s system, or acquire more 1401s. They had the same problem with all vendors’ equipment. None of these options was attractive. To change systems to get more computing power required rewriting software, since what they had would not work on a different type of machine. The cost of rewriting could easily exceed the financial benefits of moving to a bigger machine. Such a change also called for retraining or hiring new staff familiar with the new system. The alternative of adding more units of the same system was equally unattractive because each required duplicate staff, equipment, and maintenance of hardware and software.
None offered the benefits of a bigger system. For example, if a company had a customer file of names and addresses used by two or more applications, such as order taking and billing, that file could not be shared across both uses of computers. If the two applications sat in two different machines, staff needed to create and maintain two sets of customer files, ensuring errors caused by redundant data and uncoordinated updates to the records. It was not uncommon for data processing personnel costs to consume 80 percent of a data center’s budget, with rental of equipment making up most of the rest. Customers wanted systems that were “upgradable” or “compatible”; that is, computers of different sizes such that as the customer’s needs grew, they could bring in larger machines and run the same software and peripheral equipment as with their smaller processor. In the 1950s and early 1960s, it was a wish, and for vendors an aspiration.
IBM had worse problems than its customers did. The 1401s were proving so popular that engineering in Endicott, which had developed the system, resisted attempts by Poughkeepsie to build large-end computers, leading to growing rivalry between the two organizations. One engineer recalled that “so intense was it that sometimes it seemed to exceed the rivalry with external competitors.”12 Systems made by Poughkeepsie would not run programs written for the 1400s, the fastest-growing part of IBM’s product line, most frequently replacing old tabulating systems. Customers wanted to move to the more expensive Poughkeepsie machines without having to go through conversions from the less-extensive small 1400s. They put increasing pressure on IBM sales, executives, and IBMers visiting SHARE meetings to provide compatibility. Senior management faced the prospect of trying to reorder their half dozen different incompatible systems. Besides the internal political and economic costs of Endicott and Poughkeepsie competing against each other were the R&D expenses of multiple product lines and the burden of training IBMers to sell and maintain so many systems. These internal struggles were not unique to IBM, as they existed in many large enterprises and in governments, such as army versus navy.
IBM needed to get down to one system to lower its costs and be more competitive against a growing array of rivals. IBM required only one to establish dominance—a monopoly—in the market for computers. That last aspiration, better known as “lock in,” meant getting customers to use one system, so when they needed more horsepower they could swap out one computer, or type of peripheral equipment, for another from the same vendor. When this goal was finally achieved, IBM’s customers traded out machines over a weekend. IBM wanted scalability to simplify production and to reduce the cost of R&D. Competitive pressures weighed on the divisions. If a customer had to upgrade in the early 1960s, they could just as easily move to a competitor’s machine, since they would have to rewrite their software anyway. Attractive alternatives existed that sometimes were cheaper, faster, or better, such as machines from GE, Honeywell, and Sperry Rand.
In addition to these concerns, all intimately familiar to Watson Jr. and thousands of employees, he had another. He recalled in his memoirs that “paradoxically, there also was a feeling in the early ’60s that IBM had reached a plateau. We were still expanding, but less quickly than before—in the year Kennedy beat Nixon, for example [1960], we only grew nine percent. As we reached the two-billion-dollars-a-year mark people began to speculate that we’d gotten so big that naturally our growth rate had to fall.” Watson thought “that seemed illogical.” The two divisions he created to compete against each other raised up an ugly side effect: “Our product line became wildly disorganized,” he said, with eight computers by late 1960, plus the older tabulating products.13 How to solve IBM’s problems was not obvious, but many people in the company thought they could fix it. Consensus grew that the problem involved finding a way to make multiple systems compatible with each other so that a program running on one could operate on another with little or no modifications. Once the decision to proceed had been made, the central challenge for IBM’s management was, to use historian Steven W. Usselman’s words, “massing the resources.”14
THE DIFFICULT PATH TO APRIL 7, 1964
The power of compatibility was demonstrated in the fall of 1960, when IBM introduced the more powerful 1410 to replace the 1401. Software and peripheral equipment for the 1401 worked with the 1410. Customers and IBM sales loved that fact. Poughkeepsie’s engineers were close to completing work on a set of four computers that were also “upwardly compatible,” two for scientists and engineers and two for business users, all known as the 8000 series, to replace the 7000s. Compatibility of peripheral equipment between the 1401 and the 1410 suggested the path IBM would have to take.
Engineers already had nearly a decade of experience developing new computers. They knew that roughly half of any increased ability of a computer to process work resulted from advances in electronics, the rest by modifying computer designs to rebalance a combination of software and basic pieces of hardware. Did these tricks work? Between 1953 and 1965, IBM’s computers increased performance by 50 percent each year.15 Such improvements drove up IBM’s profits. Engineers were confident that a path forward existed, but internal politics got in the way of good engineering.
T. Vincent “Vin” (also known as “T. V.”) Learson (1912–1996), then in charge of future product development as the vice president of manufacturing and development, had to move quickly. Learson was a gifted problem solver. Watson described what happened: “He saw that it was time to break down the rivalry between the two divisions, and he did it by applying a management technique called ‘abrasive interaction.’ This means forcing people to swap sides: taking the top engineer from the small-computer division and making him boss of the best development team in the large-computer division. A lot of people thought this made about as much sense as electing Khrushchev president.”16 He replaced the Poughkeepsie manager in charge of the 8000 project with Bob O. “Boe” Evans (1927–2004). Evans had served as the engineering manager for the 1401 and 1410 in Endicott. Earlier, Evans had made it clear that he favored compatibility across all future products. In 90 days, Evans came back to Learson, recommending that work on the 8000s be stopped and that both sites should begin working “to develop a total cohesive product line.”17 To block competitors, since such a change in strategy would delay announcements of new computers, Evans recommended some upgrades to existing equipment. He also boldly proposed a new base technology for all future systems, Solid Logic Technology (SLT), to make IBM’s machines more competitive.
Frederick P. Brooks Jr. (b. 1931), who led the design team for the 8000, fought back. The two men (Evans and Brooks) were formidable opponents. Both were engineers, each with years of experience running engineering and product development activities at IBM, articulate, and highly respected by their staffs and senior management. Brooks was not as high ranking as Evans, so to constrain the former, Learson brought in as Brooks’s manager Jerrier A. Haddad (b. 1922). Learson knew Haddad because Haddad had spent the previous two years in charge of the Advanced Engineering Development Division. Learson asked him to study the proposed approaches of Evans and Brooks. Haddad recommended going with Evans’s ideas. Learson accepted his suggestion and killed the 8000 project in May 1961. Bob Evans had won. He immediately asked Brooks to develop the plan for a compatible family of computers. Brooks recalled, “To my utter amazement, Bob asked me to take charge of that job after we had been fighting for six months.”18 Brooks accepted the invitation and with that both engineering communities stopped feuding and began collaborating. There were still opponents in the company, but no matter, the trajectory toward a common system had been set.
Using a tactic similar to Evans bringing Brooks into the fold, Learson next turned his attention to John W. Haanstra (1926–1967), president of the General Products Division, which produced the 1400s. Haanstra wanted to continue riding that wave of attractive products. Learson wanted to further cement collaboration across IBM’s management, so he assigned Haanstra to chair a task force called SPREAD (Systems Programming, Research, Engineering, and Development), with Evans as vice chair. Brooks joined later as one of 11 other members. In December 1961, the task force presented its technical recommendations, the basis of which eventually became the System 360. Their report called for five compatible computers, labeled processors (defined as the computer, its memory, and channels to connect to peripheral equipment). Software and peripherals on one processor were to work with all other processors, a collection of machines that had a 2,000 range of speed and volume of performance using SLT circuits and myriad other innovations. Most of the recommendations were on the hardware side, less so on the software side, specifically regarding operating systems. The hardware-software systems had to be competitive in the market, not just benchmarked against existing IBM machines. The plan called for using standard hardware and software interfaces between computers and peripherals, such as between disk and tape drives connecting to computers, so these did not have to be swapped out when a new processor was installed.19
Because so much would be new, these processors would not be compatible with IBM’s existing products. That was an enormously important point. Customers moving to the new IBM machines would have to rewrite existing software, but only once, after getting on the path of the new system. Sales teams would find incompatibility with existing products an enormous risk because customers would naturally also consider machines offered by other vendors. IBM expected competitors to have systems competitive with its current product line by 1963, which made everyone nervous. In fact, in December 1963, Honeywell introduced its inexpensive H-200 to go after the 1401 and be compatible with it, meaning software on the IBM machine would supposedly run on Honeywell’s computer and its future systems. IBM’s salesmen reported that within two months of the announcement Honeywell had received 196 orders for the H-200.20 Evans urged not responding with an improved 1401 because then the 7000’s engineers would want to do the same for their system, slowing or stopping progress toward one unified system. The situation looked bad.
Then, “Almost miraculously his [Evans’s] vision of the new product line was saved by a last-minute technical accomplishment. In mid-1963 engineers in the Poughkeepsie and Endicott laboratories had begun exploring the possibility of adding special microcode to the control stores of computers to improve their performance when simulating earlier IBM computers.”21 I italicized part of the quotation from an IBM engineer to emphasize the lucky break. Microcode improved simulation of software from earlier computers so well that it became the silver bullet IBM needed, and even led to wide use of the word emulator. This function allowed software working on a 1401 to do so in the two smaller models of the proposed new system, emulating 1401 computing. To a user, it was the same as running actual 1401 software on the new machines. Pugh understood the attraction that salesmen would find in this lucky break: “It was a salesman’s dream come true. The new product line now offered computers that would execute a customer’s current programs faster than his present computer and that provided a path into the new world of compatible computers, capable of applications not possible with older equipment.”22 Sales now got on board, and its executives began pressuring R&D and manufacturing management for early introduction of the new processors.23
Watson recognized the gravity of what was at stake:
From the beginning we faced two risks, either of which was enough to keep us awake at night. First there was the task of coordinating the hardware and software design work for the new line. We had engineering teams all over America and Europe working simultaneously on six new processors and dozens of new peripherals … but in the end all of this hardware would have to plug together. The software was a bigger hurdle still. In order for System/360 to have a common personality, hundreds of programmers had to write millions of lines of computer code. Nobody had ever tackled that complex a programming job, and the engineers were under great pressure to get it done.24
A second set of problems involved IBM manufacturing its own electronic components for the new systems. The electronics industry was about to start making integrated circuits—computer chips—and the new computers were going to be filled with these components, which were new to everyone. To be independent and not rely on others, IBM had to make its own. As Al Williams told Watson, “If we are going to stay in the computer business, we’d better learn how to make these things ourselves.”25 It proved to be an expensive proposition. In those days, a computer factory cost about $40 per square foot to build, and a dust-free integrated circuit factory cost over $150 per square foot. Once the project started, the board of directors began giving Watson a hard time, much as it did to his father in the 1930s when he invested in new products. Eventually, the corporate management committee, including Watson and the board of directors, sucked in a deep breath and approved the SPREAD recommendations. Watson made the final decision on March 18. IBM was off to the races in the wildest ride of its history. The easy part was telling the world what was going on, but that, too, was high drama.
THE BIG DAY AND HOW CUSTOMERS RESPONDED
IBM could not hide what was going on. New employees flocked to Endicott, Poughkeepsie, and other labs and plants. IBM labs on both sides of the Atlantic were abuzz. Customers heard rumors, salesmen were spreading them too, the computer press was speculating, and executives at GE, Honeywell, Sperry Univac, and elsewhere were trying to anticipate what IBM would do and how they would react to it. At IBM, nobody seemed comfortable with progress on the new system, and April 7, 1964, the day the world would meet the System 360, had not even arrived yet. Engineering, manufacturing, sales, and Corporate were still debating, literally losing sleep, in many cases working 100-hour weeks to make the deadline. Engineers moved cots into their offices. One engineer yelled at Watson, who stopped in to see how programming was going, telling him to get out of his office so he could work. The chairman of IBM beat a hasty retreat.
It all became public at noon eastern time in the United States on April 7, 1964. In the United States, over 100,000 customers, reporters, and technologists met at the same time in 165 cities, while others gathered around the world over the next few days to hear the same news. Watson must have felt elation and relief that he could stand up and declare the obvious, that this was “the most important product announcement in the company’s history.” In the years it took to get to that announcement and in the several following it, he wondered whether the company could pull it off: make the equipment work, and deliver it when promised to customers. Years later, when asked whether IBM would ever engage in such a massive project again, one executive barked out, “Hell no, never again.”26 Watson tilted toward a similar reaction, commenting in 1966 that, “At our size, we can’t go 100 percent with anything new again,” meaning anything that big. He opined privately to his managers that, “There is no question that we cannot go through another announcement like 360 where we obsolete virtually our entire installed revenue base at one time and where we commit a very substantial portion of our total production to a new technology.” He made it a policy “never to announce a new technology which will require us to devote more than 25% of our production to that technology.”27 The generation that brought out the S/360 remained unique in the company, a special clan, bonded ferociously to IBM. While their experience was not entirely a positive one, it generated an endless supply of memories and company lore.
On April 7, 1964, IBM introduced a combination of six computers; dozens of items of peripheral equipment, such as tape drives, disk drives, printers, and control units, among others; and a promise to provide the software necessary to make everything work together—a mindboggling total of 150 products. The breadth of the announcement is hard to imagine today. Then as now, a computer company introduced one or a few products at a time, such as Apple does when it brings out a new generation of 3 or 4 compatible smartphones. IBM’s press packet was an inch thick, the internal announcement materials—known as Blue Letters—the thickest ever, and manuals describing all the machines, components, software, and their installation and operation filled more than 50 linear feet of bookshelves.
New technologies in the machines, in combination with existing ones, sped up processing, made it possible to handle larger volumes of information, and handled more complex business and scientific applications. The new architecture was significantly different from what had existed before. An analogy suggests the nature of that difference. Think of an architecture that might differentiate a Victorian house of 1890 from a split-level ranch in suburbia of the 1960s. Like the ranch house, the Victorian involved new building materials, rooms laid out differently, and included electricity and water heaters, and that made it possible for owners to experience different lifestyles than their grandparents had. The IBM announcement promised a 50-fold range in performance from the smallest processor to the largest, a commitment IBM was able to deliver. While this might sound less dramatic than the jump from one to another achieved by earlier systems, System 360’s capacities were much larger than those of earlier systems, and the performance range from one system to another was even greater within this family. Forty-four peripheral devices could attach to these computers. Historians sometimes confuse how many processors were introduced so, to be specific, the computers were System 360 Models 30, 40, 50, 60, 62, and 70, for a total of six. Before IBM delivered Models 60, 62, and 70, the better-performing Models 65 and 75 had replaced them.
The central feature of the System 360 was its compatibility with other models. A growing data center could install a small System 360 computer and when it needed more computing power upgrade to a larger one without rewriting software or replacing peripheral equipment. Once familiar with the System 360, its peripherals, and its software, one did not have to learn a great deal more to handle an upgrade to a new model. The name 360 was chosen to suggest an analogy with the idea of 360 degrees, covering everything; it had nothing to do with the decade of the 1960s. IBM published fisheye photographs (figure 8.1) of the new system to reinforce the point of comprehensiveness. The first deliveries of the smaller machines were promised for the third quarter of 1965 and deliveries of the larger ones in the first quarter of 1966. All could be purchased or leased; much software continued to be bundled as part of the hardware, although private software developers could sell or lease software that ran on these IBM systems.
This image of the System 360—known as the fisheye 360—was widely used in IBM’s advertising and marketing materials. Photo courtesy of IBM Corporate Archives.
As was normal practice, the sales force was briefed on the announcement the day before it was made public. Salesmen for large systems, housed in the Data Processing Division (DPD), heard that the S/360 was “one system” for “all customers.” They were told that, “Whatever your customer’s data handling requirements are, whatever they will be in the foreseeable future, the System/360 can be custom-fitted to perform his job. In fact, this amazing new system makes possible, for the first time in the industry, a truly long-range growth plan for all customers.” They were briefed about the system’s flexibility and numerous choices of configurations. They were promised that additional information would arrive in their offices that week and that they would receive training about the “one system in many sizes,” which could be customized and open up new markets, creating “new customers” while expanding “existing applications.” They were urged to “reach out for new and additional applications to be integrated with current ones for a System/360 total operating system. For everyone.”28 This was all sweet music to a salesman.
Users knew they had a potential game changer on their hands. It was the combination of compatibility and modularity that drew their attention. As one economist touting IBM’s new system put it, “The 360 system will be modular; and accordingly could be expanded, or contracted, without need for reprogramming,” adding, “The system will also be compatible with ‘most’ programs of the 1401 system—now IBM’s most popular computer.”29 Users quickly understood the significance of the product, but they prudently took a wait-and-see attitude until IBM delivered on its promised shipping dates, while all of IBM’s competitors went home to figure out how to respond. In a single stroke, IBM had wiped out demand for its current product line and ultimately that of most of its rivals. For a while, competitors sold their existing computers using the old sales trick of FUD, sowing Fear, Uncertainty, and Doubt among their customers. They had little choice, because they needed to introduce upward-compatible systems of their own and price them competitively with IBM’s. The computer industry had changed.
Initial press coverage of the System 360’s release offered muted reaction but nonetheless confirmed something big had happened. “The internal IBM reaction could be characterized as quiet, smug elation.” One competitor was quoted as saying, “The 360 looks like a clean machine and it took a great deal of courage for IBM to break with its past product line,” because the courage involved “opening the door to that chamber of expense horrors, reprogramming.” Datamation, the leading computer magazine of the day, argued that when IBM makes an announcement it “creates an automatic bandwagon effect.” The magazine reported a range of reactions from delight to criticism, with concerns about reprogramming in the air as more than a whiff, despite the attraction of compatibility. The economics seemed to work on announcement day, however. One customer said that the “rental cuts of 50 percent … will force [me] down the 360 route.” An economist explained, “The crucial advantage of the System/360 is that it will offer more computer power per dollar than any system now available. This advantage derives from IBM’s success in achieving low-cost mass production of its logic circuits and its memory cores,” with “the cost advantage” backed “by IBM’s reputation, by its superb sales and service organization—more important to IBM even than its innovating capabilities—by its library of programs available, without additional charge, to users of its computer systems.”30 The announcement was an early example of IBM’s “total cost of computing” justification to customers for moving to a new product.31 Everyone wanted more information, of course, despite the fact that IBM’s marketing announcements were excellent. “They touched all the bases” while buying themselves up to two years to make good on promises and knock competitors back on their heels.32
IBMers did not know at the time of the announcement, or in the next two to three years, how extensively their products would change the world, but we do today.33 What they knew was that careers had been made and broken, personal lives and marriages disheveled. They understood that they had to overcome monumental technological problems ranging from mechanical and electronic issues to development of frightfully complex software. Customers had to be persuaded to abandon the comfort of known technologies. The company had spent vast sums, in excess of what it made in any one year, in developing this system. Then, after the system’s release, they spent several more years fixing software, spending more billions of dollars so that it would work as advertised. For “I.B.M.’s $5,000,000,000 Gamble,” it spent every penny it could get its hands on or borrow.34 Failure meant the death of IBM, and every IBMer believed it.35 Some historians thought the fear was exaggerated, but that did not matter, because the IBMers engaged in System 360’s history acted on that belief.36
IBMers got more than an inkling of the consequences of their work when “first day orders” started coming in. In this industry, a company would announce that it would start shipping a product on a certain date, such as six to nine months later for peripheral equipment, 18 to 24 months for mainframes. That delay between announcement and shipping date gave customers and suppliers time to determine which models made sense to acquire, get them approved and budgeted, allowed time for the physical planning on where to house them, make changes to air conditioners in data centers, train staff, complete software remediation, and so forth. The prerelease announcement served as a clever way to shut off competitors or, as with the S/360, pressure them to respond with their own products.37 Placing a tentative order at announcement time reserved machines, since IBM followed the practice of shipping computers to customers in the sequence in which they received orders.
In the first month following the S/360 announcement, customers worldwide ordered over 100,000 systems!38 These were “positional orders”; that is, to get in line for such systems should they need them when their IBM salesman came around 9 to 18 months later to “confirm” their order. Most initial orders were submitted configured with more memory and with a greater number of peripherals than IBM’s sales force had forecast. To put that set of orders in perspective, in that same year in the United Kingdom, all of Western Europe, the United States, and Japan, there were slightly more than 20,000 computers installed. IBM’s first orders were five times that number. Not all would be installed, of course, but by the end of 1966, IBM had shipped some 7,700. In 1965, it had shipped 668, and it would ship more in the early 1970s.39
For now, IBM had to navigate from April 7 to when it started delivering machines to customers. That is when the company entered the most dangerous, intense, and challenging era of its history up to that time. It is also when Arthur Watson was thrust into the middle of the S/360.
HOW IBM FIXED ITS PROBLEMS
IBM rushed to announcement day to fend off competition and rumors when only some of the hardware had been built, let alone tested. On April 7, to quote Watson, “Not all of the equipment on display was real; some units were just mockups made of wood. We explained that to our guests, so there was no deception. But it was a dangerous cutting of corners—not the way I think business ought to be done—and an uncomfortable reminder to me of how far we had to go before we could call the program a success.”40 Learson had gotten the S/360 to announcement day. Watson then assigned his brother, Arthur, responsibility for managing the project going forward as a way to reintegrate him into mainstream IBM and position him to someday take over the company. He would be responsible for both engineering and manufacturing. Learson would run sales for the new system, “twisting the tails of our salesmen,” as the skilled and aggressive sales leader that he was. Tom Watson Jr. thought Learson had the more difficult task, since the engineering community had momentum on the S/360. The risk of customers converting to someone else’s machines rather than to the S/360 greatly concerned Watson.
Almost immediately, software problems slowed development. Throwing more programmers at the project did not help. Brooks famously observed years later that “the bearing of a child takes nine months, no matter how many women are assigned.”41 Based on the S/360 experience, Brooks went on to write one of the most widely read books on computing still studied today.42 Battles and concerns also existed regarding components. Palmer, McDowell, and others wanted to lead the industry in components, so IBM entered the chip manufacturing business. While developing SLTs, IBM engineers learned how to make them. Developments in digital memory and storage progressed. Individual hardware problems kept emerging, which engineers resolved.
As the number of orders for all S/360 processors and peripherals kept increasing, manufacturing became nervous, especially in 1965, when they were asked to increase production by a factor of two. One production manager said it could not be done, refused to do so, and was replaced. As one engineer recalled, “The result was a near disaster.” Quality declined and, by the end of the year, the Quality Control Department had impounded 25 percent of all SLT modules, bringing production to a halt. In 1965, some 36 million were made. Problems were solved, and manufacturing proceeded in 1966, resulting in 90 million produced. A new plant just south of Poughkeepsie at East Fishkill made more semiconductor devices than all other manufacturers worldwide combined. Production expanded to new facilities in Burlington, Vermont, and Essonnes, France. Problems in manufacturing ferrite-core memories were resolved. In 1965, IBM set up a plant in Boulder, Colorado, to work on these components also, but it took the handwork of workers in Japan to get production of memories up to required amounts and quality.
Manufacturing became a worldwide effort, causing new problems in coordinating activities and fabricating machines. Arthur Watson had some experience managing IBM’s small factories outside of IBM USA but none with resolving engineering problems, let alone massively large global problems in development and manufacturing. He could do little to break through the mounting engineering and manufacturing problems. He was out of his league, as his heritage was sales and general management. His brother increasingly challenged him to resolve these problems. Meanwhile, Learson and his sales teams wanted additional improvements to the product line. Relations between Learson and Arthur completely deteriorated. Then, in October 1964, IBM had to announce significant delays in shipping products.
Tom took his brother out of his job and turned over his responsibilities to Learson, who in turn quickly brought in four engineering managers to punch through the problems. Nicknamed the “four horsemen,” they had full authority worldwide for getting the S/360 manufactured and to customers. Their collection of problems, one noted later, was “an absolute nightmare,” “a gray blur of twenty-four hour days, seven days a week—never being home.”43 In five months, they had worked out enough of the problems to start meeting delivery dates. In January 1966, Learson became president of IBM. Al Williams retired, as he and Watson had long planned. When Tom took Arthur out of his job, the latter was shunted into the role of vice chairman. Arthur’s career was broken, and he retired in 1970.
Tom Watson Jr. left a record of the sad decline of his brother’s career. “I went after my brother in the same way I’d have gone after anybody else,” he said, and they argued over progress, or rather the lack thereof. “I castigated him” and told “him he’d better make sure the 360 wasn’t going to be obsolete before it was even delivered.”44 Some of Arthur’s managers were not collaborating. Al Williams was becoming frantic because of the increasing costs of production, forcing him to borrow money and to spend every penny IBM had on the S/360. Some $150 million in parts were moving from one plant to another as part of the production process, with accounting for their costs and whereabouts out of control. To fix that problem, John Opel (1925–2011), a young executive in the Data Products Division, was brought in. He started in sales but acquired manufacturing experience and had a fine analytical mind. In the 1980s, he became CEO of IBM. Meanwhile, back in 1965–1966, IBM sold $370 million in stock to raise cash.
Tom Watson Jr. admitted to being in a nearly continuous panic from 1964 to 1966. He was also beginning to see what was happening to his family: “By now [1964–1965] I knew that my plan for bringing him [Arthur] into the domestic company had been a horrible blunder, bad for Dick’s career and for our personal relationship.” Instead of giving him an opportunity to shine, Tom had “handed him a stacked deck. He couldn’t hold his own against the demands put on him by Learson.”45 Meanwhile, “everybody was scared” that the whole S/360 initiative was going to crash.46 Tom admitted that his father’s standards and expectations played with his mind. Nevertheless, machines shipped, and field engineering somehow kept them functioning. Tom confessed how he closed out Arthur’s involvement in the S/360: “I felt nothing but shame and frustrations at the way I’d treated him.” He should have left him in World Trade. Tom added, “As it was, we remade the computer industry with the System/360, and objectively it was the greatest triumph of my business career. But whenever I look back on it, I think about my brother I injured.”47
Intractable software problems in 1964, 1965, and 1966 kept IBMers frenetically busy. As one engineering manager described the problem, “Of all the challenges presented by the System/360 announcement, none was more difficult to accomplish than providing the promised software support.”48 Without software, especially the operating system, nothing worked. The operating system had many problems, especially in its ability to multiprocess—that is, run more than one job at a time—so essential to making the S/360 fast and productive. Other software problems surfaced with telecommunications, and even with application programs. Programming support became another contentious issue. It seemed that every aspect of software had problems, with the software development staff often described as being in “disarray” as early as 1963. But Evans and others worked out the elements of the new operating system, called OS/360, and began writing software before, during, and after its introduction on April 7, 1964. “The magnitude of the task of developing the proposed operating system was grossly underestimated,” recalled Emerson Pugh.49 Fred Brooks volunteered to help, and IBM added 1,000 people to the operating system project, costing IBM more for software in one year than had been planned for the entire development of the S/360 system. The software would take years to complete, but at last it worked well enough to keep the shipping delay to one month.
It would be safe to say that almost every initial customer and their sales team encountered problems. By then, however, branch offices had been hiring “systems engineers” (SEs) to help. These were college graduates, usually with technical degrees, systems engineers who knew about software and how to “debug” their problems. They became the technical sales arm of the “account reps,” serving in that capacity for the next several decades. SEs were the employees IBM had started to hire in 1962 to assist the sales force in selling and supporting computers. They now heroically tackled S/360’s software problems. Field engineers, who installed equipment, fixed hardware problems. Salesmen calmed their customers, while branch managers worked to keep their staffs motivated and focused. S/360s were worth a couple of thousand points, and a normal complement of peripheral equipment added thousands more, so there was much to protect and so much to sell. The only good news it seemed was that sales commissions were great, bonuses increased across IBM, and stock values rose. Nearly every account of what happened with the hardware and software between 1962–1963 and the end of 1966 cites the Fortune magazine article that carried the headline “IBM’s $5,000,000,000 Gamble,” published in September 1966.50 The gamble Fortune wrote about proved to be far greater than executives originally considered.
Tom Watson Jr. told a different version: “The expense of the project was indeed staggering. We spent three-quarters of a billion just on engineering. Then we invested another four and a half billion on factories, equipment, and the rental machines themselves. We hired more than sixty thousand new employees and opened five major new plants. It was the biggest privately financed commercial project ever undertaken.”51 The number of new hires exceeded 70,000. A business strategist looking back at the history of the project may have gotten it more right than reporters or Watson did when he argued, “The monumental project was launched with less planning than a typical company payroll system would get.”52 While harsh, it was a fair judgment, which resulted from IBMers moving into unknown technological territory. Watson later confessed that some of the products announced in April 1964 should have been introduced two years later. Bob Evans referred to the S/360 project as a “you bet your company” one but “with less risk than it would have been to do anything else, or to do nothing at all.”53 The amount spent was comparable to the total revenue IBM earned in 1967, some $5.3 billion ($38 billion in today’s dollars). William Rogers, an early historian of IBM, concluded that by sticking with the project “IBM could keep most of the market to itself.”54
Despite the costs and anxiety, in 1965—the first year IBM had committed to shipping systems to customers—it managed “by some miracle” (Watson’s words) to ship hundreds of medium-sized S/360s. Their quality did not always match the original design specifications. Shortages of parts, others that did not work, and software filled with “bugs” (problems) spread to the field and many countries. Watson was amazed, because despite these problems, “customers were still ordering 360s faster than we could build them,” forcing delivery dates out as much as three years. That delay created an opening for rivals, which worried IBM’s sales executives and branch offices.55
IBM CONQUERS THE MAINFRAME MARKET
One would expect that with the churn inside IBM it would face severe problems in the marketplace. While it did, success with the S/360 papered over enough of them to give customers confidence in acquiring System 360s. By the end of 1966, customers had taken delivery of nine models of the S/360 for a total of 5,261 systems in the United States, an enormous increase in volume for any IBM system, equaling 13 percent of all computers in the United States. IBM’s backlog of unfulfilled orders accounted for 57 percent of all machines on order from all manufacturers. IBM’s dominance was immediate and obvious.56 Its competitors responded. Honeywell, Burroughs, GE, NCR, and Sperry Rand, operating largely in the United States, CII in France, and ICT (later ICI) in Great Britain introduced systems compatible with each other, much as IBM had done, but they were not compatible with IBM’s. A second, smaller group chose to manufacture machines compatible with IBM’s, including RCA and others in Europe and Japan, relying on RCA’s licenses. This second set of firms grew in importance in the 1970s when they, IBM, and their customers were acquiring post-S/360s—the S/370. A third response involved niche players not competing against the S/360s, such as for specialized uses or supercomputers.
All vendors combined in the 1960s and 1970s were often called Snow White and the Seven Dwarfs. It was a pejorative term irritating to IBM’s competitors, but it came at a time when other nicknames were being used also, such as “Big Blue” for IBM, after the color of many of the company’s machines. IBM (a.k.a. Snow White) was followed into the compatibles market by most of its rivals, which held smaller market shares but were nonetheless capable of competing for a while against Big Blue. They included Burroughs, Control Data, GE, Honeywell, NCR, RCA, and Sperry Rand. All sold large mainframes. As of 1964, they controlled 30 percent of the computer industry. The industry was valued at about $10 billion, so they owned just over $3 billion of it, IBM the rest. IBM’s S/360 proved so successful that five years later, IBM’s worldwide inventory had grown to $24 billion, with that of the Dwarfs growing to $9 billion.57
IBM’s product single-handedly grew overall demand for computing so massively that it raised all boats, its own as well as those of old and new rivals. The industry’s annual compound growth in the second half of the 1960s was in double digits year over year, so many thousands of organizations expanded their use of computers in those five years. Thanks to the 1400s and systems from other vendors, in that first decade, big and medium-sized companies across the industrialized world now used computers. S/360 was the turning point. This growth proved so massive that the volumes of machines and users of the 1950s seemed small in comparison by the end of the 1960s and tiny by the end of the 1970s. Usselman made the salient observation that this system “introduced a useful measure of stability into computing at a time when the industry might have suffocated under a staggering array of independent approaches,” spurring that demand.58
In the long run, another group of companies produced plug-compatible machines that fit into IBM’s S/360 technological ecosystem. They did the same to some rivals of IBM. Plug-compatible companies comprised a new segment of the industry. These included Memorex and Telex on the hardware side and for software Informatics and ADR, among others. A different class of rivals included service bureaus, notably ADP and Ross Perot’s EDS, and even time-sharing firms, such as Comshare and Tymeshare. Because of the IBM consent decree of 1956, third-party leasing companies proved successful in financing new and secondhand S/360s and, later, S/370s, largely in the United States. Leasco and Greyhound were the biggest players, but many dozens of small participants, even one-person operations, also thrived. European governments entered the arena with their “national champions.” The European initiatives failed to slow IBM’s relentless dominance of their computer markets, discussed in a later chapter.59
GE had its 400, 600, and 100 series in the mid- to late 1960s but exited the mainframe market in the 1970s. IBM’s old nemesis Sperry Rand promoted its UNIVAC 1108 and 1106, later the 9000 series. It acquired RCA’s customers and machines. Honeywell brought out the 500 series and the Century 100 and 200, and when GE got out of the business, acquired its customers. The earliest producer of compatibles, RCA, dropped out of the market in the 1970s. By then, the big threat to IBM came from Amdahl, formed by Gene Amdahl (1922–2015), a highly gifted IBM engineering manager who had played a pivotal role in the design of the System 360. When IBM refused to innovate beyond the S/360 to the extent he wanted, Amdahl formed his own company, which produced S/370-like machines that competed effectively against IBM in the 1970s. Key niche players in the 1960s and 1970s included SDS, CDC, and DEC.60
Throughout the 1960s and 1970s, peripherals also grew as a business, covering all manner of equipment, such as tape drives, disk drives, printers, control units, point-of-sale terminals, remote job entry (RJE) terminals, and even cash registers. According to Campbell-Kelly and Garcia-Swartz, “By the late 1960s, IBM derived more revenues from peripherals (60 percent of the total) than from the processors themselves. IBM’s peripherals also accounted for between 75 and 80 percent of all peripheral-related revenues in the United States at that time.”61 So, it had become an important new battleground for IBM, one where fighting increased in the 1970s. These rivals displaced about 14 percent of IBM’s tape drives and nearly 5 percent of its disk drives in the United States by 1970, enough to catch the attention of IBM’s sales force.62 At the time, the United States had the most competitive market in the world. The number of salesmen from IBM and other vendors increased by the thousands worldwide, with large organizations often having to deal with dozens of them constantly visiting data processing managers. Battles were fought over every computer a customer contemplated acquiring, and over peripherals and software, too. Normal sales tactics were applied by all: leading with discussions about “feeds and speeds”—that is, technical features—largely everyone’s line; IBM’s cost justification and financial presentations, RCA’s and the leasing company’s 15 to 20 percent lower prices; and so forth.
But once in a while a threat called for particular attention. In IBM’s case, that meant Control Data Corporation (CDC), a company based in Minneapolis, Minnesota, which produced systems for the scientific and engineering communities. Formed in 1957 by former employees of Sperry Rand, by the early 1970s it accounted for some 5 percent of the value of all computers installed around the world. It also sported an outstanding team of executives and engineers. William Norris (1911–2006), a World War II Navy cryptanalyst and founder of Engineering Research Associates (ERA), a formidable postwar computer start-up targeting scientific users, which Remington Rand bought in 1952, launched CDC. Seymour Cray (1925–1996), one of the most important designers of supercomputers, led the charge in developing computers at CDC and in 1972 formed his own supercomputer company, Cray Research. In 1964, CDC introduced the 6000 series, a highly successful family of supercomputers. Watson and his management team wanted to blunt CDC’s initiatives and therefore pressed their engineers for a new high-end machine, named the System/360 Model 90, to be enhanced with new technologies. The machine was announced, and the sales force was sent off to sell it. CDC considered IBM’s actions predatory and filed an antitrust suit against it in December 1968. The two firms settled out of court, but it was one of many court cases that erupted as a result of IBM’s S/360 and later S/370, about which we devote an entire chapter. It seemed to some that “if you couldn’t compete against IBM, sue them and maybe make money that way.” For others, IBM was simply a ruthless, predatory rival.
A SUMMARY OF RESULTS
IBM was growing the size of the computer industry and its own business. In the process, it constrained its rivals, although they continued to increase in size and number simply because the entire market expanded so rapidly. Pent-up demand for computing was released by technological innovations brought forth by IBM but also by users who accumulated enough experience with computers to understand their value in driving down operating costs through automation of clerical functions, better inventory control, more effective management of complex manufacturing operations, and by performing new functions in engineering, science, and early uses of online distributed computing.
Because the S/360 was the heart of much computing by the end of the 1960s, its users constituted a world of their own. Thousands of programmers only knew how to use software that ran on S/360s. Additional thousands of data processing personnel had only worked with IBM equipment, from keypunch machines, such as the popular IBM 029, to printers, tape drives, and disk drives, and, most important, its software, which in many instances took years to master. IBM continued cultivating middle and senior management in their accounts, not just heads of data processing, to discuss business uses and costs of computing. They proved so effective in cultivating relations that it became almost a throwaway line that “nobody was fired for recommending IBM.” That bit of folk wisdom floated around until the late 1980s. User groups like SHARE grew in size and importance, such that by the early 1970s the computing space was largely an IBM world on both sides of the Atlantic, most assuredly in the emerging markets in Latin America, and Japan. IBM’s computing vernacular, approaches to computer uses, and ways of doing business seeped into other organizations to a greater extent than in Watson Sr.’s day.
It is worth revisiting the company’s numbers. From IBM’s revenue in 1964 of $3.2 billion, the firm grew to post revenues of $8.2 billion in 1971. It hired thousands of people to feed the hungry R&D and manufacturing IBM needed to get S/360 going. In 1962, IBM had 127,000 employees worldwide, a figure that increased by some 10,000 each year until 1965, when they started shipping S/360 systems to customers. In 1965, IBM hired an additional 22,000 people, another 25,000 the next year, and ended 1971 with 265,493 employees. In other words, in the decade since 1962, the company doubled in size.63 Few of those hires remembered Watson Sr. Women came into IBM in droves, especially into manufacturing. Not since World War II had so many entered IBM. While a flood of veterans joined IBM in the late 1940s, by the early 1960s new hires were largely college graduates, especially those in engineering and sales. The employee population worldwide was now younger. IBM had quickly entered a new reality. A generation of engineers, executives, and salesmen were wiped out in the transition to the S/370. Arthur Watson was only one casualty. Tom Watson Jr. suffered a serious heart attack in November 1970, leading him to step down as chief executive officer in June 1971. At that time, Learson, president since 1966, replaced Watson at the age of 58, as an interim head before placing the planned heir, Frank T. Cary (1920–2006), who was slotted to take over in 1974 if Watson made it to the mandatory retirement age of 60. Learson made it to retirement, at which time Cary took over, in January 1973, a story for another chapter.
Did the S/360 cause Tom Watson Jr.’s heart attack? It is hard to say, although photographs of him taken in the 1960s showed a fit man. In his memoirs, he repeatedly spoke of “fear of failure” as “the most powerful force in in my life,” especially of not living up to his father’s standards. It probably did not help that on January 17, 1969, at the height of his career, the U.S. Department of Justice filed an antitrust suit against IBM. The antitrust case must have weighed on him, because this one concerned IBM’s current core business, computers. This litigation consumed Frank Cary and IBM’s senior management for 13 years. During the presidency of Richard Nixon (1969–1974), a recession in the United States and the first global computer industry depression erupted; IBM’s stock dropped in value by 50 percent.
Watson had been an outstanding leader in the most difficult of times for his company, his industry, and computing technology more generally, and his financial record demonstrated his achievements (figure 8.2). For his reputation, it seemed fortuitous that Watson left IBM when he did, because the company entered a period of even more intense competition, accelerating technological changes, the antitrust case, and a global economy experiencing severe recessions in waves in the 1970s and 1980s. One observer might have gotten things just right in assessing the end of Tom Jr.’s career: “In a macabre sense, his father’s terminal illness and his own brush with death were beneficial to the company in that they forced timely transitions of the top executive. Almost every other major computer firm would be less fortunate.”64 All the same, his accomplishments proved so great and lasting that we can consider Tom Jr. the fourth founder of IBM.
Revenue, income, and employee growth under Thomas J. Watson Jr. Courtesy of Peter E. Greulich, copyright © 2017 MBI Concepts Corporation.
Watson & Co.’s greatest legacy was the world in which IBM’s products and its users operated, the form the company assumed in its Golden Age as the world’s dominant provider of computing, and in setting or reinforcing a style of managerial behavior among so many of its customers in the 1970s and 1980s. It was a magical period but dogged by disastrous stagflation in the U.S. economy and many innovations that nearly crushed the company to death when its Golden Age ended. Before that event, much happened. That is what we turn to over the next several chapters.