New Ventures: Lessons from Xerox and IBM

Blue Magic: The People, Power & Politics Behind the IBM Personal Computer, James Chposky and Ted Leonsis (New York: Facts on File Publications, 1988), 240 pages, $ 19.95. Fumbling the Future: How Xerox Invented, Then Ignored, the First Personal Computer, Robert C. Alexander and Douglas K. Smith (New York: William Morrow and Company, 1988), 274 pages, $ 18.95. Blue Magic and Fumbling the Future are technology adventure stories well worth reading. In the pattern of Tracy Kidder’s The Soul of a New Machine, the authors’ sympathies lie with small, capable venture teams that bring exciting new products into existence despite a hostile corporate bureaucracy. But as their subtitles indicate, these books devote almost as much attention to IBM’s and Xerox’s management and organization as they do to the venture teams. IBM emerges from Blue Magic as a fundamentally sound, highly disciplined organization in which advancement centers on personal visibility at corporate headquarters. Senior management is shown as informed and in control, and the authors detail the unprecedented personal support CEO John Opel gave to the PC team. In contrast, the authors of Fumbling the Future find both the Xerox organization and its chairman Peter McColough to have been inadequate to the task of commercializing a leading new technology—the Alto personal computer. Political factionalism—especially involving staff and outsider executives—is made out to be more powerful than senior management’s acumen. Personal animosities and lack of mutual respect are presented as the bases for important business decisions. In both cases, the authors’ judgments may be right on target, though it is always wise to consider the source: the primary input for both books came from former employees, particularly former team members, who were often at the low end of the organization looking up. Still, there are enough important differences between the two undertakings to make a corporate executive experienced in both starting and operating new businesses wary of such simplistic judgments of failure and success. Take Xerox’s “failure.” Spin-off products from the innovative Alto technology materially aided its expansion beyond copiers. Conversely, the successful IBM PC—a new related product—created a huge mass-market opportunity without IBM’s usual protection from broadly capable competitors. Thus the value of these books, for me, lies in the lessons they teach about successful corporate venturing and the questions they lead us to ask. Blue Magic begins in July 1980 with Bill Lowe’s presentation on the personal computer business to the IBM corporate management committee. Lowe was then the laboratory director of IBM’s Entry Level Systems Unit, which had been set up in Florida 12 years earlier to experiment with small, low-cost computers. In his presentation, the 18-year IBM veteran made two points: (1) this industry was now a viable opportunity for IBM; and (2) getting into it would mean going outside IBM’s culture. The committee gave Lowe one month to prepare a design proposal. It also gave him funding to assemble a task force of expert IBM engineers at his Boca Raton base to build a demonstration prototype. When Lowe reappeared with the prototype, he outlined the fundamental design philosophy IBM would have to follow to get into this rapidly changing market. Implementing it would mean deviating significantly from traditional IBM practices, since nearly all the computer’s elements (even the operating system) would come from outside vendors. Furthermore, the PC’s architecture would be open, which implied that other companies could connect non-IBM peripherals and software to it. On one hand, this choice would make it possible for competitors to sell clones or functional copies. But it would also mean that a high-quality unit could reach the market very fast, and that the IBM PC could more readily become the industry standard. If this could be accomplished, all personal computer users could readily trade up to other IBM products as their needs increased. Finally, Lowe proposed to sell the new product through retail stores—bypassing marketing, sales, and service organizations that were fundamental to the IBM image. The committee gave Lowe the authority to continue and told him to report back in October on the progress of what was now designated as his product development group. Undoubtedly, Opel’s strong personal support led this approval. When Lowe reported back, the committee approved further work and formally named the group an independent business unit, a venture format favored by Chairman Frank Cary. This began a process of unprecedented empowerment for the project, which included procedural shortcuts and massive funding usually unavailable to new product development activities. The target date for introduction was set at one year from the August presentation. The task force moved quickly to locate and contract with vendors so that work on hardware and software could proceed simultaneously. The Intel 8088 microprocessor chip it chose was five-year-old technology, proven in the industry. The task force rejected a new, more powerful chip, for fear that it would not be reliable or available in quantity. Perhaps more important, a chip that was much more powerful than those in existing personal computers could deliver performance that could potentially compete with more expensive IBM products. The contract for an operating system went to Microsoft; drives came from Tandon, circuit boards from SCI, printers from Epson, power supply from Zenith. Agreements to distribute through ComputerLand led to valuable market information that aided the design. Distribution through Sears was expected to open an even broader channel to the mass market. Don Estridge, the experienced program manager who would become the venture’s indispensable leader, joined the unit in the fall of 1980. When Lowe was promoted to vice president of the Information Systems Division in early 1981, Estridge moved into control. The rest of the story details how the dozen engineers, all with many years of experience at IBM, used the power entrusted to them to carry out whatever needed to be done. Fumbling the Future presents the Alto as though it were a focused business undertaking comparable to the IBM PC, but the story the book tells argues otherwise. In my definition, a business venture is inherently a short-term phase of a long-term activity—a transformation stage in which something that is not a business evolves into a self-sustaining business. From this perspective the IBM PC was a venture. Strictly speaking, the Xerox Alto computer never emerged from R&D. Xerox senior management first became interested in the computer market in the 1960s, when revenues from the Model 914 copier grew to a very profitable $ 1 billion-plus annually. Xerox’s visionary chairman, Joe Wilson, was determined to see this kind of growth continue. The computer, with its potential for rapid digital handling of information, which copiers handled optically, was a logical candidate for expansion. At the same time, Xerox’s enormous growth was seriously overloading its management capability. Peter McColough, who became president in 1966 after several years in Xerox sales and marketing, addressed this problem by hiring outside executives with big-company experience in finance and administration. In 1968, after years of active searching, Xerox purchased a small California-based computer company that served engineering and scientific customers. Shortly thereafter, McColough approved a proposal from his newly hired chief technical officer to build a Xerox center for computer research. From this center, located in Palo Alto, the concepts of personal distributed computing would evolve. Neither the chief technical officer nor the scientist he subsequently hired to head the lab was a computer expert, although the latter was familiar with the subject through a Defense Department research contract at the university where he had been a senior administrator. But the two executives did hire a person with years of experience in planning practical computer applications. This was Bob Taylor, who had at one time directed a large Defense Department research program on information processing which put him in contact with leaders in the field. Although he was not formally trained in computers, he had a broad vision of their ultimate potential that went well beyond mere numerical computation. So did the members of the team he recruited. These new hires were extremely competent computer experts, and they all shared Taylor’s elitist disdain for most of the Xerox organization, including its recently acquired computer subsidiary, Scientific Data Systems. Taylor’s blue-ribbon team was adequately funded, and in the informality and freedom of its geographical and organizational isolation, it focused more and more on ways to make a dedicated computer serve a wide range of needs for a single user. Control computers that coordinated processing and factory operations in the 1960s had already been used by individual scientists and engineers as a kind of personal computer for the professional. Their cost was well beyond what the Palo Alto team knew nonprofessional users could afford. But the team assumed that the spectacular improvements in the cost and performance of computer elements that they were seeing year after year would continue. By the time commercialization was feasible, they reasoned, their multipurpose personal computers could be both cheaper and better than existing minicomputers. From the standpoint of technical performance, they were extraordinarily successful. Within a year, 40 Altos had been built. Team members used some in their own work, but more impressive was the fact that other Xerox employees ranging from scientists to secretaries were using them too. Embodied in the team’s approach were simplifications designed to help novices learn quickly: refining a mouse input device, orienting input control to the monitor screen, devising a user-friendly word processing program. The team also developed a laser printer incorporating a document compiler (so that what was written on the screen was what would be printed), multitasking (which let the central processor handle 16 housekeeping tasks at once), and an Ethernet communications network (which allowed small, self-contained computers and peripherals to communicate). Thus the new system could functionally compete with the then prevalent practice of time-sharing on a larger centralized mainframe—and even interact with one. But the isolation of these 200 people raised formidable problems in commercializing their innovations. From 1971 on, four organizational levels separated Taylor and the chairman; and few, if any, cordial bridges existed between the Palo Alto group and the rest of the corporation. Others in Xerox saw the group’s high morale as naive arrogance and its long-term visions as impractical, if not grandiose. Although design of a manufacturing scale-up of the Alto had started, there had also been attempts to adapt parts of the Alto technology for smaller scale products: a computer with a simplified programming language, an on-line office system designed around the needs of a publishing house, and a laser printer. By 1975, Alto’s commercial use was still untested, and the customer who needed it still hypothetical. Nevertheless, the technology’s champions invested more time and money in lower level efforts to bring a version to the marketplace. Some of these efforts were stopped before they reached senior management; others were cut off at the top. In one last effort to generate a market, a personal computer appeared in a Xerox television commercial; in that same year (1979), the corporation again decided against marketing the Alto. In 1981, Xerox did introduce the Star computer office system, which incorporated most of the Alto’s functional aspects. But like the Alto, the Star was largely designed without direct knowledge of current customer needs or the competitive environment. Xerox also made little effort to integrate the system with its marketing or manufacturing capabilities. It offered neither a spreadsheet nor compatibility with other computers. Its full system hardware was slow to develop. Its programming language was kept private. And its workstation alone cost five times more than a stand-alone personal computer. Although the authors don’t mention the consequences explicitly in the text, the index lists these limitations as “Star computer office system…reasons for failure of.” The PC and the Alto are both typical of ventures in computer-related businesses in which forces outside the corporation produce extremely rapid market and technology changes. But these ventures are clearly very different from one another. For that reason, I find it less easy than the authors to call IBM an unqualified success and Xerox an absolute failure. The IBM PC was a directly targeted, speedy, free-spending project based on a detailed understanding of the customer, the competition, the technology, and the corporation. Competitors had proven all the elements of the task—materials, software, expertise, performance, distribution. Patent position was immaterial, for IBM’s greatest proprietary assets were its strong position in the computer industry and its insight into the business threats and opportunities the personal computer posed for large computers. Moreover, the chairman and at least part of the corporate management committee seem to have understood IBM’s limitations in developing the product. Suspending normal practices only long enough to get started quickly was a stroke of genius, possible largely because the PC amounted to a defeatured version of IBM’s core business products. The entire organization with its traditional strengths did not need to change radically once the beachhead had been established. The key factors in carrying this kind of venture to completion are personal commitment at the top and consistency of the venture team in meeting expectations. In this case, executive support went far beyond the PC’s funding. It was reflected in the power entrusted to Estridge and his team. It also appeared in the support given to the headstrong team in the face of large and powerful elements in the corporation: when IBM divisions wanted to supply parts, they were forced to bid against the outside vendors. The Charlotte, North Carolina plant had to be 15 % under outside vendors’ bids on a central processing unit assembly; the Endicott, New York plant was told to lower its bid for printed circuit boards. This was not just team charisma, but unwavering top management support of a very rare sort. Significantly, the one corporate group the team could not bypass was the Quality Assurance Unit; it had to satisfy the unit’s requirements because the product would bear an IBM label. This kind of venture is seldom sold to top management. Rather it is orchestrated by a senior executive with sufficient signature authority to ensure that the costs of following through on success will be forthcoming. I wonder what the Charlie Chaplin tramp advertising budget amounted to, especially as a percentage of first year PC sales. Today’s ads still try to involve “IBM’s Business Partners” (that is, the retail distributors) to bring in more sales from small business. It remains to be seen whether IBM will prosper in the long run by bringing its own worst competitor into existence in-house too early—for in the personal computer world, customer expectations are not wholly within IBM’s control. Xerox’s Alto represents a venture at the other end of the spectrum—an innovation that requires years of steady spending before customer need is clearly defined. Such ventures usually demand some form of patent protection to allow the innovator to recoup at least part of the investment. The Alto team members apparently were not able to obtain this protection because of prior art, were unwilling to ask for it because of their independent temperaments, or were ignorant of its importance. Computer journals and a 1977 Scientific American article by a principal scientist on the Xerox team indicated the nature of the work going on at Palo Alto. Steve Jobs visited the lab by invitation toward the end of 1979, when the Lisa was still in its design stage. While a new product is being built, market contacts are far less expensive than repeatedly manufacturing something that cannot be sold. Therefore, a broad but specific target like personal distributive computing ordinarily requires ongoing marketing collaboration, even if the product is formally part of R&D. Yet the Alto team had little contact with customers, nor was it encouraged to develop any. Moreover, Xerox’s compartmentalization apparently allowed money to be spent on research as long as the copier-dominated operating or staff units could control commercialization. Whereas IBM cultivated competition between complete business teams (with the market as judge), Xerox’s internal competition seemed to be carried on between administrative fiefdoms. Selling any pioneering product up the chain of command requires either spectacular commercial results or extraordinary political power and diligence. Independent commercialization was outside the charter of the Palo Alto group, but there is little record in Fumbling the Future of any direct contact between Taylor and Xerox senior managers who were outside the technology chain of command. In contrast, Blue Magic notes that Lowe spent as much time at IBM headquarters in Armonk in the 1970s as he did in Boca Raton, while Estridge was there 80 % of the time he spent away from Boca Raton. Xerox management did make supportive statements on occasion, but in vague phrases like “the architecture of information” that didn’t necessarily mesh with what the people in Palo Alto were doing. The dream of the Alto team—personal distributed computing—diverged from Xerox’s long-term strategic focus on the office of the future. But innovative technology developments have unexpected benefits to those who are alert to them. Xerox made repeated efforts to use parts of the team’s technology capability in particular products; and spin-offs of Alto technology eventually emerged in the form of laser printers, terminals, word processors, and local area networks. These incremental additions to the product line brought the breadth Xerox needed to augment copier sales—and they did not require starting afresh in computer innovation. That the corporation salvaged so much in other products, despite internal barriers and competing technologies, demonstrates basic strength in people, if not in organization. And the broadening of the corporate base that took place may make the acceptance of new technology easier the next time. Staffing of any venture is critical, and a venture team should feel like volunteers. The PC team was a select group of individuals whose track records in the IBM environment were well-known. They were junior enough to be partially motivated by ambition, and their leader, Don Estridge, was a high-potential candidate whose supportive management style was ideal for the PC adventure. The Alto team, in contrast, was selected for technical proficiency proven only in cultures that were radically different from Xerox’s. They came from—and remained in—a technical subculture in which the goals were performance and evolving capability—almost independent of business realities. If given the proper responsibility, technical people with this slant have often learned to meet customers’ needs in product design. But since they were being administered like an R&D unit, the builders of the grand Alto prototype system found that it was never more than a showcase from which the operating units would choose valuable parts. Interestingly, neither company did well in holding on to its venture participants. The Alto’s team members drifted away over the years. Some went to competitors including Apple, whose Lisa and Macintosh computers resembled the Alto. Dispersal of key team members from IBM began within two years of the PC’s introduction. To their surprise, they seem to have found that they were not the same people after experiencing the high ground of self-recognized accomplishment. No longer content to be subject to ordinary organizational discipline, many of them changed jobs repeatedly. The one exception was Estridge: he won promotion with unusual rapidity on the basis of his spectacular managerial accomplishment, becoming head of worldwide manufacturing before his untimely death in an airplane accident in 1985. Going after tomorrow’s businesses is also a lot easier to carry off if the businesses you are in today are working well. When IBM targeted the PC in 1980, it had just resolved antitrust charges with the Justice Department. Senior management was eager to regain lost business and technical momentum. Moreover, the company knew that personal computers would soon be a threat if they weren’t made a corporate opportunity: the industry was rapidly becoming a contender for markets IBM wasn’t serving and would soon be directly competitive in its capability. When Xerox fumbled its opportunity to commercialize the Alto in the mid-1970s, no long-term technical effort could have competed successfully for senior management’s undivided attention or massive funding. By 1975, the company faced a general business downturn, extensive antitrust actions, growing domestic and international competition in the copier business, and declining profits (with a consequent imperative to improve efficiency and control costs). Scientific Data Systems, meant to compete directly with IBM in mainframe commercial data processing, had just been shut down after proving a costly failure. And there was no immediate danger to Xerox sales if the Alto were delayed or scrapped. Top managements of both IBM and Xerox knew that their organizations were on dangerous ground as other companies developed the technological capability to threaten their core businesses. Part of the threat came from newcomers who had little to lose by making an existing machine base obsolete. But more important was the worry that their own people and practices were too inefficient to respond to the rapidly changing competitive environment. That thought must have led IBM’s Opel to want to create a legendary example of what a lean team could do within the corporation. Certainly McColough intended to use R&D to repeat the past glories of Chester Carlson and Joe Wilson. And this raises a more fundamental question: To what extent can the chief executive of a major corporation make an informed decision about new technology? McColough may have paid more than the cumulative budget of the Alto project to acquire Scientific Data Systems—and Fumbling the Future presents the acquisition as a quick, unilateral decision. It seems no more than a hunch, and his next decision—to authorize an expensive laboratory so distant geographically and intellectually—gives no clue to what he expected or would have done if success attended both efforts. Information, however partial, could have been available to him; but he did not even discuss the acquisition in the new technology with his chief technical officer. Opel may have been better prepared because the PC lay in his and his company’s own field of expertise. Yet his decision on how to proceed had apparent links to his predecessor’s affinity for independent business units. And there were, after all, seven such units in IBM, among them, robotics and medical systems. If we knew more about IBM’s initiatives in these areas, we might have more insight into Opel’s decision-making style. In the end, his decisions might even appear similar to McColough’s.

New Ventures: Lessons from Xerox and IBM

Partner Center