ENGINES OF PROGRESS: DESIGNING AND RUNNING ENTREPRENEURIAL VEHICLES IN ESTABLISHED COMPANIES ROSABETH MOSS KANTER, JEFFREY NORTH, ANN PIAGET BERNSTEIN, and ALISTAIR WILLIAMSON Harvard Business School INTRODUCTION TO THE CASE REPORT SERIES In order to create new revenue streams through new or improved lines of business, many companies have established formal programs to stimulate and manage innovations and venture creation. We refer to such programs as “entrepreneurial vehicles”+ngines for driving the creation of value through new ideas-and the projects they nurture as “newstream” operations in “mainstream” (ongoing, activities, to distinguish them from “mainstream” established) lines of business (Kanter 1989). The core dilemma around entrepreneurship in established companies is, of course, organizing so as to manage newstreams in the midst of mainstreams-even when there are conflicting requirements for effectiveness in each. Between 1986 and 1988 a Harvard Business School Research Team studied eight corporate venturing programs in depth, covering companies in a large range of industries and with strikingly different kinds of entrepreneurial vehicles. Burgelman (1983) identified three major levels (or tasks and roles) in the internal corporate venturing process. These levels are similar to the levels in Bower’s (1970) resource allocation model: corporate management, which provides the structure and rationale for new business development, as well as authorizing projects and monitoring performance; new venture development management, which delineates specific areas in which to seek project managers; and the venture managers themselves, who generate the ideas, champion specific products or technology, and oversee the technical work. The focus of the Harvard research was the middle level, the integrative mechanism for linking corporate strategy and innovation projects. This report covers one of eight programs examined by Rosabeth Moss Kanter’s research team in 19861988. Jeffrey North, Anne Piaget Bernstein, and Alistair Williamson collaborated on the report, and Gina Quinn edited it. This report is based on multiple rounds of interviews in 1987 and 1988 and a review of company documents and published materials. Research funding was provided by the Harvard Business School Division of Research, whose support is gratefully acknowledged. Address correspondence to Gina Quinn, Baker 400, Harvard Business School, Soldier’s Field, Boston, MA 02163. Journal of Business Venturing 5, 415-430 0 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 415 416 R.M. KANTER ET AL The research examined the entrepreneurial vehicles themselves-how companies organize to find, nurture, and use newstream activities. The eight sites were selected to maximize differences in strategic intent, and these initial differences in goals and strategy were then reflected in operational differences, such as the linkages between newstream and mainstream organizations, as well as in the payoff to the companies. The Minnesota Innovation Research Project used a similar strategy of comparing projects that were quite different in form, scope, intent, and context (Van de Ven, Angle, and Poole 1989). The strategic intent continuum is anchored on one end by entrepreneurial vehicles designed to maximize newstream ventures, for purposes of corporate diversification, and is anchored on the other end by vehicles designed to impact back on mainstream activities, by stimulating innovation through the mainstream. In short, the “ideal type” of strategic choice is whether the primary goal is to be economic (creating new sources of revenue regardless of the impact on the mainstream organization) or cultural (showing the mainstream organization how to be more innovative regardless of the magnitude of the economic payoff of newstream activities). In the latter case, newstream activities are certainly intended to be profitable, but their assumed value as role models far exceeds any immediate bottom line benefits. In practice, of course, there is a mixture of goals. When economic goals dominate, however, top management intends to be more heavily involved in defining the newstream strategy, there are more explicit biases in the selection of newstream projects, and there is often more separation between mainstream and newstream, the newstream often being organized as independent, stand-alone ventures. When cultural goals dominate, newstream activites are more likely to come out of spontaneous submissions by lower-level managers and employees, and many close linkages are maintained between newstream and mainstream. Between the two extremes, various mixes of economic and cultural goals produce other kinds of entrepreneurial vehicles. The establishment of a formal newstream channel raises significant issues. Organizations have long used “parallel organizations” to stimulate and guide activities outside of the mainstream, as an outlet of communication and decision making that bypasses the routine system and hierarchy in the interests of change (Stein and Kanter 1980; Kanter 1983). The process of turning most commonly researched “parallel systems” for the entrepreneurial research into new lines of business in the established corporation has been the new venture department (NVD). The rationale for a separate new venture department is to find a home for radically new products or technologies that need more to be developed than simply proven as a concept in a laboratory and then transplanted to an existing division (Burgelman and Sayles, 1986). In short, an NVD is used for products that are both new and different. Dzjj‘ierent means that the product or process may not fit well with the existing business flow; new implies the need for a different managerial mode than that appropriate for ongoing operations (Kanter 1986, 1989). Furthermore, autonomy-being protected from the day-today operations and reporting requirements of the parent corporation and allowed a degree of self-sufficiency until viability has been achieved (Burgelman and Sayles 1986, p. 125)has been found to be a major factor (along with top management support) in successful internal corporate ventures (von Hippel 1977; Galbraith 1982) and corporate venture capital departments (Siegel, Siegel, and MacMillan 1988). But the NVD as a stand-alone entity has been found to be an unstable organizational form. Fast (1979) studied 18 NVDs established between 1965 and 1975; half were inoperative by 1976 in the form in which they had been established. But only a few had been disbanded completely; others had retained the ventures they started and become an operating division managing those then-established businesses or had become a staff department, generally in strategic planning. ANALOG DEVICES ENTERPRISES. 1980-1988 417 This evolution is not surprising; it may even be inherent in the very definition of a newstream channel. If a new venture organization is set up because there is no other way to do things that are “different,” it may not produce results for the parent because it is resource-starved, skill-starved, or increasingly detached in its operation. The economic evidence is mounting that relatedness is a key ingredient in successful diversification, whether through acquisition or internal venturing. For example, unrelated acquisitions tend to be sold at a higher rate (Montgomery and Wilson 1986; Porter 1987) than related acquisitions. So if “difference” is a problem, then stand-alone vehicles may either dissolve or gradually migrate towards “ownership” by the established businesses. The closer they come to the mainstream, however, the greater the domain conflicts, the fewer the justifications for standalone new venture operation, and the greater the pressures to be servicing the mainstream with modest, incremental improvements. Furthermore, there are numerous tensions and conflicts that arise from the establishment of a parallel system or a newstream channel. Some of these are inherent in the difference between mature management practices suitable for routinized, established businesses and those appropriate for new ventures; others stem from the simple fact of establishing two different channels side by side. Among the conflicts that are identified in the literature are strategic conflicts of interest involving domain and synergy; administrative conflicts involving the unwillingness of other departments to share resources with the new venture or the unwillingness of the venture to use the policies and systems of the established organizations; “culture” clashes because of the more chaotic nature of innovation; and measurement and reward issues, because it is often misleading to measure new venture performance in the same way an established business is measured (Burgelman and Sayles 1986; Kanter 1986, 1988, 1989; Quinn 1985; Miller, Wilson, and Adams 1988; Sykes and Block 1989). The success of innovation and entrepreneurship in established companies, therefore, depends on more than effective management of each innovation project. It also depends on the effective management of the newstream channel that is producing ideas, nurturing them to the point at which they are transferable, and linking them to the mainstream. Studying the design of the new venture vehicle itself can illuminate the key considerations and tradeoffs involved in different mechanisms for producing new ventures in the corporate context. Thus, entrepreneurial vehicles are dynamic and evolve over time as a function of the strategies of their parent companies, the results achieved, and the experiences gained in establishing and managing them. Since results are a matter of expectations, it is important to assess the success or failure of a venturing program against the goals that are set for it. Ventures with more modest goals are sometimes better able to produce results than those established with greater expectations. Furthermore, longevity is only one measure of the success of an entrepreneurial vehicle. The eight venturing programs studied in detail range in age from almost twenty years to less than three years, but all of them have changed in major respects over time, and they were still changing when the field research was completed in early 1989. Changes in the nature of the entrepreneurial vehicle over time are a function not only of internal organizational evolution, but also of external developments, especially for those established in response to the environment. The impetus for some companies to start a new venture program is reactive: to capture rather than lose people and ideas, to have a means for commercializing inventions produced with company resources whether or not they were envisioned in top-management strategy, and to retain potentially creative people. This was often a response to a venture-capital-induced brain drain in particular companies and communities at particular times (e.g., Oregon in the mid-1890s). It was also a response to the perceived success of start-ups that had used technology developed in a large corporation. 418 R.M. KANTER ET AL But, as the external environment changed (venture capital availability declined; high-technology companies took nosedives), companies whose entrepreneurial vehicles were defined in response to that environment and that, meanwhile, had experience under their belts, came to rethink their new venture approaches. The eight research case reports, taken together, provide an opportunity to compare and contrast venturing programs of different types and track records, to view their evolution over time, and to access the conditions that make entrepreneurial vehicles able to serve as effective “engines of progress” for their parent companies. The experience of Analog Devices Enterprises, the first in the series, offers insight into the rise and fall of a corporate venture capital group. ANALOG DEVICES ENTERPRISES, 1980-1988 The mission of ADE was growth. We seek to invest thrusts. Our objective is to option to eventually expand to invest in young high technology companies with potential for high in firms whose technology and products support our five diversification make these investments in such a manner that Analog Devices has the its technological base. -Analog Devices Enterprises, Progress Report, September 1983 Analog Devices Enterprises (ADE) was a venturing arm of Analog Devices Inc. (ADI). Formed in 1980 as a vehicle for investing in emerging new technologies, by 1987 ADE consisted of eleven companies in which ADI had an equity stake of between 10% and 70%. Since 1980, ADI had invested $41 million in this portfolio in an effort to diversify Analog’s technological base and eventually to contribute to the company’s technology-driven growth. But by 1988, the company was withdrawing its commi~ent to ADE and rethinking its newstream approach. Founded in 1965 and headquartered in Norwood, MA, Analog Devices Inc. is a leading producer of integrated circuits, data acquisition components, and subsystems used in computerized measurement and control systems, including medical instruments, industrial automation controls, avionic systems, and scientific instruments. In 1986, when the research began, Analog Devices employed about 3,000 people in the United States and about 2,000 people abroad, where 40% of the firm’s sales originated. In 1988, net sales were $439 million and net income was $38 million. AD1 served more than 5,000 customers; the biggest, IBM, accounted for 7.5% of Analog’s annual sales. ADI’s product line contained over 600 items with 1,000 variations. Technological innovation was key to the company’s past success and to its strategy for future growth. Like many newer technology-based companies, Analog had a humble beginning. The first production facility was located in the basement of the Cambridge apartment building where founder Ray Stata and his co-founder and MIT roommate, Matthew Lorber, lived. The pair started the firm with bank loans secured with their stock in Kollmorgen Corporation as collateral. Kollmorgen had bought their first start-up, Solid State Instruments, a company they had created for the production of testing devices for gyroscopes. ADI’s first products were operational amplifiers, which process and strengthen electrical signals. In 1969, Analog purchased Pastoriza Electronics Inc., a small firm with patents on analog-to-digital conversion devices. This jnvestment of $1 million was at the beginning of what was to become Analog’s main product line, which contributed $75 million in sales in 1986. In 1971 Stata became president and, in 1973, chairman of ADI. Stata attempted to establish a work environment that would attract and retain the ANALOG DEVICES ENTERPRISES, 1980-1988 419 technical talent he needed. For example, an English engineer, Gilbert, was given the choice of working in Oregon, because it was his wife’s home state and the climate reminded him of England. Analog provided Gilbert with a lab, a secretary, and the appropriate equipment in a location 3,000 miles from Analog’s headquarters. At the same time, Stata placed technology bets. In addition to developing analog-to-digital products, he believed that integrated circuit (IC) technology would become a major industry in the years ahead. Rather than reinventing the wheel by starting at the beginning to develop what other IC pioneers had already discovered, or buying a company that possessed state-of-the-art IC technology, Stata devised an arrangement to obtain access to the technology with a more modest financial and managerial commitment. In 1970, he proposed that AD1 provide $2.5 million in venture capital to a group of engineers to start Nova Devices, which would produce silicon chips designed to replace transistors in analog-to-digital conversion devices. In return for seed capital and marketing assistance, Analog received access to Nova Device’s developments and the option of acquiring 100% of Nova at a later date if the start-up proved successful. To gain the approval of ADI’s board of directors, which was divided over the decision to get into the IC industry, Stata put up his own AD1 stock as collateral to secure a $1.5 million loan to Nova Devices along with ADI’s $1 million in equity investment. The gamble paid off. In 1972, AD1 exercised its option to acquire Nova Devices. Nova Devices became Analog Devices Semiconductor (ADS), which currently provides 70% of ADI’s sales and 80% of its profits. Despite this success, Analog’s core product lines and its research and development activities were centered on only a few narrow fields of technology. Furthermore, ADI’s marketing focus was toward producers of complete systems-AD1 made only the components. Top management believed that Analog’s customers soon would be demanding complete systems, not simply the components and the chips that composed Analog’s basic business. Realizing how fast the electronics industry was growing, AD1 sought a way to keep abreast of technological developments in industries close to those Analog had already entered or might wish to enter in the future. During the winter of 1979-1980, Analog’s top people, including officers, marketing staff, and key technologists formed a diversification task force. They met on Saturday mornings to discuss the future and to develop a list of fields into which AD1 should be moving. Among promising areas for Analog’s future, they identified digital signal processing, high-speed computers, high-speed automatic test systems, industrial image processing, automation, and emerging high-speed semiconductor processes such as gallium arsenide, which could prove to be important to Analog in the long-run. But Analog did not have the resources to pursue all of these technologies alone. In addition, there was a shortage of experts in several fields, a problem that money could not solve. In the mid-1970s, when sales were nearly $100 million, Analog attempted to develop electronic converters that could process high-frequency video information for use in radar systems and other ultrafast instruments, but discovered a lack of internal expertise. The gap was filled in 1978 through the acquisition of Computer Labs, a $3 million North Carolina company. The successful venture capital seeding and subsequent integration of Nova Devices, along with the purchase of Computer Labs, probably made Stata such a firm believer in owning interests in start-ups and early-stage companies. By 1978, after the established success of Nova Devices’s other deals, AD1 was getting a reputation as a company willing to invest in partnerships with smaller high-tech firms. Analog started to receive unsolicited propositions for funding or acquisition from young technology companies. Many of these 420 R.M. KANTER ET AL. were legitimate and potentially rewarding opportunities for Analog to participate in the creation of blossoming technologies and fuel for its own long-term growth. Describing the small firm’s view of the potential relationship, George Wilson, then president of Bipolar Integrated Technology (BIT), a company in which AD1 initially invested $3OO,ooO, indicated the allure of a high-tech company like ADI acting in a venture capitalist capacity: “Ours was such an advanced technology that it was not realistic to go to venture capitalists and expect them to understand it. We were going to need a company within the industry that would understand our technology.” At this time, all of Analog’s financial resources were committed to its existing operations and lines of business. Furthermore, AD1 was working toward some internal diversifications that further constrained available resources; there was no spare capital for venturing. Analog turned to another source, one of its own institutional shareholders, Amoco. Analog had done business with Amoco since 1977. In the summer of 1977, Stata and Larry Sullivan, the senior vice president of Corporate Development, had arranged a deal in which Amoco (Standard Oil of Indiana) acquired 15% of Analog stock for $5 million. The executive officers approached Amoco for funding because, as Sullivan explained, “the public markets were dead, so we looked for private sources. The oil industry was awash in cash, so we got into a deal with Amoco.” Gordon McKeague, president of Amoco Technology Company, a Standard Oil subsidiary, and manager of corporate development for Amoco, became a member of Analog’s board of directors and is still serving in that capacity. In June 1980, when venturing opportunities were creating more of a need for new capital, Analog approached Amoco again. The result was a novel financial agreement. Amoco and Analog entered into a five-year agreement whereby Amoco would receive up to 10,000 shares of preferred stock annually, priced at $1,000 per share, in return for up to $10 million in cash paid to Analog for use in investments in the selected ventures. At the end of the five years, Amoco would have up to twelve months to convert the $1,000 preferred stock into $2,000 common shares at the prevailing market price. Amoco would not share in the profits or losses of the portfolio investments, nor would it have any active participation in the management of the portfolio apart from Gordon McKeague’s seat on Analog’s sevenmember board. On July 31, 1980, ADI officially announced the formation of Analog Devices Enterprises (ADE), to augment internal expansion by affiliation with entrepreneurial companies in selected growth markets. ADE would supply expansion capital to young companies and also share with them Analog’s experience and management tools. Amoco had its own reasons for underwriting ADE. Amoco Technologies Company was a subsidiary with its own mandate for diversification and venturing. Amoco technologies already owned 21% of Cetus Corporation, a genetic engineering company, and 21% of Solarex, a solar technology firm. Analog’s signal processors would aid Amoco’s exploration and refining efforts. Analog now had the financial resources to begin assembling a portfolio of investments that would give the companies windows on new technologies. Each of the agreements between Analog and the portfolio companies was different. In general, however, these agreements gave Analog an equity position and one or more of the following rights: the option to acquire the company at a later date; access to the new technology or processes; joint R & D agreements; or the option to market the new firm’s products. ADI had an agreement with its auditors that losses in ADE would not be reflected on the ADI profit and loss sheet. This would be important to insulate AD1 from the vicissitudes ANALOG DEVICES ENTERPRISES, 1980-1988 421 of risky start-up companies, and to avoid competition for resources between Analog’s mainstream and the new external ventures. There were three specific reasons that ADE was organized as a corporate venture capital group investing in a portfolio of stand-alone computers. First, to reduce cost. Second, to minimize structural changes. Outright acquisitions would require frequent changes in Analog’s financial and management structure; and increase in the portion of the company consisting of acquired operations would cause Analog to lose focus on its core products and marketing strategies. Finally, to contain risk. Blossoming technologies are extremely risky. Often a product or a process looks both feasible and elegant on the drawing board, but finding marketable applications can be extremely difficult. In the 196Os, Stata could afford to bet the company on the gamble that integrated circuits would be the wave of the future, but the situation had changed by 1980, owing to Analog’s responsibilities as a public company. Furthermore, if a portfolio company fails, the loss is localized, having a lesser effect on the management and the morale of the core company. By September 1983, there were eleven companies in the ADE portfolio. Analog produced a report to shareholders describing its portfolio and venturing objectives. Analog stopped adding companies to the portfolio in 1984. The strategy of ADE was then changed to concentrate on “follow-on investments in those companies which continue to be strategically compatible with the Company and to sell its interests in the companies whose strategies and future direction diverge from those of Analog,” according to ADI’s 1986, 10-K report. In 1985, at the end of the five-year Analog-Amoco agreement, Amoco did not renew the agreement. Analog’s explanation was that a $30 billion company like Amoco needed larger investments to aid its growth. MANAGING THE NEW VENTURE INVESTMENTS Sullivan served as the general manager of ADE, with primary responsibility for managing the ADE portfolio companies. Sullivan was assisted by an internal, ad hoc board of advisors consisting of Robert Boole, ADE’s Director of Venture Analysis, Ray Stata, and Sullivan himself. John Hudson, ADE’s controller, was frequently included as a member of this board, as were various operating managers, senior technologists, corporate financial staff, and marketing staff from ADI. The idea was to assemble the most qualified and knowledgeable group of people possible to make portfolio decisions drawing on ADI’s experience, skills, and creativity, yet to retain the flexibility of ADE. The ADE portfolio was selected after studies by ADI’s corporate marketing staff, who used as a criterion areas that appeared to be relevant for the future, to offer good opportunities for complementarity with existing core businesses, and to have a distinct competitive edge. Specific screening factors included the following: l The experience, accomplishments l The strengths, l The size and potential l The presence of distinctive, innovative for success relative to competitors. creativity, and balance of the management team. and caliber of the technologists. growth-rate of the market for the company’s products. factors in the products and the requirements 422 R.M. KANTER ET AL l The ability of the company customer base. l A credible business to develop, and financial market, and direct new products to their plan. The portfolio was assembled in a carefully focused process. First, Robert Boole, who was reporting to Sullivan, would find the companies and make initial evaluations from a venturing standpoint, along with Larry Sullivan. A special emphasis was placed on the technical and marketing staff of the proposed portfolio company. Since it was felt that the necessary technical expertise was to be found inside Analog, outside help was rarely sought in evaluating the ventures. By the summer of 1983, “the peak of VC madness,” approximately seven companies were in the portfolio, and closings were occurring every two or three months. Although the extremely favorable valuations on many companies were, at this time, attractive to the majority of investors, Analog followed a different investment strategy. They demanded strategic links with the potential portfolio companies and refused several attractive companies because they disagreed with the method of valuation used and the unrealistically high projections. The process of negotiation with potential portfolio firms was often a lengthy, openended process, taking months to reach an arrangement in line with Analog’s strategic objectives and the entrepreneur’s goals. Analog was not the only investor in many of these deals. In the case of BIT (see Appendix I), Analog took a small equity position, and other needed capital was sought from venture capitalists and banks. ADE helped the founders of BIT to locate and to negotiate additional funding from venture capitalists and banks. Venture capital groups, which did not take BIT seriously at first, were eventually impressed when BIT conducted market research, using ADI customers as a survey pool, and formulated a complete business plan with help from Analog officers. Once the candidate company underwent the screening process and the decision was made to invest, an agreement would be drafted that called for monthly profit and loss statements and balance sheets to be reported to ADI, supplemented with quarterly reports containing greater detail. An Analog vice-president was chosen to sit on the young company’s board, and Analog retained absolute visitation rights. However, Analog did not interfere with the company’s day-to-day management or decision-making. Rather, an emphasis was placed upon building and maintaining effective lines of communication. Regular board meetings and frequent telephone conversations-when Sullivan found a reason to call-were the norm. John Hudson, ADE’s controller, frequently met with or had telephone contact with the CEO of each company. Whether or not this worked in practice depended on the openness and the integrity of the start-up’s CEO. The flow of information was critical to monitoring the companies. Boole sat on the boards of directors of three portfolio companies, and at least one Analog representative sat on the board of every portfolio company. Sullivan served on seven boards. Analog’s philosophy toward management of the portfolio companies was “handsoff.” But ADE was very active in maintaining communication links between Analog and portfolio company managers, as well as in giving advice and attempting to influence strategies and policies whenever the entrepreneurs seemed to ADE to be “going off into left field.” But power was limited: Sometimes the advice was taken; sometimes not. Over the years, however, ADE grew and its focus shifted from building portfolios to making strategic decisions about investments that were already in the portfolio, specifically, ANALOG DEVICES ENTERPRISES, 1980-1988 423 deciding whether to stay with an investment or divest it. According to Boole,. ADE did not have much luck divesting companies. Although ADE did not attempt to manage the day-to-day activities of the portfolio companies, ADE officers did step in to make changes in portfolio company management, -that is, if the company team was not working effectively in their estimation, and especially if they felt that their investment was being threatened and the company was in danger of failure. Sullivan was instrumental in removing the CEOs in four of the portfolio companies, and altering the management structure in another. In one case, in which he was able to move quickly, the change of management was swift and fairly painless. In another situation, however, it took a year to convince two of the founders that the CEO was incompetent. By then, the damage had been done and the investment was subsequently lost. There were also attempts to develop relationships among the portfolio companies, to pursue possible synergies. In the fall of 1983, Analog brought together people from the portfolio companies at a conference in Monterey, California. Although the companies were diverse in their missions and operations, the entrepreneurs discovered a range of overlapping interests as they met and talked. IS and Numerix decided to share the cost of hiring a salesrepresentative to penetrate the oil industry. Other companies began discussions that led to a purchasing pool to get volume discounts on memory chips and other components. Some companies discussed buying and selling with one another. They exchanged ideas on how to reward and motivate employees, and how to buy test-equipment. As always, there were opportunities to refer entrepreneurs to various experts within Analog for technical and marketing advice. Some of the benefits have also been specific financial cooperation. For example, BIT and ADS worked out a margin split agreement, whereby the gross margin of a product after production and marketing costs is split equally between producer and developer. This provides both an incentive and a degree of fairness to the financial rewards of a venture. However, it may not have been possible if BIT and ADS were not “part of the same family.” Analog Devices, Inc. became the exclusive marketing agent for BIT products sold in Europe. This was a significant boost for BIT, which would find it difficult to maintain its own foreign marketing program because of its small size. “It makes sense,” said Boole. “We had an excellent marketing organization in Europe, and they fit in well with products that we make internally.” In this example, managers at BIT were able to learn a great deal about marketing in Europe and the Far East by sharing Analog’s experience. In the United States it is not uncommon for a young firm to contact potential customers during early stages of product development. In foreign markets, where credibility is sometimes a more delicate issue, BIT pushed Analog to do the same, but reluctantly backed off when they learned from Analog that it would not be advisable to aggressively pursue customers until their products were ready to ship. ORGANIZATIONAL AND MANAGERIAL CHALLENGES In addition to the risks inherent in new technologies, the uncertainties of the marketplace, the ambiguities of power, and of the (theoretically) hands-off ADE approach, there were also growing problems at the interface between ADE and its parent cqmpany operations. The distinction between Analog divisions and external start-up companies posed a challenge for the ADE newstream managers, when mainstream division managers saw the outside companies getting investments while the internal divisions were operating under strained 424 R.M. KANTER ET AL. budgets. The solution was further separation. ADE managers explained, to mainstream employees who expressed concern, that the money came from different places for different things, and that the venture employees were risking their careers in ways that Analog regulars were not, since mainstream employees would not lose their jobs if their projects did not succeed. Furthermore, Analog tried its own internal start-up, Digital Signalling Processing Division. But DSPD, as an internal venture, had much tighter control and faced conflicts with the ADE ventures. DSP was established in 1984. By 1986, it had introduced ten new products, including a single-chip digital processor that performs the functions previously done by an entire circuit board, and, soon, had established itself as an industry leader. DSP experienced its own problems of fit within the mainstream organization; as a start-up venture, it was expected to perform like an aggressive, fast-moving, independent entrepreneurial business. Yet it had to contend with the same constraints as any other Analog division. However, as an internal start-up, DSPD did not have the freedom of the ADE venture to establish its own systems. The original general manager complained that his group was trapped in a compensation system designed for the rest of Analog but insufficient for a startup. The bonus system was tied to a company and not the division’s performance, and although there was some discussion of a bonus of 10% to 20% of base salary tied to division performance, it was not considered enough to keep people working long hours and weekends. In portfolio companies, however, employees had equity and a share in the profits. Thus, DSPD had problems attracting and retaining qualified people, and many positions at DSP remained unfilled. The result was an overloaded and undermotivated staff, especially as they compared themselves with the ADE ventures. Finally, the general manager was asked to work with portfolio start-ups on joint ventures; the result was more strain on his divisions resources. Ultimately, he left the company. His conclusion was that internal start-ups face challenges that do not put them in a favorable position to support or take advantage of ventures with external start-ups. Some of Dintersmith’s comments further indicate that there is great difficulty in coordinating the development of a number of complex new technologies simultaneously and also in arranging synergistic cooperative arrangements between different teams of technologists. For example, joint definition can be a problem when two teams do not agree. Each team sees the other as a distraction that hinders them from achieving their separate goals. The resulting effort becomes frustratingly weak, and is often far less than would be possible if all energies were focused toward the same definition and goal. ANALOG’S ASSESSMENT OF ADE A rule of thumb for new businesses is that 2 of 10 might be expected to succeed. Indeed, U.S. government statistics support this, indicating that 80% of new businesses fail within the first five years. Against this probability for success, ADE performed adequately, but below Analog’s expectations, which led to an increased emphasis on financial results and a proposal to end ADE. Since the portfolio investments we studied began in 1980, through 1988, 6 of 13 investments have yielded positive resultspither technological advances or financial gains or both. The initial intent of these investments was always more centered on the strategic benefits that could be achieved from them. In the absence of technical benefits, however, Analog would opt to buy low and sell high to profit from the investments, if possible. Photodyne, the best ADE investment financially, was sold to 3M in 1988 for a 35% compound rate of return. Unfortunately, Photodyne did not yield the technical benefits and synergies ANALOG DEVICES ENTERPRISES, 1980-1988 425 that were hoped for. On the other end of the scale, Gigabit has had sales increases of over 100% every year for the past three years, and it produces all of the microchips on a Gray 3 supercomputer. Gigabit was considered a financial success, possibly becoming the best investment in the portfolio. Analog expected to market Gigabit products in 1991 and 1992. BIT was also a success strategically, producing chips that are contained in ten Analog products and also boosting the value of the ADE investment by 50% in dollar terms. But if new ventures need “patient” money, ADI’s results in 1986 showed that patience is a luxury. An external change put more pressure on Analog to emphasize the financial side of its newstream investments. In 1986, AD1 experienced its first year of stalled growth. The 1986 annual report warned that “so far we have been able to maintain our investments without unduly comprising our bottom line, but looking ahead we know we cannot continue to increase expenses without a commensurate increase in sales.” The financial reports for 1986 and 1985 showed expenses of $4,356,000 and $3,016,000, respectively, related to ADE investments. These expenses came about as a result of a change in accounting, dictated by the SEC. Analog Devices had been employing the investment method of accounting for the ADE investments. Though the accounting rules for corporate venture capital investment are vague, the SEC felt that a modified form of the equity method would be more appropriate. This requirement introduced an unpredictable and uncontrollable element into Analog’s income statement. To reduce this effect, Analog began to seek more investors in each deal. A consequence of this move was a likely reduction in the ability of Analog to realize strategic benefits from its investments. Furthermore, Stata stated, if the SEC’s position had been known before ADE was founded, ADE would never have been born, for the new requirements put the long-term goals of ADE directly in conflict with the short-term goals of ADI. It was only a matter of time before the existence of ADE was questioned. In 1987, Analog’s CFO concluded that “from a financial viewpoint, ADE has been very painful financially and will continue to be for some time,” though he pointed out that “if all we ever do is make money on this by buying low and selling high, then the program will not have met its goals.” If technical benefits cannot be produced, a financial gain can relieve some of the disappointment. However, by 1988, the future of ADE was very uncertain, and Analog was retreating from its commitment to the venturing arm of the company. Each key actor derived a different lesson from the poor performance of ADE, ranging from the strategy itself to the level of expectations set forth. Some felt that the problem was not ineffective technology or market forecasting, but rather, management. Another thought that perhaps the matter of diversification would be better left to the divisions, though he later said that divisions may not want to diversify because of their focus on the base business. Still another cautioned patience, stating that “you’d better be prepared to go ten times longer funding than you expected.” Finally, Sullivan stressed relatedness. In his eyes, the deals that will prove workable are those in which investments are made in companies with complementary technologies that can be exploited by Analog’s core business. The future focus, he contended, should be on obtaining leverage through marketing rights, joint R & D rights, and technology transfer options. A final thought was that it might be concluded that Analog’s investments into divergent businesses just don’t work. IMPLICATIONS The very factors that create success for a venture capital operation can compromise its ability to serve the parent company’s diversification objectives (Hardymon, DeNino, and Salter 426 R.M. KANTER ET AL 1983). For example, there can be a conflict between the organizational independence required to make good investments and the ~imi~t~ons on the kinds of companies to be invested in that are presented by the parent company’s business strategy (i.e. I the desire to find synergies with existing businesses). Some anafysts have argued that me objective of a corporate venture capital program is generally strategic rather than financial, and that the possibilities for synergy include acquisition possibilities, marketing rights, and technology windows for the established corporation as well as capital, credibility, and marketing channels for the new venture (e.g., Winters and Murfin 1988). Indeed, the promise of such synergies could serve as an argument for the greater potential of corporate venture capital over inde~nd~nt venture capital to nurture successful new ventures, as corporations can bring more to the party than simply financing. However, the Analog case and related research suggests a different conclusion. Achieving synergies requires active management of the interface between the funding coloration and the new venture, not simply a statement of the theoretical “fit” (Kanter 1989). However, autonomy is often found to be a condition for effective start-ups as well as for effective corporate venture operations. A survey of 52 corporate venture capital groups found that higher performance was achieved by groups run by managers with greater autonomy than by groups highly controlled by corporate management (Siegel, Siegel, and MacMillan 1988). Autonomy tends to undermine the connections, relationships, and trade-offs required to make synergies work, as the tensions between the internal Analog division and the ADE portfolio companies showed. Thus, as in the case of Analog Devices Enterprises, an established corporation that begins by viewing its venture capital investments as a part of corporate strategy, expecting transferable benefits, may find venture autonomy an unstable situation. As a venture takes its own course, diverging from the direction of the co~oration’s mainstream, or as the absence of hoped-for synergies make financial returns loom larger as the rationale for investment, the company may respond by moving in one af two directions. It could pull the ventures closer to the corporation (through acquisition or tighter management, presenting the risk of harming the fragile venture, if not appropriate to its embryonic state, but also the potential for improved performance-with new management and greater mainstream linkages); in that case, the corporate venture capita1 group has changed its focus. Or, the company could loosen its strategic objectives and tighten its purely financial ones, moving toward becoming only an arn~‘s-length investor. Then, as other investments look more at&active, because of changes in industry conditions (as happened to Analog’s hightechnology investments in the 198Os), or as external changes influence strategic choices (e.g., the accounting rules changes Analog faced), the company can move away from venture capital attogethcr as a source for news~eam d~ve~opnl~nf. REFERENCES Block, Z. 1982. Can corporate venturing succeed? Journal of Business Strategy 3(Fall):21-33. Bower, J.A. 1970. managing the Resource Mlocation Process. Boston: Harvard Business School, Division of Research. Burg&man. R.A., and Styles, L.R. hide Coupoyate huwation: &me&v, Structam, and Manager&t Skills. New York: Free Press. Fast, N.D. 1979. The future of industrial new venture departments. ~~a,str~alMarketing ~anuge~ent X(November):264-273. Galbraith, J. 1982. Designing the innovating organization. Organizational Dynamics 10(Summer):5- l‘z ANALOG DEVICES ENTERPRISES, 1980-1988 427 Hardymon, G.F., DeNino, M.J., and Salter, MS. 1983. When corporate venture capital doesn’t work. Harvard Business Review 61(May-June):l14-120. Kanter, R.M. 1983. The Change Masters: Innovation and Entrepreneurship in the American Corporation. New York: Simon and Schuster. Kanter, R.M. 1985. Supporting innovation and venture development in established companies. Journal of Business Venturing 1:47-60. Kanter, R.M. 1988. When a thousand flowers bloom: Structural, social, and collective conditions for innovation in organizations. Research in Organizational Behavior 10. Kanter, R.M. 1989. When Giants Learn to Dance: Mastering the Challenges of Strategy, Management, and Careers in the 1990s. New York: Simon and Schuster. Kanter, R.M. 1990. Improving the acceptance and use of new technology: Organizational and interorganizational challenges. In B. Guile, E. Laumann, and G. Nadler, eds., Improving the Use of New Technology. Washington: National Academy Press. Miller, A., Wilson, B., and Adams, M. 1988. Financial patterns of new corporate ventures: An alternative to traditional measures. Journal of Business Venturing 4:287-298. Montgomery, C.A., and Wilson, V.A. 1986. Mergers that last: A predictable pattern? Strategic Management Journal 7~91-96. Porter, M.A. 1987. From competitive advantage to corporate strategy. Harvard Business Review 65(May-June):43-59. Quinn, J.B. 1985. Managing innovation: Controlled chaos. Harvard Business Review 63(May-June):7384. Siegel, R., Siegel, E., and MacMillan, I. 1988. Corporate venture capitalists: Autonomy, obstacles, and performance. Journal of Business Venturing 3:233-248. Singh, H., and Montgomery, CA. 1987. Corporate acquisition strategy and economic performance. Strategic Management Journal 8:377-386. Smith, D.K., and Alexander, R.C. 1988. Fumbling the Future: How Xerox Invented, Then Ignored, the First Personal Computer. New York: Morrow. Stein, B.A., and Kanter, R.M. 1980. Building the parallel organization: Toward mechanisms for permanent quality of work life. Journal of Applied Behavioral Science 16:371-388. Stevenson, H., and Gumpert, D. 1985. The heart of entrepreneurship. Harvard Business Review 64(March-April):84-94. Sykes, H.B. 1986. Lessons from a new ventures program. Harvard Business Review 64(May-June):6974. Sykes, H.B., and Block, Z. 1989. Corporate venturing obstacles: Sources and solutions. Journal of Business Venturing 3:159-168. Van den Ven, A.H. 1986. Central problems in the management of innovation. Management Science 32:590-607. Van de Ven, A.H. 1990. The process of adopting innovations in organizations: Three cases of hospital innovations. In B. Guile, E. Laumann, and G. Nadler, eds., Improving the Use of New Technology. Washington: National Academy Press. Van de Ven, A.H., Angle, H., and Poole, MS., eds., 1989. Research on the Management of Innovation: The Minnesota Studies. New York: Ballinger/Harper and Row. von Hippel, E.V. 1977. Successful and failing internal corporate ventures: An empirical analysis. Industrial Marketing Management 6. Winters, T.E., and Murlin, D.L. 1988. Venture capital investing for corporate development objectives. Journal of Business Venturing 3:207-219. 428 R.M. KANTER ET AL. APPENDIX 1. THE PORTFOLIO COMPANIES (AS OF OCTOBER 31, 1986) TAU TRON, INC. An early investment of $1,547,000 in Tau Tron, Inc., a manufacturer of high-speed digital signal communications test-equipment, was sold to General Signal for $1,933,750 in June 1982. The proceeds were reinvested in the portfolio. SIGNAL PROCESSING CIRCUITS INC. (SPCI) The first ADE investment occurred in 1980 after a group of Utah engineers approached ADE for seed capital. ADE quickly invested $700,000 after a panel of Analog marketing people and engineers reviewed the venture. The agreement gave Analog the option of later acquisition, which was actually done in October 1983. This investment-along with an internal investment to develop digital signal-processing circuits, based on the high-speed process developed by SPCI-led to the current Digital Signal Processing division of Analog Devices. NUMERIX CORPORATION Larry Sullivan and Robert Boole searched nationwide for a maker of array processors, computer components that increase the speed of number crunching for minicomputers. After a year of searching, they saw an advertisement in a trade periodical for a firm called CNR, Inc. CNR was developing an array processor that looked very promising. Sullivan conceived of a very complex deal under which the key people and the array processor project were spun out of CNR into a new company, Numerix. Analog invested $1.5 million in Numerix. The chief technologist (who was also the founder) had been working on array processors that might have significant applications for Analog’s customers. For example, in May 1983, Numerix began marketing a 32-bit array processor that is three times faster than the market leader and nearest equivalent. This device can be used to improve computer-aided design (CAD) systems by making complex sets of calculations almost instantly. As part of its investment deal, ADE picked up the right to acquire Numerix. This option effectively made AD1 the sole supplier of Numerix capital needs. Numerix’s subsequent fast growth and its need to fund follow-on products to its first products required AD1 to invest another $8 million by the end of 1986. INTERNATIONAL IMAGING SYSTEMS (IzS) In 1981 Boole and Sullivan sought a means of participating in digital image processingusing computers to interpret medical images, natural resource mapping, and landsat photographs. A phone call to Milpitas, California, put Boole in touch with the president of International Imaging Systems, Inc., a six-year-old, six-million-dollar company. The president was trying to figure out how venture capital could solve the cash flow problems for this producer of image-processing hardware and software. ADE invested $3 million in 12S in June 198 1, receiving about 45% of the equity in return and a long-range option to acquire 12S. ANALOG DEVICES ENTERPRISES, 1980-1988 429 JUPITER SYSTEMS ADE learned about the importance of high-re~lution graphics in its dealings with IS. Boole and Sullivan therefore went in search of complementary businesses. At a Boston computer graphics trade show in the summer of 1982, Boole discovered Jupiter Systems. At the same time, Jupiter was getting short of funds and the president was seeking new sources of financing. ADE invested $2 million in Jupiter Systems, receiving about 20% of its equity and the option to acquire the company in the future. It was believed that this product line would complement Charles River Data System’s micr~ompu~rs and 12Simage processors. ADE hoped that Jupiter Systems could provide sales op~~unities for AD1 DSP products, and that Jupiter products could eventually be used by ADI’s Measurement Control Systems Division. CHARLES RIVER DATA SYSTEMS, INC. This firm was founded in 1973 as a supplier of DEC peripherals. The firm produced the Universe 68, a microcomputer system used in scientific, engineering, and industrial data transaction applications. “We decided to get involved in Charles River Data Systems simply to learn about where the industry was going,” said Sullivan. “We knew this type of computer would have an impact on our business. ” ADE invested $1 million in 1982 and invested $1.9 million in subsequent rounds. GIGABIT LOGIC, INC. Gigabit was a start-up working on gallium arsenide chips, a material that can be used to make chips three times faster than silicon chips. Analog invested $1.5 million in seed capital, for which it got the rights to market some gallium arsenide chips in the future and the right to produce gallium arsenide chips for use in its own products that do not compete with GigaBit’s. ADE had invested another $1.9 million in GigaBit by the end of 1986. QU~~AT~ TECHNOLOGY CORPORATION (QTC) Boole and Sullivan discovered that users of array processors had special software needs. Synergy in array processors and digital signal processing would be found if ADE could help to satisfy this demand niche. ADE thus invested $500,000 in QTC, a start-up located in Portland, Oregon, in 1982. To fund QTC’s growth and help launch a new product line, ADE invested another $600,000 in QTC in 1985. ADI then picked up the right to acquire QTC in the future. BIPOLAR INTEGRATED TECHNOLOGY, INC (BIT) ADE invested $300,000 in this start-up in July 1983. In return, Analog got a joint-marketing agreement and a technology-tmnsfer agreement. The seed capital, only a fraction of the $10-11 million BIT would require to begin full-scale manufacm~ng, was treated as a loan, with an interest rate of 10%. This would later be converted to common stock. The rest of the funding was to be provided in a complex plan for three rounds of financing that would carry the firm from bare bones start-up to expanded operations and first shipping to fullscale production. ADE invested $2.8 million in the subsequent rounds. 430 R.M. KANTER ET AL. The goal of BIT founders was to produce the fastest bipolar chips for the digital signalprocessing and the data-processing markets. The goal for further development of bipolar technology, like complementary metal oxide semiconductor (CMOS) and gallium arsenide (GaAs), the competing technologies for chip manufacture, was to lower the power consumption and increase the speed of chips. Miniaturization was also a goal, and BIT has developed the LBT (Little Bity Transistor), the smallest bipolar transistor. George Wilson and several colleagues, the founders of BIT, had worked at Tektronix, where they knew Barry Gilbert, ADI’s technologist in Oregon (hired away from Tektronix). While Tektronix was committed to gallium arsenide technology, Wilson and his colleagues spun off to work on their ideas for bipolar processes, which they believed could be used to make a fast and dense chip. Although Analog discussed the possibilities of acquiring the firm at some time in the future, Wilson et al. were definitely not interested. They preferred independence and the chance to take the company public. TEST SYSTEMS, INC. Test Systems in Tempe, Arizona, was developing the market for inexpensive test equipment used to check circuit boards. This area complements ADI’s Component Test Systems Division, which works on methods of testing integrated circuits. Both seek to offer the market more cost-effective methods of testing. Test Systems sought management and organizational advice from ADE as well as help in shaping its market image and launching a sales program. ADE invested $9 million in Test Systems in 1983 and received an option to acquire the company at a subsequent date. PHOTODYNE ADE invested $1 million in this maker of fiber optic test equipment in 1982. The company pioneered the use of fiber-optic sensing heads and specialized signal-processing circuits for use in their equipment. From its Newbury Park, California, headquarters, Photodyne has achieved impressive sales in the United States and abroad, where 35%-40% of its sales are generated. To fund Photodyne’s fast growth, ADE invested an additional $1.5 million in Photodyne in 1985 and received an option to acquire Photodyne in 1989. TIGRE ADE discovered Tigre, a French firm, through I’S (Tigre and I* maintained agreement). A $200,000 investment was made in 1984. a joint marketing INDUSTRIE IMAGERIE SYSTEME ADE discovered this French company with the help of venture capitalists during visits to several European countries to scout investments. An investment of $500,000 was made in 1984. AD1 worked out joint R & D and U.S. marketing agreements with the company as part of the investment. This portfolio company has been financially successful and has recently started selling its products in the United States.