Data General Corporation Business Information, Profile, and History

History of Data General Corporation

Data General Corporation is a manufacturer of multi-user computer systems: minicomputers, workstations, and servers. During the minicomputer boom in the 1970s, Data General was one of the fastest growing U.S. companies and was considered one of the leaders in minicomputers. Since minicomputers began losing sales to personal computers in the 1980s, however, Data General had difficulties adjusting its market focus. In the early 1990s the company began to establish itself in the area of workstations, servers, and data storage systems. With 33 subsidiaries and more than 250 sales and service offices in 60 countries, Data General is a significant international company, earning about half its revenue from foreign sales.

The minicomputer, a medium-scale, cabinet-sized computer that functions either as a single workstation or as a multi-user system with numerous terminals, was first introduced in 1959 by Digital Equipment Corporation. In 1968 Esdon de Castro and two other young engineers, quit their jobs at Digital to start their own company, Data General. At Digital they had been working together on a project for a new 16-bit computer, faster than the existing model that functioned at a rate of only 12-bit units. It is unclear whether they had planned ahead to start their own business with technology they were developing at Digital, or whether they quit because management had turned down a proposal of de Castro's to build a whole new series of computers that would have made much of Digital's line obsolete. The end result was that Data General's first product, the 16-bit NOVA minicomputer sold very well by filling a gap in Digital's product line. Ten years later Digital's president told Fortune magazine, "What they did was so bad we're still upset about it." But Digital never sued the new company for any theft of intellectual property.

De Castro and his two colleagues, joined by salesperson Herb Richman of Fairchild Semiconductor, obtained financial backing from lawyer Frederick Adler. To start the company Adler put up $50,000 of his own money and raised much of the rest of the founding capital, which totaled $800,000. With that Data General was incorporated on April 15, 1968, and set up operations in what had been a beauty parlor in Hudson, Massachusetts. A year later the company moved to Southboro. De Castro was president and CEO, Richman became vice-president of marketing and sales, and Adler served as secretary and a board member in addition to managing the finances for the first few years.

Data General's first machine, the NOVA, unveiled at the annual national Computer Conference in 1969, was a great success. Besides having several features that Digital's comparable computers lacked, the Data General multi-purpose minicomputer was produced very cheaply and thus could be sold at a relatively low price, averaging $26,000. The NOVA had the advantage of specially designed large circuit boards, which reduced the amount of hardware in the computer. The company shipped more than 200 of these minicomputers in its first year.

At the time, the minicomputer market, unlike that for mainframes, primarily consisted of engineers, scientists, and purchasing agents of original equipment manufacturers (OEMs). Therefore, it was relatively easy for a new company such as Data General to break in with inexpensive advertisements in key trade journals. These clients were more interested in a bargain than a famous brand name, and Data General offered volume discounts as high as 40 percent. Furthermore, the company did not have to provide a service organization as the big computer manufacturers did, since its clients were technically capable of taking care of the computers themselves. Soon 70 percent of Data General's clients were OEMs, which packaged custom software and peripherals with core computers and resold them to the final customer.

Shrewd business management from the beginning also aided Data General's successful growth. It went public after only a year of business, in 1969, and, raising money by offering stockholder equity shares, did not have to go into debt. The owners used the stock as a means of growth rather than hoarding a majority for themselves. When they made a second public offering several years later, the founders followed the advice of Adler and each sold some of their own stock in the process. Data General stock was listed on the New York stock exchange on December 28, 1973.

The company was also known for keeping down its overhead costs. It had low receivables and inventories as a percentage of sales. A relatively large proportion of the sales crew's pay was from commissions. With all its computer lines simple and compatible with each other, Data General did not have to develop new systems software. Its R&D expenditures were not put into risky new technology areas, but rather spent on ways to improve and cut costs on existing products. Finally, its offices, even after moving its headquarters a second time in 1977 to the present location in Westboro, have been kept simple, and there have been almost no executive perks.

Data General's early advertising style was aggressive, compatible with the business behavior of its OEMs and in sharp contrast to the polite customer support offered by the mainframe companies such as IBM. Data General thus sought to differentiate itself from its chief competitor, Digital, precisely by the aggressiveness of its sales staff. Although it may have seemed that Data General had a long way to go the take on Digital, which held 85 percent of the minicomputer market, there was no company in firm second place. In a similarly aggressive style, Data General was quick to take copiers of its designs to court.

Data General quickly followed the new trends in technology, learning from the leader's mistakes, and providing cheaper machines. The company kept its prices down by such techniques as using plug-in printed circuit boards in its computers, instead of hand-wired ones, the first minicomputer manufacturer to do so. In 1973, Data General was the first major company to introduce a minicomputer with a new core memory design permitting twice as much memory on a single circuit board as was then standard. Thus, the NOVA 2, at the same price as its predecessors, could support more complex software. This enabled them to run programming languages such as FORTRAN, which previously had been used only on mainframe computers. In the mid-1970s Data General introduced its second product line, the more powerful ECLIPSE scientific and commercial computers. These new minicomputers did not differ greatly in their architecture from the NOVA and were compatible with the NOVA in their system software.

In 1978, just ten years after its founding, Data General, as the fastest growing computer company to date, was listed 500th in the Fortune 500 rankings with sales at $380 million. Yet it ranked much higher than that in profits. With a goal of providing stockholders a high short-term rate of return on their investments, Data General was able to maintain its 30-40 percent growth rate. Expansion was manifested in the hiring of 7,000 new employees between 1974 and 1978. Managerial complications associated with such rapid growth caused the company to fail only once in fulfilling customer orders, in 1973. However, it was unable to substantially increase its 8-11 percent share in the minicomputer market, because the market was growing even faster at the incredible rate of 40-45 percent annually.

Meanwhile, Data General had grown in other directions, as well. It established its Canadian, European, and Asian operations, and created a technical service organization. It also began manufacturing its own computer components, setting up a semiconductor operation in Sunnyvale, California, to produce its own microprocessors and other chips. Around 1976 de Castro decided to fully integrate the company by also producing peripherals.

Eventually the company's physical growth became too complex to be run in the same manner it had been. For too long de Castro had tried to manage the whole company himself. Lack of coordination led to a costly proliferation of research projects. In one case two competing research groups were funded to develop a 32-bit "supermini," a minicomputer powerful enough to compete with a mainframe. Internal disagreements in 1976-77 among engineers, concerning whose project team would design the new computer, postponed execution of the project, while Digital's 32-bit VAX, introduced in 1977, was becoming very popular. After a frantic year and a half of work, chronicled in Tracy Kidder's best-selling book The Soul of a New Machine, Data General finally announced its 32-bit MV/8000 in April 1980. This was the upgraded version of the ECLIPSE line. But the company's delay resulted eventually in a smaller share of the new supermini market. Meanwhile, during the fourth quarter of 1979, the hiring and training of many employees for its new service department resulted in a decline in earning for the first time.

In response to managerial deficiencies, in 1980 the founders decided to reorganize the company into a divisional structure. Already in 1979 the company had started a transition away from a strictly functional organization, by establishing product-oriented groups. Now new executives were brought in from outside, and decision making was decentralized among three divisions, each oriented to a different target market. In addition to new middle level managers, a new senior vice-president, Robert C. Miller, a veteran of IBM, was appointed in 1982 to head the company's three business divisions, relieving de Castro from day-to-day operations.

However, these changes were not enough to put Data General back on the fast track. The real problem the company faced in the 1980s was that the minicomputer boom had slowed and so had Data General's sales. Earnings in fiscal 1981 fell from $55 million to $41 million. In fiscal 1982 profits dropped an additional 51 percent to $24.6 million, while revenues grew that year by only 9.3 percent to $806 million. Since then, Data General had not been able to regain for long the rate of growth it had seen in the 1970s. Basically, it was unable to adjust itself quickly enough to serve the new growth segment of the market: personal computers and workstations.

In 1981, the year Data General shipped its 100,000th computer, it made its first entry into the lower-end personal computer market with its "Enterprise." Unfortunately, this product failed due to its high price and limited software ability, reflecting the company's lack of familiarity with the new mass market for personal computers. Data General's next personal computer, released in 1983, had the ability to run both standard PC software and some of the software that ran on the company's minicomputers. But by then the competition from Apple, IBM, and others was overwhelming, and sales tended to be only to existing minicomputer clients. A laptop computer released the following year was also unsuccessful, because its display was not satisfactorily legible, and demand was overestimated.

Meanwhile, under new management, Data General began selling direct, in IBM fashion, to end-user clients, rather than to OEMs. It went after these new, commercial markets with its superminis by incorporating more software, peripherals, and support. The company developed an office automation system, combining functions such as word processing, electronic mail, and filing, which it named CEO (Comprehensive Electronic Office), and which ran on the new MV/series of superminis. Its customers tended to be large corporations that had previously been served by IBM or Wang Laboratories.

However, this change of client required a different kind of sales force focusing on the computer's abilities to solve office problems rather than its abilities to process data and support programming languages. "Its hard selling tactics were increasingly out of step with a changed marketplace," observed a 1982 Forbes article. Many of Data General's most successfully aggressive salespeople left the company at this time. Another obstacle that faced the company in this new market was a preference by customers who already had computer systems to stick with their previous supplier. It was one thing for Data General to try to take on Digital in the minicomputer wholesale market of the 1970s, but, as later became clear, it was a wasted effort to try to take established end-user clients away from IBM.

Sales temporarily rebounded in early 1984, up 40 percent, bringing revenues over the $1 billion mark. The organizational restructuring of 1980 was finally paying off, and retrained sales crews brought in a few big accounts with the company's office automation system, including the U.S. Forest Service and E.F. Hutton. Data General's latest high-end supermini, the MV/10000, in 1983 was indeed superior to the competition.

The spurt was short-lived, however. Expecting continued high growth, de Castro had added nearly 3,000 new employees and increased capital spending by 78 percent to meet the expected new demand. But it did not materialize. With its only successful products in the slowing minicomputer market and with the computer industry as a whole in a slump by the mid-1980s, high growth could not be sustained. Profits plummeted 70 percent in 1985. The following year, Data General suffered its first losses, which accelerated to $127 million in fiscal 1987.

The company reacted by downsizing. Already in 1986 it had closed its semiconductor plant in Sunnyvale, laying off 75 employees. It also closed two plants that made computer terminals and printers and merged two internal computing operations, eliminating 400 more jobs. The following year 1,000 were laid off and three more facilities were shut down. In cutting back on marketing and service, de Castro, once again fully in charge following Miller's resignation in 1987, decided to focus sales for the time being on its traditional, more familiar market of OEMs.

A major reason minicomputers were becoming less popular than personal computers and workstations was that the latter were built with standard "off-the-shelf" microprocessors for use with common systems software, a concept referred to as "open systems." Minicomputers, on the other hand, were built with the manufacturer's own unique hardware and software. Open systems gave the user the ability to exchange software and data between various brands of computers. Realizing this, de Castro decided to base Data General's next generation of computers on the simpler, yet faster RISC (reduced instruction set computing) microprocessors made by Motorola and run them on a UNIX-based operating system. This was the AViiON line, introduced in April, 1989. With the experience and technology to produce at low cost, Data General had a potential advantage in a market of standard computer systems. Furthermore, purchasing existing microprocessor chips and operating systems software licenses would save the company the engineering costs in these areas. In 1988 Data General was continuing to close plants and lay off workers and was losing what share it had in the declining minicomputer market. Market share had fallen to 0.9 percent of worldwide revenues for 1989, according to market researcher Info Corp.

In December 1990, before the AViiON could prove itself, the board of directors, apparently fed up with mounting losses for five straight years, removed de Castro from the chair and replaced him with Ronald Skates, who had already been CEO for two years. Skates proceeded to pare costs and plan Data General's future around the AViiON line. Skates cut research and developed expenditures to eight percent of annual sales from de Castro's 12 percent. He withdrew funding from major joint development projects with other companies, and he accelerated the layoffs, reducing the work force by 50 percent between 1986 and 1991. To help pay off debts, in March 1991 he sold off the Japanese subsidiary Nippon Data General for about $46 million.

The AViiON was relatively successful, grossing $200 million in revenue in fiscal 1991 to help make that year profitable. Most AViiON sales were of its high-end server, a computer designed to store data used by other computers linked with it in a network. However, the market for such server computers was still relatively small. AViiON sales were unable to offset declining sales of minicomputers, and Data General fell back into the red in 1992.

The company's latest product strategy introduced in late 1992 was data storage systems. It formed a new division dedicated to producing storage products based on the high-reliability technology of redundant arrays of inexpensive disks (RAID). The system, called CLARiiON, was to be used especially in conjunction with the UNIX-based computers of other manufacturers, such as IBM, Sun, and Hewlett-Packard. With CLARiiON, Data General was the first large company to provide RAID for UNIX. With a maximum capacity of 24 G (billion) bytes, the desk-side CLARiiON was able to store more data than most companies needed, but it was hoped that it would be used to provide the support for new memory-intensive applications, such as imaging and voice recognition. For clients who missed the high storage capacity of mainframes, but wanted open systems, RAID was the answer.

By early 1993 the company had completed two straight quarters of profits, and sales of its AViiON and CLARiiON systems were steadily growing. While other large computer companies, such as Digital and IBM, were beginning to cut back, Data General had already completed its downsizing. From its high of 18,000 employees in 1985 it had reduced itself to a lean 6,900. Data General looked like it was ready to move ahead.

Principal Subsidiaries: Data General Europe, Inc. (France); Data General, Ltd. (United Kingdom); Data General GmbH (Germany); Data General Australia Pty, Ltd.; Data General (Canada) Ltd.; Data General de Mexico, S.A. de C.V.