In 1974, the stock market was low, capital gains taxes were high, and venture capital money was very scarce. Seed money was even scarcer. But Thomas Perkins and Eugene Kleiner were an unusual pair of venture capitalists. Both helped found companies of their own (University Labs, which later merged with Spectra-Physics, and Fairchild Semiconductor, respectively), and they encouraged Treybig to keep working. Two of the first people he talked to were old Hewlett-Packard associates, Michael Green and James Katzman. Green had designed the minicomputer industry's first time-sharing system for the H-P 2000 in 1968, and Katzman had designed the H-P 3000.
Green and Katzman liked the sound of the technical challenge and the idea of a new business adventure. Katzman had already left Hewlett-Packard to work at Amdahl Corp., and Green was tired of reorganizations and canceled projects at Hewlett-Packard. Though Treybig's concept of the system was a little sketchy, "something you could draw on the blackboard in about 10 seconds," says Green, he was attracted by Treybig's enthusiasm and the sense that he seemed to know what he was talking about.
At the same time Perkins was working at University Labs, he was also the general manager of the computer division and the director of corporate development at Hewlett-Packard. There he'd had the chance to observe Treybig, Green, and Katzman at work. He liked what he had seen enough to decide in mid-1974 to put up $50,000 to recruit Green and Katzman to start working on the definition of a system that could fit the market application that Treybig perceived.
Over three or four months, Treybig, Green, Katzman, Perkins, and another Kleiner & Perkins partner, John Loustanou, had an opportunity fairly rare in start-ups -- to think out the company's development in detail before starting."We worked out all aspects in advance -- people, finance, manufacturing, everything," says Treybig. "We could afford to do it because we weren't starving to death. One reason we have been able to grow so fast and not have problems was that we thought through most potential problems before we started."
They worked together exceedingly well. "We didn't have to discover each other," says Perkins. "We had all worked together. There were no problems about who was boss, no politicking, no three vice-presidents who wanted to be president. Many of the troubles that can plague a new venture weren't there."
From the beginning Tandem's strategy was clear and, Treybig claims, simple. The goal was to produce a price-competitive computer system that didn't fail, that didn't destroy data, that could be expanded from the power of a minicomputer to that of the biggest mainframe, and also networked with as many as 255 systems in other locations, without changing software or hardware. The challenge was to do it technically, and to grow fast enough to be able to fend off competitors and provide opportunities for the talented and aggressive people they expected the venture to attract.
At the end of four months, progress seemed good, and the group was ready to go out after more money.
"Who needs another minicomputer company?" was the reaction of most potential investors Kleiner and Perkins approached. Treybig and the rest of the team had never started a company before, and no one was sure that the technical problems of building a fail-safe, modularly expandable computer could be solved.
But Perkins and Kleiner never doubted the plan or the management, and they decided the odds on the technical risk were good enough to put in an additional million of their own -- a sizable investment for an $8-million fund. And John Loustanou decided to join the Tandem team as chief financial officer.
A year later, with a formal plan, 10 employees, and a much more detailed design of the new system, the venture had built up enough momentum to raise another million from outside investors, though neither the market nor the capital gains tax rates had improved. One of the investors and now a member of the board, Franklin "Pitch" Johnson of Asset Management Co., a Palo Alto, Calif., venture capital firm, points out that the track records of Perkins's and Kleiner's other ventures added to the attraction of the deal. Johnson, whose company's initial $60,000 investment has increased in value more than 5,000%, has yet to sell a share.
Tandem delivered its first system in May 1976, to Citibank, and except for a brief hiccup in the last quarter of 1976, when the company had no orders in October, November, or the first half of December, growth seemed steady and almost uneventful. Though the first year's business was about half of what had been expected, after that revenues and income caught up to the original projections, sales of $7.7 million in '77, $24.3 million in '78, $56 million in '79 and $109 million in 1980. Expenses came in as projected, and partially because of the support of Kleiner & Perkins, cash crises never materialized. Tandem was one of six small companies to have an initial public offering in December 1977, raising almost $9 million for about 21% of the company. A total of 770,000 shares were sold at $11.50 per share. The stock price is now in the 80s, with almost 12 million shares outstanding.
Today Tandem occupies 302,000 square feet in three connected cement-and-glass buildings next to a shopping center in Cupertino. Floor-to-ceiling windows in the cafeteria look out on the company pool and volleyball courts. In Silicon Valley, the company is recognized as a leader not only in growth, but in attracting and keeping talented people. Turnover is low, about 8%, less than a fourth of the American Electronics Association median. And productivity is high, sales of $98,400 per employee in fiscal year 1980, about twice the AEA median.
The door to Treybig's office is usually open, and his calendar for the days he spends at Tandem is mostly blank. He likes to reserve his time for thinking about company strategy, for drop-ins -- managers who would like his help selling the company to a prospective customer or employee, employees who have a problem their manager can't solve -- and for projects like putting "everything you need to know about running a company" and Tandem's five-year plan on one piece of paper.
