One More Year or Bust

 SANDRA KURTZIG

In the waning months of 1975, Sandy Kurtzig would never have been able to imagine success on Atari’s scale. After ASK failed to write a program for Hewlett-Packard’s 2100 minicomputer, HP had given her a chance to develop a version of MANMAN for a different machine—but the effort was not going well. ASK’s two programmers occasionally lugged their sleeping bags to Hewlett-Packard, where they figured they might end up spending the night in the low-ceilinged offices, sleeping in shifts in one of the supply closets. HP gave them free access to the flagship HP 3000 minicomputer, the machine they were writing for, but only after 6 p.m., when the HP engineers had gone home for the night.

The ASK programmers worked at HP’s sprawling hundred-acre campus in Cupertino, headquarters of the company’s commercial minicomputer business. Two miles south was the Rancho Rinconada neighborhood where Fawn Alvarez had grown up. Three miles west, almost a straight shot down Homestead Road, Steve Jobs and Steve Wozniak were trying to build a personal computer in Jobs’s parents’ garage.^IV

On this particular night, Sandy Kurtzig had joined the programmers. She was aware of the stakes as she watched the two young men tap away at the computer, their twentysomething faces vaguely green in the screen’s reflected glow. ASK needed to write a version of MANMAN that could run on a minicomputer. The new machines were threatening the time-sharing services that provided the bulk of ASK’s income. Minicomputer prices were dropping, and customers were beginning to realize that they could afford a dedicated machine of their own.¹ Writing a version of MANMAN for the HP 3000 would also strengthen ASK’s relationship with HP. Kurtzig had blown it with the computing giant once already. She could not let that happen again. The customer who wanted the HP machine was the defense giant Boeing. Boeing’s Electronic Support Division had agreed to pay ASK $50,400 for a MANMAN program on the HP 3000. The deal was more than twice the size of the Powertec job that had been canceled.²

Marty Browne, the rock-and-roll-loving programmer, and Roger Bottarini, a computer scientist Kurtzig had hired from UC Berkeley, had been working on the Boeing program every night for months. The plan was to write something that, with a few modifications, could be sold as a generic product for anyone buying an HP 3000 minicomputer. Browne and Bottarini were writing MANMAN as a series of text prompts, intended to be easy to follow, even for people who had never used computers. “Enter part number,” the program would command, and a person at the keyboard could do just that.

Occasionally Kurtzig, who had moved herself out of programming and into a more traditional CEO role, joined Browne and Bottarini, the three of them alone at night in the HP building, aside from the security guards. “She was mostly there for moral support,” Browne says. Kurtzig rarely coded, and for that Browne was privately grateful. She had written serviceable code when the company had launched, but as the MANMAN program had grown more complex and moved to new platforms, he and Bottarini found it easiest to rewrite what Browne called “Sandy code.”³ Kurtzig, meanwhile, was discovering that her strength was translating customers’ needs to the technical staff.

Kurtzig used the time at HP’s offices to absorb details about its corporate culture. She knew about the company’s flexible work hours and traditional twice-a-day coffee breaks, complete with doughnuts, designed to promote informal exchanges among employees at all levels. She knew about the founders’ famed “management by walking around,” visiting employees at their desks or on the line, inquiring how work was going. Now she noticed smaller things: how, for example, every security guard’s desk displayed a photo of Jimmy Treybig, who had left HP to launch Tandem Computers with backing from the young venture capital firm called Kleiner & Perkins, which had an office near Don Valentine’s Sequoia Capital. The security guards feared that Treybig might return to try to steal secrets.^III

Browne and Bottarini were writing good code, but Kurtzig began to worry that her company would fail again to write a program for an HP machine. Boeing kept changing the requirements for its version of MANMAN: demanding, for example, twenty-four alphanumeric characters for part names instead of MANMAN’s standard fifteen characters. Kurtzig’s primary contact at Boeing, who had signed many of his early memos to her “Love,” passed along change after change, the notes becoming brusquer as Kurtzig pushed back. Browne, who remembers the entire summer of 1975 as “a nightmare,” was convinced that there were more people writing change orders at Boeing than writing code for ASK.⁴

Kurtzig embraced a wide range of oportunities for ASK in the years following the company’s funding. (“The only long-range plan I’d had was lunch,” she liked to recall years later.)⁵ At a time when having more than one customer in a single industry made a young software company an expert in that field, ASK had become a manufacturing expert because Kurtzig had taken on multiple manufacturers as customers. Saying yes to Tymshare had moved ASK into time sharing, and saying yes to HP was bringing MANMAN to minicomputers.

ASK had only a half-dozen employees. With this small team, the company was either already working on, or considering working on, projects for seven different customers on five different types of minicomputers made by three different manufacturers (HP, DEC, and Data General). Computers from different manufacturers required different programs. Even different models of computers made by the same company had little in common.

ASK was still selling MANMAN on the Tymshare network, and there were even a few companies still using ASK’s batch-processing service operated by the punch-card ladies in Palo Alto’s Town and Country Village. Tymshare and the service bureau were not cutting edge, but those businesses required little attention from ASK and together brought in enough money to fund the development work at HP. Kurtzig loathed debt.

Despite the multipronged efforts, ASK was barely profitable. The company had $143,000 in sales but only $2,900 in profits in the first six months of 1975.⁶ As the year neared its end, Kurtzig decided that she would have to focus ASK for it to survive. A project for Hughes Aircraft, written for yet a third HP minicomputer, seemed headed for success. She would retain that job, continue to collect from Tymshare and the batch-processing service—and jettison everything else.

The decision to streamline meant denying the requests to write versions of MANMAN for DEC and Data General platforms. The decision would also cost Kurtzig her administrative second-in-command, who had brought in the DEC deal and resigned after she turned it down. But most significant, focusing meant pulling out of the Boeing deal that kept the ASK team awake through the night at Hewlett-Packard.

When Kurtzig told Boeing that she wanted out, the aerospace giant threatened to sue. Kurtzig countered that “Boeing Screws Small Company” might make a great newspaper headline.⁷ In the end, Kurtzig extricated ASK and even negotiated a payment from Boeing that she estimates at nearly $100,000. Streamlining was painful, but she believed it was ASK’s only hope for survival.

In January 1976, after almost two years and two false starts, ASK delivered a functional version of MANMAN on an HP minicomputer to a customer who paid for it. Kurtzig’s decision to focus all of ASK’s minicomputer efforts on a single machine had yielded success. The program, written in FORTRAN with an easy-to-follow interface, would be the model on which future versions of MANMAN would be based. ASK had become one of the few software companies to transition from time sharing to minicomputers.⁸

The customer was the Industrial Products Division of Hughes Aircraft, the same company where Mike Markkula had his first engineering job. It is not surprising that two of the first three customers that wanted MANMAN on HP computers—Hughes and Boeing—were divisions of major defense contractors. Defense-based industries, with their lack of price sensitivity, had driven technological change in Silicon Valley since before the Second World War, when FMC (originally Food Machinery Corporation) had begun selling the tread technology used for orchard tractors to the US government for use in tanks.

The Hughes Industrial Products Division, based in Carlsbad, California, built laser cutters for the apparel industry and wire-bonding equipment, among other products. For $150,000, Hughes purchased an HP 21MX computer loaded with MANMAN software as well as tape drives, discs, two printers, and four terminals that the company spread among the stockroom, purchasing department, and receiving department. Employees in these departments had no previous computer experience, but by responding to the MANMAN prompts that scrolled on the screen, they could enter and retrieve up-to-the-minute information about inventory and ordering.^II MANMAN “looks at the quantities we are after in our bill of materials, looks at the stock, purchase orders and work-in-process, and then comes back and tells us what we have to buy, when we have to buy it, when it should be in the stockrooms and when we can ship to the customer,” explained Hughes’s data processing manager.⁹

Soon Hughes reported that whereas it had once taken three full-time employees to enter stockroom transactions on paper, and three weeks for those transactions to post, with MANMAN, one person could do the same job in three hours, with instantaneous posting.¹⁰

Hughes, HP, and ASK jointly promoted the sale in press releases, advertising, and conference presentations. The collaboration benefited all three companies: Hughes appeared a forward-thinking business that embraced cutting-edge technology; Hewlett-Packard could announce a foray into commercial computing; and tiny ASK appeared a partner capable of working with giant companies. Kurtzig, who felt that the press release sent out by HP could be more effective, rewrote it and sent it to manufacturing trade publications. When she still did not receive the response she wanted, she sent the identical press release again.¹¹

Kurtzig highlighted the Hughes sale at the 1976 American Production and Inventory Control Society show in Atlanta, where she was mistaken repeatedly for a “booth babe,” a professional model hired by a company to lure visitors to its display. The mistake was understandable at a conference with a separate spouses’ program (featuring a crafts fair), opened by a beautiful blonde in a Gone with the Wind–era flouncy dress who had been hired to greet attendees as a “Southern belle.”¹²

At roughly the same time that ASK delivered the computer loaded with MANMAN software to Hughes, Kurtzig gave birth to her second son, Ken. Like his older brother, Andy, now three, Ken was born on a weekend. Kurtzig explained a few years later, mostly in jest, “I told God I had a project that had to be finished, and He’d just have to hold the baby until Saturday.”¹³

The only downside to the delivery of the machine to Hughes was that it left ASK without a computer. A software company could not write software without a computer.

There was hardly money to buy one. ASK had profits of less than $3,000, and the machine Kurtzig wanted, an HP 3000, cost more than twenty times that. Unwilling to burden ASK with bank debt, Kurtzig obtained a three-year, $25,000 loan from her father at 8 percent interest.¹⁴ She also negotiated a deal with HP to buy the computer with a relatively small down payment and pay the balance with one copy of MANMAN, which retailed for $35,000.¹⁵

ASK installed the computer in the company’s new rented office space. The low-slung building on El Camino Real in Los Altos was an easy drive from Kurtzig’s home. For $714 monthly rent, Kurtzig got what she called “1200 square feet of plush office space” and a receptionist she shared with other tenants. She had the entry painted a vibrant orange.

Kurtzig had her own office, as did Marty Browne, whom Kurtzig had named operations manager. Everyone else worked at wooden desks in a large open space, enjoying a special sense of camaraderie. “I spent several years of my life after I left ASK trying to find that again in a company,” one early employee recalls.

“Sandy was basically hiring one employee per year when I got there,” explains Liz Seckler, who joined ASK in 1976 as employee number six.¹⁶ The employees were nearly all fresh graduates from Stanford or Berkeley, armed with degrees in computer science (from Berkeley) or math (from Stanford, which did not have an undergraduate computer science major until 1986).¹⁷

They went to lunch at the Boardwalk, a new beer-and-burger joint on El Camino resplendent with stained-glass windows, an oak bar, and a full lineup of pinball and video games. At the height of the national fascination with jogging in the 1970s, the ASK group donned their short shorts and sweatbands to run together at a local park. They visited each other’s houses for dinner or on weekends. They held low-key Friday-afternoon beer busts, the pressure to leave their desks beginning to build around 4:30 p.m. “There was no real difference between night and day, whether you were working or not,” recalls Marty Browne, who adds that as ASK grew and organizational structures became more formal, the group put a printer on top of the beer-filled minifridge so that they could claim that the fridge was a printer stand.

“It was an extension of the college experience,” says Seckler. “It wasn’t artificial fun—‘Look at our Ping-Pong table and our pool table!’—it was organic; more like a dorm, where there were people a little older than you and people a little younger, and everyone was working together.

“Sandy was like the RA,” she adds.¹⁸ Kurtzig was only twenty-nine, but to a recent college graduate, she was the adult in charge that every small startup eventually needs.

To find more programmers, Kurtzig conducted on-campus interviews, which always included a programming problem that applicants had to work through on paper after they spoke with her. Her interviews were unconventionally intimate. After she told Seckler that she needed to find people who knew her better to write her recommendation letters, the two women spent much of the interview chatting about their childhoods in Southern California. Kurtzig ended Howard Klein’s interview by inviting him to meet the rest of the team. She then informed Stanford’s Career Placement Center that “something came up,” and she would not be able to conduct the other interviews scheduled for that day. When Kurtzig thought a candidate would make a good addition to the team, she had the person come to the office to be vetted by everyone else. “Make sure you don’t scare her away,” she counseled the group before leaving them to interview Seckler.

“Sandy kind of fostered this collaborative environment that felt like ‘If one does well, we all do well,’ ” explains Howard Klein, one of Kurtzig’s first hires. “She enabled that environment, and we perpetuated it.” Kurtzig could also take an outsized interest in her employees’ personal lives. One day she called Klein into her office to take a phone call. On the other end was Kurtzig’s freshman roommate, whom she thought would make a good match. She was right. They have been married since 1989.¹⁹

Throughout the workday, Kurtzig would throw open her office door, telephone pressed to her ear—she was almost always on the phone, it seemed—and ask someone to locate a report or answer a question. “She was zipping through all the time,” Seckler says. “It was exciting to be around her.”

Kurtzig was not aware of it when she was working through the night at Hewlett-Packard, but the computing giant needed ASK and other small software companies.^I HP, still focused on scientific instrumentation, computers, calculators, and other hardware products for engineers, had almost no experience writing or reselling commercial software.

In 1975, ASK entered a cooperative sales agreement with HP that allowed an ASK representative to go on HP sales calls. HP relied on ASK to help teach nontechnical businesses what computers could do. It was difficult for an HP representative to show up alone at the offices of a small business and ask if the owner would like to buy a $60,000 computer, when most businesses had no idea why they would ever want a computer. But the person from ASK could explain that an HP computer loaded with MANMAN software could turn a small manufacturer into an “intelligent factory” in which anyone, regardless of technical skill or job category, could search inventory or schedule work flow with a few keystrokes.²⁰

When Kurtzig proposed scrapping the sales agreement for a royalty structure in which HP would pay ASK a percentage on every system sold, she was surprised to receive a counteroffer in the mail: HP wanted to buy ASK for $1 million. Kurtzig was tempted. “I wasn’t even thirty; there were plenty of other things I could do,” she recalls thinking. She began to imagine what it would be like to sell the company and relax. She went to her meeting with the vice president of HP’s computer division, excited but determined not to sell for less than $2 million.²¹

The meeting was a disaster, in Kurtzig’s recollection. The terms of what she had thought was a firm offer never came up. Instead, there was talk of Kurtzig and her team joining HP as employees. When she demurred, someone mentioned that HP had plenty of programmers who ought to be able to write a materials requirements planning package to compete with MANMAN. Soon, Kurtzig says, the vice president who had called the meeting was irate, shouting that HP would write the program for the flagship HP 3000 computer within nine months “whether you help or not.”

Kurtzig, by now yelling herself, retorted that HP would never be able to do it—and even if it somehow succeeded, ASK would get there first.

She left the meeting seething. “Damn them!” she swore. She vowed “to show them all what ASK could do.”²²

Despite Kurtzig’s bravado, ASK remained precarious. Although the company had sold the version of MANMAN written for Hughes to other customers, it nonetheless ended the 1976 fiscal year with a loss of $2,664. Kurtzig had doubled her salary at the beginning of the year, but she had not taken the raise, writing herself an IOU instead. The $17,692 “due to president” at the end of the 1976 fiscal year accounted for nearly two-thirds of ASK’s total liability.²³

Kurtzig again needed to retool ASK. She drafted a formal business plan, adding accounting components to MANMAN, selling the service bureau, and repairing the relationship with Hewlett-Packard. She also assembled a four-man board of advisers that gave her some important guidance but also indicates how far outside the typical Silicon Valley culture she was. Every adviser was a small manufacturer, and three were former or current ASK customers. There were no venture capitalists, no computer manufacturers, no other software executives. The closest ASK came to a representative from the Silicon Valley tech industry was the associate from Wilson Mosher & Sonsini, the multifaceted law firm that had worked with Tymshare and now handled ASK’s legal work.

The venture capital community in Silicon Valley did not yet appreciate the importance of software. Today software accounts for more than half of venture capital investments in the Valley, but in 1977, the figure was 7 percent.²⁴ Venture capital was for hardware: computers, semiconductors, disk drives, and other tangible, capital-intensive businesses. In 1977, when Larry Ellison and two cofounders tried to raise funds to launch Oracle—today one of the largest companies in the United States—venture capitalists were not interested. “When they heard the investment was about software, they wouldn’t even see me. In fact, their receptionists would search my briefcase to make sure I hadn’t taken a current copy of BusinessWeek with me when I left the room,” Ellison says.²⁵

Kurtzig, who never used the word “software” in her 1976 business plan, would say she was in the “applications” business—in part because “software” was such an unfamiliar term that when she used it, people assumed that she sold lingerie. Software developers told rueful jokes about clueless customers who wanted to know how much the software in their computers weighed.

Even computer companies could be slow to acknowledge the value of the software that ran on their machines. Kurtzig repaired her relationship with HP, but she nonetheless had a terrible time convincing the company to designate ASK an “original equipment manufacturer,” or OEM. HP sold computers at a discount to its OEMs, which then incorporated the machines into their own products. An OEM that built assembly equipment controlled by HP computers, for example, could buy the computers at a discount and then fold the cost, plus a markup, into the price of the equipment.

Kurtzig argued that the software ASK loaded onto HP’s computers should count as a feature that added value to the machine in the same way that assembly equipment did. ASK should be able to buy an HP computer at a discount, load MANMAN, and then sell the combined hardware/software package to customers as a “turnkey system.” Kurtzig recalls having to pull out HP’s own manual and read aloud the section on OEMs to a skeptical executive before she prevailed. It would take years, but ASK would become the largest original equipment manufacturer of, and a vital partner in popularizing, the HP 3000 minicomputer.²⁶

As Kurtzig went over ASK’s books on New Year’s Day 1978—it was typical for her to work on a holiday—she was gratified to see that her efforts over the previous months had paid off. ASK had earned $33,000 in profits, a huge increase over the 1976 losses.

At the same time, she had to admit that this profit, born of what felt like round-the-clock hours in the office, amounted to scarcely more than the $24,000 salary that she paid herself. It seemed meager compensation for the strain of running her own business. She recalled, “I made a New Year’s resolution: One more year or bust.”²⁷

---

I. HP allowed several small local software companies free use of its machines afterhours; the companies included Abacus, Allegro Software Development Corporation, American Management Systems, LARC Computing, Quasar, and SDG.

II. A classic materials requirements planning (MRP) package, MANMAN consisted of six modules: inventory control, bill of materials processor, material requirements planning, processing, work in process, and management reporting/product costing.

III. Many people who know the highly ethical Treybig have said that the fear was absurd.

IV. In 2010, Apple, the company Jobs and Wozniak launched, would buy HP’s Cupertino campus and raze it to build Apple’s new ring-shaped headquarters.