Software Arts Wrote The First Best-seller
It's Friday afternoon," says Julian Lange, checking his watch. "Last time I looked, it was Sunday night." Lange is the chief operating officer of Software Arts Inc. and manages the day-to-day operations of the Cambridge, Mass., company that developed VisiCalc, the hottest-selling business application software for personal computers. As at many of the fledgling companies in the personal software industry, events move so fast at Software Arts that today is often gone before Lange has a chance to deal with it.
As a result, Lange spends most of his time worrying about the future, six months or a year down the road. If he doesn't think about what kind of management the company is going to need, or how much office space, or what marketing strategies are right, the future will pass by too fast and the company will spin out of control. And things keep speeding up at Software Arts.
The company was founded just three years ago by Daniel Bricklin, now 30 and chairman of the board, and Robert Frankston, now 32 and president. In 1980, it had revenues of $1 million. When 1981 started, the three top managers thought they might have revenues of $2 million for the year. In June, they revised that estimate to $2.5 million. In October, they realized they would have over $3 million.
The reason for the exponential growth at Software Arts is that VisiCalc was the first program that proved personal computers could be useful for businesspeople. And that has sparked an explosion of growth for both hardware and software producers. Forecasts say the industry will grow 50% a year in the next decade, from about $500 million in 1981 to as much as $20 billion in 1990.
For all the speed of events these days, Software Arts had very quiet beginnings. In the spring of 1978, Dan Bricklin, an intense fellow who says things like "that's neat," was working on an M.B.A. at Harvard Business School. With pencil, paper, and calculator, he was hacking his way through the mind-numbing spreadsheet projections for the cases the school is famous for. He had an idea for a better way to crunch numbers: Program a computer to do all the grunt work. "I visualized an electronic blackboard and electronic chalk in a classroom," says Bricklin. (See "How VisiCalc Works," INC., November 1981, page 104.)
That summer, he doodled around on paper with images of the programming problem and decided that it would work. He thought he might be able to sell the program door-to-door to hightech executives around Boston's Route 128. But the first working version was limited to 5 columns and 20 rows (compared to the current version's 63 columns and 254 rows) and was difficult to use without some knowledge of programming. Bricklin realized he needed someone who knew better how to commit his idea to machine code.
He turned to Bob Frankston, with whom he had become friends at the Massachusetts Institute of Technology's Computer Sciences Laboratory in 1970. Frankston, who speaks casually of such programming concepts as orthogonality, took to the idea immediately and started writing the code for their program. Working with Bricklin and Steve Lawrence, their first employees, Frankston was eventually able to design VisiCalc's mass of code into a mere 20,000 bytes of memory, quite a technical feat. "We weren't writing a program as much as we were making a product," says Frankston. "It was a matter of being very sensitive to how the program would be used and what level of quality was necessary."
At the suggestion of another MIT and Harvard Business School graduate, Daniel Fylstra, Frankston coded VisiCalc for an Apple computer.Fylstra was looking for a home-budgeting program to continue building a catalog of programs for his nascent publishing company, Personal Software Inc. From what Bricklin told him, he thought VisiCalc might do. He already knew there were a lot of Apple owners because he was selling computer-chess programs for the Apple with some success.
Bricklin and Frankston had been talking about starting a company for most of the last 10 years. When Fylstra expressed an interest in selling the still-unnamed VisiCalc for them, they decided the time was right and incorporated Software Arts in January 1979.
As Bricklin started his last semester at the B-school in the winter of 1979, he and Frankston settled into a regular routine. During the day, Bricklin went to school and Frankston slept. In the evening, the two of them would talk about the development of their program. Occasionally, they would stop in at Fylstra's "office" -- his third-floor apartment -- and speculate about the future. At night, because the rates were cheapter, Frankston worked on the code for VisiCalc on a time-sharing terminal in his attic while Bricklin slept.
In April, Fylstra came up with the name, which stands for Visible Calculator and for which Personal Software still owns the rights. He and Frankston were having breakfast in a greasy spoon in North Cambridge when he thought of the name. "It wasn't a great name," says Frankston, "but everything else we came up with was worse."
As Frankston made progress on the code, each successive version of VisiCalc created a little more excitement in the small group of entrepreneurs. Unfortunately, few people shared their enthusiasm. In May, Bricklin and Frankston showed a prototype of VisiCalc to computer dealers, who weren't impressed. In June, after Bricklin graduated with his M.B.A., Frankston gave a presentation at the National Computer Conference in New York City. About 20 people showed up for his talk. All but two were family or friends, and those two walked out before he was finished.
In July, they scraped up $20,000 in cash, pledged a loan for $65,000, and bought a Prime 550 minicomputer for the development work on VisiCalc. That month, they got their first encouraging sign: Benjamin Rosen, a respected computer industry analyst, gave VisiCalc a rave review in his newsletter. He called it "the software tail that might wag the personal computer dog."
They finally introduced the program to the public in October, but even then not much happened. "We knew it was a good product and were using it ourselves," says Bricklin, "but we were caught between thinking it was great and getting a lukewarm reception. You know, we said, if each computer store sells one copy every other week, then we can do it. But we couldn't believe the numbers that added up to. Bob and I are very insecure about these things."
The next year, however, word of VisiCalc's powers started to spread through office grapevines, and people began showing up in computer stores to buy Apple computers just so they could use VisiCalc. Because Software Arts didn't adapt VisiCalc to another machine until the fall of 1980, the program is even credited by some observers with giving Apple Computer Inc. a push into first place in revenues in the personal-computer market. Since then, sales of VisiCalc have added up to nearly 200,000 copies and Bricklin's and Frankston's problem hasn't been so much to keep up with the growth in revenues as to keep up with the growth in the growth rate.
Software Arts has so far been able to keep a firm hand on its spectacular growth by having people like Julian Lange spend their waking hours in the future, anticipating the needs of the company. And Bricklin and Frankston complement each other in numerous ways, with Bricklin leaning to the business side and Frankston leaning to the technical side. When the company started, Bricklin was fresh out of business school and set it up as a textbook case of professional management. "This company is like my third year of business school," he says. "I get to try out everything here that I learned there."
The company's growth is fast enough to upset daily routines frequently. The organizational chart changes every three months or so. The number of employees grew from 4 to 35 in 12 months, and this month is around 50. And the company will be moving into its own building and quadrupling its space next spring. "We're trying to grow in a controlled way," says Lange. "That means always being aware of where we're going in the next six months, or year, or five years.Right now we have the management to get up to about 100 employees, but we've got to worry about the next quantum leap, probably in about a year."
It's difficult to forecast what's going to happen in an industry that's only five years old and a company that's only three years old, particularly when both are growing so fast. So the reins on Software Arts's growth are held very loosely. When Bricklin or Frankston or Lange talks about the company's future, he speaks vaguely. "We have a vision of the company as the industry leader," says Lange. "And we have sincere commitment to quality because that's the only way to maintain our leadership."
Nobody at Software Arts will talk about sales projections or future products. Everyone the company deals with is required to sign a nondisclosure agreement in order to find out what it's up to. "We've been burned before by media hype that promised more than we could deliver," says Bricklin.
Hype they've had. The company has been written up in Business Week, Fortune, Time, the New York Times, and the Wall Street Journal, not to mention numerous computer magazines. And Bricklin was recently presented the Grace Murray Hopper Award by the Association for Computing Machinery, a kind of Pulitzer prize for significant accomplishment in computers by people under 30. "Was it Andy Warhol who said everybody would be famous for 15 minutes?" asks Bricklin. "I've had my 15 minutes. Now I can get back to work."
Working at Software Arts is in many ways like living in a college dormitory.The average age of the employees is around 30, and they are crammed into tiny offices two at a time. Blue jeans and work shirts are common, many of the men wear beards and sideburns, and the company refrigerator is stocked with Coke, Pepsi, and an assortment of natural fruit juices.
There are a few differences. Everyone, including the receptionists, has a computer terminal. All the interoffice mail is sent through the company's newest Prime minicomputer. Dozens of personal computers, Apples, Radio Shacks, IBMs, and others are scattered throughout the offices, most with their innards permanently exposed.
These offices aren't quite as relaxed as a dormitroy, either. Nobody talks about it, but there's an underlying apprehension about how the company's next product will will be received. How do you beat the record of a product as successful as VisiCalc? "For a while there, we were afraid to do anything else because it might not be as good," says Bricklin. "But that passed. Now we're more concerned about building a company than about whether one product is better than another." Adds Frankston, "Who has time to worry?"
There are other questions about the future for Software Arts. A number of programs similar to VisiCalc have been introduced in the last year with names like SuperCalc, Execuplan, T/Maker II, and Supercomp-Twenty. "Eventually VisiCalc will become a generic name," says Bricklin, "Or at least someone will come along with KillerCalc."
And the fantastic success of both Software Arts and Personal Software has strained the relationship between the two companies. Fylstra moved his company to California in 1979 and has since built a "family" of Visi-products that use many of the same commands and much the same logic as VisiCalc. His company, which had revenues of $12 million last year, is frequently credited with creating VisiCalc, a confusion that Fylstra would prefer not to clear up and that annoys Bricklin and Frankston to no end. Software Arts will be trying to market its own products in the future.
Despite the problems, people are enormously confident at Software Arts. They know they're good, they know they're on the leading edge of new industry, and they see a future with unlimited opportunities. "It's nice to have a hand in the company that produced VisiCalc," says Lange. "It's the most fun thing I've ever done."