The Past Is Prologue: Carol Bartz and Autodesk in 1992=Yahoo Now

Autodesk? If the name barely registers, you’re not alone–even though, remarkably, the London Business School last year calculated that Autodesk was by one measure the most profitable company of the 1980s, based on the school’s survey of 2,000 public companies world-wide.

Though the world’s sixth largest PC software company, Autodesk is hardly a household name for a couple of reasons. One is that it dominates a niche: software that allows relatively inexpensive personal computers to produce powerful models for engineers, architects and other professional designers.

He relinquished the top spot in 1986 to Alvar Green, formerly Autodesk’s chief financial officer, to return to programming. But the real power still rested with Mr. Walker, Autodesk’s biggest shareholder, and an elite group of programmers called “Core,” who had either helped Mr. Walker found the company in 1982 or led its most important projects.

Core members are contentious, eccentric free-thinkers who have had a way of devouring professional managers. They have often attacked each other and company executives, usually by sending “flame mail”–biting electronic letters. The outbursts sometimes have led to changes, and sometimes brought work to a halt. “The whole company is a theocracy of hackers,” says Charles M. Foundyller, president of Daratech Inc., a market research firm in Cambridge, Mass.

A year ago, Mr. Walker issued the ultimate in flame mail, a 44-page letter brutally attacking Mr. Green for allegedly trying to bolster short-term profits by neglecting investment in new products and marketing. Mr. Green later decided to resign, but stayed on until the selection of Ms. Bartz, 43, who formerly ran worldwide field operations for Sun Microsystems Inc.

She is regarded as a canny pick, particularly because she has experience managing rapid growth. She’s also a tough manager who got her first big promotion at Sun when she convinced top management that she could do a better job than her boss, who was on vacation. “I am not coming to Autodesk as a dictator,” she says. “But I am not a consensus manager in the extreme. I do not believe the best decision is a group grope.”

That, however, is largely how Autodesk has been managed until now. It was founded by Mr. Walker and a dozen programmer pals just as International Business Machines Corp. revolutionized the computer industry with its original PC. Mr. Walker saw Autodesk as a diversified supplier of PC software with a can’t-miss future. “We should consider ourselves extremely lucky to be in this business at this time in history,” Mr. Walker wrote in 1982, egging on his co-founders. “I cannot imagine any scenario other than the total collapse of society in which the sales of microcomputer application software will not grow by a factor of 10 in the next five years.”

Actually, Autodesk’s own sales did better than that, jumping nearly tenfold to $9.8 million in the fiscal year ending Jan. 31, 1985, only its second full year of operation. Sales reached $100 million, another factor of 10, four years later; for the year ended Jan. 31, it earned $57.8 million on revenues of $284.9 million.

Mr. Walker didn’t invent the program that drove all this phenomenal success. Instead, Autodesk’s hit product proved to be a computer-aided-design (CAD) program that Mr. Walker purchased from an outside programmer named Michael Riddle. The program, which became AutoCad, did for designers of buildings, interiors and machines what VisiCalc’s spreadsheet did for the accountant: It made the personal computer an essential tool, where once pencil and paper reigned.

Mr. Walker quickly grasped the promise of AutoCad when Mr. Riddle gave him a presentation in 1982. “When I showed him the program, he was quiet for 45 minutes. It was the first time he’s ever been quiet with me,” says Mr. Riddle. Then, “he says, ‘You’ve got a fortune here.'” Before long, Autodesk dropped virtually all of its other work to concentrate on AutoCad. The program, which now sells for about $3,500, was a runaway success, cutting deeply into the sales of computer companies that blended software and hardware into systems that might cost 10 times the price of a PC with AutoCad, and yet afford only somewhat better performance.

While established CAD leaders ignored the threat from the PC, Autodesk began to entrench itself with customers. The company signed up dealers by the hundreds; many were architects and draftsmen themselves who sold it to their colleagues. And Mr. Walker created an AutoCad language so that consultants or customers could take the program and modify it to handle specific tasks. Today, thousands of AutoCad applications exist.

Mr. Walker gave Mr. Riddle an extraordinarily generous royalty agreement that eventually amounted to more than $10 million. The payment may be a record for an outside programmer in the PC business, but Mr. Walker has always operated by a different set of rules. He doesn’t care a whit about office decorum or hierarchy, so Autodesk was always casual and libertarian, even by techie standards.

“I realized this was a different place when at my first staff meeting, I was licked by a dog,” says one manager. But within this unstructured setting, Mr. Walker for the first four years ruled by charisma, relentless memo writing and sheer force of will.

Mr. Walker has unusual interests, which he imposed on Autodesk. When he grew intrigued with outer space, Autodesk invested in a company that salvages used fuel tanks from the Space Shuttle with the idea of sending them back into orbit, carrying the concept of recycling about as far as it can go. When he grew enamored of cellular automata and chaos theory, arcane fields at the intersection of computing, mathematics and biology, the company released a family of products based on those concepts that are essentially video games for brainy adults.

While Mr. Walker is intensely private about his personal life, he has had no qualms about airing Autodesk’s dirty laundry–or effusively describing his technical ideas. He published a book containing scores of confidential Autodesk memos, many written by himself. And he once unsuccessfully tried to interest journalist Hunter Thompson in chronicling the company’s rise. An obsessive writer who often revises a memo dozens of times before releasing it, he also has written a manuscript for a diet book, based on his experience of losing (and keeping off) about 100 pounds.

He is prone to making unexpected pronouncements. In a rare public appearance in March, Mr. Walker interrupted the description of a new product with this observation: “We are living on a small blue sphere in an endless black void.”

Besides programming, Mr. Walker wrote press releases and ad copy and even pitched the product at trade shows in the company’s early years. But wearing so many hats frayed his nerves. He began to show increasing impatience with co-workers. “The only way he knows how to deal with people is to bluster,” says Mr. Riddle, who argued with him about the technical direction of AutoCad.

These fits of impatience dove-tailed with Mr. Walker’s continuing suspicion of professional managers, shared by other members of Core. In early 1986, he forced out John G. Ford Jr., the vice president for marketing and sales, who built the dealer network that many observers say is still Autodesk’s most valuable asset. Neither Mr. Walker nor Mr. Ford will comment.

Despite Mr. Walker’s rough edges, employees were, and still are, drawn to him the way kids admire the baddest boy in class. He “is the cult hero of Autodesk,” says Joe Oakey, who directs the company’s charitable foundation. “He could stand up before a company meeting and say ‘I hate you,’ and everyone would cheer.”

It was also in 1986 that Mr. Walker tired of management and handed daily responsibilities to Mr. Green. Two years later, he resigned as chairman to devote himself fully to writing software from his nearby home. But he still held a huge stake in the company. (He currently owns 869,000 shares–less than 4% of the shares outstanding, but still worth more than $30 million.)

Mr. Green was ill-suited to ride herd on the rambunctious Core. Trained in finance, Mr. Green didn’t even keep a computer on his desk, so he missed the electronic chatter that went on behind his back. When he needed to send an electronic message–the preferred means of discourse at Autodesk–he asked his secretary to do it.

Meanwhile, disputes kept breaking out among programmers and managers, usually about the technical direction of the company. “Over time, Autodesk became almost unmanageable,” says Mr. Foundyller, the analyst. “Why? Autodesk was run very democratically. People met. They discussed things. Many flowers bloomed. But nobody harvested.”

Sometimes, the paralysis was relatively innocuous, as when employees voted to delay the company’s move into a new office complex because they preferred an alternative site opposed by management. Other times, disagreements led to debates over how to lessen the company’s dependence on its AutoCad cash-cow–or even whether the company should try to diversify. The need for consensus led to many organizational quirks. Last year, for instance, the critical AutoCad division was assigned two general managers–one from the business side and one from Core–because neither was believed to have the experience to run it alone. Ms. Bartz has already changed that, appointing a new head of the division to whom the former co-general managers report.

The most bitter disputes arose between programmers and the company’s marketing and sales executives. “A tremendous schism” has existed for years between the two sides, says Mark Macgillivray, who has consulted for Autodesk on marketing issues. Core members and other programmers have simply refused to work on certain new products because they found them boring. Sometimes these are products that customers are clamoring for, such as a more memory-efficient version of AutoCad, which “the techies fought us tooth and nail on,” recalls one marketing executive.

At many software companies, a product manager balances the interests of sales, which wants to satisfy customer demand for certain product features, and the interests of engineers, who push certain features because they are possible. At Autodesk, when products were being conceived “it became very difficult to get features agreed to,” says Tim Cox, who was a product manager for two years until departing last November. “The problem we kept running into–everything needed to be the programming group’s idea.”

Indeed, opposing Core was tantamount to “attempting to butcher the sacred cow,” says Roger Clay, who left Autodesk last year to form a software company.

For a long time, the financial results didn’t reflect the conflict. From 1986 to 1990, net income nearly quintupled and sales jumped more than five-fold. But in the fourth quarter of fiscal 1991 ended Jan. 31, earnings fell about 25 percent below expectations on an unexpected slowdown in growth. The stock fell 22% in a single day. Financial analysts blamed Mr. Green for not keeping them informed, and Mr. Walker blamed him, too–for catering too much to analysts. He asserted that Mr. Green kept profit margins high at the expense of much-needed investments in new products and marketing, which “is how a company dies from making too much money,” he said in his memorable broadside in April 1991, entitled “The Final Days.”

Writing from his new home in Neuchatel, Switzerland, where he had recently moved to find more seclusion, Mr. Walker observed that it pained him “watching Autodesk squander everything I’ve been working 16 hours a day for since 1982.” He accused Mr. Green of “taking his marching orders from the accounting rules rather than the real world” and said he “was so appalled by what I heard at one management meeting that I vowed never to attend another management meeting and I never have.”

The memo was regarded as overwrought by some of Mr. Walker’s own Core associates. “That letter caused me a lot of pain to read,” says Gregory P. Lutz, one of two Core programmers on the board of directors. “He was right about a lot of things and I hadn’t done anything about it. But I thought some of it was unfair and a little exaggerated.”

Moreover, the broadside didn’t mention that Mr. Walker himself had left the company and picked Mr. Green as his successor. “When somebody like Mr. Walker isn’t there, it leaves a vacuum,” observes Microsoft’s chief, Mr. Gates. “People are free to do what they want. It’s just a damn shame that he hasn’t chosen to stay in management or even within the mainstream of software development at the company.”

But Mr. Walker’s pungent analysis (termed “just brilliant” by Mr. Gates) succeeded in getting everyone’s attention at Autodesk. “It was like when you fire a shotgun in an aviary,” recalls Mr. Walker of the letter’s effect. “It caused everybody to say: What is our strategy? What are we doing out there?”

Mr. Green dutifully decided to do most of what Mr. Walker wanted: invest more in marketing and advertising of new products Mr. Walker said were being “abandoned” after introduction; push harder on a new version of AutoCad for Windows, Microsoft’s emerging standard for controlling PC software; and back forays into new areas such as software tools for do-it-yourself designers and scientists. But Mr. Walker’s attack destroyed the credibility of Mr. Green, who in October disclosed plans to resign as chairman and chief executive.

Mr. Green defends his record and downplays Mr. Walker’s memo as a factor in his demise, saying he intended for some time to step down. But he adds that Mr. Walker “perpetually” criticized him. “To a great degree, he was right,” Mr. Green says of Mr. Walker’s memo, adding that “John’s batting average is pretty high” when it comes to picking strategies.

Mr. Green protests that it never was clear to anyone how best to diversify Autodesk’s revenue. He concedes that he approved the company’s issuance of an extraordinary $1.50 a share special dividend because retained profits were mounting so rapidly. “We were asked by people, ‘Don’t you have anything better to do with your cash,'” recalls Mr. Green. “Well, no we didn’t.”

To get a closer look at operations–and to help select a new chief executive–Mr. Walker invited himself back to Autodesk this year for three months as “manager of technology.”

The timing of Ms. Bartz’s ascension to the top job appears auspicious. There are signs that managers and programmers are starting to cooperate better. In March, for instance, Core and other AutoCad development teams for the first time agreed to compile a single to-do list required to complete the next version, which is due out by midyear. The list, which used to be kept in the heads of various developers, consists of 7,000 items. Each is assigned a completion deadline and the name of a person responsible for it.

Moreover, new products are starting to flow. In March, the company removed one monkey from its back by shipping its first Windows version of AutoCad. Autodesk also introduced its first scientific product in March, a program called HyperChem that allows chemists to create molecular models. The scientific market is “the hidden iceberg in software,” says Joel Voelz, the product manager. All together, Autodesk expects to release 25 new products in the fiscal year ending next Jan. 31.

“Very few public companies have been encouraged to report disappointing short-term results to make the investments that are necessary to adapt to a changing market,” Mr. Walker says. “Autodesk had that problem. Autodesk doesn’t have that problem any more.”

Indeed, last week, Autodesk reported that first-quarter profit fell 41%, citing heavy investment in new products. But investors seem willing to give Autodesk some time: The stock, which is trading at about half of its year-ago all-time high, went up on the news.

Another point in Ms. Bartz’s favor: Mr. Walker seems content for the moment. He returned to Switzerland on April 16, but not before pledging his “total and unqualified” support for Ms. Bartz and promising that neither he nor other Core members will “step in and prevent change at the last minute,” as has occurred in the past.