Autodesk Faces Dicey Task of Broadening Line : Software: To prove it’s no one-trick pony, the Sausalito company is pursuing some esoteric technologies that most established firms won’t go near.
When John Walker and some fellow computer programmers launched Autodesk Inc. in 1982, they had a clear plan for every aspect of the business except one: the products.
Walker figured that with top-flight programming talent and the voracious demand for software spurred by the success of the International Business Machines personal computer, the company could just try a bunch of things and hope one worked.
And one, a computer-aided design package called Autocad, not only worked, but also quietly became one of the most successful software products in history. Sausalito-based Autodesk earned a remarkable $43.6 million on sales of $175.6 million for the nine months ended Oct. 31, and Autocad has spawned a $1-billion industry of software developers, equipment vendors and dealers devoted to making personal computers better tools for designing everything from houses to helicopters.
But now, Autodesk faces the dicey task of proving that it’s not just another one-trick pony in an industry that already has a string of them. And in trying to move beyond near-total dependence on a single product--a goal that has eluded most software powerhouses, including Lotus Development, Ashton-Tate and Wordstar International--Autodesk has pursued some esoteric technologies that most established companies won’t go near.
Although Autodesk’s diversification strategy differs dramatically from the acquisition-oriented approach of Lotus and other firms, it must overcome the same basic problem.
“Software companies get mesmerized by their market niche,” says John Rossi, a PC software analyst at Robertson, Stephens & Co., a San Francisco investment bank. “All their efforts are very focused, very specialized, so it’s hard to send the whole thing down a different track.”
But if any software company is in a position to do things differently, it might be idiosyncratic Autodesk. Founded as a cooperative, Autodesk never accepted venture capital funding, and it thrives on an informal, egalitarian management style that’s almost unheard of in American business.
Alvar J. Green, the modest and cheerful British native who serves as chairman and chief executive, occupies his post because Walker and the other company founders prefer to focus on programming.
“It’s not as if I tell people what to do or anything,” he declaims, sitting in a barren 10-by-12-foot office with a cheap wooden desk and parking lot views. “You have to rally people around an idea here.”
It’s fitting, perhaps, that Autodesk’s brand of consensus management has yielded an approach to new markets that is thoroughly consistent with Walker’s original theory: try a few things and see what works. “We’re looking for technologies that have unlimited upside potential, but where we know what the downside will be,” Green says.
The company even launched a new-business development unit this month and invited “interesting project concepts and business plans from any legitimate source.”
Thus far, Autodesk has identified a handful of such technologies. The most mainstream of them is multimedia computing, the integration of sound and video into the personal computer. Many personal computer companies expect multimedia to be a major growth area, and Autodesk has set up a new division to market its multimedia products, which so far have met with limited success.
On the more daring side is Xanadu, conceived some 20 years ago by computer visionary Ted Nelson as a new way of navigating through a forest of information. In a Xanadu “hypertext” database, the information--including text, video or any other type of data--would be linked conceptually by the system, allowing the user to skip freely from one information source to another without using keyword searches and other traditional database methods.
Xanadu Operating Co., based in Palo Alto, had struggled for years to get a product out the door, and few in the business took it seriously. But Autodesk did and purchased the company in 1988 for an undisclosed sum.
Malcolm L. Davies, Autodesk’s senior vice president for marketing, jokes that when Xanadu products are finally shipped--and the goal is sometime next year--the company will apply to the Guinness Book of World Records for the longest-running software development project in history. But if Xanadu succeeds, the potential market is enormous, because it will vastly simplify and humanize the task of finding computerized information.
Similarly out on the edge is Autodesk’s cyberspace, or artificial reality, project. The concept, being pursued in varying degrees of intensity by many large firms, is to create a completely computer-generated environment. A current version includes a mask with two small television screens and a set of electronic gloves; one can actually navigate around a seemingly three-dimensional room and use one’s hands to pick up and move objects.
Autodesk aims to produce a basic artificial reality operating system, and then software developers would use it to create “environments” for entertainment, physical fitness or education. Such a system could also have links to Autocad, allowing an architect, for example, to walk around inside a building that exists only as an Autocad design.
“It’s a simulation technique that complements Autocad,” said project director Randal Walser. “We believe an industry is about to emerge.”
Autodesk has also purchased minority interests in several small software companies with interesting product concepts and last year acquired a competitor in the low-end computer-aided design business. But even with $150 million in cash, Autodesk is not interested in buying established firms to get into mainstream software markets, such as word processing or spreadsheets. That’s in distinct contrast with other big software firms, such as Lotus, Ashton-Tate and Wordstar, which have tried--and mostly failed--to diversify by acquiring going businesses.
Ill-advised acquisitions, however, are hardly the only reason that most big personal computer software companies have only one major product. W. Frank King, senior vice president at Lotus, notes that dependence on a single product--Lotus gets 65% to 70% of its revenue from the 1-2-3 spreadsheet package--tends to be self-perpetuating. “When you have that kind of franchise, you spend heavily to protect it,” he said.
Lotus, he added, has at times been too focused on diversification. And with 1-2-3 products bringing in some $500 million a year in revenue, any new business has to get awfully big to make a dent.
To meet its goal of gaining only half its revenue from computer-aided design products within five years--the figure is over 90%--Autodesk will need new businesses that generate hundreds of millions a year.
Scott Cook, president of Intuit Software, suggests another, more disquieting reason that software companies tend to be dependent on one product.
“They got lucky the first time around,” he suggests. “Instead of victory through internal process and people doing things right, they got lucky, got there first and got a dominant share.”
The key ingredients to systematic success, Cook emphasizes, is being close to the customer and close to the market--something it’s intrinsically difficult to do in a new business. Autodesk has been famously successful in developing a dealer channel for Autocad and grasping the needs of the computer-aided design user, but Green admits that the company made key mistakes in pricing and targeting the multimedia products.
Interestingly, the two big software companies that have succeeded in spreading their wings--Microsoft and Borland International--are firms that made their initial splash with system software and programming languages, respectively, rather than applications software for performing particular tasks.
Mike Hallman, chief operating officer at Microsoft, attributed some of his company’s success in building many types of products to the fact that “Microsoft always took a broader view of itself, with a mission to be a broad-based software company,” and thus didn’t get trapped into over-specialization.
Now, he maintains, the company is aided by the “uniting theme” of the graphics interface, which governs its popular Windows program--which Microsoft considers an operating system--and its applications programs, such as Word and Excel. But Hallman couldn’t say whether operating system expertise had proved more “transportable” to other products than specialization in particular applications.
Although Autocad is an applications program, it has through the years developed certain resemblances to a systems product. A primary reason that Autocad has been so successful, in fact, is because of an “open” architecture that allows third parties to customize the program with specialized hardware and software.
And Autodesk, from its early days, has also viewed itself as a broad-based software company, although most industry analysts consider it a computer-aided design company. Over the next few years it will become clear who was right.
AUTODESK FINANCIAL DATA
REVENUE: (millions)
‘86: $29.5
‘87: $52.4
‘88: $79.3
‘89: $117.3
‘90: $178.6
NET INCOME
‘86: $6.5
‘87: $11.6
‘88: $20.5
‘89: $32.7
‘90: $46.4
