By Nicholas G. Carr
Originally published in The New York Times on July 23, 2004.
Microsoft's decision to return $32 billion to its shareholders may be a wise business move, but it is also an admission of defeat: The company is confessing that despite years of trying, it has not found an attractive way to invest its cash reserves.
Microsoft may be the biggest name in software, but its problem is not unique. In recent weeks, many of the largest suppliers of business software, like Computer Associates, Seibel Systems and Veritas, have announced that their growth will fall short of investors' expectations.
The software industry's sluggishness is not just a reflection of the vagaries of the economic cycle. It is a manifestation of a fundamental, if often overlooked, characteristic of the industry's product: Software never decays. Machinery breaks down, parts wear out, supplies get depleted. But software code remains unchanged by time or use.
For software companies to grow, therefore, they have to give buyers good reasons to throw out perfectly serviceable versions of programs and install new ones in their place. Until recently, that hasn't been a problem. The rapid growth in the power of microprocessors, combined with ever-shifting computing standards, forced companies to replace or upgrade their existing programs at a breakneck pace.
As long as software quickly became obsolete, it didn't matter that it didn't wear out. Indeed, much of Microsoft's growth over the years has been fueled by upgrades to its two core products: the Windows operating system and the Office suite of basic business applications.
But now the software upgrade cycle is slowing. Computers have become so powerful that companies no longer need to rush to buy new models with the latest chips. At the same time, the spread of the Internet has solidified many computing standards.
As a result, the need to upgrade software has greatly diminished. In many cases, software purchases have reached what economists call the point of diminishing returns.
The change can be seen most clearly in personal computers. The majority of business users of PC's rely on a well-established set of programs - e-mail, word processing, Web browsing and spreadsheets - that use only a fraction of the computing power of today's desktops and laptops. The case for continuing to upgrade these programs is weak and getting weaker.
But the same trend is playing out in complex and expensive enterprise applications - the programs that underpin business processes like accounting, customer service and purchasing. Over the last 10 years, most big companies spent a great deal of money to install these programs, and they see no compelling reasons to upgrade them.
In fact, they are rebelling against software makers' requirements that they always run the latest versions of programs. One recent survey of corporate software buyers showed that nearly 75 percent want to see less frequent upgrades, and more than 20 percent plan to stop buying upgrades.
The prospects for the consumer market seem brighter in some ways. As more people come to employ their PCs as entertainment centers - using them to edit movies, touch up photographs, and play games and music - they need to buy more powerful processors and install sophisticated new software. But the consumer market plays by different rules, and it is hard for software-makers to charge much for their products. Apple Computer gives away its most popular software program, the iTunes jukebox, making up the loss through sales of hardware and songs.
Since its founding less than half a century ago, the software industry has become one of the largest in the United States. More than that, it is the icon of the modern information economy, a symbol of growth, prosperity and excitement. But now the industry is changing, and for the first time its managers, employees and investors are confronting an era of limits, with heightened challenges and lowered expectations.
Software companies are smart and inventive, and they will continue to come up with new, if ever more specialized, products. The industry will remain a large and important one, but it seems fated to resemble more and more a traditional, mature sector like manufacturing.
It is no longer unthinkable to say that software's glory days lie in the past, not the future.