The Romance Ends for IBM and Microsoft as Experts Speculate the Partnership is Over

The Economist

February 23, 1991

Young love, great riches, first tiff.

Microsoft's partnership with IBM has made Bill Gates, Microsoft's founder, a billionaire and helped IBM become the biggest personal-computer maker in the world. Is the romance over?

They were the glamour couple of the 1980s. The unlikely marriage between tousled Microsoft and buttoned-down IBM created the hottest products in the decade's hottest industry, personal computers. Celebrated everywhere, the duo had no reason to question their good fortune, and no need to wonder which their friends really liked best: IBM's earnest reliability or Microsoft's youthful creativity. Now a setback has brought long-buried differences to the surface. Computer-industry gossips wonder if the marriage can survive.

Premature though speculation of divorce may be - both Microsoft and IBM are quick to pledge their commitment to each other - even minor flaws in their domestic harmony will be closely watched. Together, IBM and Microsoft set the technical standards that helped fuel the explosive growth of personal computers in the 1980s: "IBM-compatible" hardware and Microsoft's DOS operating-system software, the basic program needed to run any computer. Together, both companies are trying to set a new standard for the more complex computing world of the 1990s: a jointly developed operating system called OS/2. But customers have been slow to embrace OS/2. The frustration is driving Microsoft and IBM apart. The eventual result may be new leadership for the personal-computer industry, or no clear leadership at all.

Since they first began holding hands, Microsoft and IBM have shared barely a care. The men from Big Blue originally came to Microsoft's founder, Bill Gates, to license the programming languages that he had written for the early personal-computer enthusiasts, most of whom had built their own machines. IBM also needed an operating system. Gary Kildall, whose Digital Research created the CP/M operating system for the first generation of personal computers, did not jump at IBM's offer. So in 1980 IBM came back to Mr Gates, then 25, who quickly bought a rudimentary operating system from another young company, renamed it DOS, and fixed it up for IBM.

DOS now sells about 17m copies a year, and Mr Gates's 40% stake in Microsoft is worth about $4 billion. Today's success was created largely by yesterday's success. Because it was IBM jumping into the computer market with Microsoft in 1981 - as opposed to yet - another bunch of blue-jeaned entrepreneurs - would-be competitors reckoned the couple's technology might be worth copying. Microsoft's willingness to license DOS helped rivals to copy it and then use that technology as the basis for their own innovation. The more people jumped on to the bandwagon, the more others wanted to join them.

Almost everybody in the computer industry expected something similar to happen with OS/2. It hasn't. Since its launch in 1988, sales of OS/2 have grown to only about 300,000 a year. Nobody knows why. Perhaps the relatively high price of the more powerful personal computers needed to run OS/2 has put off budget-minded companies. Perhaps buyers are put off by the complexity of learning a new technology. Or perhaps OS/2 simply solved problems that do not trouble most customers. Whatever the reason, OS/2's adversity is pushing Microsoft and IBM apart.

Microsoft responds to slow OS/2 sales with a youthful eagerness to do better. "We were wrong," admits Steve Ballmer, a top Microsoft executive. He says both he and Mr Gates thought OS/2 would outsell DOS by 1992 at the latest. It will not. So Microsoft is now concentrating its efforts on a product customers do want to buy: Windows.

Windows does the most visible part of the job OS/2 was designed to do. It helps users to juggle a variety of programs at once, each in its own window on the screen. It is nowhere near as good as OS/2 at managing complex tasks, like getting computers to co-operate across a network. But customers do not seem to care. In nine months on the market, the latest version of Windows has sold 3m copies. Microsoft is now trying to market Windows as a stepping stone to OS/2, by making it easy for users of Windows and writers of programs for Windows to migrate to OS/2 eventually.

Microsoft's enthusiasm for Windows annoys IBM. For Big Blue, OS/2 is just one piece of a grand vision about how all of a company's computers can work together. Any gains it might make by fiddling with its strategy for OS/2 would be offset - perhaps more than offset - by the confusion such fiddling would engender about IBM's overall strategy. So IBM is publicly sticking to OS/2, though it is also quietly hedging its bets by incorporating into its grand scheme other products with many of the same capabilities.

Piece by small technical piece, these strategies are eating away at the consensus that originally made OS/2 seem a certainty as the computing world's next-generation technical standard. Microsoft has set up Windows as a rival to a component of OS/2 called Presentation Manager, which is still backed by IBM. Perhaps in retaliation, IBM has signed an agreement with Novell, a maker of personal-computer networks. This could make Novell's products an alternative to Microsoft's networking software, called LAN Manager. And IBM is financing an ambitious project by a young Californian company, Patriot Partners; it would let software companies write programs compatible with OS/2 and Unix, an operating system with similar capabilities originally created by AT and T's Bell Labs.

With more and more options to choose from, customers and software developers are increasingly reluctant to plump for just one. What makes this a particular problem for Microsoft and IBM is that, while the consensus on OS/2 is unravelling, its chief rival, Unix, is at last making headway. When OS/2 was announced, Unix was dismissed as a flop because there were almost as many versions of Unix as people selling it. Now various versions of Unix have been boiled down to two: one from the Open Systems Foundation, a consortium that includes IBM, Hewlett Packard and DEC, and one from Unix International, a consortium involving AT and T and Sun Microsystems.

About 1m copies of Unix were sold in 1990. A recent poll by DMR, a Canadian information-technology consultancy, shows that nearly as many American companies are planning to adopt Unix as OS/2. This raises the possibility that OS/2-and, with it, IBM's partnership with Microsoft-will get caught in a vicious circle, much like DOS got caught in a virtuous one. Slow sales will encourage Microsoft to innovate and IBM to hedge, which will make customers cautious, which will slow sales, which will encourage further tiffs between the two companies, and so on. It could all end in tears.