By Jessica Twentyman
Last summer, a group of executives from IBM Corp’s Personal Software Products division, led by general manager John W Thomson, descended on the company’s headquarters in Armonk, New York. Their alleged mission, according to sources close to the company, was to urge chief executive Lou Gerstner to phase out the OS/2 desktop operating system. Despite massive investment in the product, reportedly in the region of $400-$500m each year, and multiple attempts to repackage the technology, OS/2 was still losing ground to Microsoft’s Windows NT operating system they said. Continuing to invest in OS/2, they argued, amounted to folly. Gerstner reportedly refused point blank to consider the suggestion. Closing down OS/2, he said, would constitute an admittance of failure, and would deal a hefty blow to IBM’s reputation. There was also his professional prestige to consider – one of Gerstner’s first moves on joining IBM in 1993 was to publicly pledge his support for the product. As a result of Gerstner’s persistence, OS/2 lives on, but many of those involved in that deputation to Armonk have since departed. Thompson left in December to become general manager of IBM’s North American operation, and according to renowned IBM watcher, Bob Djurdjevic of Annex Research, the resumes of the remaining top OS/2 executives are in regular circulation around the company. The dilemma facing OS/2 is unchanged. The operating system continues to be praised for its superior functionality and robust performance, but confidence in the product has been seriously dented by numerous strategic U-turns and diversions over the past few years. Publicly, IBM continues to claim that OS/2 is stronger than ever.
Privately, some are asking if OS/2 can ever earn IBM any money or if it is simply an expensive indulgence. One major problem that IBM seems to have been unable to solve is OS/2’s pitifully small market share in the desktop and low-end Intel Corp-based server market. According to market analysts Dataquest Inc, Microsoft Corp controls around 90% of the personal computer operating system market with its Windows 3.1, Windows 95 and Windows NT products. By contrast, OS/2, which is widely regarded as technically superior to the Windows products, holds a meager 2% share. Furthermore, the rise of Windows NT seems set to continue as doubts about its scalability and reliability are addressed. According to Microsoft, Windows NT Server sales doubled in the second quarter of 1997 from a year earlier. Analysts predict that combined NT and Back Office, the software suite that runs on NT, sales will grow by 100% in fiscal 1997 to $2.3bn, and will double again in 1998 to $4.2bn. OS/2 sales, meanwhile, remain, at best, static. Although IBM frequently puts the number of OS/2 users worldwide at 15 million, many analysts believe it is closer to 10m. IBM’s figure almost certainly includes copies of the software which were distributed free of charge, and the many copies delivered through its dual boot program, which saw both OS/2 and Windows 3.1 pre-loaded on IBM personal computers. This practice, now abandoned, was simply a means of stuffing OS/2 down users’ throats, says Djurdjevic, and he believes that hundreds and thousands of preloaded copies of OS/2 have never been used. If information about the OS/2 user base is sketchy, the financial details concerning OS/2 are even harder to ascertain. Donn Atkins, vice president of marketing for the Personal Software Products division of IBM, puts revenues from OS/2 related business in the range of $750m per annum. But this figure includes large amounts of revenue from other business areas, such as hardware. Even at this level, analysts question whether IBM derives sufficient sales to justify the amount which is poured into OS/2 development – over $1bn since the product’s inception. IBM is unrepentant, saying that development of OS/2 will continue throughout 1997, with a major new release scheduled for 1998. IBM’s key aim is to position the operating system as central to its network computing strategy. For example, in OS/2 Warp 4, released last September, IBM added a number of Internet and intranet features, including an integrated Netscape Navigator browser, native support for Java applications, firewalls and Lotus Domino 4.5 groupware. Analysts point out, however, that adding new capabilities is unlikely to improve its competitive standing. Ultimately, they say, new features will not increase OS/2’s market share. One area that does need attention is the lack of applications for OS/2. IBM has long tried, unsuccessfully, to stimulate native OS/2 application development, an exercise made more important by the fact that OS/2 does not support the growing band of Windows 95 and NT applications.
IBM is now looking to Sun Microsystems Inc’s Java Internet programming language to bolster its application portfolio, but any success will be dependent, say analysts, on OS/2 surviving long enough to see Java application development gain industry- wide acceptance and support. In the long-term, say analysts, the outlook for OS/2 is gloomy. Michael Gartenberg of the Gartner Group is adamant that the market has rejected OS/2, and says that most of the questions he receives about OS/2 relate to migration away from the system, not deployment issues. Gartenberg predicts that Warp 4 will prove to be IBM’s last OS/2 release. John Oltsik of Forrester Research Inc agrees that the OS/2 client software has seen its last release, and gives OS/2 Server just a couple more releases, attributing the slightly longer lifespan merely to higher transition costs. In a few years, he says, OS/2 will be thought of in just the same way as we now think of Wang VS – the now defunct minicomputer operating system. Djurdjevic believes that IBM, in its continued support for OS/2 is guilty of a massive cover-up designed to mislead investors: Whichever way you look at it, IBM is trying to resuscitate a dead body. My prognosis is that the patient will remain dead. Despite these derogatory comments, most of the, albeit small, band of OS/2 users remain committed to the system. Right now, we can run everything we need, says Martin Brampton, chairman of the EurOS/2 user group.
From the March 1997 issue of Computer Business Review.