The fact that Siemens Data Systems is now only about the same size as Nixdorf Computer AG underlines what a pedestrian performance the Siemens business has put in in the past five years or so, when growth has averaged only 4.7% a year: a decade ago, Siemens’ computer business was five or six times the […]
The fact that Siemens Data Systems is now only about the same size as Nixdorf Computer AG underlines what a pedestrian performance the Siemens business has put in in the past five years or so, when growth has averaged only 4.7% a year: a decade ago, Siemens’ computer business was five or six times the size of Nixdorf. Does the fact that Siemens is in a position to rescue Nixdorf mean that Siemens has avoided the mistakes that drove Nixdorf to surrender its independence? Hardly. The company’s computer business is overwhelmingly in what currently look like some of the most unattractive mainframes in the market. They are slavishly IBM-like – so close to 370 architecture that Fujitsu’s IBMulators simply need remicrocoding to run Siemens’ flagship BS2000 operating system, the company’s answer to MVS – yet they are not IBM-compatible. Even if they were, they would have the aura of a wasting asset because IBM is going to have to fight harder and harder for every mainframe dollar in the 1990s: what growth there is is likely to dip below 5% within two or three years, and IBM will need to grow its mainframe business faster than that to continue to justify the development costs. Happily for Siemens, the typical West German user is so conservative that there is likely to be a time lag of two or three years before what is happening to IBM’s business in more sceptical markets like the US, the UK, France and Holland begin to ravage its mainframe base – but the fact that that conservatism is not unshakeable is made very clear by Nixdorf’s fate: enough customers at the lower end of the market, where such decisions are less difficult to take, have deserted Nixdorf’s proprietary 8870 family to drive the company to penuary, and no-one in Munich or Paderborn should doubt that the same will surely happen to Siemens’ mainframe base in a dauntingly short time. Although Siemens builds under licence the original Sequent Computer Systems Inc National Semiconductor NS32000-based multiprocessor Unix machines, its Unix business has been overwhelmingly right at the bottom end of the market with its own NS32000-based uniprocessors, while Nixdorf has put together a Unix product line that provides a comprehensive open alternative to all but Siemens’ very largest mainframes. Thus it can be argued that Siemens Data needs Nixdorf – and its enormous mid-range customer base – at least as much as Nixdorf needs Siemens. But Siemens’ stewardship of its own computer business – scarcely a year of profit for 20 years before the early 1980s – suggests that the new Siemens-Nixdorf Information Systems will need a large injection of new management if it is to become the force in the world computer industry that its size implies – and a substantial – well-managed – US acquisition would be a good place to start.