In the first 10 or 15 years of the computer industry, the vast majority of companies bought computers not because their business no longer functioned efficiently or effect ively without one but because their competit ors had installed one and was able to give customers a more timely service – which they didn’t really need, or supplementary informat ion that while useful was not actually vital. It often cost and arm and a leg and in many cases didn’t do more than create an impress ion of improved effici ency – but once one company in an industry had a computer, all its competitors had to get one too or go out of business. The same harsh rule applies to the move to open syst ems for computer manu facturers – but the fact that there is a real price to be paid for making the move is underlined by a new report. John Abbott has been reading it.

Despite any of the technical difficulties that may have been cited, the real reason why companies such as Atari Corp and Commodore International Ltd have not yet entered the Unix systems business are all to do with marketing. Both companies found their niche by bucking the trend of IBM compatibles, and while they now both offer MS-DOS machines as a sideline, they are understandably thinking twice before moving into the highly competitive world of Unix workstations with their near commodity Motorola 68000 family lines, even though that is what they will have to do if they wish to break into the mainstream business world. Elsewhere the prospect of the reduced margins that have become associated with Unix can cause even greater problems.

Big gamble

As has been highlighted here several times, IBM took a big gamble when it made its RS/6000 Unix box competitive with the rest of the industry, thereby attracting business away from its AS/400 line, where margins and lock-ins make the sale so much more valuable. Now International Data Corp analyst – and Computergram alumnus – Kate Oakley has produced a report that looks at the same issue. In a survey of some 15 US-based companies (excluding IBM), the report takes publicly available financial data to compare gross and operating margins over the period 1985-89, and finds that Unix high flyers such as Sun are indeed achieving lower operating margins than vendors still successfully holding on to their proprietary business, such as Apple and DEC. Compaq heads the table, a fact attributed to its huge volumes and strong channels strategy in a standards-based market, with Apple second and DEC fourth. As for the Unix vendors, NCR Corp is the surprise number three – but then it achieves high added-value with its Unix-based application-specific offerings for the banking and retail worlds – while Sun only manages eighth position, followed by Sequent Computer Systems, Silicon Graphics, Pyramid Technology, Unisys and Altos Computer Systems. But as a warning of just how fast the market can turn, the table is footed by Data General and Wang Laboratobies, two vendors who relied for too long on proprietary business, only to find it evaporated almost overnight.

One of the problems is that Unix is still a half way house. Although hardware vendors are beginning to suffer from the cut-throat competition found between personal computer vendors, Unix has not yet achieved its eventual goal of becoming a shrink-wrapped commodity market, meaning that research and development costs are still high while volumes are still fairly low. Although a hardware manufacturer no longer has to produce his own operating system, and can offload some development onto organisations such as the Open Software Foundation, there is always another hot new technology around the corner, such as networking, parallel processing or distributed computing, that must be investigated. For in the world of lower margins, timing is key. Ms Oakley cites the example of Apollo Computer, which paid a high price for pioneering the technical workstation before Unix was the obvious operating system choice. Other examples are numerous. Should Hewlett-Packard have developed its own RISC chip, or waited, like

DEC, until merchant RISC processors could have saved it the effort? Would Altos now be as successful as Compaq if it had used its commodity hardware and sales channels to go after the Xenix personal computer marketplace? Of course it’s all too easy with the benefit of hindsight. Anticipating which horse to back is one way of differentiating your product from the run-of-the-mill, and IDC identifies security, transaction processing and networking as likely areas for adding value to Unix-based systems, while tuning applications such as databases for specific hardware can also bring the margins back up. Applications availability and a focus on niche markets can also do the trick – owning a niche market has been key to for both Apple and NCR.

Ruthless

Paying software companies to do the Unix implementation for your system is becoming more common, something which both IBM and Unisys have been doing recently. Rationalising product lines can help cut costs – Sun had to be fairly ruthless towards its customer base in its fast move from Motorola to Sparc machines, but its single architecture should stand it in good stead when competing against the likes of Hewlett-Packard, which now has four different types of workstation. And as the market matures, traditional areas of additional revenues from areas such as service and systems integration should also increase. Buddy, Can You Spare a Dime? Or How to Make Money in the Open Systems Business by analyst Kate Oakley is available from IDC in Boston.