The history of Digital Equipment Corp over the last decade is testimony to the fact that although the computer industry may be fast-moving, making up for past strategic mistakes can take years. Chief executive Robert Palmer, installed in 1992, says that following losses in four of the last six years, three re-organizations and the garage sale of entire divisions, DEC is at last in the right shape to profitably face the next millennium. Palmer was commenting on figures for the year to June which showed the company bouncing back from a loss of $111.8m in fiscal 1996 to profits of $140.8m. But a cursory glance at DEC’s figures suggests caution and indicates Palmer and DEC’s long-suffering investors should wait awhile before cracking open the champagne. There are several warning signs: Revenues declined 10.4% from $14.6bn to $13.1bn; some $123.9m of the companys profits of $140.8m were achieved in the fourth quarter; and of those profits, $116m was derived from interest and investment income.

Poorly led

While most commentators have concentrated on the company’s hardware difficulties – such as the unprofitability of the PC division and continuing worries over the acceptance of the Alpha processor as a serious alternative to Intel – the performance of DEC’s services division was moribund. Its revenues declined 5.7% over the year and compared badly to the 21% average growth achieved by services companies in last years CBR Services top 50 table. Part, but not all, of the blame can be attributed to falling service revenues from its 1980s success story, the Vax/VMS computer. In the 1980s, growth in services revenues outstripped hardware sales, hardware revenues grew a solid 60% but services grew 93%. As those proprietary computers have been replaced with open systems, DEC has faced more competition for contracts. So why is DEC’s services division not competing more aggressively for new business, especially given DEC’s expertise with the NT operating system and its relationship with Microsoft? Anecdotal evidence from users suggests a poorly led, ill-focused company with low morale among staff. Customers complain of a lack of urgency and, according to analysts, receive a message of indifference from DEC when tendering services contracts. If you want to give us your business we’ll take it but if you don’t, that’s okay. DEC is taking steps to arrest the alarming decline in its services division, but the action being taken is only of a piecemeal, country-by-country nature and is not part of an overall strategy dictated from the boardroom. They need to decide whether the services division should support the mainstream box-shifting business of DEC or be a services business entity on its own, says another analyst. Such strategic decisions need to be made by Palmer and the board who should now be able to turn their attentions away from hardware, which is showing some signs of life. In the fourth quarter, sales of Windows NT servers grew 50% compared to the quarter a year earlier and, according to Jon Oltsik at Forrester Research, DEC is in a good position to gain from its large base of VMS users who are poised to move to Microsoft Windows NT. A further unique selling point DEC enjoys is its highly regarded, but poorly selling, Alpha processor which easily beats Intel’s best for power. In spite of much negative comment, the importance of Alpha should not be underestimated: It has brought in more than $12bn in revenues since its launch in 1992, even though DEC has been criticized for not being able to produce the Alpha in large enough volumes to make it a profitable line in its own right. And DEC’s alliance with Microsoft, combined with the fact that the Seattle-based software company only supports Alpha as an alternative to Intel, means that for high performance users, Alpha systems are the only choice. When the Alpha-optimized 64-bit version of NT becomes available, DEC will be further out-in-front. IDC analyst Jay Bretzmann, like Oltsik, believes DEC’s customer base is its biggest strength and that it will benefit from selling Microsoft Windows NT servers to them as the product matures. DEC’s advantage lies in the expertise it has built-up selling to large accounts, says Bretzmann. The company is also strong as a provider of web servers and equipment to internet service providers, an expertise its services side could leverage as internet services and trading booms. Bretzmann believes DEC still has a three year window of opportunity with Alpha because it can offer power, particularly for Windows NT, which no other vendor can match. Until Intel’s Merced processor comes out in 1999, and hits volume in 2000, there’s plenty of opportunity for DEC, says Bretzmann.

Fab-less

Nevertheless, DEC’s Prioris range of Intel-based NT servers, which Bretzmann also rates highly, means less power-hungry users are also catered for. The costs of producing the Alpha processor have been mitigated by DEC’s licensing agreement with Advanced Risc Machines. DEC’s raw performance expertise, allied to ARM’s ability at designing low cost, low-power consumption microprocessors, has produced the StrongARM, which runs at up to 185 dhrystone MIPS (million instructions per second) yet costs under $50 a unit. The processor has found favor for mobile applications, such as handheld computers from both Newton and Psion, and for use as an embedded processor. By the time its Hudson, Massachusetts fabrication facility can no longer support the next generation, around the turn of the century, it is likely to be sold off leaving DEC fab-less. But by then, DEC will know whether Alpha and StrongARM are here to stay.