Information Processing

DEC Discovers it Can't Live by VAX Alone

So it's branching out from the venerable mini strategy

Leslie Helm in Boston
Business Week

November 21, 1988

It was a sign of how quickly things can change. The scene was Digital Equipment Corp.'s annual meeting on Nov. 3. DEC founder Kenneth H. Olsen was assuring investors that all was well despite a 17% drop in profits in the Oct. 1. first quarter on a 16% gain in sales. Then he offered a wry assessment: There are problems only in one product line and one country, he said. ''They just happen to be a big product and a big country.'' The product is DEC's largest, most profitable computer. The country is the U. S.

Just a year ago, Olsen was basking in the overwhelming success of a brilliant strategy: Stick with one computer design, using a single set of software, and make it in many different sizes. That way, no matter how big or small, any DEC computer could share information with any other. It was an easy idea to sell, especially against IBM, which had one or more architectures and sets of software for each size of computer it sold. Between 1984 and 1988, DEC's sales doubled, and its earnings quadrupled.

Today things are a lot more complicated at DEC. Its VAX computer design, in all of its permutations, remains the core of the business. But suddenly, VAX can't do it all. Customers are shifting more computing jobs to networks of low-cost desktop machines, reducing the need for large minicomputers. And they're demanding hardware and software that can connect with many different sizes and brands of computers. Though it recently has announced a spate of joint-development deals and alliances to ride this trend (table), DEC has been slow to react. And Olsen insists that a large VAX, powering hundreds of terminals, is still the most cost-effective way to deliver computing power to desktops.


While he may be technically right, that thinking helped DEC misread the market in 1988, a year Olsen had viewed as a chance ''to seize market share.'' DEC geared up to sell big VAXes, doubling its capital budget to $ 1.5 billion and increasing its 111,000-employee payroll by 10%. But the demand never materialized -- in part, Olsen says, because financial-services companies stopped buying after the stock market crash. Now analysts think that DEC will bottom out in fiscal 1989, ending June 30. The consensus: an earnings drop of 9%, to $ 1.2 billion, on sales of $ 13 billion.

That has caused a change in plans. Now, Olsen is relying for growth on new products that have little in common with his beloved VAX. In 1989, DEC will sell clones of the IBM PC/AT as well as engineering workstations that use a version of AT&T's Unix operating system. While IBM is still a threat, Olsen may have to worry more about such spry competitors as Sun Microsystems, Apollo Computer, Apple, and Compaq, against whom the VAX isn't much of a weapon. Even if Olsen does well, he'll be competing in markets where 45% gross margins are the norm -- vs. the 58% that DEC usually gets. ''The change is going to be gut-wrenching,'' says Dale Kutnick, an industry consultant.

Investors have already figured that out. The company's stock, which hit 199 in mid-1987, fell to 87 last month, giving DEC a multiple of less than 10 times earnings -- a valuation 10% less than the S&P 500 price-earnings average. Only a 10-million-share buyback helped push the stock back above 90.


The stock may not return to favor soon, given the cloudy future of the minicomputer business. In recent years, overall growth in minicomputer sales has slowed to around 6% annually, according to market researcher International Data Corp., vs. 30% for competing products such as workstations. For a while, DEC took minicomputer market share from IBM faster than workstation makers stole business from DEC. But IBM's new AS/400 minicomputer is keeping its customers in the fold. And that's forcing DEC to compete in an arena that it would rather avoid: the market for open systems.

These are systems built around widely accepted software and communications standards, often including some version of the Unix operating system, a computer's primary software. DEC sells about $ 1 billion worth of machines each year that use its version of Unix, called Ultrix. It also is backing the Open Software Foundation, a group pushing a standard based on IBM's version of Unix. But until recently, DEC treated Unix like a stepchild. At the annual meeting, Olsen again belittled it as a system inferior to VMS, the primary VAX software.

His motives are understandable, given his investment in the more profitable VMS. ''Anybody who has a dominant market share really wants closed systems,'' notes W. Ford Calhoun, a vice-president of information science at Smith Kline & French Laboratories. Olsen argues that he can sell VAX side by side with Ultrix and other non-VAX products, though he concedes that his sales pitch will be more confusing than when VAX was king. ''We may have to work on the phrasing a bit,'' he says.

Confusing message or not, there's no doubt within DEC about what has to be done: Work feverishly to succeed in open systems. The company has established a special department with 400 software engineers to focus on open systems. DEC is also beefing up Ultrix: The catalog of applications software such as spreadsheets or data base managers for Ultrix has swollen from 150 to 1,000 in just nine months.

KEY TEST. By far the boldest move, however, is DEC's joint development with MIPS Computer Systems Inc. After its own efforts to design a so-called RISC (reduced instruction set computer) chip fell short, DEC turned to MIPS for the critical component. RISC machines achieve higher speeds at lower cost by eliminating all but the most important instructions. Sun Microsystems Inc. and Hewlett-Packard Co. are already selling RISC-based machines, and companies ranging from Apple Computer to Xerox Corp. are designing them -- all able to do many jobs VAXes now do, more cheaply. DEC's first RISC machine, an Ultrix workstation aimed at the engineering market, is due in January. And DEC officials say an entire line of RISC machines will follow. Eventually, the RISC line will extend from desktop machines to mainframe-caliber systems, the way the VAX line now does.

The RISC workstation will be a key test of how well this strategy works. DEC sold about $ 750 million worth of VAX workstations this year, but it didn't reach its goal of displacing Sun, whose fiscal 1988 revenues hit $ 1 billion. The company is tuning into other desktop markets, too. It is working with PC makers such as Apple, and has contracted with Tandy Corp. to make an IBM PC/AT clone -- replacing the VAXmate, an awkward combination of a VAX terminal and a PC. VAX customers also will get their lowest-priced computer yet, a ''PVax'' that will reportedly list for below $ 5,000.

A big win with these new machines would solve many of DEC's problems -- but also create new ones. For example, a successful MIPS-based workstation, while shoring up DEC's faltering position in engineering departments, could hurt earnings. To compete in the Unix market, these machines will have to be priced far below the proprietary VAX line -- as much as 75% lower in processing power per dollar, analysts say. That ''has the potential of wreaking havoc on their pricing structure,'' says Robert Randolph of TFS Inc., a Boston area consultant.


The danger is that DEC's own customers will switch from the VAX to the Unix system. Already, analysts say, earnings are down because customers are buying smaller VAXes, which have lower margins. ''If we don't do a good job, we'll have a mess on our hands,'' says Don McInnes, vice-president of DEC's Engineering Systems group.

It won't help that DEC will simultaneously step up the VAX assault on IBM. DEC is perfecting new software to make VAX more competitive in the $ 27 billion transaction-processing market, such as bank teller machines. ''We will be going after customers we haven't gone after before,'' says William D. Strecker, vice-president for product strategies and architecture. A recent marketing deal with Allen-Bradley Co. should also boost DEC's prospects in factory automation. Although one project to beef up the top end of the VAX line was killed, a second is ahead of schedule. Due out next fall, it's expected to have four times the performance of today's largest VAX -- for about the same price.

The good news is that DEC's challenges pale in comparison to those facing other minicomputer makers (page 106). DEC has 30,000 VAX customers, most of whom will keep buying the proprietary design that currently runs their software. Kutnick predicts that DEC will boost its minicomputer market share from 27% now to 31% by 1993. And, even after the stock buyback, DEC will have $ 2 billion in cash to invest in maintaining VAX while developing new businesses.

If a crisis occurs, moreover, DEC still has an amazing resource in Ken Olsen. Every six or seven years, the founder has taken a more direct role in DEC's management, often to clarify a new direction. The last time was when he laid down the VAX-only rule. Lately, he has left most day-to-day operations to John J. Shields and John F. Smith, senior vice-presidents. But says IDC analyst Don Bellomy: ''It may be necessary again for Ken to come down from the mountain and impose a solution.''

Joint ventures, alliances, and marketing deals are helping DEC fill holes in its product line and strategic plan. Some will boost VAX minicomputer sales. Others create non-VAX businesses. Here are the latest deals:
DEC has bought a 5% stake in MIPS Computer and will use MIPS microchips to build high-performance Unix workstations, due out in early 1989
DEC will sell Tandy-made personal computers under the DEC label, giving it a competitively priced clone of the IBM PC/AT.
Allen-Bradley will sell DEC computers in factory automation systems, potentially giving DEC a major boost in a key market
A major supplier of software for mainframes, Computer Associates has agreed to adapt several key products for use on VAX systems
These PC suppliers and DEC are smoothing the connections between their machines. DEC will service the PCs when they're connected to VAXes
Ashton-Tate will adapt its popular PC data-base program, dBase, so that VAXes and PCs can share information more easily. DEC will sell the program


Copyright 1988 McGraw-Hill, Inc.