Conference 7.286::digital

386.0. "DEC: The Challenge Ahead" by BUSY::KLEINBERGER (MAXCIMize your efforts) Mon Sep 21 1987 13:28

                            DEC: The Challenge Ahead
                                       by
                                Marc G. Schulman

                                    MAY 1987

Preface

This is a transcript of a speech delivered to the DEC Financial Leadership
Forum  in May 1987. It points many of the reasons for the Corporation's
success over  recent years and illustrates the challenges that lie ahead.


                                                                     Page 2

Good evening. It's a pleasure and an honour to have this opportunity to present
my thoughts on the critical trends and issues in the computer industry and
their implications for DEC. It's a pleasure because my career as a securities
analyst, which dates from 1972, began with an analysis of Digital Equipment
Corporation. Being selected to address you tonight suggests that those years
have not been for nought. It's an honour because DEC's accomplishments,
particularly over the last two years, have been remarkable. Product expertise,
marketing expertise, and financial expertise have combined to produce a surge
in the company's fortunes that, I am sure, will be viewed as a legendary
chapter in the industry's history. While the past two years have been a
momentous period in DEC's history, its important not to lose sight of what
might be an even more extraordinary accomplishment. Since I wrote the first of
what now totals 60 reports on DEC, the company's revenues have grown at a
compound annual rate of 30%, and despite a more than doubling in the number of
shares outstanding, DEC's earnings per share have risen at at 26% compound
annual rate. From this perspective, the past two years represent the latest
instalment in a long-term success story. You are all deserving of praise, but
your success rests upon the foundation built be your predecessors. 

My approach this evening will be to identify the circumstances and forces that
conspired to produce the surge in DEC's growth and profitability that began in
early fiscal 1986, to delineate the circumstances and forces that I believe
will characterise the industry over the next few years, and then to assess the
implications of those circumstances and forces for DEC's future. 

The resurrection of DEC from the dark days of 1983-84 is attributable to many
factors. Some of them -- the increased attention to expense controls, the
rationalisation of manufacturing, the steady stream of new, cost-reduced VAXes,
the redirection of the sales and marketing effort toward large, end-user,
service sector accounts, and the "Digital has it now" advertising campaign --
are internal. While they are testimonials to the wisdom and capabilities of DEC
management, they are not what I will focus on tonight. Instead, I will
concentrate on the industry environment in which these decisions were made.
Changes in what computer users wanted and the situation of DEC's competitors
contributed to DEC's turnaround. These external factors -- variables over which
management had no control -- were, in my view, at least as important as were
the internal factors. Because my purpose is to identify industrial trends and
forces and to assess their impact on DEC's future, it seems appropriate to
focus on the external factors that have aided and abetted the company. 

Although no one knew it at the time, a computer industry revolution began in
August 1981. When IBM introduced its personal computer (PC), it believed that
it would sell 250,000 units by the end of 1986. Instead, it sold more than
3,000,000. Today, the world wide installed base of IBM compatible PCs is
roughly 10,000,000. The revolution created by the huge and unanticipated
success of the PC was a revolution that led to the proliferation and
legitimasation of end-user computing. In so doing, the PC undermined IBM's
competitive position and strengthened DEC's. Just as political revolutions have
frequently swallowed their perpetrators, so did the revolution in computing. 
                                                                            
In 1983-84 -- when the PC boom was in full flower - IBM was the beneficiary and
DEC, along with many others, was the victim. IBM's revenue growth was boosted
by the PC's overwhelming acceptance, and equally more important, IBM -- which

                                                                     Page 3

had previously been perceived by users as a conservative force -- came to be
viewed as an innovator. For DEC, the short-term effect of the PC was
disastrous, as those of you who were here will attest. There was no gold at the
end of DEC's Rainbow. Instead, there was concern among both users and investors
that the company -- which was also having difficulties with its Venus and
MicroVAX development efforts -- had seen its best days. 

During those difficult times, few, if any, industry analysts made note of the
fact that the PC, by shifting the industry's centre of gravity from data
processing professionals to end-users, was rapidly eroding IBM's power. In late
1984, IBM's PC shipments suddenly levelled off. This event, which precipitated
the deep slump that would plague all sectors of technology for two years, was
precipitated by corporate users' recognition that the huge quantities of PCs
that they had purchased had little effect on organisational productivity. What
was needed -- methods to allow PCs to participate in networks -- was not
available. 

The absence of integrative products hurt IBM more than DEC because it was IBM,
not DEC, that had populated the world with PCs. IBM lacked integrative products
because it had not anticipated that a user requirement for such products would
develop. IBM could not respond to the growing demand for peer-to-peer networks
unleashed by the boom in end-user computing that its PC had created. 

In April 1985, DEC enhanced DECnet to provide for the attachment of PCs. During
the next 12 months, further IBM connectivity announcements, growing user
awareness of DEC's capabilities and intensifying customer disenchantment with
IBM combined to produce the mid-1986 pickup in DEC's orders and the slump in
IBM's orders. 

DEC had the right kind of networking because of its early advocacy of
distributed processing. End-user computing is the 1980s name for distributed
processing. The success of the IBM PC created a world in DEC's image. It also
destroyed IBM's strategy of product segmentation. Users did not view the
incompatibilities among IBM's various mid-range systems as a liability, and
users did not fully appreciate the absence of incompatibilities in DEC's
product line until demand for peer-to-peer networking exploded. Now, IBM is
being forced to homogenise its mid-range systems. This would not have necessary
if the market had not demanded peer-to-peer networks; the market would not have
demanded peer-to-peer networks if end-user computing had not been proliferated
and legitimised; and end-user computing would not have been proliferated and
legitimised if IBM had not introduced its PC. IBM was, indeed, an agent of
change, but a change that benefited DEC, not IBM. 

If it was the PC that produces the environmental change that eroded IBM's
hegemony, then the next question is why DEC has benefited so
disproportionately. After all, there are several other major computer companies
that share DEC's belief in distributed processing, and only one of them --
Tandem Computers -- is sharing DEC's prosperity. The explanation, I think, is
as follows: 
    
    1. The irony of IBM's -- the bastion of centralised computing -- 
       being injured by its successful PC effort is matched by the irony 
       of DEC's -- the bastion of distributed processing -- being helped 

                                                                     Page 4

       by its unsuccessful PC effort. Had the Rainbow and Professional 
       been successful, DEC probably would have been less aggressive in 
       its support of IBM's PCs on DECnet, because support if IBM's PCs 
       would have resulted in lower demand for DEC's PCs. By conceding 
       the desktop to IBM, DEC facilitated the absorption of the desktop 
       into the VAX computing environment. The vendor that most 
       strenuously resisted the IBM PC -- Wang -- has been among its 
       primary victims.
    
    2. The PC, by stimulating demand for networks and involving 
       end-users, has changed the criteria used by major corporations in 
       reaching purchasing decisions. Prior to the mid-1980s, users 
       purchased computers for use in stand-alone environments and did 
       not require that vendors supply application software. In such an 
       environment, price/performance played a significant role in 
       purchasing decisions. Now, network functionality and application 
       software have taken precedence over price/performance. So have 
       customer support and perceived vendor viability -- the former 
       because network nodes are installed in many geographic locations 
       and the latter because corporations become highly dependent on 
       their networks. The vendor that stressed price/performance in 
       competing against DEC -- Data General -- has fallen on hard 
       times.
    
    3. The vendor, which, by virtue of its size and advocacy of an 
       operating system (UNIX) that offered to extinguish software 
       lock-in, presumably should have benefited greatly from IBM's 
       malaise, did not. AT$T's effort to become a major computer vendor 
       has been a miserable failure. As a result, no major corporation 
       has served as a champion of UNIX.
    
    4. The only other company that could have challenged DEC during the 
       immediate post-PC boom period -- Hewlett-Packard -- has been 
       preoccupied with its effort to bring a product line with a 
       radically different architecture to the market.

By virtue of the forces unleashed by the PC, the soundness of its strategies,
the capabilities of its products, and the problems of its competitors, DEC's
power has now reached new heights. This power is evidenced by the company's
ability to price its products on the basis of their value, rather than their
cost, the company's ability to resist the spread of UNIX, and most importantly,
the company's ability to gain widespread acceptance of networks that, with the
exception of IBM PCs, consist of VAXes, and only VAXes. 

The balance of competitive power has shifted in DEC's favour because of the
superiority of DEC's networking and DEC's success in redefining the locus of
account control to the network from the mainframe computer. In attempting to
maintain and improve upon the new balance of power, DEC will face two primary
threats, the first from IBM, and the second from other industry participants.
IBM will develop a networking capability that will be positioned as an
alternative to DEC's. Other industry participants will attempt to transform
networking into a commodity. 


                                                                     Page 5

In the early 1980s, IBM identified "Japan Inc." as the primary threat to its
continued industry dominance. Because the Japanese threat was based on low
manufacturing costs, IBM embarked on a massive capital spending program in
order to become the world's low-cost producer. The effort may have been
successful, but it didn't matter, because DEC, not the Japanese, emerged as
IBM's chief nemesis, and the DEC threat reflects the functionality of its
products, not their manufacturing costs. 

While IBM's concern about DEC may have started to rise at about the same time
as its customers began to voice their displeasure regarding IBM's ability to
network its PCs, it was about a year ago -- when DEC began to penetrate IBM's
traditional strongholds -- that IBM publicly admitted, albeit, tacitly, that
DEC had seen the future more clearly that it had. "Cross-system consistency"
entered the IBM lexicon at this time last year. The phrase was recently replace
by System Application Architecture, or SAA, which represents IBM's multiyear
objective of achieving application software commonality across its product
lines. SAA, the implementation of peer-to-peer capabilities within SNA, the
endorsement of departmental computing signified by the 9370 series processor,
and the support of Ethernet -- a de facto DEC standard -- on the 9370, together
represent IBM's mea culpa and the first prong of a two-pronged strategy
designed to shift the balance of power back to Armonk. 

The first prong of IBM's strategy is not likely to pose a significant risk for
DEC. It is a defensive strategy aimed at stabilising the competitive situation
pending the arrival of OS/2 and, with it, a partial implementation of SAA. In
both networking and midrange processors, DEC will sustain a sizeable
competitive edge. An SNA retrofited to operate in peer-to-peer mode will have
to cope with a hierarchical overhead not found in DECnet. When DEC announces
the MicroVAX III, it will be two generations ahead of IBM in midrange hardware
technology, and it will have a chip-based processor whose performance will
rival that of the most powerful 9370. 

The second prong of IBM's strategy is the Personal System/2, a product line
that, in IBM's words, will have a key role in the implementation of SAA. When
the extended version of OS/2 becomes available -- probably late 1988 -- IBM
will presumably have achieved a seamless integration between those of its PCs
running that operating system and its System/370 and System/3X processors. By
that time, the System/36 and System/38 will have been merged into a single
product line -- the "Silverlake" processors. Co-processors running in the
PS/2's expansion slots will provide the seamless integration. The recent
agreement between IBM and Lotus that provide for a version of 1-2-3 that will
run on System/370s may be a preview of a key element of IBM's strategy -- to
replicate in mainframe form the most popular PC application software. 

The level of competitive risk rill rise in late 1988. With the arrival of
extended OS/2, IBM will be able to offer two alternatives to DEC: a three-tier
approach involving 9370 as departmental processors, and a two-tier approach
involving only mainframes and PCs. When this happens, the old debate over
whether departmental processors are necessary will be reopened. It will be
reopened because IBM is strong at the high end, while DEC is strong in the
midrange. To the extent that popular PC software are available, the case for a
two-tier approach will be strengthened, and DEC's applications commonality
advantage will be weakened. 

                                                                     Page 6


In order to head off these risks, DEC needs to accomplish the following by
mid-1988: 
    
    oo	It should have achieved volume shipments of a quadratic 
       multiprocessor based on the VAX 8800. Such a product, which would 
       provide 25-30 MIPS of processing power in a single system image, 
       would establish DEC as a mainframe vendor. So established, DEC's 
       competitive position would not erode if the two-tier approach 
       becomes the dominant computing structure.
    
    oo	Versions of Lotus 1-2-3 and other key PC software packages should 
       be available on VAXes. That they have not been has not worked to 
       DEC's disadvantage, because thus far, they have not been 
       available on any of IBM's products except the PC, and DEC does 
       not compete against the PC. When they are available on the 
       System/370s, they will pose a risk because DEC does compete 
       against System/370s.
    
    oo	DEC should develop application software that will not be 
       available on IBM PCs. A logical candidate is a distributed 
       relational database management system (DBMS). Statements from IBM 
       officials indicate that it will be several years before IBM will 
       have this capability. In view of the likelihood that availability 
       of a distributed relational DBMS would result in the migration of 
       data from mainframes to PCs, it is not clear how IBM will choose 
       to employ this capability when it is available. A second 
       consideration arguing in favour of a distributed relational DBMS 
       is that it would boost the company's competitive position in the 
       on-line transaction processing market. Tandem's NonStop SQL is a 
       distributed relational DBMS.

While coping with IBM, DEC must also deal with a number of other vendors whose
primary objective will be to turn networking into a commodity. This process had
already started: 
    
    oo	At the physical level, the "commoditisation" process is 
       essentially complete, as virtually every vendor supports the 
       attachment of its products to Ethernet.
    
    oo	At the logical level, commoditisation is under way. DECnet is now 
       supported by workstation vendors -- Sun Microsystems, for 
       instance -- and minisupercomputer vendors, such as Convex.

It is only a matter of time before DECnet's rapidly growing installed base
causes the general-purpose systems vendors to swallow their pride. Prime
Computer and Data General are known to be considering taking the plunge. When
this happens, DEC will be in the same position as IBM, whose SNA was subject to
commoditisation beginning in the early 1980s. 

The commoditisation of DECnet, not an awakened IBM, represents the principal
risk confronting DEC. As commoditisation proceeds, DEC will lose what is today
a primary source of product differentiation, and its account control will slip.

                                                                     Page 7

Its competitors will find price to be a more effective competitive weapon than
it is today. If DEC does not lower its prices to reflect the lessened
differentiation of its products, its growth will slow and it will lose market
share. If it does lower prices, its profit margins will contract. 

Of these two possibilities, maintaining market share by accepting lower margins
is preferable. Losing market share would mean that a rising percentage of the
processors attached to DECnets would run operating systems other than VMS. Over
time, it is highly likely that the predominant operation system on DECnet would
become UNIX, for which the availability of application software is rapidly
growing. If this scenario were to unfold, DEC's crown jewel -- VMS -- would
lose its lustre. IBM, which seems to be warming up to UNIX, would announce its
support of DECnet and move in for the kill. 

Is it true that network commoditisation holds such dire consequences for DEC?
No, but only if DEC can identify and develop new forms of product
differentiation to replace what it will lose when its competitors attach their
products to DEC's networks. These new forms of differentiation must cause DEC's
customers to conclude that, although other vendors' products can be attached,
it is in their interest to continue to attach DEC's products. Proprietary
applications that are better than anyone else's will be necessary to prevent
the storm clouds from gathering. President and Chief Executive Ken Olsen has
said that DEC is a software company. If DEC truly believes what Ken said and
dedicates itself to being the leader in the software of the 1990s -- artificial
intelligence and expert systems -- the miracle of Maynard will continue.