Chapter
1
Losing the Leadership: How the Information Technology Leaders of the 1980’s lost their Leadership by 2001 How can large technology companies with the
resources, money, power, global reach, talent, patent portfolios, brands,
production capacities, and market leadership positions that Lucent, Xerox, IBM
and others enjoyed lose out to new companies starting with almost nothing and
very little capital? It seems impossible. Yet it did happen and is continuing to
happen. In the 1980’s and 1990’s, every leading computer company lost out to
startups created during that era. By 2001, IBM was the only company of the
computer market leaders in the 1980’s that was still in the general purpose
computing business, and even IBM’s worldwide market share had dropped from a
dominant 66% in 1975 to a mere 17% by 1992[1].
The market share winners were startups created in the 1970’s and 1980’s who
often were able to use the leader’s own technology to beat them. In the late
1990’s and beginning 2000’s, a very similar scenario is being played out in
the communications industry (more on that in Chapter 7). I call this pattern the “Slingshot Syndrome.”
The syndrome appears to be particularly successful when the markets involved are
being disrupted by rapid technological change, putting a premium on rapid
commercialization of new technology. The term is derived from the biblical story
of David and Goliath where David slayed the arrogant, heavily armored and
unsuspecting giant with a mere slingshot, and then decapitated Goliath with the
giant’s own sword. The startups with their slingshots of new-to-the-world
products and rapid commercialization are going to battle with the Goliaths of
the information technology industry — and are winning. How is this possible? One is tempted to blame bad management of the
losing giant companies for their troubles, and a number of authors have written
analyses along that line; however the facts strongly indicate a much more
systemic set of problems that are apparently faced by all. But are they
inescapable? Why did each of these companies, especially the
ones who held the future technologies and paradigms in the palm of their hand,
not get to the future first? The answer
in each case is that they were distracted by the need to protect their current
businesses, product lines, revenue streams, profit streams, personnel, and
manufacturing facilities. They were so distracted with the present that they
lost sight of the future, and in many cases, they destroyed their own future
with the discontinuation of projects that would become the next generation
technologies. These new technologies and paradigms were threatening their
existing sources of revenue and profits, so the corporation moved to protect
them rather than invest in what was threatening them. In doing so, they were
unable to invest in their own future.
Let me tell you an illustrative story. In 1981, as a young Member of
Technical Staff at AT&T’s Bell Labs, I was given the assignment to predict
the future of distributed computing, and to propose a strategy for the Bell Labs
Computing and Networking Technology Department that I was later to head. I had
the good fortune of receiving an invitation to visit Xerox’s Palo Alto
Research Lab, known as PARC, where I was shown their experimental Alto
workstations networked together with their experimental Ethernet. The Alto
workstation had icons, overlapping windows, drag-and-drop, a WYSIWYG (“What
You See Is What You Get”) text editor (similar to the current Microsoft Word),
a bit-mapped screen, pop-up menus, was controlled by a mouse, and was networked
over Ethernet to print servers running laser printers. This may all seem very familiar and quite obvious
today, but at that time, IBM had just introduced the PC, which included one
floppy drive, 18 kilobytes of memory,
no hard drive, and no mouse. A 32-employee company named Microsoft had just
introduced MS-DOS. The Apple Macintosh was still three years in the future, and
computer terminals and PC’s had command-style (think of MS DOS) interfaces and
were connected to mainframes or minicomputers with telephone wires running 300
to 1200 bits per second. Printers were still of the impact printing variety,
either slow desktops that were essentially modified electric typewriters or
large, faster, noisy chain or drum printers in computer rooms. Bell Labs’ computing environment, which was
state-of-the-art, consisted of IBM, Amdahl, Univac, and Honeywell mainframes, a
Cray supercomputer, and scores of Digital Equipment Corporation minicomputers
hooked to “dumb” terminals over telephone lines. This
may sound like a lot of technical gibberish, but what I saw at PARC was a
glimpse of the future that became the computing paradigm of today and now runs
in our offices and homes. When the PARC innovations were later combined with the
PC and the Internet, it changed the world and created the modern information
age. I was very impressed with what I saw, so impressed
in fact, that I went back to Bell Labs to write a memorandum informing my
supervisors that Xerox had invented the future of computing and all computing
would be organized that way in the future. I further proposed that AT&T (the
then owner of what is now Lucent and Bell Labs) should partner closely with
Xerox since it was clear to me that Xerox was in a position to take over the
computer industry from the companies that currently reigned: IBM, Digital
Equipment (now part of Compaq), Univac (now part of Unisys), Control Data
Corporation, Honeywell, Burroughs (now part of Unisys), Wang, Data General, and
NCR. Finally, I proposed that Bell Labs should be wired with Ethernet
immediately, in an effort to deploy this new form of computing as soon as
possible as well as understand the implications on communications technology. At that time, AT&T’s Western Electric (later
renamed Lucent) communications products were completely dominated by circuit
switching technology rather than the packet switching technology that Ethernet
and the Internet were beginning to demonstrate, although Bell Labs and Western
Electric’s internal computer network was already running a early version of
TCP/IP, the packet switching protocol that later powered the Internet. I included in my recommendation a request for
sufficient budget to cover the expenses for a Bell Labs–wide Intranet. The
number was large enough and the plan was grand enough to require the approval of
a Vice President. That was my first personal experience with a Bell Labs Vice
President. I recall him carefully examining the samples of Ethernet coaxial
cable that I brought along, as well as the “fang” connector. He then told me
that I had the right idea, and that he would approve the large budget I was
requesting! But I would have to do it with Bell Labs technology, not anything
developed by Xerox, and he sent me to scour the research department for a
competitive technology. I did just that and discovered that within Bell Labs
Research, a promising version of packet switching was being developed that was
later to be called “cell switching.” Cell switching, which involved
switching short, fixed-length packets, appeared to have a number of advantages
over both circuit switching and the versions of packet switching being used in
Ethernet and elsewhere. I worked with the researchers and developers to build
enough prototypes of their switch (called Datakit) to serve the 20,000 computer
terminals in Bell Labs. We interconnected the cell switches with light fibers,
making it the first cell-switched light fiber packet network in operation in the
world. A cell-switched packet network connected by light fiber is the
telecommunications industry’s current (2001) view of the future and Bell Labs
had it running internally in 1986. By 1986 (the same year a small startup named Cisco
shipped its first product), I had overseen the deployment within Bell Labs of
not only the largest operational cell-switched fiber data network on earth, but
also the first intranet, protected by the first firewall (invented in Bell Labs
Research), and the first large corporate single-directory email system. Bell
Labs’ and Western Electric’s (later to be renamed Lucent) backbone corporate
network (named BLN or the Bell Labs Network) was a packet switched design based
on an early version of TCP/IP (which when introduced by DARPA[2]
into the public domain in 1982 created the Internet) and had been operational in
Bell Labs since 1979. I know this to be a fact because I authored the BLN
Internet Protocol (IP) and Transaction Control Protocol (TCP) modules myself, as
well as led the BLN project through its first release. I had been inspired by a presentation given by Vint
Cerf (who deserves the title of “the father of the Internet”) at a data
communications conference at Snowbird, Utah in 1977, where he described TCP and
IP standards that he and others were working on. I went back to Bell Labs and
coded it into BLN, and had it running as a production network in 1979, 3 years
before the TCP/IP standard was released into the public domain by the Defense
Advanced Research Project Agency (DARPA). Note that BLN was routing IP-like
packets around Western Electric and Bell Labs 7 years before Cisco released
their first router. By 1986 the BLN and Datakit prototype networks were the
backbone networks of Bell Labs and Western Electric, interconnecting two
continents via the first undersea light fiber cable, linking over 20,000 nodes
and 28 locations, and carrying a very heavy traffic load every day. Because the Internet then was still a very open
community of researchers, the Bell Labs Network (BLN) in the latter 1980’s
actually was carrying much of the bulletin board and email backbone traffic of
the Internet. AT&T converted the cell switching prototypes that we had
deployed in Bell Labs into two data switching product lines (Datakit VCS and ISN),
converted the Bell Labs email system (POST) into the AT&Tmail email service,
and merged some of the prototype networking features into its Unix and Computer
Systems products, but the BLN itself was never commercialized. Two other pieces of the puzzle of what became the
modern computer paradigm also came from Bell Labs in this time frame: the UNIX
operating system and the C (as well as the later C++) programming language. UNIX
was the operating system of choice for the Internet developers, with the effect
that to this day, servers running the UNIX operating system handle the majority
of web traffic. The C and C++ programming languages came to dominate to an even
larger extent. By the early 1990’s, almost everything related to
communications systems, the Internet, and PC’s was written in either C or C++,
replacing the dozens of programming languages that had preceded them. The Java
programming language, which caused a great stir when it was introduced in the
mid-1990’s, was nothing more than a clever derivation of Bell Labs’ C++
programming language. The dominance achieved by UNIX and C were so great that
the leading computer companies formed the Open Systems Foundation (OSF)
consortium in the late 1980’s to leverage their combined strength in order to
prevent AT&T from monopolizing the future computing paradigm! As a
result of these events, by the early to mid 1980’s Xerox and Bell Labs had
every technology they needed to become the dominant forces in the new computing
and networking paradigm that was to unfold, and were running them in prototype
form inside of their research labs. Despite the fact that both companies did
succeed at turning some of their internal prototypes into products, they lost
the leadership almost immediately to new startups. The Apple Macintosh,
introduced in 1984, took the Xerox Alto workstation concepts and made them
affordable to consumers, something Xerox never did. In 1979, 3Com was founded to
commercialize Ethernet. In 1982, SUN Microsystems was founded and merged the
Alto workstation concept with UNIX, the Motorola 68000 microprocessor, and
DARPA’s TCP/IP Internet protocols to commercialize the concept of a
professional workstation. In 1984, H-P formed an internal startup to
commercialize laser printers. Finally, in 1985 Microsoft commercialized the Alto
iconic interface into their first release of Windows, which transformed it from
a mediocre software company to the dominant player it is today. Meanwhile, Xerox chose to focus its resources on
its survival in the copier business, where it faced competitive attack from
Japanese manufactures. Lucent (then part of AT&T) focused its R&D
resources on the digital circuit switching of voice communications in order to
protect its Electronic Switching Systems business from competitive threats from
Nortel, ITT (now Alcatel) and Ericson. Both succeeded at fending off the
attackers to their core businesses, although not without loss of market share
and loss of access to the new developing markets. By 2001, Lucent was telling its customers that
packet switching is the future of all communications, replacing circuit
switching. In Lucent’s 2000 Annual Report, a year in which the Lucent stock
fell from $82 to $18 per share and the Chairman and CEO was fired by the board
of directors, Lucent noted that “lower revenues in switching products
primarily due to the shift in customers away from circuit switching …
negatively impacted growth.”[3]
In the same report Henry Schacht, Lucent’s Chairman and CEO, said that Lucent
would “fully focus [the] business on delivering the next-generation
Internet.”[4]
By 2001, Lucent admitted that it was for sale and was fighting off rumors of
impending bankruptcy after losing $3.7 billion in the first quarter of that
year. Standard & Poor’s cut Lucent’s once stellar debt rating to that of
junk bonds (BB-plus)[5]
in the middle of 2001, while Xerox bonds were cut to a junk rating in 2000.
Xerox was also fighting rumors of impending bankruptcy,[6]
having fired its CEO in May 2000 and losing $292 million in 2000[7]
and $289 million in the first three-quarters of 2001.[8]
Despite these events, the startups created from Lucent and Xerox’s inventions
had developed a total market value of around 1 trillion dollars! If Xerox had
managed to maintain as little as 1% ownership in each of the companies formed
around its inventions, it would have more than four times its 2001 market
capitalization! Cisco, the company that was founded in 1986 to commercialize
network technology similar to what Bell Labs had been running internally since
1979, briefly became the most valuable company on earth in 2000, surpassing both
Lucent’s and AT&T’s valuations. Lucent and Xerox were not alone in this bloodbath
— Palo Alto Research Center’s (PARC) historic breakthroughs were known
during this period to the technical community within other large information
technology companies as well. The computer industry leaders in the early
1980’s consisted of IBM, Univac, Burroughs, Honeywell, Digital Equipment,
Wang, Data General, Amdahl, NCR, AT&T, Prime, Perkin Elmer, Gould, Cray and
Control Data. None of these companies moved into a leadership role in
commercializing the Xerox breakthroughs either, although Digital deserves an
honorable mention for playing a significant role in the Ethernet
commercialization. Instead, they left the field wide-open to new startups such
as SUN, Apollo (later bought by Hewlett Packard), 3Com, Cisco, Apple, and then
Dell, Compaq, and Microsoft to introduce these breakthrough technologies into
the market. Note that of the industry leaders of early to mid 1980’s, only IBM
remains as an independent general purpose computer manufacturer 15 years later.
Univac and Burroughs merged into Unisys, Digital Equipment was bought by one of
the startups (Compaq), and Honeywell sold its computer operation to the French
company Group Bull. Wang, Data General, Cray, and Control Data are either no
longer in the general-purpose computer hardware business or no longer exist. NCR
is still in the computer business, but only with special purpose data warehouse
machines known as Teradata. Why was IBM the exception to this almost complete
turnover of computer manufacturers? One answer is that IBM did something in the
early 1980’s that none of its competitors did.
It was a move that IBM had never done before, nor has done since. IBM
created within their corporate walls a startup operation that was so removed
from the usual IBM controls and culture that it was kept a secret from the other
IBM divisions. They basically hid the entire project in order to give it the
nimbleness required to create a new market without any interference from their
existing business units or functions. That startup was the PC division, located
in Boca Raton, Florida, miles from any other IBM presence. The removal of the PC
division from the controls of the corporate environment certainly contributed to
its enormous success; its open architecture was of equal significance to the
modern information industry as the Xerox PARC inventions, the Internet, or the
Bell Labs contributions. IBM would probably still be the leading PC
manufacturer in the world if it had not attempted to fold the PC division back
into the company, rather than leaving it to operate as an independent startup
with its own culture and controls. What happened was that in the early 1990’s
IBM’s most profitable product line, the mainframe computers, came under severe
attack, threatening the viability of the whole company. IBM’s attention was
diverted to that issue and what to do about it. While its attention was
diverted, the 1980’s startups Compaq and Dell gained market share and
eventually passed IBM as market share leaders in the mid-1990’s. By 2001,
Dell, a company that started operating in a dorm room in 1984, was the largest
PC maker on earth and rumors were spreading that IBM may exit the PC business
after a number of unprofitable years and sagging sales. In the communications equipment business, things
played out differently, but no better for the entrenched companies. In 1984,
AT&T settled the government antitrust suit that had been filed in 1969 by
breaking itself up into the Regional Bell Operating Companies (RBOCs) and
retaining only the long distance company and the equipment manufacturing company
(Western Electric) and R&D (Bell Labs). The Regional Bell Operating
Companies for the first time were completely free to buy their equipment from
any company. They immediately started courting ITT, GTE, Nortel, Ericson, and
Siemens. The split-up of AT&T came at the worst possible
moment for Western Electric, because telephone switching was just starting to
convert from analog to digital switches, and Western Electric was behind.
AT&T responded to this threat by handing Bell Labs a virtual blank check to
do whatever it took to build the best digital switch in the world, and to do it
fast. With so much at stake to protect its existing equipment business,
AT&T, Bell Labs, and Western Electric had no time to pay any attention to
new computer paradigms from Xerox’s PARC, the emerging packet switching
protocols from DARPA, cellular communications which it had co-invented with
Motorola, or the new cell-switching technologies from its own research lab. The
resulting digital switch, the 5ESS, was the largest project ever undertaken by
Bell Labs, eclipsing its earlier work on the Anti Ballistic Missile system
(SAFEGUARD), Bell Labs contribution to NASA’s Apollo Program moon shot, or
PicturePhone. But it was a success, and by the late 1980’s two of 5ESS’s
competitors had exited (ITT and GTE), Ericson and Siemens had decided to compete
in other markets, leaving only Nortel as a serious competitor in the United
States. In the 1970’s, like Xerox, AT&T together with
its Bell Labs and Western Electric subsidiaries had actually developed what
turned out to be a very good vision of the future Internet services. They
foresaw a world in which intelligent services would appear to be inside of the
network, and customers would be able to access these services without knowledge
of where the computers were physically located. They attempted to implement such
a network and called it ACS and later NET1000. One of the first services was to
be travel services. The technology chosen to make it work was not suitable or
ready, and AT&T gave up in the early 1980’s after spending more than $1
billion. One of the reasons that AT&T invested in
NET1000 to begin with was the purchase by IBM of Satellite Business Systems, a
satellite carrier that allowed corporate customers to bypass AT&T’s long
distance telephone network. AT&T was afraid that IBM would create a data
communications infrastructure that had no need for AT&T’s telephone
network. So AT&T countered by attempting to create a computing service that
had no need for the customers to buy IBM’s mainframes. The same year that
AT&T announced that it was discontinuing NET1000, IBM sold SBS to MCI.
Presumably, both companies thought that their nightmares were over once the
other had ceded incursion into each other’s market spaces. What both of them
missed was that the nightmare was real — there would be a bypass intelligent
data network that would make owning mainframe computers superfluous. It would be
called the Internet, and it would be built by hundreds of companies that IBM and
AT&T had never heard of or who did not exist at the time. It would, just
like they both worried, challenge both of their leadership positions in computer
manufacturing and communications.
In 1995, in anticipation of the effects of the US Telecommunications Act
of 1996, AT&T CEO Bob Allen decided to break up AT&T once again,
splitting equipment manufacturing (Lucent and NCR) from the long distance
operations (AT&T). A result was that Bell Labs stayed with Lucent. An
unexpected side effect of the Telecommunications Act of 1996 was that it created
a huge, but temporary demand for telecommunications equipment. The Regional Bell
Operating Companies, suddenly aware that they would soon face competition in the
local services monopolies from many sides (AT&T, Cable Companies, and new
Competitive Local Exchange Carriers), decided en masse that they should use their monopoly revenues, while they
still had them, to upgrade their antiquated networks to meet the certain
competition.
Since most of their installed equipment was either made by Lucent or
Nortel, this created an enormous boom for both companies just as Lucent was set
free from AT&T. Although the Internet age as we know it now was created in
1994 by the introduction of the first browser and the allowance by the US
government of commercial activity on “the net,” Lucent was far too busy
meeting the demand for voice network upgrades from Regional Bell Operating
Companies to notice or think about it. Instead of taking advantage of its
leadership and understanding of packet switching, cell switching and optical
interconnects that it had been building up internally for over a decade, Lucent
let a new raft of startups take the lead in all three categories because it were
too busy with temporary demand. By the end of the 1990’s, having realized its
mistake, Lucent feverishly used its highly inflated stock to buy up startups in
these categories, and used its excellent bond ratings to buy market share with
the Competitive Local Exchange Carriers (CLEC’s) by advancing them more than
enough credit to buy their equipment. But it was too little, too late: in 2000,
as the Regional Bell Operating Companies (RBOC’s) decided that their network
upgrades were completed, Lucent stock fell from $81 per share at the end of 1999
to $5 per share in 2001. The Lucent CEO was fired, and serious enough liquidity
problems surfaced that the bond rating was cut to junk. By early 2001, rumors of
liquidity problems and potential bankruptcy surfaced, and Lucent was attempting
to sell off enough assets to remain a going concern. By the spring of 2001, the
CFO had been fired and it was clear that Lucent as a whole was for sale, if they
could find an interested buyer. When Lucent’s fiscal year closed at the end of
September, 2001, they reported a mind-numbing loss of $16 billion ($11 billion
of which was due to restructuring charges and other one-time charges). If Nortel was faring
better, it could not be determined from the financial reports. Both companies
announced layoffs and divestitures that essentially cut their total number of
employees in half during 2001 alone. Nortel blew almost everyone’s mind by
reporting a $22 billion net loss in the first half of 2001 ($14 billion of which
were due to write-downs of acquisitions and inventory), which helped bring its
stock price down 94% to under $5 from its previous year high of over $82. The
possibility that the new generation of communications equipment startups,
ranging from optical switching to cell switching to soft switching technology,
would grab the market lead from the likes of Lucent and Nortel had become a
virtual certainty. [1] International Data Corporation, 1993. [2] The US Defense Department’s Advanced Research Project Agency [3] Lucent Technologies, 2000 Annual Report, pg. 17. [4] Lucent Technologies, “Chairman’s Message”, 2000 Annual Report. [5] Reuter’s, June 12, 2001. [6] Business Week, March 5, 2001. [7] Xerox Corporation, 2000 10-K Report, as amended in June 27, 2001. [8] Xerox Corporation, “Third Quarter Results,” www.xerox.com, October 23, 2001.
|
Copyright © 2001 Reid McRae Watts, Progeny Ventures LLC and its licensors. All rights reserved. Last Revised: December 05, 2002 |