United States. Congress. Senate. Committee on the.

The Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) online

. (page 78 of 140)
Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 78 of 140)
Font size
QR-code for this ebook


ing it.

IBM embarked upon a campaign, perhaps unequalled in the history of com-
petition, to block Control Data from marketing more of its extra-large computers,
tbe only ones actually available and in use, to the university max'ket, using the
IBM Paper Machine, "bait and switch" tactics, thi-eats, disparagement, "buy-
backs" and discriminatory discounts. This resulted in extreme confusion in the
market place and the immediate loss of much otherwise profitable busine.ss to
Control Data.

IBM openly proclaimed that it was not going to lose any more "university
business" and customers were so advised. Also. IBM's chief executive officer,
Thomas J, Watson, Jr., in December of 1964, stated that his company "is going
to compete at whatever price level is necessary to maintain our position in the
industry . . .". (See Appendix 4 attached for a more complete description of
this quotation taken from the December 1964 issue of DATAMATION magazine. )
The discount effort, controlled by IBM top management, was of such magnitude
that a separate part of the pricing practices section of the Report is devoted to
its description. Despite giving lip-service to a standard "educational" discount,
IBM gave varying discounts, from 20% to 100%, to various institutions, the
amounts being tailored to whatever appeared necessary to beat out Control Data,
Greatest discounts were given in the case of large computers to the most in-
fluential institutions and to universities Avhich had large money contracts with
the United States Government, Purchasers of small computers or other etpup-
ment received little or no discounts. This program severely injured competition.
Should it be continued, IBM will drive all competitors from the market.

Because of its great size and dominant position, IBM controls the market
structure in the data processing industry. If IB]M induces customers to lease
rather than purchase equipment, other manufacturers must also lease if they

40-027 — 75 46



5550

wish to be competitive. IBM has eucouraged leasing of computer systems in the
following ways :

1. Rapid changes by IBM in model numbers, thus preying on customers'
fears of rapid technological obsolescence, have made many customers un-
willing to invest large sums of money in computer purchases, especially when
these new models have improved performance/price ratios.

2. IBM has steadily reduced the cost of leasing in relation to the purchase
price of its systems by doubling maintenance charges on purchased com-
puters over the last three years without correspondingly increasing rental
rates (see Appendix 5) and has provided additional incentives to lease in the
form of reduced rental for extra -shift use (explained in Appendix 6) and
allowance of rentals ]jaid for one computer as a credit on the subsequent pur-
chase of other equipment (.see Page 27 infra) .

The ultimate effect of these practices is identical to the effect of the practices
attacked in the 1952 Government suit against IBM. Because many customers
have felt compelled to lease, IBM's competitors have been required to over-
extend themselves financially in funding an increasing percentage of leasing.
IBM, on the other hand, with its enormous financial resources, can easily afford to
postpone revenues and profits over a longer period.

On the othc-r hand, when leasing is not to the competitive advantage of IBM,
as where IBM is supplying equipment to its competitors, the pattern is reversed
and customers are required to purchase. Thus with the market trend toward
leasing firmly established and growing. IBM very recently altered its policy so
as to require its competitors to purchase. These competitors are required on many
occasions to provide compatibility with IBM systems and, therefore, to offer
IBM equipment as part of their own systems. (The equipment may be of a
type manufactured only by IBM or it may be desired as a result of IBM's hard-
ware, software or programming activities). Competing manufacturers who
need ir>M enupiment as part of their systems are ?iow i-eriuired to ninke sub-
stantial capital outlays even though their own customers are demanding to lease
the same equipment.

The Miscellaneous Unfair Trade Practices section of this Report contains a
brief description of IBM's '"software" (computer programs or instructions)
efforts, l)y which IBM ties its customers to its entire line of computers through
making available to them at little or no expense vast "libraries" of programs
helpful for facilitating greater utilization of computers at reduced cost. Having,
during the industry's lifetime, by virtue of its market dominance, been in the
position of establishing de facto standards, IBM now seeks to set new standards
and to exclude competitors from the formation of these standards and to delay
competitors from meeting the standards. IB:\I has also changed equipment stand-
ards with the same anticompetitive effect. The Miscellaneous Unfair Trade
Practices section also describes IBM's anticompetitive activities in connection
with its service bureaus and patent policy.

The foregoing practices are only briefly summarized at this point, as an aid
to understanding the overall competitive situation. The details are set forth
in the following sections.

F. Appropriate Relief

IBM's practices have substantially injured all competition in the computer
industry and have substantially impaired the likelihood that other members
of the industry will be able to compete effectively in the future. As already
indicated, several firms in the industry have gone out of busine>s or sold tli'e
computer part of their operations. Moreover, we are confident that no major
competitor, despite public relations efforts to convey a contrary impression, is
able to make a profit from computer sales. Although Control Data was able to
make a profit for several early years in its short history, its most recent financial
statements show virtually no profit and reveal the repressive effect of IBM's
anticompetitive practices. Immediate and effective relief is imperatively needed
to create conditions in the industry which will permit other producers' to chal-
lenge IBM's dominance. Indeed, if IBM's practices are not enjoined, it is highly
unlikely that the few competitors left in the industry will remain viable^ or
that they will be able to engage in the research and development and marketing
endeavors which are the lifeblood of their continued existence.

Damages are of course recoverable in a private action but tliese cannot fully
compensate for the severe and long-term injury to competition which IBM's
practices have wrought. Prior efforts at restraining or confining IBM's anti-
competitive efforts have demonstrably failed, such as civil suits by competitors



5551

and the Consent Decree entered iu 1956 following a civil suit brought by the
Justice Department. Notwithstanding these prior suits, the Consent Decree
and various investigations, IBM top management consciously resolved to proceed
with the present course of conduct. In fact, in April 19U5, Thomas J. Watson, Jr.,
Chairman of the Board of IBM. stated in substance, "We have been living under
the letter of the Consent Decree until now. We are going to be competitive in
the large-scale area."

With .such baclvground, it is readily apparent that the only effective remedy
is the fragmentation of IBM into a number of smaller independent and unrelated
companies incapable of wreaking havoc on the scale of such a gigantic corpora-
tion as IBM. The concluding section of this Report is accordingly devoted to
appropriate relief, not only dismemberment btit additional forms of relief.

II. IBM's ANTICOMPETITIVE PRACTICES

A. The 360 Story

IBM announced its 360 Series, a group of six supposedly completely compatible
system.s, to the trade on April T, lih'll. Since tliat time. IBM has extensively
advertised and aggres.sively marketed these systems (or systems subsequently
announced to replace them). We believe that both the original announcement
of the 360 Series and IBM's subsequent marketing program operated — and were
intenti(mally designed — to delay sales of existing computer systems by IBM's
competitors until such time as IBM could correct deficiencies and incorporate more
(•.iur,iatibility in the systems it had hepv. marketing. More specifically, we believe
that a thorough investigation will produce evidence that IBM employed the
following tactics. IB]\I announced its 360 Series to the trade before the Series'
technological development had been accomplished in order to delay — and ulti-
mately prevent — purchases of its competitors' computers ; after the original
announcement, IBM frequently changed model numbers, quoted inaccurate and
misleading delivery dates, promoted and sold computers which had not been
developed and which IBM had no intention of delivering — all to confuse pur-
chasers, to disguise the prematurity of its original announcement, and to delay —
and ultimately prevent — purchases of its competitors' computers.

-/. The Computer Industry

As would any sensible monopolist. IB.M hesitated to introduce innovations
in computer technology. By the early 19(;0s. however. IBM became seriously
concerned that its commanding position in the industry might be threatened:
customers, scientific and business, were beginning to re(piire more powerful
computer systems; IBM's small competitors, notably Control Data, seemed
prepared to meet the demand : IBISI's effort, in partnership with the U.S. govern-
uient. to develop a "large-scale" system (the "STRETCH" project) as a standard
ju-oduct had been unsuccessful. IB]M aware that it would be able to retain
its share of the market only by developing new and better computer technology
quickly, established a task force — "Project Spread" — to examine current tech-
nology and to recommend a new line of computers. The "Spread" report, issued
in January, 1962. suggested a series of five compatible computers. The original
design specifications for these computers, designated tentatively as the 8000
Series, provided for "printed" circuits and predicted that the computers could
l)e delivered during the third quarter of 1963. After some internal dispute it
was decided that the performance/price ratio of these computers would be
in;idequate with printed circuitry: the decision was made to use a different
design ulilizing "chip" circuitry. Since these design modifications meant that
the new computers (now the 300 Series) could not be delivered before the
third quarter of 1965. company ofllcials decided in 19^3 to introduce stop-gap
computers to retain IBM's position until its technological problems could be
overcome, and as a way of encouraging customers not to purchase competitors'
products during the development period of the new Series. IBM had every reason
to be confident that delaying competitors' sales would be sufficient to protect
its position in the market : once the technological gap is closed many computer
purchasers will prefer IBM simply because they are confident that IBM will
never be foi-cAd out of the industry.

Thus, IBM had a substantial interest in delaying sales by its competitors
for at least a two-year period. Any IBM effort to accomplish .'s'uch delays would
be aided by facets of competition in the computer industry of which i'B'Sl was
certainly aware. When the industry leader announces a new product line,
prospective purchasers cannot simply ignore the product — even if another com-



5552

pany is currently delivering a product wliicli might be just as good or even
better. Prudent purchasers must delay at least long enough to study and compare
the specifications of the new product. Consider, for example, the pre-\iction in
ELECTRONIC NEWS one week after the 360 Series was announced (April 13,
1964. page 1) : "Sales of computers would virtually dry up until customers
had analyzed the capabilities of the new computer generation, but once this
is done sales would jump."

Computer purchasers are especially likely to delay when a new series of
compatible systems is announced. Compatibility features are prized by many
computer users. ^ company may require a small computer for some tasks and
a much larger one for others : the company would naturally want to use pro-
grams designed for the small computer in the larger .system's operations as
well. Thus, Standard Oil Company of California has adopted the following
policy for computer purchases :

All systems used will be totally compatible, both upward and downward,
so that as the work loads cliange in given areas equipment can lie changed
to meet these needs.
Compatibility features are also important to a company planning its initial
investment in a computer. Many companies purchase a small machine and "grow""
into a larger computer ; significant savings can be achieved if programs developed
for use with the initial computer will not have to be scrapped when a larger
one takes its place.

Finally, IBM was aware that its sales would not be seriously injured if
customers learned later that IBM had bought technological development time
by marketing "paper machines". Once committed to an IBM computer, a pur-
chaser is unlikely to cancel his order even if the computer is not delivered as
promised. Indeed, we are informed that all IBM salesmen are trained to bend
every effort, as soon as the customer places his order, to get the customer to start
spending money ; the purchaser's computer operators are sent to IBM's trai)iiiig
school ; IBM instri;ctors begin to train tlie purchaser's programmers ; the pur-
cha.ser's executives are invited to IBM's executive '^rainin-:;- school • the jmrchaser
is encouraged to begin to design the room or l)uilding to hold the IBM computer
when it is delivered. Most important, IBM encourages the purchaser to begin
to develop programs. All of these preliminary steps are expensive ; it is common
in the industry to estimate preliminary expenses as equal to the cost of the
.system. And, of course, the larger the computer, the greater is the initial invest-
ment. Under these circumstances, many customers would and do hesitate to
cancel their contract with IBM a year or tv/o after the initial order simply
because IBM has slipped delivery — possibly wasting their initial investment
and possibly incurring additional delays. Note, for example, the following story
reported in DATAMATION (December. 1965) :

IBM will pay LA city an estimated $20-30K penalty for a three-month
delivery delay of third level software on the 360-30F due in mid-December.
With $300K invested, the City didn't feel justified in reopening bids to other
vendors, will accept .$17K plus $100 a day until everything is in, with no
rent until it is all working.
IBM's financial condition is s\ich that it need not be concerned about such
liquidated damages — so long as competitors are being deprived of sales.

It is appropriate to point out that delaying competitors' sales would have
incidental advantages for IBM. Any producer with limited capital is at a con-
siderable disadvantage in the computer industry. Large sums must be spent on
research and development : production facilities and each unit produced are
expen.sive. Extensive bank credit is a necessity^ — but is available only if the
borrower's product, as measured by his success in marketing it. warrants the
risk. If IBM can delay a competitor's sales, the competitor's financial position
may be significantly injured.

2. IBM's Methods

(a) Announcement of the 860 Series. The 360 Series was designed :nd devel-
oped in an atmosphere of haste and concern that customers might turn to
competitors' products before 360 computers could be marketed. We believe that
the Series was announced in the same almosphere and with the purpose of
delaying sales by IBM's competitors.

For some time prior to IBM's official announcement of the 360 Series. IBM
salesmen advised customers, in situations where tliey encountered difficulty in .



5553

selling an existing computer and were fearful of a competitor breakthrough,
that the customer should forestall any immediate procurement because announce-
ment of a new "third generation" of computers was imminent. The "rumors"
of the 360 Series' imminence were obviously inaccurate; indeed, the Series was
publicly announced before models in the Series had undergone normal company
developmental testing. We are reliably informed that, until 1963, IBM subjected
new products to exhaustive testing at three or four stages of their development,
lu what was called the ''A" test, a prototype of the product was constructed
ro determine whether it met specifications ; the "B" test followed re-design
and correction of the prototype. If the product passed the "B" test, a manufac-
turing run of the product was arranged ; units from this run were then sub-
jected to the ''C" test. IBM normally delayed announcement of a new product
until it had pa.ssed at least the "B" test. The 360 Series, however, was treated
atypically by IBM. Although some standard components of the lower performance
computers in the 360 Series had passed the "B" test by April 7, 1964— when the
prevalent trade •rumors" were for the first time' iml)licly confirmed by 1B:M with
great fanfare — the basic overall system had not been tested. Moreover. IBM
had originally planned to announce individual models of the 360 Series one at
a time, "as they reached the appropriate developmental stage. It was obvious,
hr.wever, that announcement of a complete series of compatible systems would
maximize the delay of customer orders. After substantial disagreement within
IBM, a decision was made to announce all System 360 models on April 7, 1964 ;
in fact, one IBM executive was demoted, we are informed, because he disagreed
with this decision.

IBM's behavior after the April 7, 1964 announcement provides evidence that
the 360 Series had not been completely developed when originally announced
and that IBM was anxious to hide this fact from its customers. One device
adopted by the company was frequent announcements of new models and with-
drawal of original, and even subsequently announced, models in the Series. Often,
sui>posedly new models have been near-copies of unperfected old models with
slightly modified performance and/or price specifications. These announcements
are described in greater detail below. It sufficies here to quote a WALL STREET
JOURNAL reporter's reaction to IBM's press release of April 26, 196-5, which
announced the addition of three new models for the 360 Series and the cancella-
tion of five previous models (two of these had been announced only the month
before) : "Was its announcement on April 7, 1964 premature and made for com-
petitive reasons before development was completed?" (The WALL STREET
JOURNAL, Monday, April 26, 196.5, page 4, column 1 ; the entire article is at-
tached as Appendix 7.) IBM's public response to the widespread speculation, of
course, was an emphatic denial.

Another device IBM adopted to hide from its customers the fact that the 360
Series was announced prior to completion of its technological development was
frequent modification of delivery schedules. Although the original delivery sched-
ules provided for lengthy delays, in December, 1964, IBM announced accelerated
delivery schedules "made possible by accelerated production at IBM manufactur-
ing facilities in response to customer demand". (BUSINESS AUTOMATION
NEWS REPORT, December 7, 1964). Nine months later, however, the company
was singing a quite different tune. In October, 1965, the Chairman of IBM's
Board of Directors announced that delivery schedules were being postponed for
from two to four months, this time attributing the delay to "problems in build-
ing up the rate of production as rapidly as necessary to meet the unprecedented
demand for the new equipment." (WALL STREET JOURNAL, October 27, 1965,
page 32, column 1).

In fact, we believe that IBM never had the ability, or the intention, to abide
by its original delivery schedules. We are informed that, prior to the 360 Series,
IBM scheduled manufacturing and delivery of computers in relation to orders
received. Delivery dates were widely disseminated internally; every salesman
knew almost instantaneously the probable delivery date of any system he was
trying to sell. In addition, IBM had meticulously programmed its production con-
trol. Illustrative of IBM's ability to predict delivery prior to the 360 Series is
the experience at the company's Rochester, Minnesota plant in 1962 ; approxi-
mately 11,000 units were shipped from the plant, and only one was shipped after
the specified delivery date — that one was late by only six hours !

The evidence available compels the inference that IBM decided to announce
a nonexistent series of computers in order to prevent, or at least delay, sales by
its competitors— all to retain its dominant position while overcoming its market-
ing and technological handicaps.



5554

(b) Marketing the 360 Series. The 360 Series, as orieinallv announced April 7,
1964. included the following models : 360/30, 360/40, 360/50, 360/60, 360/62 and
360/70 (higher postscript numbers in the Series denote machines of greater
computing power). As we have indicated, IBM's subsequent behavior was designed
to mask the fact that the Series contained many "paper machines" as of the
announcement date. In addition, the companys marketing tactics were designed
to cause confusion among customers and further delay of competitors' sales —
giving IBM more time to overcome technological handicaps. These tactics were
successful.

We have already mentioned two of IBM's favored techniques: (1) frequent
changes in delivery schedules, and (2) continual announcements of new models
and cancellation of previous models even after orders have been taken for them.
We have attached as Appendix 8 a chart indicating the model number changes
and their dates, and a detailed description by Auerbach Associates, an independ-
ent analyst of computer systems, of changes in the 360 Series followiiig April 7.
1964. There is every reason to believe that many of the model numbers introduced
after the original announcement were "paper machines", some of which IBM
never intended to produce. Many of the new models, for example, were supposedly
time-sharing systems, which would require design modifications In the basic 360
main frame. It is believed that at the time these were announced, their develop-
ment had not even progressed as far as that of the basic 360 System at the time
it was announced.

In individual competitive situations as well, IBM marketed "paper machines"
to delay and ultimately prevent sales by its competitors. In August, 1965, for
example, IBM was trying to sell its 360/30 model to Aluminum Corporation of
America in competition with Control Data's 3100 System (which was somevvhat
higher priced). When Alcoa requested comparative "benchmark" tests of the two
systems, IBM's salesmen indicated that the test would have to be run on the 360/
40, a more expensive system. Control Data ran the test successfully on the 3100.
The IBM salesman then recommended a 360/44. which had not yet been an-
nounced, claiming that the system was three times faster than the 360/50, a more
expensive system, but with a lease price between the 360/30 and the 360/40.
Public announcement of the 360/44 was not made until a week later; at the time
of the negotiations with Alcoa. IBM's Pittsburgh office had no knowledge of the
360/44 System — personnel from the Cleveland oiffice were sent to make the presen-
tation. Alcoa ordered the 360/44 to be delivered in January. 1967. and agreed to
accept the 360/30 as an interim system. Thus, IBM was able to prevent Control
JData's sale of an existing system with better performance and a higher perform-
ance/price ratio ill an IBM's announced system and gain at least a year and a
half of additional development time — simply by creating a new model number.

The purposes underlying IBM's marketing strategy are most clearly seen in
connection with the extra-large system market — a market in which IBM most
feared losing orders because of Control Data's competition. The original 360
Series did not contain any extra-large computers ; and Mr. Watson admitted at a
shareholders' meeting in April. 1964 — following announcement of the 360 Series —
that IBM had no computer as powerful as Control Data's already-announced
and soon to be delivered 6600. In August. 1964. IBM described to the trade the
360/90 (or /92 or /91 or /94 or /95— all of which numbers IBM salesmen have



Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 78 of 140)