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 17 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 17 of 140)
Font size
QR-code for this ebook


pete, they hold a share equal to what they hold in the United States,
and in some cases more.



5028

The foreign market for computers, dominated by U.S. suppliers,
has for several years been the fastest growing segment, and it is ex-
pected to remain so for the foreseeable future.

I want to now turn to something that I have called the total service
concept. I think you gentlemen are looking at some capital goods in-
dustries. Many cai^ital goods industries generate a very close relation-
ship between suppliers and users of equipment, but none of them
seem to have generated any greater degree of cooperation between the
supplier and the user than that which exists in the computer field.

Much of this pattern was adopted from the punched-card business
in which companies such as IBM and Remington Rand had developed
a highly service-oriented combination of products and services long
before the advent of computers.

The computer with its need for a high degree of specific knowledge
and expertise in installation, maintenance and programing seems to
have heightened the desire of the full service concept of user support
by suppliers.

As the sophistication, particularly the sophistication of computer
programing, has increased, this desire has become stronger until today
literally everyone in the computer business is providing a combination
of equipment and services to its users.

This close cooperation develops a following, or a brand associa-
tion, in the specific user for what he believes is the most compatible and
useful combination of products anc^ services suitable to his own use.

That brand association also serves as a pow rful stimulus to the pro-
vision of customer satisfaction by the various suppliers.

In recent years the market for the larger computers in this class has
been primarily a replacement and upgrade market rather than the
forming of new associations with first -time users of computers. Per-
haps as much as 90 percent of the dollar value of the ])roduct delivered
in recent years in the United States has either replaced or augmented
computers already in use. Satisfied present customers, therefore, rep-
resent the largest market for additional revenue.

In addition to product characteristics and services provided, a great
deal of the attitude of users is based on a kind of charisma about the
supplier. His intent to remain in the business long term ; the direction
of his product and serWce orientation, and the apparent technical
capability of his organization as determined by day-to-day contacts
with his various sales, maintenance, and technical support personnel.

From the point of view of the supplier this charisma is extremely im-
portant. It represents the franchise on his present market share, and
it is an important contributor to his expectation of any increase in
market share.

I would like to then dwell for a moment on the ingredients of suc-
cess in this business for the supplier. Now I am looking at it from the
user's point of view. AAHiat does a supplier have to have in order to con-
vince me, as a user, that I ought to do business with him?

The closeness of relationship that I have mentioned here between
supplier and user and the tendency for many computer systems to be
leased rather than purchased places stringent demands on the orga-
nization which aims to succeed and continue to succeed in the com-
puter systems business.



5029

First, he must have the necessary ingredients — desirable product
characteristics, high product quality, and competitive prices ; they are
all a necessity.

Next, the charisma that I noted earlier must be satisfied. In order
to do these things the succesful supplier of computer systems must
engage in a very high level of research and development activity.

This activity must be applied not only to the equipment, but to soft-
ware, the complex computer programs that are provided by the sup-
plier to enable the user to get the most out of his system.

"While research and development work on the computer equipment
was tlie most important part of this activity for some time, more re-
cently research and development on software has become most impor-
tant to the user, since the potential performance of the computer
system has been improved to the point that the ability of the user
to make use of it is limited by software rather than by hardware
characteristics.

Today, to hold a franchise in the computer systems market, a sup-
plier must spend enormous sums of money and engage very sophisti-
cated professional talent to develop and train users in the use of his
software offerings.

Selections of computer suppliers by users are now often made on the
basis of software characteristics i-ather than on equipment character-
istics.

Once a supplier has fulfilled these requirements he needs large
amounts of money to finance the equipment that is leased from him by
his users and he needs specialized financial skills to understand the
financial implications of a mixed sale-lease business. I think the pre-
vious speaker made that quite clear to you.

And last, he needs a large installed base of systems in the hands of
customers to support the extensive maintenance and technical support
services demanded by the total service concept.

Attempts have been made to break up this package of services tra-
ditionally provided by the computer system supplier, and some of
these attempts have been successful.

Today, after all, when peripheral devices are obtained from in-
dependent suppliers, you must make some kind of a maintenance ar-
rangement with that supplier or with an independent computer
maintenance company.

A growing amount of software, developed by firms which do not
supply computer systems, is being acquired by users, and independent
consultants and programing firms are employed to augment or re-
place the services that are offered by the computer systems supplier.

These are a couple of examples of penetration of this package, con-
cepts that are already in place. While these businesses have been mod-
erately successful on a very small scale they do represent a potential
source of competition for at least some portions of the total service
traditionally provided by the computer system suppliers.

The largest grouj) of users that we contact, however, still appears to
wish to retain a single point of responsibility for computer equipment,
software, and maintenance, to avoid the necessity to determine for
themselves whether equipment failure, faulty maintenance, or faulty
software has caused his svstem to fail.



5030

Further separation of these services from the supply of the prod-
uct — absent legislation — is a feasible course in many cases, but a con-
siderable economic or service advantage must be provided to convince
the large body of users that the resulting dilution of responsibility
for the total service is worthwhile.

Now, I would like to turn to aspects of competition as they exist
presently. By any measure it is clear that IBM holds the major share of
this market. That, obviously, raises the question, are the firms now in
the market — and I am speaking of computer system suppliers — a
strong enough force to operate as effective competition for IBM ?

At the moment it would appear that the computer system suppliers
are not very effective. After all, a market in which the second place
supplier has about 15 percent of the volume of the largest, hardly
qualifies even as an oligopoly, and the two firms who hold the second
position were helped to that market share by recent acquisitions of
other firms who were leaving the business.

More effective competition of a different sort seems to have been
provided by the plug-compatible peripheral equipment suppliers. I
think you understand by now what plug-compatible peripheral equip-
ment is and why it is there.

What may not have been said is that as computers have developed,
increasingly the investment in computers has swung in favor of periph-
eral equipment. I don't know of any large system user today that I
can think of who doesn't spend more money on jDeripheral equipment
than he does on his main frame.

Now, I am sure that somebody can cite an exception or two, but I
have even challenged a few of our user clients themselves to analyze
their own bills, and had them come back and say, "Hey, you knoAv, you
are right. I didn't think about it, but that's the way it is.''

But a number of companies have marketed this equipment, which is
directly interchangeable with IBM products and could replace the
IBM equivalent by simply unplugging the IBM product and plugging
in the competing product. And, of course, that is where the name
"plugcompatible" came from.

There is an important observation to make at this point, however.
Practically without exception these devices are substitutable in IBM
systems Avithout changing the software. That is one of the reasons that
they have had the kind of user success that they have had.

You simply plug in memory, you plug in tape, you plug in disk.
These equipments are designed in such a way that "they are not only
able to be plugged into IBM sockets, but they also operate under the
umbrella of IBM software, and nobody has to make any changes.
Clients of ours have, in fact, removed IBM tapes, put in independent
peripheral supplier's equipment, not liked them, taken them out, and
put back in IBM. There are all sorts of combinations of that sort going
on, because it operates entirely under the umbrella of the software that
is supplied by IBM. That is the sacrosanct part, nobody touches that.
Anything you can do without modifying the software is fine. If it is
economic, I Avill try it. But when you get to modifying with the soft-
ware it becomes a very different problem.

While these companies are relatively small by standards of the prin-
cipal computer systems competition, they concentrated on the market
for only one or a very few specific peripheral device products, and they



5031

have been a more effective competitive force to IBM in these smaller
market segments than have an}- one of the principal computer system
siippliei-s in the overall systems market.

But the efforts of these competitors have resulted in IBM's reducing
prices, improving equipment performance for the same price, and
develoi^ing more advantageous leasing terms for equipment.

Clearly, then, competition, when it becomes noticeable, influences
the larger supplier to offer more advantageous terms to users. So I
believe that more effective competition by the competing computer
systems suppliers, so that at least two of tliem have 20 or 30 percent
of the market apiece, would result in better value for the user, and it
should be encouraged for the long-term benefit of the industry.

But the largest supplier in this business, in my observation, holds
his market share through the free choice of the computer users. These
users must be persuaded that a new supplier will give them a better
deal.

The sole restraint on expansion in the industry api^ears to be the
cost and risk of obtaining a more significant market sliare.

For any of the present competitors to obtain a more meaningful
market share will require a minimum investment of literally several
billions of dollars, much of it at high risk, and a return to unprofitable
operation for several years.

In addition, it will require outstanding equipment and software
and incredible evidence that he is committed to build and maintain
that market share by matching, or hopefully exceeding, the service
and technical support that is provided by IBM.

Such a feat is not impossible, but at least two of the largest indus-
trial firms in the country— I am referring to GE and RCA — have
found it too much to contemplate within the last couple of years, and
any others who try it have certainly got to find forebearing inves-
tors who are willing to take the long view. Despite this difficulty, I
think what we have heard about the expansion of the industry would
indicate that the potential reward is enormous, and I can only hope
that to improve the value to the user, several organizations will see
the reward and will somehow find a way to make the necessary effort.

There are a few observations that I have on the subject of trends
in computer use. They are a little different from the ones that were
made earlier.

First, we have heard about networking computers and things of
that sort, but I think we ought to recognize that two conflicting trends
are really apparent in the market for systems of this size.

The first and the oldest one is a trend toward centralizing the data
processing resources of an organization into a large center designed
to perform data processing work on an efficient computer system which
is designed to do many joh« simultaneously.

That is about where we are. That is" where these larger systems
come from today. In connection Avith this trend, input data prepara-
tion — keypunching, if you will — is being removed as a function of
the data processing center and is more and more being performed by
the data processing center's users — and here I mean the accounting
department, the manufacturing department, or the branches of an
agency in government — and coupled to the central computer site by
communications lines.



5032

This centralization was undertaken in the belief that large com-
puter centers exhibited a considerable increasing efficiency with size.

While there has always been an increasing efficiency with the size
of computers, the margin over small computers has shrunk in recent
years as inexpensive computers, and particularly the inexpensive
peripheral devices associated with them, have become available.

This trend toward centralization of computer resources has run
counter to a basic management philosophy of many organizations
toward decentralization of authority and responsibility to obtain
more effectiveness in their organization.

The centralization of computer resource in opposition to their basic
organizational philosophy was justified as a necessary step to improve
efficiency.

With the prospect of improved efficiency in smaller computers, how-
ever, many top managements are responding to the requests of sub-
sidiary managers to operate their own computer systems in direct
opposition to the computer centralization trend still evident.

Nobody expects the larger computer centers to disappear, but in
many organizations we expect to see a leveling off in the growth rate of
the larger centers and the displacement of some of their functions
by smaller, less complicated computers coupled together in a network
with each other and with the large central computers.

The distribution of input data preparation to the users and the use
of networks of smaller com))uters both will increase the use of electrical
communications to transmit computer data.

Data communications, probably the fastest growing segment of the
computer field, has already raised serious public policy issues not yet
completely resolved by the Federal Communications Commission and
the State communications regulatory bodies.

The problem is that the extensive telephone communications network
in place in the United States and most of the rest of the world can bo
used to transmit computer data, but it was designed for an entirely
different purpose, and it is quite inefficient and expensive for data
communications use.

Despite its high cost, however, computer users have already found
it effective for many years to couple computers to communication lines
to perform their work.

If data communications were less expensive it would be used much
more extensively. This had led the Federal Communications Commis-
sion to conclude that competition rather than controlled monopoly
should be encouraged to drive down the cost and improve performance
of data communications.

The results so far have been encouraging. Data communications
costs seem headed for a 50-percent reduction quickly, and more later.

Even at present communication costs, as many as 75 percent of the
larger computer systems now being delivered are equipped to use
some kind of data communication facility.

The expected availability of inexpensive means of data communi-
cation has led to this conception of computer networking that you
have heard so much about here today.

Networks which o])eratc this way arc already in use in law enforce-
ment and credit validation, to name a couple. The national banking
svstem is headed toward it in order to handle monev transactions.



5033

So far, the revenue to communications carriers from this source,
that is from computer users to communications carriers, is only about
3 — or, I think, IDC says 4 — percent of the cost of operating comput-
ers. It is probably a roughly equal proportion of the income of the
communication carriers. So far, it is not large.

The impact of data communication used in conjunction with com-
puters is growing at a very rapid rate, however. It is my estimate that
it will at least triple in the next 5 years, which would require that it
grow about twice as fast as the total market for computers.

The volumes of data transmission that this growth will generate
will be adequate to support specialized data communication services
to most locations in the country.

We can expect to see the development of these communication
services proceed at a rapid pace to match the needs of computer
users.

I beg the question of who will develop them, but that is a question
for the FCC. One of the interesting features of this new form of dis-
tributed computing is the real possibility, however — already demon-
strated — that large numbers of computers built by various suppliers
may be connected together through data communications networks.

This development should sen'e to permit a wider choice by users
from among the several suppliei-s with less difficulty than they now
encounter in changing suppliers for large computers, and thus provide
a spur to competition in this new and growing segment of the business.

As a closing remark, it is obvious from earlier speakers that com-
puter technology continues to improve the price performance of com-
puting equipment, and we expect that trend to continue. I would like
to underscore, however, again, that the most costly part of computer
use is the people. There are people that prepare programs, people that
operate equipment, and people who prepare input data.

You have seen some evidence of the work that has gone on in the
data input field. The game there is to eliminate input data preparation
as a human function and take input data from the transaction.

There are other developments afoot that will reduce the work of
the operator of computers.

But in the expensive activity of preparing computer programs, the
key to improved effectiveness and efficiency is the development of im-
proved software aids for computer programers.

Really, the objective in programing aids is to eliminate the pro-
gramer entirely. There is no fundamental reason, even now, why the
person who is able to describe the problem to be solved and to postu-
late the procedure to be used in solving it — and I refer here to the
systems analyst or the user of the computer service — should have to
learn a foreign language which is undei*stood by the computer.

The user should be able to describe the problem and the procedure
in language familiar to him and let the computer make the transla-
tion. In relatively crude form, I think you have heard already today,
such aids presently exist. But ultimately, we are looking toward the
programer — I shouldn't call liim the programer. but the man with the
problem — being able to connnunicate the problem to the computer lit-
erally by speaking to it rather than by putting instructions in writing.

Work to reach this objective is proceeding in a number of locations.
I wouldn't want to say that a clearcut answer is at hand. Also, looking



5034

back at the rate at which developments of this sort have caught hold, I
guess we are talking about the mid-1980's.

But the fact that it is even feasible. to consider such a solution, I
think, serves to underscore the fact that computer technology and ap-
plications are only now beginning to be developed, and there is still
a very long way to go before we find out where the end is.

Senator Hart, The mid-1980"s sounds pretty close for almost an
unbelievable feat.

]\Ir. Wallace. In 1958 I was employed at RCA, which was then
in the computer business, and I was trying to convince our manage-
ment that if we didn't have good, solid data commmiications products
in 2 years we were really going to be out of the market. It is now 1974,
and you see they are now coming. So, these things do take time.

Senator Hart. The staff has developed a number of questions.

Mr. Nash. Toward the end of your statement, Mr. Wallace you
talk about trends in computer use. You say that you expect to see a
leveling off in the growth rate of large centers and displacement of
their functions by smaller, less complicated computers coupled to-
gether in a network with larger central computers.

You referred to that again a moment ago. Given that prognostica-
tion of the future, we would be interested in your reaction — if you
have one — ^to the significance of IBM's acquisition of a significant
share of CML, the domestic satellite subsidiary of Comsat.

Mr. Wallace. Well, I liaven't really thought much about it. It is a
fairly new development. It certainly does show promise to raise other
questions concerning integration of communications and computers,
which is a subject which the FCC spent some years trying to untangle
and has only recently been able to even make utterances on.

IBM's anounced jjolicy was that it was going to hold this property
long enough to get it going and then withdraw, as I recall it.

The simultaneous ownei-ship of a communications common carrier
and a computer company of the magnitude of IBM would raise seri-
ous questions that Avould be clearly inconsistent with the policy pres-
ently in force by the FCC to try to separate these controlled and uncon-
trolled segments of quasi-communications activities.

I think you are probably familiar with their distinction. That is a
long Avay from being finally adjudicated yet. This seems to fly directly
into the face of that, if they are, in fact, integrated into networks.

Mr. Nash. From an industry structure, if technology is moving
toward commimication orientation with respect to data transmission
and networks of computers, do you view control of a communications
netAvork as a requii'ement to peri)ctuate, say, present market shares
by computer companies ?

Mr. Wallace. No, I don^t thiuk so. Wliile we are talking about
these things like they are in the future, there are in fact a number of
these networks already in operation. Practically every major time-
sharing company has a network of this kind of its oavu. General Elec-
tric has one that is worldwide, for example.

I came from the communications industry, and I am in favor of a
clean interface, as they say, between communications facilities and the
communicants. I think that it has been achieved. It is necessary, really,
ill order to intercommunicate, for example, between different com-
puters, that you get some kind of a commonality between them in the



5035

communications mediimi. But I don't think it is necessary, nor do I
think it is desirable, for computrr companies to acquire control over
the communications facilities that are used to interconnect these
networks.

Mr. Nash. The other point I would like to touch on relates to your
comments respectinc; problems confronting the computer companies
trving to increase their market share. I think you put two values to
that, the capital requirements due to the leasing nature of the business,
and second, the total service concept, putting in charisma or brand
loyalty.

Thinking out loud, assuming there would be adopted a national pol-
icy of promoting more competition in the computer industr^^, I would
be interested in your reaction to the feasibility of some restructuring
along the following lines.

One possibility I heard about relates to separating out of tlie total
service concept, "so we have main-frame manufacturers barred from
sendcing or maintaining the eauipment. It has been argued- this would
sort of break the lock of dependence on the manufacturei-s and better
equipment would then be purchased by the user.

Would such an approach, in your opinion, be feasible?



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 17 of 140)