James W Driscoll.

Office automation : the dynamics of a technological boondoggle online

. (page 1 of 2)
Online LibraryJames W DriscollOffice automation : the dynamics of a technological boondoggle → online text (page 1 of 2)
Font size
QR-code for this ebook



James W. Driscoll*

WP 1126.80

May 1980






James W. Driscoll*

WP 1126.80 May 1980

*The author is an Assistant Professor in the Industrial Relations
Section of the Sloan School of Management at M.I.T. This paper was
presented at the International Office Automation Symposium held at
Stanford University in March 1980.


For anyone who has taken the time to read these words given the title
of the paper, the proposition that office automation will dramatically
reshape American society within twenty years is probably assumed.
Likewise, few would disagree with the assertion that the direction of
office automation is largely driven by technological developments rather
than responding to specific organizational or societal needs. (While most
users are currently interested in office automation to reduce labor costs
and improve productivity, they are seeking to take advantage of
technological developments rather than guiding the form of automation.)
The second fact was recently confirmed by a recent study here at M.I.T. of
the current office automation efforts in nine large users (Driscoll, Sirbu,
Alloway, Hammer, Harper, and Khalil 1980).

Less well-understood is presence of choice within any technology.
There is no such thing as a technological inevitability. David Noble, a
colleague here at M.I.T., has made that argument persuasively, and
documented his contention in a series of compelling studies. More
recently, Wendy Mela and Richard Walton at Harvard Business School have

supported the presence of technological choice in the specific area of

concern here, namely advanced office technology.

If the direction of technological development is not a given, and if

office automation is certainly being driven by technological developments,

then who is making the choices among technological alternatives? Noble

believes that technology is a means of social control in the conflict

between classes within the society. The ruling (capitalist) class selects

specific technologies to maximize its control over the working class. My

own analysis explores some additional explanations. The purpose of this

paper is to analyze the current technological path of office automation,

and to demonstrate the possibility of an alternative, more humanistic path.

1 2fl5fc>£>

Then, I will explore the reasons for our current path and suggest actions

for interested groups in the society who prefer my proposed humanistic


The Future of Office Automation

Because of my graduate work with Larry Williams and Tom Lodahl at
Cornell in the early 70's, I have been studying office automation since the
early days of word processing. More recently, I have conducted my own
studies of electronic mail systems and the state of the office-automation
art in large users. Additionally, I have participated occasionally as
M.I.T.'s representative on the Office Automation Roundtable and I conduct a
weekly seminar here at M.I.T. on recent developments in office automation.
As my presence in this Stanford symposium suggests, I am one of the "usual
suspects" when somebody rounds up a crowd of experts to discuss office
automation .

Based on this reasonably informed perspective, I detest what I now
foresee as the future of office automation.

The smart money in the technological sweepstakes now rests on two
entries: decision-support systems and "true" automation . Again and again,
experts tout the advantages of these "totally integrated" approaches to
office automation and contrast them with the incremental advantages of word
processing. But, what is the inevitable result of these particular
technologies which are currently being "chosen"?

Decision-support systems , as the first wave of the future, represent
little more than the extension of the use of computers to managers (a
clarification suggested by Kenan Saheen of the University of
Massachusetts). A decision-support system focuses on the key decisions
made by a manager or professional and tailors a computer-based system to
support those decisions. Its components may vary, but typically they


include a regression-based forecasting tool for analyzing alternative
decisions and direct access to organizational and outside data bases for
input to the analysis. A well-known system developed here at M.I.T. by
John Little and his colleagues, Brandaid, helps, as I understand it, a
product manager select a marketing strategy for a given product.

Decision-support systems emphasize computer support for a few key
managers in an organization and help them to make decisions as individuals .
The complex communications links currently being developed for boards of
directors can also be conceptualized as decision-support systems. Of
course, such systems neglect current research on how decisions are actually
made in organizations. Decisions, most scholars now agree, are the outcome
of a complex social and political process involving many people and a
variety of special interests. (Graham Allison's, The Essence of Decision
(1971) best demonstrates this position). To the extent that decision-
support systems are an attempt to rationalize this process, their implicit
model of an organization emphasizes a few individuals, at high levels in
organizations, making decisions with the support of technological wizardry
to direct the actions of all the other people in the organization.

"True" office automation , the second wave of the future, arises from a
distinction popularized by Michael Zisman, yet another M.I.T. affiliate,
based on his work with Henry Morgan at Wharton (Zisman, 1978). Zisman
correctly criticizes word processing and electronic mail as mere
"mechanization" or the replacement of human labor with machine power.
Automation, by contrast, is the exercise of discretion by machines. The
computer system in true office automation controls office activities and
exercises judgement in performing tasks according to its programmed logic.
Research currently underway here at M.I.T. by Michael Hammer, Jay Kunin,
Sandy Schoiket and others at the Laboratory for Computer Sciences (along


with other projects across the country) are now attempting to develop
computer languages to supply the logic for such "true" automation. They
are attempting to discover all the structured work in offices which could
be more efficiently performed by a computer system. Zisman's program for
running the manuscript-review process for a journal editor is the best
known example of such automation, but the current efforts are driving for
the computerization of all structured tasks which possess some
generalizability across offices.

What is the implication of such "true" office automation? The office
of the future would maximize on machine efficiency by using the computer to
gobble up the structured tasks in any office and leave people in only two
roles: bosses and garbage collectors . The boss decides what tasks must be
done (perhaps with the help of a decision-support system) and asks the
systems analyst to prepare the program. The rest of the workforce picks up
the garbage which is left over at the edges of the programmed tasks (Marvin
Sirbu of the Center for Policy Alternatives at M.I.T. has elaborated this
point in a recent paper) . Such leftovers have no internal coherence since
their sole determining characteristic is that the machine couldn't do them.
They do not form an integrated, purposeful whole which would engage the
interest and attention of a human being. The only human control in the
system resides at the top of the organization in the systems analyst or
programmer and whatever collective action the lower level people can take
to sabotage the system by letting the garbage pile up.

Figure 1 portrays the "office of the future" resulting from the
current technological path. My assumption is that you will either like or
dislike the picture dependina on whether your present position makes it
likely that you will be a boss/ systems analyst or a garbage collector.




D.S.S. D.S.S. D.S.S.








An Aside on Word Processing and Electronic Mail

The reader may quibble at my neglect of the two most popular current
applications in office automation: word processing and electronic mail. I
omit them because my intent is to project the future of the office given
the logic of current technological development.

By the standard of future importance, word processing is widely
considered irrelevant. While the largest sales volume is currently in
stand-alone word processing systems to support secretaries, every single
expert in the field (and indeed the current massive advertising campaigns
from the vendors themselves) emphasizes the importance of integrated
computer-based systems to support office workers. Text editing and
retyping are downplayed as only one, minor feature of such a system.

Likewise electronic mail within organizations simply substitutes
computer systems for existing communications media such as the telephone.
To the extent electronic mail incorporates control decisions about storing
messages, forwarding them, automatic addressing and other functions, it is
evolving into precisely the "true" automation envisioned by Zisman. While
the mere mechanization of some office tasks such as recording may save more
time and money than current word processing systems, they do not represent
the automation of the office which has excited so many of us.

With all that said, the evolution of word processing and electronic
mail foreshadow precisely the nightmarish vision described in Figure 1.
The logic of word processing has always been specialization and
centralization. While the decline in product cost has made decentralized
systems feasible, the vast majority of vendors and users talk about how to
get as much typing as possible loaded onto a word-processing machine in
order to reduce the number of secretaries in an office. This logic was


clear in the early days of large word processing centers, but remains today
despite the pioneering efforts of Lodahl , Williams and Williams to point
out the ineff iciences of such specialized systems (1979).

The consequence of such specialization in word processing has been to
increase the separation between boss and secretary in the office and to
create a new breed of even more menial office workers. Little improvement
in the jobs of non-word-processing secretaries has resulted, despite the
advertising claims of the vendors about career paths for women.

This separation of secretary from boss will dominate future systems as
can be seen from current discussions about inputting text to automated
systems. One way or another, product designers are seeking ways to keep
the boss from having to type. Menues of preprogrammed commands are
presented requiring only a single keystroke for action, or an electronic
"mouse" is moved across a pad to convey information, or the user simply

touches the screen. In the most dramatic example, developed by Richard
Bolt and his colleagues here in M.I.T.'s Architecture group, the user sits
in an easy chair with both hands on pressure-sensitive armrests allowing
the user to "zoom" across a wall-sized projection of a desk top by pressing
down with either hand.

The standard form of input for text from such high-status users, be
they hunt-and-peckers, mice, pointers, or zoomers, is by dictation to a

Likewise, the separation of high-skill and low-skill workers is
increased by the more advanced versions of electronic mail. The major
applications outside the research community, where we found the most widely
used systems in our recent study (Bellinger, 1980), emphasize multi-media
communications among key decision makers or in support of the board of

directors. Electronic mail does not appear likely to emerge as a
substitute for the telephone for all office workers, rather it will only
support key managers and professionals.

While these trends can be justified based on ease of user acceptance
and the current cost of hardware, the leading edge of users and vendors in
both word processing and electronic mail would clearly create an
organization with two distinct social classes as portrayed in Figure 1.

Before exploring the reasons for that future, a few words are in order
to appraise its value.
A Nightmare of Cost Ineffectiveness

An organizational innovation can always be appraised from two
perspectives: the goals of the organization and the interests of the
organization's members. Despite (or perhaps because of) their confusing
connotations, the first perspective is usually called "rationality" or
"economics" and the second "politics".

A "rational" perspective . There is little reason to believe the
proposed office of the future will either save money or advance
organizational objectives. A recent review of word processing in the
federal government painted a negative picture of the economic impact of
such systems. More generally, our own recent analysis of some forty
studies conducted by large office automation users revealed jto single post-
implementation, economic evaluation of an office-automation system
(Driscoll et al . , 1980). Despite the heavy emphasis that such users gave
to cost savings in proposing new systems, they had never gone back to
validate their claims.

Some recent attempts to cost justify office-automation systems are
embarrassing from both a methodological and a managerial perspective.

Frequently the analyst will ask people on a questionnaire how much time
they have saved from a new system and then multiply their response by
current salary levels to estimate savings. A competent researcher would
ask at least for a second, corroborating source of information and some
comparison of time spent against a control group. A practical manager
would want to know whether those projections ever turned up as hard dollars
in a budget account which he or she could spend on something else. "Soft
dollars" are viewed skeptically.

I have suggested the reasons for this disappointing economic
performance of the current trend in office-automation systems in an earlier
paper (Driscoll, 1979). Most generally, these systems neglect the
interdependent nature of office work as the product of many people and they
simultaneously fail to provide any motivation for most people to work any
smarter or harder in pursuit of organizational objectives. For example,
decision-making is a small (and some would argue insignificant) part of a
manager's job (Mintzberg, 1974). Therefore the potential impact of
decision-support systems is limited. Their likely impact is further eroded
by their neglect of the social and political component of decisions.
Similarly, "true" office automation promises to decrease the motivation of
office workers. Their motivation springs in large part from the nature of
the work itself as well as from their social contacts. An emphasis on
maximizing machine efficiency, specialization, and centralization destroys
these two mainsprings of worker motivation.

A political perspective . The social consequences of the current
technological path are, if anything, even more negative than the economics.
As Figure ] illustrates, office automation is likely to increase the
distance between people at the top and the bottom of organizations. The

few decision makers and systems analysts will command more in salaries,
benefits, and differential treatment than do current managers. By
contrast, the lower-level workers will find much of their work less
interesting, have less chance for promotion (since the jump to decision
maker is vastly more difficult), and have less power to demand reasonable
salaries since many of their fragmented tasks can be performed by a newly
hired worker with little or no training. Unless the reader advocates the
overthrow of the current system of government in the United States by
violent means, such a picture is distressing. It increases the likelihood
of a revolution, but it contradicts our values of egual opportunity and
individual mobility in the economic realm.
Alternative Technological Paths to the Office of the Future:

My explanation for the likely evolution of office automation takes two
parts. First, a coherence and an inertia to the current trend that
provides it with stability. Second, a series of external causes first
established and now maintains the direction. I will address these two
points in order .

Technological path is the term I have adopted to capture the notion of


internal structure. A path is not a chance stroll through the forest of
technological innovation. One step follows another because of the contour
of the ground, the color of the soil, and markers along the way. Even so,
there seems a certain direction, inevitability, and sometimes human
leadership that characterizes developments in office automation.

Systems analytic is the label I use to describe the current trend in
office automation. Table 1 summarizes the features of the current path and
contrasts it with a largely hypothetical alternative which I have labeled
"humanistic." There is little original to these distinctions. What


deserves attention is the apparent clustering in the present technological
path of so many wrong answers to the major guestions about organizational
behavior. A series of narrow, short-term perspectives characterizes the
current approach.

For example, McGregor (1960) popularized the difference between
optimistic and pessimistic assumptions about human nature held by managers.
Successful managers tended to hold optimistic views, and his term for that
constellation of assumptions has become rooted in the management literature
as "Theory Y". By contrast, less successful managers more often hold
pessimistic views of their fellow beings, characterized by McGregor as
"Theory X". If there is an organizing framework for the present
technological path, it is "Theory X" assumptions about human nature.

Designers of current office automation systems in the systems analytic
path assume that people are lazy and cannot be trusted. Therefore, their
systems seek to reduce skill levels reguired by the organization and to
generate information by which operators can be controlled by higher level
managers. The line-counter on word processors is the most obvious example.

This emphasis on measurement and control leads inevitably to the
statement of objectives for the system in terms of efficiency, cost
savings, and personnel reductions. Also implicit in this negative world
view is the importance of the few motivated individuals at the top of the
organization in setting its direction. This elite needs information, both
to control the untrustworthy lot below them, and to enable the few to make
the major decisions which impact organizational goals.

Unfortunately, this obsession with efficiency, cost savings, and
reduction of inputs flies directly in the face of current wisdom about


managerial control. Some years ago Anthony and Dearden at Harvard Business

School made two critical distinctions within the management process.

First, they argued that managers ought to concern themselves with

effectiveness as well as efficiency . Effectiveness assesses progress

towards objectives while efficiency merely assesses the number of inputs

reguired for a given level of performance. Managers too often sacrifice

effectiveness in the pursuit of efficiency. This trend was apparent in our

recent evaluation of office automation efforts in large users.

The second distinction advanced by Anthony and Dearden separated
operational control from strategic planning. The vital role for top
management is setting strategic direction, not monitoring performance of
lower-level workers.

The current path of office automation thus contradicts not only the
best judgement of organizational psychologists about effective management,
but the dominant conception of management control as well:

The most devastating shortcoming of the current path from my
perspective, however, is its repressive political overtones. Clearly, the
systems currently under development cater to the approval of a few key
decision makers, since that is the current decision-making structure of
most large organizations. The systems designers pay little heed to the
needs and interests of the large number of lower-level participants whose
working lives will be affected by this technology.

In addition, current systems ignore the external effects of office
automation beyond the boundaries of the user organization. For example, at
several conferences I have heard suggestions that much of the non-automated
work, such as text input, might better be performed by part-time employees
in their homes or subcontracted to "service bureaus." At a time when labor


economists are raising the possibility that the large supply of such lower-
paid, temporary, dead-end jobs in our economy is the major cause of our
chronically high unemployment, such an external effect is unconscionable
for national employment policy.

Likewise there are possible negative impacts on physical health from
prolonged use of a cathode ray tube and unfortunate mental health
implications of low-skilled, high- turnover, meaningless work. However,
if reactions at conferences on office automation are any indication, these
health effects are among the few subjects guaranteed to induce boredom
among current vendors and users.

The humanistic path , in contrast to the systems-analytical
technological path, is marked by different initial assumptions about human
nature and leads to quite different office-automation systems. Since
workers are now assumed by systems designers in the humanistic path to have
the potential for self-motivation and control, the immediate purpose of an
automated system is to increase the flow of information to the system
operators in order to allow them to utilize and increase their skills and
knowledge. Decisions are spread as much as possible throughout the
organization rather than being concentrated at the top.

Wendy Mela and Richard Walton provided a delightful example of a
humanistic alternative from their own research at one of my recent
seminars. The designer of a product information system for a large
retailer of consumer goods assumed that the purpose of the new computer
system was to provide information on stock levels, advertising campaigns,
and the like to the national sales manager. That key decision maker could
then better deploy his sales force and advertising budget. However, the
local sales managers desired a system which provided them with the same


information so that these lower-level participants could make deployment
decisions locally. Clearly, the computer can serve either group. A
humanistic path would suggest providing the information to the local
manager on the assumption that he or she wanted to act in the company's
interest without the need for centralized control by the national sales

A humanistic path also includes a focus on organizational objectives
rather than inputs, since progress towards such goals is the focus of
feedback to operators. Such an orientation towards goals , rather than
inputs, implies and reinforces the humanistic path's attention to groups of
people rather than individuals. Goal orientation emphasizes groups because
objectives usually apply to some organizational unit rather than
individuals. The focus on individuals in the systems-analytic path results
only from the need assumed by the designer to control individual behavior,
not from an intrinsic need by top managers to know how people accomplish
their objectives.

And finally, a humanistic path gives explicit recognition to the
guality- of- work- life issues neglected by systems analysis because a wider
range of motives is attributed to individuals. People are assumed, in many
cases, to desire meaningful work, training, and the opportunity for
advancement. Therefore an effective system must provide such potential if
it is to increase productivity.

On balance, then, the humanistic path reflects current wisdom about


Online LibraryJames W DriscollOffice automation : the dynamics of a technological boondoggle → online text (page 1 of 2)