Robert M Alloway.

User managers' systems needs : it appears that tomorrow is here today, we need to change the rules of the game and play catch-up ball online

. (page 1 of 2)
Online LibraryRobert M AllowayUser managers' systems needs : it appears that tomorrow is here today, we need to change the rules of the game and play catch-up ball → online text (page 1 of 2)
Font size
QR-code for this ebook

Center for Information Systems Research

Massachusetts institute of Technology

Alfred P, Sloan School of Management

50 Memorial Drive

Cambridge, Massachusetts, 02139

617 253-1000


It appears that tomorrow is here today.

We need to change the rules of the game

and play catch-up ball.

Robert M. Alloway

May, 1980

CISR No. 56
Sloan WP No. 1125-80


User Managers' Systems Needs *
Managers need information not computer-based information systems. The
latter is a means to the former end. Managers need flexible access to
relevant data and the ability to analyze that data. Computer systems are
intended to provide the means to make this possible, economic, and easy for

This paper assesses user managers' systems needs, the means, with an eye
toward their appropriateness for various managerial ends. We have asked and
answered the questions, how many systems do users currently have, how
appropriate are those systems, and how many new systems do users want?
Systems were classified by type into monitor, exception, inquiry, and analysis.
The results are dramatic. The differences in managerial appropriateness
of systems types has already caused a significant shift in users' demand mix
by systems type. This presents serious management challenges to both DP and
user departments. Moreover, the level of user demand for new systems, for
each type and collectively, is simply overwhelming. This increases the
managerial challenge; necessitating improved processes and criteria for
prioritization of systems needs.

As part of a larger research project 114 user and DP managers in six
industrial firms completed an extensive questionnaire. A stratified sample
of senior, middle, and junior managers was selected from the manufacturing,
finance, and DP departments. Exhibits 1 and 2 provide some basic information
about the firms and respondents studied.

*I wish to thank the anonymous companies and managers who participated in this
research, Christine Bullen for her management of the data gathering process,
Jerome Nolte for his statistical analysis support, and the Center for
Information Systems Research of M.I.T. for partial funding.

"User N'eeds Survey: Preliminary Report", Robert M. Alloway, et al , M.I.T. ,
Sloan Working Paper 1096-79, December 1979.


To establish the base case, user managers listed and classified by type
the systems they already have. A straightforward classification of
application syster s type was used in the questionnaire:


Monitor The system monitors daily detail activity producing standard
reports on a fixed schedule (daily, weekly, or monthly).

Exception The system processes daily detail activity but produces
exception reports where the definition of exception
conditions is fixed .

Inquiry The system provides a database with flexible inquiry
capability, enabling managers to design and change their own
monitoring and exception reports.

Analysis The system provides powerful data analysis capabilities
(modeling, simulation, optimization, or statistical
routines) and the appropriate database to support managerial
decision making.

The first two types, monitor and exception, fall into the category of
applications traditionally called transaction processors. These systems
have been the bread and butter of DP, helping to capture, store, manipulate
and report the structured, high volume activities of daily operations.
Transaction processing systems generate management reports by successive


stages of increasingly summarized detail activity. The lowest level
summaries are provided to supervisors and managers directly responsible for
daily operations. Successively higher level summaries are distributed to
successively higher levels of management. There is an implicit assunption
in this traditional approach to management information — summarized daily
activity, which _is appropriate for first line managers, further summarized

is appropriate for higher levels of management. In general, this is not

true. To the limited extent that this is true, transaction processors do

provide some relevant information to higher level managers.

Inquiry and analysis systems, however, are more managerially oriented
in their intention, design, and use than transaction processors. It is the
difference between starting with the data and sending summary reports to
the managers most likely to find them relevant and starting with a
managers' needs and working down to the data and analysis necessary to
support those needs.

Flexible inquiry systems were originally developed to provide ad-hoc
inquiry into transaction processing data. They have been enhanced to also
provide flexible monitor and exception reporting. They require specialized
software (database management system and high level inquiry language),
hardware (disks and terminals), and are generally limited to accessing one
database at a time (eg., order entry or purchasing) although progress is
being made in linking databases together.

Analysis systems include a diverse mix of approaches to supporting
judgemental decision-making, from problem- finding and contingency planning
to selecting "best" alternatives. They are necessarily customized for a
particular set of decisions (eg., financial forecasting or production
scheduling) and include flexible access to the required database.


"A Framework for Management Information Systems", Garry and Scott Morton,

Sloan Manaoement Review, Fall 1971.


What Systems do Users have ?

Eighty-one us';r managers in six companies listed the systems which they
personally use on a regular basis. They then classified each system as
monitor, exception, inquiry or analysis. This provides the breakdown of
the installed base of application systems shown in Exhibit 3. There are
277 installed systems in the sample, of which 228 or 82 . 3?o are monitor type
systems. In contrast only 3.2% of the installed systems are of the
analysis type.

There are no real surprises here; rather there is confirmation of most
managers' expectations. The average manager regularly uses 3.4 systems, of
which approximately 90?c are transaction processors and 10?c are managerially
oriented. I almost said "only 10?o" but it wasn't too long ago that 100?o
were transaction processors. Improvements in technological capability have
been pushing inquiry systems, and improvements in the "DP-literacy" of user
managers have been pushing analysis type applications.

Ninety percent of the collective experience of DP and users is in
transaction processors. This simple observation of systems mix of the
installed base has seme interesting implications.

DP ' s policies and procedures, depth of skills, organizational
structure, evaluation/reward criteria, and expertise in developing
applications are dominated by transaction processors. Thus, there is a
strong tendency for any systems request that goes into DP to come back out
implemented as a transaction processor.

Users' perceived use of computers, expectations of systems development
procedures, anticipated benefits, and justification criteria are similarly
dominated by their transaction processing experience. Thus, there is a

strong tendency for any systems requests that go into DP to already look
like transaction processors. The interaction of DP and user biases toward
transaction processors has the same effect as a conscious conspiracy —
many inquiry and especially analysis systems needs are implemented as
transaction processors instead.

Moreover, in most companies the established standard procedures for
needs identification, project selection, and systems development are the
result of institutionalized transaction processing experience. This
further increases the probability that an inquiry or anlaysis systems need
will be implemented as a transaction processor.

The creation and use of inquiry and analysis systems are different from
transaction processors. Recognition of this simple fact has been hindered
by a superficial appearance of systems similarity — all four systems types
are indeed computer-based. The development of inquiry and analysis systems
is adversely affected by the policy requirement to follow established
standard procedures which are appropriate only for transaction processors.

Companies have refined their experiences into institutionalized
standard procedures and self-fulfilling expectations oriented exclusively
to their predominate transaction processing past. Recognition of this as a
problem is organizationally difficult and ackward. Reassessing institu-
tionalized standard procedures and expectations to facilitate the
development of additional systems types is a true management challenge.


How Appropriate is the Installed Base ?

In a separate set of questions, user managers listed their most
important tasks or decisions which they felt could be supported by a
computer system. The support actually available for these important tasks
was then classified by systems type. These results are shown in Exhibit 4.
Users listed 229 important tasks or decisions of which 79 or 35?o had no
systems support of any type.

Clearly the installed base is not appropriate for the 35?o which are
unsupported. This alone is an important source of demand for new systems
and user dissatisfaction given DP ' s backlog in creating new systems.
Monitor systems provided support for 120 or 53?o of the important managerial
activities although, as will be demonstrated, not necessarily

The second portion of Exhibit 4 omits the 79 important managerial
activities which are unsupported and re-computes the percent distribution
by systems type. This distribution is not very different from the
distribution by systems type of the total installed base in Exhibit 2.
Apparently whether a system is intended to support a managerially important
activity or not makes little if any difference in the type of system which
is developed. Presumably, this is because of the institutionalized
standard procedures for systems development which reinforces DP and users'
biases resulting in transaction processors irrespective of the intended use
of the system.

Exhibit 5 compares the total installed systems base from Exhibit 2 with
those systems cited as supporting users' important activities in Exhibit 4.
Of the 277 total installed systems only 150 or 55% were cited in


conjunction with users' important activities. In other words, while 35% of
managers' important activities go completely unsupported, 45% of the
systems v^ich ars installed do not relate to managers' most important

The 127 installed systems v^ich were not cited are probably necessary
for daily operations and fulfill some of the respondents secondary needs.
However, to presume all of the 228 installed monitor systems provide
management information is apparently in error — 108 of them were not even
mentioned in conJLnction with important managerial activities. Transaction
processors should be recognized for the valuable functions they do perform
and not tarnished by being labeled MIS, a function 1x1% of them do not even

There is a definite pattern in Exhibit 5 by system type. All of the
installed analysis systems were cited whereas only 53% of the monitor
systems were cited. For important managerial needs, inquiry and analysis
systems appear to be more relevant.

This was investigated in greater depth by asking user managers to
designate their desired systems type for each important activity they had
listed. In Exhibit 6 the systems type is considered to be appropriate when
the installed and desired systems type are the same.

Only 46, or 31%, of the systems v^ich support users' important
activities are of the appropriate type. In other words, 69% of the 150
currently installed systems which do relate to important managerial
activities do so only partially or in an inappropriate fashion -
inconvenient, inflexible or incomplete.

This is a reality which incites many user managers — unjustly I think.
In fact, if the "managerially unimportant" 45% of the installed base did
not exist, many managers' most important tasks would be managing those
daily operations which these systems support. These systems have enabled
managers to spend more time managing.


Again we see that inquiry and analysis systems fare significantly
better than monitor and exception systems. The vast majority of inquiry
and analysis systems, 66?o and 18% respectively, are considered by their
users to be of the appropriate type. In contrast to the low

appropriateness percentages for monitor (25?o) and exception (ll?c') systems
this makes a very strong statement. User managers consider inquiry and
analysis systems to be significantly more appropriate for their important
needs than monitor and exception systems.

Exception systems are seen as particularly inappropriate for important
managerial needs because, I think, of the fixed nature of the definition of
exception conditions embedded in the application software. Managerial re-
definitions of exception conditions necessitate software maintenance,
thereby constantly involving users in DP backlog delays and red tape in a
thankless quest to update inevitably obsolete reports.

Of the 120 monitor systems that relate to important managerial
activities the breakout of desired type is shown in Exhibit 7. Thirty of
these systems are considered by their users to be of the appropriate type.
The remaining 75?o should have been a different systems type to appropri-
ately address their users' important needs.

Sixty-four percent of these systems (38% inquiry plus 28?o analysis) are
of a basically inappropriate type. It is difficult for user managers to
consider the projects which created those inappropriate monitor systems to
be responsive to their needs. No matter how hard the project team worked,
no matter how technically competent they were, no matter the project came
in on budget and schedule, and no matter how bug-free and efficient the
system; a monitor system cannot behave like an inquiry or analysis system.
A well-designed and implemented, but inappropriate, monitor system


contributes to the generally recognized user feeling of DP unresponsiveness
to managerial needs. When this occurs frequently the reputation of DP as
unresponsive and :omputer systems as inappropriate to important managerial
needs is confirmed and spread.

Exhibit 8 juxtaposes lines from previous exhibits to emphasize their
cumulative effect on user managers' perceptions. The difference in
managerial appropriateness between monitor/ exception and inquiry/ analysis
system typss is dramatic. Of the total base of installed systems 5h% are
cited in conjunction with managerially important activities. Of those
systems which do relate to important managerial activities 31?i are of the
appropriate systems type. The cumulative effect is devastating. User
managers, see only 17?o of the installed base of application systems to be
appropriate to their important needs.

This does not imply that the remaining 83?o of the installed base is
inappropriate for other necessary corporate functions. These systems were
not necessarily intended to provide support for middle level managers'
important activities. This exhibit does not reflect the recent mix of DP
system development by type inasmuch as most monitor systems were developed
years ago and all inquiry and analysis systems are fairly recent.
Moreover, when originally developed, many monitor systems directly and
appropriately addressed important managerial needs, such as payroll and
accounting, so successfully that they are no longer important managerial

Even accepting this literary of extenuating circumstances, I am still
convinced that this is how user managers perceive DP systems and why they
believe DP to be unresponsive to their managerial needs. Millions have
been spent on developing, running and maintaining existing systems. Only


17% of the installed base is relevant and appropriate to important managerial
activities. Although inquiry and especially analysis systems are dramat-
ically more relevant and appropriate to managerial needs, these types are
such a small percent of the total installed base as to be considered atypical
DP applications. Moreover, 35?o of important managerial activities have no
systems support of any type.

A search for the cause of user dissatisfaction with DP need go no further
than this mismatch between user managers' important needs and the installed
base of applications.

What System Types do Users Want ?

The preceeding comments foreshadow a shift in user demand away from
transaction processors and toward inquiry and analysis system types. The
question for many DP departments is when will this shift occur and when
should they begin preparing to meet this often predicted mix shift in user
demand. The answer, according to Exhibit 9, is last year.

The known backlog of user requested systems development projects already
exhibits a dramatic shift in the proportion of system types from the
currently • installed base. Whereas 90°o of the installed base is transaction
processors, 40?o of the known backlog is requests for managerially-oriented
inquiry and analysis systems. The shift in user demand mix has already

Users are expecting increased implementation of managerially-oriented
systems types in the near term and DP should already be able to perceive this
shift. If standardized procedures and ingrained transaction processing
attitudes result in the implementation of monitor systems where inquiry or

"Defining Success for DP", Robert M. Alloway, M.I.T., CISR Working Paper 32 ,

March 1980.


analysis systems are needed, requested, and expected user dissatisfaction
will surely increase.

The invisible backlog (desired systems not yet requested of DP) has
basically the same proportional mix as the known backlog. For the
foreseeable future user demand will continue to emphasize analysis and
inquiry systems over transaction processors.

The total demand mix summarizes the planning horizon for DP in terms of
the relative importance by systems type. This obvious mix shift emphasizes
the necessity of modifying historically derived standard systems development
procedures. In order to be successful and responsive to managerial needs,
DP's attention, priorities, and proportion of effort must be cut in half for
monitor systems and increased by a factor of three for inquiry and six for
analysis systems.

This dramatic shift in user demand must be matched by an equally
dramatic shift in DP procedures and capabilities. Most DP departments are
currently perceived by user managers to be managerially unresponsive. To the
extent that DP departments have failed to recognize this mix shift in user
demand early enough to match supply to demand, these perceptions are well

DP's corresponding shift in supply requires pervasive changes to its
internal structure and procedures — from training programs and personnel
evaluation criteria, to project design and procedures for different system
types, to defining success and strategic planning for DP. Recognition of

"Planning Skill Development for Systems Analysts", Robert M. Alloway and
Jerome T. Nolte, M.I.T., CISR Working Paper 31 , November 1979.

"Temporary Management Systems", Robert M. Alloway, Stockholm School of
Economics, Institute of International Business Working Paper , September

^"Defining Success for DP", Robert M. Alloway, M.I.T., CISR Working Paper 32 ,
March, 1980.


the magnitude of change necessary should not induce general paralysis, but
rather, the explicit realization of the managerial nature of this
challenge, the significant benefits to be gained, and the undeniable
necessity for a flat-out effort to play catch-up ball.

What Systems Types will Users Want?

If you will allow me considerable latitude with cross-sectional data
to "kludge" longitudinal implications, Exhibit 10 can be interpreted as
demand trends for systems types over the next five years.

Implementation of managerially relevant and appropriate systems has
triggered a disproportionate increase in demand for these systems. Today's
small installed base of inquiry and analysis systems has already resulted
in a significant mix shift in demand. As DP departments are successful in
fulfilling today's demand the known backlog becomes part of tomorrow's
installed base. Tomorrow's demand for new systems will be similarly
reflective of the relevance and appropriateness of tomorrow's installed

The best predictor available of tomorrow's demand mix are the 35°i of
the managerially important tasks which are today completely unsupported.
The percentage breakdown of desired system type for Lnsupported important
managerial activities was used for tomorrow's demand in Exhibit 10.

The chronology by systems type is thus traced and forecast. The
proportion of monitor systems as a percent of the installed base has
dropped from 100% yesterday to 82.3?o today. Demand for monitor systems is
forecast to drop from 40?o today to 22?o tomorrow. By contrast, analysis
systems constituted 0% of the installed base yesterday and 3.2?o today.


Today's demand has surged to 20?o and tomorrow's is forecast to continue to
grow to 40?o.

DP used to play a leadership role vis-a-vis users in the application of
computer technology to organizational needs. DP has lost this leadership
position by failure to shift the mix of delivery capability early enough to
meet emerging demand. Today DP must play catch-up-ball just to rebalance
delivery capability and demand.

Theoretically it is possible for DP to regain its leadership role by
providing a mix of systems types appropriate for tomorrow's demand.
However, this is such a dramatic change as to be impossible in the near
term for most, if not all, DP departments.

To be practical, DP and users should target their change planning to
achieve today's demand mix as soon as possible, yesterday preferably, and
to create policies, organizational structures and procedures facilitative
of growing into the position to deliver tomorrow's demand mix within 3

How Many New Systems do Users Want ?

The previous sections of this paper discussed the proportions or mix of
user demand by systems type. This section introduces the quantities of new
systems needed to canplete the picture of user demand. In Exhibit 11 no
attempt has been made to standardize new systems by size or number of
man-months required to create them. Just like the installed base of 277
systems, some are larger or more complex than others. The new systems
counts also include systems replacements but not enhancements or


The known backlog of all systems, 188, is so large (68?o of the
currently installed base) that it results in the commonly observed 2-3 year
backlog in most DP departments. Systems development personnel are already
flat-out dealing with this known backlog.

The invisible backlog of all systems, 309, is staggering (164% of the
known backlog or 115^0 of the installed base) . The invisible backlog number
is not as reliable as the known backlog because these desired systems have
not passed the "rigors" of proposal preparation and approval. However,
this figure is probably in the right ballpark.

The size of the invisible backlog implies that the known backlog will
never get any shorter. No matter how fast DP can create new systems the
users will keep the known backlog full. The length of a DP department's
known backlog is not an indication of anything other than the users
planning horizon and perceived stability of need.

Total demand for all systems, 497, is simply overwhelming — 179?o of
the installed base, which, by the way, took 10 to 15 years to create.
There is no way that any DP department can actually fulfill this level of
demand. Priority setting is as important in determining success as the
number and quality of systems developed.

Conversely, user departments will be hard pressed to fulfill their
needs even using a consortium of suppliers — DP, own DP staff, and
software vendors (package and custom). Users' prioritization of needs is
as important in fulfilling their information needs as the number and
quality of systems implemented.

User management must fulfill its responsibilities in prioritizing
individual systems, levels of effort, and type mix. It is simply unfair to


Online LibraryRobert M AllowayUser managers' systems needs : it appears that tomorrow is here today, we need to change the rules of the game and play catch-up ball → online text (page 1 of 2)