Professional Documents
Culture Documents
Chief Executives Define Their Own Data Needs
Chief Executives Define Their Own Data Needs
Chief executives
define their
own data needs
What are the real informa- Mr. Rockart is director of He could have been the president of any one of a
tion needs of the chief the Center for Information number of successful and growing medium-sized
executive officer or any Systems Research, Sloan companies in the electronics industry. He had spent
other top executive of a School of Management, the previous day working to salt away the acquisi-
company? If presented in Massachusetts Institute of tion of a small company that fitted an important
the form of computer Technology. He is also position in the product line strategy he had evolved
printouts, often the reports senior lecturer at the Sloan
for his organization. Most of this day had been spent
submitted are what some School, specializing in
suhordinate thinks con- management information
discussing problems and opportunities with key
tain pertinent and useful systems, planning, and managers. During both days he had lived up to his
information. Then again, control. reputation of being an able, aggressive, action-ori-
if presented hy word of ented chief executive of a leading company in its
mouth, the informal ap- segment of the electronics field.
proach overlooks the kind Unfortunately, the president had chosen the late
of routine data (often afternoon and early evening to work through the
computer-hased) that papers massed on his desk. His thoughts were not
should he supplied reg- pleasant. His emotions ranged from amusement to
ularly to the top executive.
This article examines the
anger as he plowed through the papers. "Why," he
several methods of provid- thought, "do I have to have dozens of reports a
ing information to top month and yet very little of the real information
management now in use, I need to manage this company? There must be a
discusses the advantages way to get the information I need to run this com-
and disadvantages of each, pany!"
and offers a new approach In effect, he was expressing the thoughts of many
that focuses on individual other general managers—and especially chief execu-
managers and their hard tive officers—whose needs for information are not
and soft information needs. as clearly determined as are those of many func-
The discussion includes
five illustrative examples
tional managers and first-line supervisors. Once one
from which the author gets above the functional level, there is a wide
draws some generalizations variety of information that one might possibly need,
ahout the method and and each functional specialty has an interest in
the chief executive's "feeding" particular data to a general manager. As
data needs. in this case, therefore, a massive information flow
occurs. This syndrome is spelled out with differing
82 Harvard Business Review March-April 1979
emphases by the recent comments of two other tion process is centered on the development of
chief executives: operational systems that perform the required paper-
"The first thing about information systems that work processing for the company. Attention is fo-
strikes me is that one gets too much information. cused, therefore, on systems that process payroll,
The information explosion crosses and criss-crosses accounts payable, billing, inventory, accounts re-
executive desks with a great deal of data. Much of ceivable, and so on.
this is only partly digested and much of it is ir- The information by-products of these transaction-
relevant. . . ." ^ processing systems are often made available to all
interested executives, and some of the data (e.g.,
"I think the problem with management information summary sales reports and year-to-date budget re-
systems in the past in many companies has been ports) are passed on to top management. The by-
that they're overwhelming as far as the executive products that reach the top are most often either
is concerned. He has to go through reams of reports heavily aggregated (e.g., budgeted/actual for major
and try to determine for himself what are the most divisions) or they are exception reports of significant
critical pieces of information contained in the re- interest (e.g., certain jobs now critical by some preset
ports so that he can take the necessary action and standard). All reports, however, are essentially by-
correct any problems that have arisen." ^ products of a particular system designed primarily
to perform routine paperwork processing.
It is clear that a problem exists with defining exactly Where the information subsystem is not com-
what data the chief executive (or any other general puter-based, the reports reaching the top are often
manager) needs. My experience in working with ex- typed versions of what a lower level feels is useful.
ecutives for the past decade or more is that the Alternatively, they may be the ongoing, periodically
problem is universally felt—with individual frustra- forthcoming result of a previous one-time request
tion levels varying, but most often high. for information concerning a particular matter in-
In this article, I will first discuss four current itiated by the chief executive in the dim past.
major approaches to defining managerial informa- Of the five methods discussed herein, the by-
tion needs. Next, I will discuss a new approach de- product approach is undoubtedly the predominant
veloped by a research team at MIT's Sloan School method. It leads to the welter of reports noted in the
of Management. Termed the "critical success factor introductory paragraphs of this article. It has the
(CSF) method," this approach is being actively re- paper-processing tail wagging the information dog.
searched and applied today at the MIT center. Final- The approach is, however, understandable. Paper-
ly, I will describe in detail this method's use in one work must be done and clerical savings can be made
major case as well as provide summary descriptions by focusing on automating paper-processing sys-
of its use in four other cases. tems. It is necessary to develop this class of data
processing system to handle day-to-day paperwork.
However, other approaches are also necessary to
provide more useful management information.
executive are not very useful, and (2) he, therefore, concepts, two of which are necessary and the third
relies very heavily on oral communication, advo- of. which provides the glamour (as well as a few
cates of this approach then conclude that all com- tangible benefits).
puter-based reports—no matter how they are devel- The first concept is the selection of a set of key
oped—will be useless. They look at inadequately indicators of the health of the business. Information
designed information systems and curse all com- is collected on each of these indicators. The second
puter-based systems. concept is exception reporting—that is, the ability
Proponents of the null approach see managerial to make available to the manager, if desired, only
use of information as Henry Mintzberg does: those indicators where performance is significantly
"... it is interesting to look at the content of different (with significance levels necessarily pre-
managers' information, and at what they do with it. defined) from expected results. The executive may
The evidence here is that a great deal of the man- thus peruse all the data available or focus only on
ager's inputs are soft and speculative—impressions those areas where performance is significantly dif-
and feelings about other people, hearsay, gossip, and ferent from planned.
so on. Furthermore, the very analytical inputs—re- The third concept is the expanding availability
ports, documents, and hard data in general—seem to of better, cheaper, and more flexible visual display
be of relatively little importance to many managers. techniques. These range from computer consoles
(After a steady diet of soft information, one chief (often with color displays) to wall-size visual displays
executive came across the first piece of hard data of computer-generated digital or graphic material. A
he had seen all week—an accounting report—and paradigm of these systems is the one developed at
put it aside with the comment, T never look at Gould, Inc. under the direction of William T. Yl-
this.')" =• visaker, chairman and chief executive officer. As
Business Week reports:
To some extent, this school of thought is correct. "Gould is combining the visual display board,
There is a great deal of information used by top which has now become a fixture in many board-
executives that must be dynamically gathered as rooms, with a computer information system. In-
new situations arise. And, most certainly, there are formation on everything from inventories to re-
data that affect top management which are not ceivables will come directly from the computer in
computer-based and which must be communicated an assortment of charts and tables that will make
in informal, oral, and subjective conversations. comparisons easy and lend instant perspective.
There are, however, also data that can and should "Starting this week Ylvisaker will be able to tap
be supplied regularly to the chief executive through three-digit codes into a 12-button box resembling
the computer system. More significantly, as I will the keyboard of a telephone. SEX will get him sales
note later on, it is also important to clearly define figures. GIN will call up a balance sheet. MUD is
what informal fnot computer-based) information the keyword for inventory.
should be supplied to a top executive on a regular "About 75 such categories will be available, and
basis. the details will be displayed for the company as a
whole, for divisions, for product lines, and for other
breakdowns, which will also be specified by simple
Key indicator system digital codes." *
A clear contender today for the fastest growing At Gould, this information is displayed on a large
school of thought concerning the "best" approach screen in the boardroom, and is also available at
to the provision of executive information is the key computer terminals. The data are available in full,
indicator system. This procedure is based on three by exception, and graphically if desired.
As in most similar key indicator systems I have
1. Interview with Anthony J.F. O'Reilly, president of H.J. Heinz Co., M.LS. seen, the emphasis at Gould is on financial data.
Quarter/y, March 1977, p. 7. Daniel T. Carroll, reporting on Gould's system in
2. Interview with William Dougherty, president of North Carolina Bank
Corporation, M.l.S. Quarterly, March 1977, p. r.
mid-1976, described the system's "core report." ^ The
3. See Henry Mintzherg, "Planning on the Left Side and Managing on the
report, available for each of Gould's 37 divisions,
Right," HBR July-August r976, p. 54. provides data on more than 40 operating factors.
4. "Corporate 'War Rooms' Plug into the Computer," Business Week, August For each factor, current data are compared with
23, 1976, P- 65. budget and prior year figures on a monthly and
5. Daniel T. Carroll, "How the President Satisfies His Information Systems year-to-date basis. The report, as noted by Garroll,
Requirements," puhlished in Society for Management Infoimation
Systems Proceedings, 1976. is ever changing, but its orientation toward "profit
84 Harvard Business Review March-April 1979
and loss" and "balance sheet" data, as well as ratios Each of the four current procedures just discussed
drawn from these financial data, is evident. has its advantages and disadvantages. The by-
product technique focuses on getting paperwork
processed inexpensively, but it is far less useful
Total study process with regard to managerial information. It too often
results in a manager's considering data from a
In this fourth approach to information needs, a single paperwork function (e.g., payroll) in isolation
widespread sample of managers are queried about from other meaningful data (e.g., factory output
their total information needs, and the results are versus payroll dollars).
compared with the existing information systems. The null approach, with its emphasis on the
The subsystems necessary to provide the informa- changeability, diversity, and soft environmental in-
tion currently unavailable are identified and as- formation needs of a top executive, has probably
signed priorities. This approach, clearly, is a reaction saved many organizations from building useless in-
to two decades of data processing during which formation systems. It, however, places too much
single systems have been developed for particular stress on the executive's strategic and person-to-
uses in relative isolation from each other and with person roles. It overlooks the management control
little attention to management information needs. role of the chief executive, which can be, at least
In effect, this approach was developed by IBM and partially, served by means of routine, often com-
others to counter the by-product method previously puter-based, reporting.^
noted. The key indicator system provides a significant
The most widely used formal procedure to ac- amount of useful information. By itself, however,
complish the total study is IBM's Business Systems this method often results in many undifferentiated
Planning (BSP) methodology. BSP is aimed at a top- financial variables being presented to a management
down analysis of the information needs of an or- team. It tends to be financially all-inclusive rather
ganization. In a two-phase approach, many man- than on-target to a particular executive's specific
agers are interviewed (usually from 40 to 100) to needs. The information provided is objective, quan-
determine their environment, objectives, key deci- tifiable, and computer stored. Thus in the key in-
sions, and information needs. Several IBM-suggested dicator approach the perspective of the information
network design methods and matrix notations are needs of the executive is a partial one—oriented
used to present the results in an easily visualized toward hard data needs alone. More significantly, in
manner. its "cafeteria" approach to presenting an extensive
The objectives of the process are to develop an information base, it fails to provide assistance to
overall understanding of the business, the informa- executives in thinking through their real informa-
tion necessary to manage the business, and the tion needs.
existing information systems. Gaps between infor- The total study process is comprehensive and can
mation systems that are needed and those currently pinpoint missing systems. However, it suffers, as
in place are noted. A plan for implementing new noted, from all of the problems of total approaches.
systems to fill the observed gaps is then developed. There are problems concerning expense, the huge
This total understanding process is expensive in amount of data collected (making it difficult to dif-
terms of manpower and all-inclusive in terms of ferentiate the forest from the trees), designer bias,
scope. The amount of data and opinions .gathered and difficulty in devising reporting systems that
is staggering. Analysis of all this input is a high art serve any individual manager well.
form. It is difficult, at best, to determine the correct
level of aggregation of decision making, data gather-
ing, and analysis at which to work.
Yet the top-down process tends to be highly useful
in most cases. The exact focus of the results, how-
ever, can be biased either toward top management New CSF method
information and functional management informa-
tion or toward paperwork processing, depending on
the bias of the study team. I have not seen a BSP The MIT research team's experience in the past two
study that gives priority to top executive informa- years with the critical success factors (GSF) approach
tion in the study's output. The design, cleaning up, suggests that it is highly effective in helping execu-
and extension of the paper-processing information tives to define their significant information needs.
network is too often the focus of the study team. Equally important, it has proved efficient in terms
CEOs' data needs 85
of the interview time needed (from three to six lished work of Robert N. Anthony, John Dearden,
hours) to explain the method and to focus atten- and Richard F. Vancil.*
tion on information needs. Most important, execu- Daniel, in introducing the concept, cited three
tive response to this new method has been excellent examples of major corporations whose information
in terms of both the process and its outcome. systems produced an extensive amount of infor-
The actual CSF interviews are usually conducted mation. Very little of the information, however,
in two or three separate sessions. In the first, the appeared useful in assisting managers to better per-
executive's goals are initially recorded and the CSFs' form their jobs.
that underlie the goals are discussed. The interrela- To draw attention to the type of information
tionships of the CSFs and the goals are then talked actually needed to support managerial activities,
about for further clarification and for determination Daniel turned to the concept of critical success
of which recorded CSFs should be combined, elim- factors. He stated,
inated, or restated. An initial cut at measures is also "... a company's information system must be dis-
taken in this first interview. criminating and selective. It should focus on 'success
The second session is used to review the results factors.' In most industries there are usually three
of the first, after the analyst has had a chance to to six factors that determnie success; these key jobs
think about them and to suggest "sharpening up" must be done exceedingly well for a company to be
some factors. In addition, measures and possible successful. Here are some examples from several
reports are discussed in depth. Sometimes, a third major industries:
session may be necessary to obtain final agreement D In the automobile industry, styling, an efficient
on the CSF measures-and-reporting sequence. dealer organization, and tight control of manufac-
turing cost are paramount.
D In food processing, new product development,
Conceptual antecedents good distribution, and effective advertising are the
major success factors.
In an attempt to overcome some of the shortcom- D In life insurance, the development of agency
ings of the four major approaches discussed earlier, management personnel, effective control of clerical
the CSF method focuses on individual managers and personnel, and innovation in creating new types of
on each manager's current information needs—both policies spell the difference." ^
hard and soft. It provides for identifying managerial
information needs in a clear and meaningful way. Critical success factors thus are, for any business,
Moreover, it takes into consideration the fact that the limited number of areas in which results, if they
information needs will vary from manager to man- are satisfactory, will ensure successful competitive
ager and that these needs will change with time performance for the organization. They are the few
for a particular manager. key areas where "things must go right" fdr the busi-
The approach is based on the concept of the "suc- ness to flourish. If results in these areas are not ad-
cess factors" first discussed in the management lit- equate, the organization's efforts for the period will
erature in 1961 by D. Ronald Daniel, now managing be less than desired.
director of McKinsey & Company.^ Although a As a result, the critical success factors are areas
powerful concept in itself for other than informa- of activity that should receive constant and careful
tion systems' thinking, it has been heavily obscured attention from management. The current status of
in the outpouring of managerial wisdom in the performance in each area should be continually
past two decades. It has been focused on and clari- measured, and that information should be made
fied to the best of my knowledge only in the pub- available.
6. Management control is the process of (a| long-range planning of the activi-
As Exhibit 1 notes, critical success factors support
ties of the organization, (b) short-term planning (usually one year), and the attainment of organizational goals. Goals repre-
(c) monitoring activities to ensure the accomplishment of the desired results.
The management control process thus follows the development of major
sent the end points that an organization hopes to
strategic directions that are set in the strategic planning process. This reach. Critical success factors, however, are the
definition roughly follows the framework of Rohert N. Anthony, Planning areas in which good performance is necessary to
and Control: A Framework /or Analysis (Boston: Division of Research,
Harvard Business School, r965). ensure attainment of those goals.
7. See D. Ronald Daniel, "Management Information Crisis," HBR Septemher- Daniel focused on those critical success factors
October r96r, p. rrr. that are relevant for any company in a particular
8. See Robert N. Anthony, John Dearden, and Riehard F. Vancil, "Key
Economic Variables," in Management Controls Systems (Homewood, 111.: Irwin,
industry. Exhibit I updates Daniel's automobile in-
r972|, p. 147. dustry CSFs and provides another set of CSFs-from
9. Daniel, "Management Information Crisis," p. ir6. the supermarket industry and a nonprofit hospital.
86 Harvard Business Review March-April 1979
tion. Indeed, that is the major focus of the MIT improve existing areas of business can be more
research team's current work. It is, however, clear readily defined.
from studies now going on that CSFs, as might be
expected, can be useful at each level of general
management (managers to whom multiple functions
report). There are significant benefits of taking the
necessary time to think through—and to record—the
critical success factors for each general manager in Illustrative CSF example
an organization. Consider:
D The process helps the manager to determine
those factors on which he or she should focus man- Let us now turn to an example of the use of this
agement attention. It also helps to ensure that those approach. The president referred to at the start of
significant factors will receive careful and contin- this article is real. He is Larry Gould, former presi-
uous management scrutiny. dent of Microwave Associates, a $6o-million sales or-
• The process forces the manager to develop good ganization serving several aspects of the microwave
measures for those factors and to seek reports on communication industry.^^ When he first looked
each of the measures. carefully at the "information" he was receiving,
D The identification of CSFs allows a clear defini- Gould found that some 97 "reports" crossed his desk
tion of the amount of information that must be in a typical month. Almost all were originally de-
collected by the organization and limits the,costly signed by someone else who felt that he "should be
collection of more data than necessary. receiving this vital data."
D The identification of CSFs moves an organiza- However, the reports provided him with virtually
tion away from the trap of building its reporting nothing he could use. A few gave him some "score-
and information system primarily around data that keeping data," such as the monthly profit state-
are "easy to collect." Rather, it focuses attention on ment. One or two others provided him with hits
those data that might otherwise not be collected but and pieces of data he wanted, but even these left
are significant for the success of the particular man- major things unsaid. The data were either unrelated
agement level involved. to other key facts or related in a way that was not
• The process acknowledges that some factors ' meaningful to him.
are temporal and that CSFs are manager specific. The concept of critical success factors sounded to
This suggests that the information system should be him like one way out of this dilemma. He therefore,
in constant flux with new reports being developed with the MIT research analyst, invested two two-
as needed to accommodate changes in the organiza- and-a-half-hour periods in working through his
tion's strategy, environment, or organization struc- goals, critical success factors, and measures. First, he
ture. Rather than changes in an information system noted the objectives of the company and the cur-
being looked on as an indication of "inadequate rent year's goals. Then, he went to work to assess
design," they must be viewed as an inevitable and what factors were critical to accomplish these ob-
productive part of information systems develop- jectives.
ment.
n The CSF concept itself is useful for more than
information systems design. Current studies suggest Factors &. measures
several additional areas of assistance to the manage-
ment process. For example, an area that can be im- The seven critical success factors Gould developed
proved through the use of CSFs is the planning are shown in Exhibit III, along with from one to
process. CSFs can be arrayed hierarchically and used three prime measures for each factor (although he
as an important vehicle of communication for man- also developed some additional measures). The read-
agement, either as an informal planning aid or as er should note that this specific set of CSFs emerged
a part of the formal planning process. only after intensive analysis and discussion. At the
end of the first meeting, nine factors were on Gould's
Let me stress that the CSF approach does not attempt list. By the end of the second meeting, two had been
to deal with information needs for strategic plan- combined into one, and one had been dropped as
ning. Data needs for this management role are
almost impossible to preplan. The CSF method
centers,. rather, on information needs for manage- 12. Since this was originally written, Gould has assumed the position of
chairman of the board at M/A-COM, Inc., a holding company of which
ment control where data needed to monitor and Microwave Associates is.a suhsidiary.
CEOs' data needs 89
tion). In general, no single job is crucially important. route, however, focused on the data needed for the
Rather, it is the profile of performance across major ongoing "management control" process, and this
jobs that is significant. need was significant at Microwave.
D The development of priorities for information
systems development. It was clear that informa-
Reports &. subsystems tion needed for control purposes by the chief execu-
tive should have some priority. (It also highlighted
Given the foregoing CSFs and measures, the next priorities for other management levels.)
step was to design a set of report formats. This step • The provision of a means of hierarchical com-
required examination of both existing information munication among executives as to what the critical
systems and data sources. factors were for the success of the company. (Too
For the soft, informal, subjective measures, this often, only goals provide a major communication
process was straightforward. Forms to record facts link to enhance shared understanding of the com-
and impressions were designed so as to scale (where pany and its environment among management
possible) perception and highlight significant soft levels.) This hierarchical approach provided another
factors. —and we believe more pragmatic and action-oriented
For some of the hard computer-based measures, —means of communication. At Microwave, there is
existing information systems and data bases supplied a current project aimed at developing and sharing
most of the necessary data. However, in every case- CSFs at the top four management levels.
even where all data were available—existing report
forms were inadequate and new reports had to be
designed. Other case examples
Most important, however, two completely new in-
formation subsystems were needed to support the The critical success factors developed in four other
president's CSFs. These were a "bidding" system cases provide useful additional background for draw-
and a vastly different automated "project budgeting ing some generalizations about the method and ex-
and control" system. (Significantly, each of these ecutive information needs. These CSFs are arrayed
subsystems had been requested many times by lower- in Exhibit IV.
level personnel, who needed them for more detailed
planning and control of job bidding and monitor- Major oil company
ing at the product-line manager and manufacturing The chief executive of this centralized organization
levels.) Subsequently, these subsystems were placed responded quickly and unhesitatingly concerning
at the top of the priority list for data processing. his critical success factors. His goal structure was
oriented toward such traditional measures as in-
In summarizing the Microwave case, it is clear creasing return on investment, increasing earnings
that the exercise of discovering information needs per share, and so forth. Yet he felt there were two
through examination of the chief executive's critical major keys to profitability in the future. One was
success factors had a number of specific benefits. to improve relationships with society as a whole
All of the seven general advantages of the CSF and with the federal government in particular. The
method for information systems development pre- other was the urgent need to provide a broader base
viously noted applied to some extent. However, the of earnings assets in petroleum-shy future decades.
importance of each of these varies from organiza- As a result of this view of the world, the CEO
tion to organization. At Microwave, the most strik- had initiated major programs to develop new ven-
ing advantages were: tures and to decentralize the organization. To fa-
• The conscious listing (or bringing to the sur- cilitate the acquisition process, emphasis was placed
face) of the most significant areas on which attention on cash fiow (liquidity) as opposed to reported earn-
needed to be focused. The process of making these ings. In addition, prime attention was given to
areas explicit provided insights not only into in- understanding and improving external relationships.
formation needs, but also into several other aspects All of these efforts are refiected in the company's
of the company's managerial systems. critical success factors shown in Exhibit IV. Progress
D The design of a useful set of reports to provide in each of these areas is monitored weekly. CSFs # 1 ,
the information needed for monitoring ongoing #3, and # 4 are reported on with regard to both
operations at the executive level. (There clearly were actions taken and the appropriate executive's sub-
other data needed—i.e., for developing strategy, deal- jective assessment of results attained. Liquidity
ing with special situations, and so on.) The CSF measures are provided by computer output. New
CEOs' data needs 91
Exhibit IV
CSFs in four cases
Chief executive of a major oil President of a store furnishings Director of a government Division chief executive of an
company manufacturer hospital electronics company
1. Decentralize organization. 1. Expand foreign saies for product 1. Devise method for obtaining valid 1. Support field sales force.
lines B and C. data on current status of hospital 2. Strengthen customer relations.
2. Improve liquidity position.
operations.
2. Improve mari<et understanding of 3. Improve productivity.
3. improve government/business
product line A. 2. Devise method for resource
relationships. 4. Obtain government R&D support.
3. Redesign sales compensation allocation.
4. Create better societal image. 5. Develop new products.
structure in three-product lines. 3. Manage external relationships.
5. Develop new ventures. 6. AcqjJire new technological
4. Improve production scheduling. 4. Get acceptance of concept of
regionalization by all hospital capability.
5. Mechanize production facilities.
directors. 7. Improve facilities.
6. Strengthen management team.
5. Develop method for managing
regionalization in government
hospital group.
6. Strengthen management
support, capability, and
capacity.
7. Improve relationship with
government department central
office.
8. Meet budgetary constraints.
Antony Jay,
Ghairman,
Video Arts Ltd.
Harvard Business Review Notice of Use Restrictions, May 2009
Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for
the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material
in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may
not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any
other means of incorporating the content into course resources. Business licensees may not host this content on
learning management systems or use persistent linking or other means to incorporate the content into learning
management systems. Harvard Business Publishing will be pleased to grant permission to make this content
available through such means. For rates and permission, contact permissions@harvardbusiness.org.