Implementing an Executive Information System (EIS)

by Floyd Kelly

An EIS is a tool that provides direct on-line access to relevant information about aspects of a business that are of particular interest to the senior manager.


Many senior managers find that direct on-line access to organizational data is helpful. For example, Paul Frech, president of Lockheed-Georgia, monitored employee contributions to company-sponsored programs (United Way, blood drives) as a surrogate measure of employee morale (Houdeshel and Watson, 1987). C. Robert Kidder, CEO of Duracell, found that productivity problems were due to salespeople in Germany wasting time calling on small stores and took corrective action (Main, 1989).

Information systems have long been used to gather and store information, to produce specific reports for workers, and to produce aggregate reports for managers. However, senior managers rarely use these systems directly, and often find the aggregate information to be of little use without the ability to explore underlying details (Watson & Rainer, 1991, Crockett, 1992).

An Executive Information System (EIS) is a tool that provides direct on-line access to relevant information in a useful and navigable format. Relevant information is timely, accurate, and actionable information about aspects of a business that are of particular interest to the senior manager. The useful and navigable format of the system means that it is specifically designed to be used by individuals with limited time, limited keyboarding skills, and little direct experience with computers. An EIS is easy to navigate so that managers can identify broad strategic issues, and then explore the information to find the root causes of those issues.

Executive Information Systems differ from traditional information systems in the following ways:

  • are specifically tailored to executive's information needs
  • are able to access data about specific issues and problems as well as aggregate reports
  • provide extensive on-line analysis tools including trend analysis, exception reporting & "drill-down" capability
  • access a broad range of internal and external data
  • are particularly easy to use (typically mouse or touchscreen driven)
  • are used directly by executives without assistance
  • present information in a graphical form

Purpose of EIS

The primary purpose of an Executive Information System is to support managerial learning about an organization, its work processes, and its interaction with the external environment. Informed managers can ask better questions and make better decisions. Vandenbosch and Huff (1992) from the University of Western Ontario found that Canadian firms using an EIS achieved better business results if their EIS promoted managerial learning. Firms with an EIS designed to maintain managers' "mental models" were less effective than firms with an EIS designed to build or enhance managers' knowledge.

This distinction is supported by Peter Senge in The Fifth Dimension. He illustrates the benefits of learning about the behaviour of systems versus simply learning more about their states. Learning more about the state of a system leads to reactive management fixes. Typically these reactions feed into the underlying system behaviour and contribute to a downward spiral. Learning more about system behaviour and how various system inputs and actions interrelate will allow managers to make more proactive changes to create long-term improvement.

A secondary purpose for an EIS is to allow timely access to information. All of the information contained in an EIS can typically be obtained by a manager through traditional methods. However, the resources and time required to manually compile information in a wide variety of formats, and in response to ever changing and ever more specific questions usually inhibit managers from obtaining this information. Often, by the time a useful report can be compiled, the strategic issues facing the manager have changed, and the report is never fully utilized.

Timely access also influences learning. When a manager obtains the answer to a question, that answer typically sparks other related questions in the manager's mind. If those questions can be posed immediately, and the next answer retrieved, the learning cycle continues unbroken. Using traditional methods, by the time the answer is produced, the context of the question may be lost, and the learning cycle will not continue. An executive in Rockart & Treacy's 1982 study noted that:

Your staff really can't help you think. The problem with giving a question to the staff is that they provide you with the answer. You learn the nature of the real question you should have asked when you muck around in the data (p. 9).

A third purpose of an EIS is commonly misperceived. An EIS has a powerful ability to direct management attention to specific areas of the organization or specific business problems. Some managers see this as an opportunity to discipline subordinates. Some subordinates fear the directive nature of the system and spend a great deal of time trying to outwit or discredit it. Neither of these behaviours is appropriate or productive. Rather, managers and subordinates can work together to determine the root causes of issues highlighted by the EIS.

The powerful focus of an EIS is due to the maxim "what gets measured gets done." Managers are particularly attentive to concrete information about their performance when it is available to their superiors. This focus is very valuable to an organization if the information reported is actually important and represents a balanced view of the organization's objectives.

Misaligned reporting systems can result in inordinate management attention to things that are not important or to things which are important but to the exclusion of other equally important things. For example, a production reporting system might lead managers to emphasize volume of work done rather than quality of work. Worse yet, productivity might have little to do with the organization's overriding customer service objectives.

Contents of EIS

A general answer to the question of what data is appropriate for inclusion in an Executive Information System is "whatever is interesting to executives." While this advice is rather simplistic, it does reflect the variety of systems currently in use. Executive Information Systems in government have been constructed to track data about Ministerial correspondence, case management, worker productivity, finances, and human resources to name only a few. Other sectors use EIS implementations to monitor information about competitors in the news media and databases of public information in addition to the traditional revenue, cost, volume, sales, market share and quality applications.

Frequently, EIS implementations begin with just a few measures that are clearly of interest to senior managers, and then expand in response to questions asked by those managers as they use the system. Over time, the presentation of this information becomes stale, and the information diverges from what is strategically important for the organization. A "Critical Success Factors" approach is recommended by many management theorists (Daniel, 1961, Crockett, 1992, Watson and Frolick, 1992). Practitioners such as Vandenbosch (1993) found that:

While our efforts usually met with initial success, we often found that after six months to a year, executives were almost as bored with the new information as they had been with the old. A strategy we developed to rectify this problem required organizations to create a report of the month. That is, in addition to the regular information provided for management committee meetings, the CEO was charged with selecting a different indicator to focus on each month (Vandenbosch, 1993, pp. 8-9).

While the above indicates that selection of data for inclusion in an EIS is difficult, there are several guidelines that help to make that assessment. A practical set of principles to guide the design of measures and indicators to be included in an EIS is presented below (Kelly, 1992b). For a more detailed discussion of methods for selecting measures that reflect organizational objectives, see the section "EIS and Organizational Objectives."

  1. EIS measures must be easy to understand and collect. Wherever possible, data should be collected naturally as part of the process of work. An EIS should not add substantially to the workload of managers or staff.
  2. EIS measures must be based on a balanced view of the organization's objective. Data in the system should reflect the objectives of the organization in the areas of productivity, resource management, quality and customer service.
  3. Performance indicators in an EIS must reflect everyone's contribution in a fair and consistent manner. Indicators should be as independent as possible from variables outside the control of managers.
  4. EIS measures must encourage management and staff to share ownership of the organization's objectives. Performance indicators must promote both team-work and friendly competition. Measures will be meaningful for all staff; people must feel that they, as individuals, can contribute to improving the performance of the organization.
  5. EIS information must be available to everyone in the organization. The objective is to provide everyone with useful information about the organization's performance. Information that must remain confidential should not be part of the EIS or the management system of the organization.
  6. EIS measures must evolve to meet the changing needs of the organization.

Barriers to Effectiveness

There are many ways in which an EIS can fail. Dozens of high profile, high cost EIS projects have been cancelled, implemented and rarely used, or implemented and used with negative results. An EIS is a high risk project precisely because it is intended for use by the most powerful people in an organization. Senior managers can easily misuse the information in the system with strongly detrimental effects on the organization. Senior managers can refuse to use a system if it does not respond to their immediate personal needs or is too difficult to learn and use.

Unproductive Organizational Behaviour Norms

Issues of organizational behaviour and culture are perhaps the most deadly barriers to effective Executive Information Systems. Because an EIS is typically positioned at the top of an organization, it can create powerful learning experiences and lead to drastic changes in organizational direction. However, there is also great potential for misuse of the information. Green, Higgins and Irving (1988) found that performance monitoring can promote bureaucratic and unproductive behaviour, can unduly focus organizational attention to the point where other important aspects are ignored, and can have a strongly negative impact on morale.

The key barrier to EIS effectiveness, therefore, is the way in which the organization uses the information in the system. Managers must be aware of the dangers of statistical data, and be skilled at interpreting and using data in an effective way. Even more important is the manager's ability to communicate with others about statistical data in a non-defensive, trustworthy, and constructive manner. Argyris (1991) suggests a universal human tendency towards strategies that avoid embarrassment or threat, and towards feelings of vulnerability or incompetence. These strategies include:

  • stating criticism of others in a way that you feel is valid but also in a way that prevents others from deciding for themselves
  • failing to include any data that others could use to objectively evaluate your criticism
  • stating your conclusions in ways that disguise their logical implications and denying those implications if they are suggested

To make effective use of an EIS, mangers must have the self-confidence to accept negative results and focus on the resolution of problems rather than on denial and blame. Since organizations with limited exposure to planning and targeting, data-based decision-making, statistical process control, and team-based work models may not have dealt with these behavioural issues in the past, they are more likely to react defensively and reject an EIS.

Technical Excellence

An interesting result from the Vandenbosch & Huff (1988) study was that the technical excellence of an EIS has an inverse relationship with effectiveness. Systems that are technical masterpieces tend to be inflexible, and thus discourage innovation, experimentation and mental model development.

Flexibility is important because an EIS has such a powerful ability to direct attention to specific issues in an organization. A technical masterpiece may accurately direct management attention when the system is first implemented, but continue to direct attention to issues that were important a year ago on its first anniversary. There is substantial danger that the exploration of issues necessary for managerial learning will be limited to those subjects that were important when the EIS was first developed. Managers must understand that as the organization and its work changes, an EIS must continually be updated to address the strategic issues of the day.

A number of explanations as to why technical masterpieces tend to be less flexible are possible. Developers who create a masterpiece EIS may become attached to the system and consciously or unconsciously dissuade managers from asking for changes. Managers who are uncertain that the benefits outweigh the initial cost of a masterpiece EIS may not want to spend more on system maintenance and improvements. The time required to create a masterpiece EIS may mean that it is outdated before it is implemented.

While usability and response time are important factors in determining whether executives will use a system, cost and flexibility are paramount. A senior manager will be more accepting of an inexpensive system that provides 20% of the needed information within a month or two than with an expensive system that provides 80% of the needed information after a year of development. The manager may also find that the inexpensive system is easier to change and adapt to the evolving needs of the business. Changing a large system would involve throwing away parts of a substantial investment. Changing the inexpensive system means losing a few weeks of work. As a result, fast, cheap, incremental approaches to developing an EIS increase the chance of success.

Technical Problems

Paradoxically, technical problems are also frequently reported as a significant barrier to EIS success. The most difficult technical problem -- that of integrating data from a wide range of data sources both inside and outside the organization -- is also one of the most critical issues for EIS users. A marketing vice-president, who had spent several hundred thousand dollars on an EIS, attended a final briefing on the system. The technical experts demonstrated the many graphs and charts of sales results, market share and profitability. However, when the vice-president asked for a graph of market share and advertising expense over the past ten years, the system was unable to access historical data. The project was cancelled in that meeting.

The ability to integrate data from many different systems is important because it allows managerial learning that is unavailable in other ways. The president of a manufacturing company can easily get information about sales and manufacturing from the relevant VPs. Unfortunately, the information the president receives will likely be incompatible, and learning about the ways in which sales and manufacturing processes influence each other will not be easy. An EIS will be particularly effective if it can overcome this challenge, allowing executives to learn about business processes that cross organizational boundaries and to compare business results in disparate functions.

Another technical problem that can kill EIS projects is usability. Senior managers simply have the choice to stop using a system if they find it too difficult to learn or use. They have very little time to invest in learning the system, a low tolerance for errors, and initially may have very little incentive to use it. Even if the information in the system is useful, a difficult interface will quickly result in the manager assigning an analyst to manipulate the system and print out the required reports. This is counter-productive because managerial learning is enhanced by the immediacy of the question - answer learning cycle provided by an EIS. If an analyst is interacting with the system, the analyst will acquire more learning than the manager, but will not be in a position to put that learning to its most effective use.

Usability of Executive Information Systems can be enhanced through the use of prototyping and usability evaluation methods. These methods ensure that clear communication occurs between the developers of the system and its users. Managers have an opportunity to interact with systems that closely resemble the functionality of the final system and thus can offer more constructive criticism than they might be able to after reading an abstract specification document. Systems developers also are in a position to listen more openly to criticisms of a system since a prototype is expected to be disposable. Several evaluation protocols are available including observation and monitoring, software logging, experiments and benchmarking, etc. (Preece et al, 1994). The most appropriate methods for EIS design are those with an ethnographic flavour because the experience base of system developers is typically so different from that of their user population (senior executives).

Misalignment Between Objectives & EIS

A final barrier to EIS effectiveness was mentioned earlier in the section on purpose. As noted there, the powerful ability of an EIS to direct organizational attention can be destructive if the system directs attention to the wrong variables. There are many examples of this sort of destructive reporting. Grant, Higgins and Irving (1988) report the account of an employee working under a misaligned reporting system.

I like the challenge of solving customer problems, but they get in the way of hitting my quota. I'd like to get rid of the telephone work. If (the company) thought dealing with customers was important, I'd keep it; but if it's just going to be production that matters, I'd gladly give all the calls to somebody else.

Traditional cost accounting systems are also often misaligned with organizational objectives, and placing these measures in an EIS will continue to draw attention to the wrong things. Cost accounting allocates overhead costs to direct labour hours. In some cases the overhead burden on each direct labour hour is as much as 1000%. A manager operating under this system might decide to sub-contract 100 hours of direct labor at $20 per hour. On the books, this $2,000 saving is accompanied by $20,000 of savings in overhead. If the sub-contractor charges $5,000 for the work, the book savings are $2,000 + $20,000 - $5,000 = $17,000. In reality, however, the overhead costs for an idle machine in a factory do not go down much at all. The sub-contract actually ends up costing $5,000 - $2,000 = $3,000. (Peters, 1987)

Characteristics of Successful EIS Implementations

Find an Appropriate Executive Champion

EIS projects that succeed do so because at least one member of the senior management team agrees to champion the project. The executive champion need not fully understand the technical issues, but must be a person who works closely with all of the senior management team and understands their needs, work styles and their current methods of obtaining organizational information. The champion's commitment must include a willingness to set aside time for reviewing prototypes and implementation plans, influencing and coaching other members of the senior management team, and suggesting modifications and enhancements to the system.

Deliver a Simple Prototype Quickly

Executives judge a new EIS on the basis of how easy it is to use and how relevant the information in the system is to the current strategic issues in the organization. As a result, the best EIS projects begin as a simple prototype, delivered quickly, that provides data about at least one critical issue. If the information delivered is worth the hassle of learning the system, a flurry of requirements will shortly be generated by executives who like what they see, but want more. These requests are the best way to plan an EIS that truly supports the organization, and are more valuable than months of planning by a consultant or analyst.

One caveat concerning the simple prototype approach is that executive requests will quickly scatter to questions of curiosity rather than strategy in an organization where strategic direction and objectives are not clearly defined. A number of methods are available to support executives in defining business objectives and linking them to performance monitors in an EIS. These are discussed further in the section on EIS and Organizational Objectives below.

Involve Your Information Systems Department

In some organizations, the motivation for an EIS project arises in the business units quite apart from the traditional information systems (IS) organization. Consultants may be called in, or managers and analysts in the business units may take the project on without consulting or involving IS. This is a serious mistake. Executive Information Systems rely entirely on the information contained in the systems created and maintained by this department. IS professionals know best what information is available in an organization's systems and how to get it. They must be involved in the team. Involvement in such a project can also be beneficial to IS by giving them a more strategic perspective on how their work influences the organization.

Communicate & Train to Overcome Resistance

A final characteristic of successful EIS implementations is that of communication. Executive Information Systems have the potential to drastically alter the prevailing patterns of organizational communication and thus will typically be met with resistance. Some of this resistance is simply a matter of a lack of knowledge. Training on how to use statistics and performance measures can help. However, resistance can also be rooted in the feelings of fear, insecurity and cynicism experienced by individuals throughout the organization. These attitudes can only be influenced by a strong and vocal executive champion who consistently reinforces the purpose of the system and directs the attention of the executive group away from unproductive and punitive behaviours.

EIS and Organizational Culture

Henry Mintzberg (1972) has argued that impersonal statistical data is irrelevant to managers. John Dearden (1966) argued that the promise of real-time management information systems was a myth and would never be of use to top managers. Grant, Higgins, and Irving (1988) argue that computerized performance monitors undermine trust, reduce autonomy and fail to illuminate the most important issues.

Many of these arguments against EISs have objective merit. Manager's really do value the tangible tidbits of detail they encounter in their daily interactions more highly than abstract numerical reports. Rumours suggest a future, while numbers describe a past. Conversations are rich in detail and continuously probe the reasons for the situation, while statistics are vague approximations of reality. When these vague approximations are used to intimidate or control behaviour rather than to guide learning, they really do have a negative impact on the organization.

Yet both of these objections point to a deeper set of problems -- the assumptions, beliefs, values and behaviours that people in the organization hold and use to respond to their environment. Perhaps senior managers find statistical data to be irrelevant because they have found too many errors in previous reports?  Perhaps people in the organization prefer to assign blame rather than discover the true root cause of problems.  The culture of an organization can have a dramatic influence on the adoption and use of an Executive Information System.  The following cultural characteristics will contribute directly to the success or failure of an EIS project.

Learning vs Blaming

A learning organization is one that seeks first to understand why a problem occurred, and not who is to blame. It is a common and natural response for managers to try to deflect responsibility for a problem on to someone else. An EIS can help to do this by indicating very specifically who failed to meet a statistical target, and by how much. A senior manager, armed with EIS data, can intimidate and blame the appropriate person. The blamed person can respond by questioning the integrity of the system, blaming someone else, or even reacting in frustration by slowing work down further.

In a learning organization, any unusual result is seen as an opportunity to learn more about the business and its processes. Managers who find an unusual statistic explore it further, breaking it down to understand its components and comparing it with other numbers to establish cause and effect relationships. Together as a team, management uses numerical results to focus learning and improve business processes across the organization. An EIS facilitates this approach by allowing instant exploration of a number, its components and its relationship to other numbers.

Continuous Improvement vs Crisis Management

Some organizations find themselves constantly reacting to crises, with little time for any proactive measures. Others have managed to respond to each individual crisis with an approach that prevents other similar problems in the future. They are engaged in a continual cycle of improving business practices and finding ways to avoid crisis.

Crises in government are frequently caused by questions about organizational performance raised by an auditor, the Minister, or members of the Opposition. An EIS can be helpful in responding to this sort of crisis by providing instant data about the actual facts of the situation. However, this use of the EIS does little to prevent future crises.

An organizational culture in which continual improvement is the norm can use the EIS as an early warning system pointing to issues that have not yet reached the crisis point, but are perhaps the most important areas on which to focus management attention and learning. Organizations with a culture of continuous improvement already have an appetite for the sort of data an EIS can provide, and thus will exhibit less resistance.

Team Work vs Hierarchy

An EIS has the potential to substantially disrupt an organization that relies upon adherence to a strict chain of command. The EIS provides senior managers with the ability to micro-manage details at the lowest levels in the organization. A senior manger with an EIS report who is surprised at the individual results of a front-line worker might call that person directly to understand why the result is unusual. This could be very threatening for the managers between the senior manager and the front-line worker. An EIS can also provide lower level managers with access to information about peer performance and even the performance of their superiors.

Organizations that are familiar with work teams, matrix managed projects and other forms of interaction outside the chain of command will find an EIS less disruptive. Senior managers in these organizations have learned when micro-management is appropriate and when it is not. Middle managers have learned that most interactions between their superiors and their staff are not threatening to their position. Workers are more comfortable interacting with senior managers when the need arises, and know what their supervisor expects from them in such an interaction.

Data-based Decisions vs Decisions in a Vacuum

The total quality movement, popular in many organizations today, emphasizes a set of tools referred to as Statistical Process Control (SPC). These analytical tools provide managers and workers with methods of understanding a problem and finding solutions rather than allocating blame and passing the buck. Organizations with training and exposure to SPC and analytical tools will be more open to an EIS than those who are suspicious of numerical measures and the motives of those who use them.

It should be noted that data-based decision making does not deny the role of intuition, experience, or negotiation amongst a group. Rather, it encourages decision-makers to probe the facts of a situation further before coming to a decision. Even if the final decision contradicts the data, chances are that an exploration of the data will help the decision-maker to understand the situation better before a decision is reached. An EIS can help with this decision-making process.

Information Sharing vs Information Hoarding

Information is power in many organizations, and managers are motivated to hoard information rather than to share it widely. For example, managers may hide information about their own organizational performance, but jump at any chance to see information about performance of their peers.

A properly designed EIS promotes information sharing throughout the organization. Peers have access to information about each other's domain; junior managers have information about how their performance contributes to overall organizational performance. An organization that is comfortable with information sharing will have developed a set of "good manners" for dealing with this broad access to information. These behavioural norms are key to the success of an EIS.

Specific Objectives vs Vague Directions

An organization that has experience developing and working toward Specific, Measurable, Achievable and Consistent (SMAC) objectives will also find an EIS to be less threatening. Many organizations are uncomfortable with specific performance measures and targets because they believe their work to be too specialized or unpredictable. Managers in these organizations tend to adopt vague generalizations and statements of the exceedingly obvious in place of SMAC objectives that actually focus and direct organizational performance. In a few cases, it may actually be true that numerical measures are completely inappropriate for a few aspects of the business. In most cases, managers with this attitude have a poor understanding of the purpose of objective and target-setting exercises. Some business processes are more difficult to measure and set targets for than others. Yet almost all business processes have at least a few characteristics that can be measured and improved through conscientious objective setting. (See the following section on EIS and Organizational Objectives.)

EIS and Organizational Objectives

A number of writers have discovered that one of the major difficulties with EIS implementations is that the information contained in the EIS either does not meet executive requirements, or meets executive requirements, but fails to guide the organization towards its objectives. As discussed earlier, organizations that are comfortable in establishing and working towards Specific, Measurable, Achievable, and Consistent (SMAC) objectives will find it easier to create an EIS that actually drives organizational performance. Yet even these organizations may have difficulty because their stated objectives do not represent all of the things that are important.

Crockett (1992) suggests a four step process for developing EIS information requirements based on a broader understanding of organizational objectives. The steps are: (1) identify critical success factors and stakeholder expectations, (2) document performance measures that monitor the critical success factors and stakeholder expectations, (3) determine reporting formats and frequency, and (4) outline information flows and how information can be used. Crockett begins with stakeholders to ensure that all relevant objectives and critical success factors are reflected in the EIS.

Kaplan and Norton (1992) suggest that goals and measures need to be developed from each of four perspectives: financial, customer, internal business, and innovation and learning. These perspectives help managers to achieve a balance in setting objectives, and presenting them in a unified report exposes the tough tradeoffs in any management system. An EIS built on this basis will not promote productivity while ignoring quality, or customer satisfaction while ignoring cost.

Meyer (1994) raises several questions that should be asked about measurement systems for teams. Four are appropriate for evaluating objectives and measures represented in an EIS. They are:

  • Are all critical organizational outcomes tracked?
  • Are all "out-of-bounds" conditions tracked? (Conditions that are serious enough to trigger a management review.)
  • Are all the critical variables required to reach each outcome tracked?
  • Is there any measure that would not cause the organization to change its behaviour?

In summary, proper definition of organizational objectives and measures is a helpful precondition for reducing organizational resistance to an EIS and is the root of effective EIS use. The benefits of an EIS will be fully realized only when it helps to focus management attention on issues of true importance to the organization.


Implementation of an effective EIS requires clear consensus on the objectives and measures to be monitored in the system and a plan for obtaining the data on which those measures are based. The sections below outline a methodology for achieving these two results. As noted earlier, successful EIS implementations generally begin with a simple prototype rather than a detailed planning process. For that reason, the proposed planning methodologies are as simple and scope-limited as possible.

EIS Project Team

The process of establishing organizational objectives and measures is intimately linked with the task of locating relevant data in existing computer systems to support those measures. Objectives must be specific and measurable, and data availability is critical to measuring progress against objectives.

Since there is little use in defining measures for which data is not available, it is recommended that an EIS project team including technical staff be established at the outset. This cross-functional team can provide early warning if data is not available to support objectives or if senior manager's expectations for the system are impractical.

A preliminary EIS project team might consist of as few as three people. An EIS Project Leader organizes and directs the project. An Executive Sponsor promotes the project in the organization, contributes senior management requirements on behalf of the senior management team, and reviews project progress regularly. A Technical Leader participates in requirements gathering, reviewing plans, and ensuring technical feasibility of all proposals during EIS definition.

As the focus of the project becomes more technical, the EIS project team may be complemented by additional technical staff who will be directly involved in extracting data from legacy systems and constructing the EIS data repository and user interface.

Establishing Measures & EIS Requirements

Most organizations have a number of high-level objectives and direction statements that help to shape organizational behaviour and priorities. In many cases, however, these direction statements have not yet been linked to performance measures and targets. As well, senior managers may have other critical information requirements that would not be reflected in a simple analysis of existing direction statements. Therefore it is essential that EIS requirements be derived directly from interaction with the senior managers who will use the systems. It is also essential that practical measures of progress towards organizational objectives be established during these interactions.

Measures and EIS requirements are best established through a three-stage process. First, the EIS team solicits the input of the most senior executives in the organization in order to establish a broad, top-down perspective on EIS requirements. Second, interviews are conducted with the managers who will be most directly involved in the collection, analysis, and monitoring of data in the system to assess bottom-up requirements. Third, a summary of results and recommendations is presented to senior executives and operational managers in a workshop where final decisions are made.

Interview Format

The focus of the interviews would be to establish all of the measures managers require in the EIS. Questions would include the following:

  1. What are the five most important pieces of information you need to do your job?
  2. What expectations does the Board of Directors have for you?
  3. What results do you think the general public expects you to accomplish?
  4. On what basis would consumers and customers judge your effectiveness?
  5. What expectations do other stakeholders impose on you?
  6. What is it that you have to accomplish in your current position?

Senior Management Workshop

Since considerable variability is expected in the results of these interviews, analysis and synthesis are required to identify recurring themes and important differences of opinion. This information is brought forward to a senior management workshop.

The purpose of the senior management workshop is twofold. First, the workshop will be an opportunity to educate senior management on appropriate use of Executive Information Systems, to address some of the cultural issues raised earlier and to deal directly with resistance to the system. Second, managers at the workshop will be asked to reach agreement on an initial set of measures to be included in the EIS. The education component of the workshop is most effective if integrated with the work of creating measures.

The initial set of measures will be established within a framework derived from the interview process. Three to five categories of measures will be established prior to the workshop, and managers will be asked to select or create three to five technically feasible measures for each category. Each of the proposed measures will be subjected to the questions proposed by Meyer (see EIS and Organizational Objectives above) to determine if they are appropriate.

Technical staff will attend to respond to feasibility questions as they arise, and to improve their understanding of the EIS requirements.

Obtaining Critical Data Linking EIS Measures to Data Sources

Data to support the information requirements of senior managers will likely be dispersed across the organization's information systems and external sources. Some data may not be currently available at all, and collection mechanisms will have to be constructed.

The EIS project team, augmented by technical experts, and working from the requirements established in the senior management workshop will develop a list of required data elements and link them with appropriate data sources. The team will then establish requirements for data extraction from each of these systems and spin off appropriate systems development projects.

EIS Design, Prototyping & Evaluation

After information sources have been established, and projects are underway to permit ongoing extraction of that information, attention will turn to the design of the EIS itself. There are several components to consider.


First, an inventory of computers used by executives must be taken to determine what upgrades are necessary and what hardware limitations will be imposed on the EIS design. Included in this inventory will be an assessment of network storage and communication facilities.

Data Repository

The second component is the design of the data repository in which summary data from all sources will be stored. The design of this repository is critical because it must allow managers to easily extract and explore data along numerous dimensions. Standard relational designs may not be sufficient or practical for this application.

EIS Interface Prototype

A third component is the design of the actual EIS interface that senior managers will interact with. Screens and commands must be exceedingly obvious and easy to use so that senior managers can quickly access the benefits of the system without wasting a lot of time learning how to use it. Ease of use can be ensured by developing a prototype system with "sample" data, and watching senior managers as they interact with the prototype. Two to three iterations of prototype redesign and testing with four senior managers would be sufficient to ensure that the system is easy to use.


Agor, Weston H. (1986). How top executives use their intuition to make important decisions. Business Horizons, January-February 1986, pp. 49-53.

Anderson, Joseph V. (1993). Mind mapping: A tool for creative thinking. Business Horizons, January-February 1993, pp. 41-46.

Argyris, Chris (1991). Teaching Smart People How to Learn. Harvard Business Review, May-June 1991, pp. 99-109.

Barrow, Craig (1990). Implementing an executive information system: seven steps for success. Journal of Information Systems Management, Spring 1990, pp. 41-46.

Bird, Jane (1993). A system for the perplexed. Management Today, April 1993, pp. 71-74.

Bird, Jane (1991). Let your fingers do the walking. Management Today, November 1991, pp. 97-104,.

Carroll, Chris and Larkin, Chris (1992). Executive information technology: A strategic necessity at Motorola Codex. Information Systems Management, Summer 1992, pp. 21-29.

Cook, Gary J. and Swain, Monte R. (1993). A computerized approach to decision process tracing for decision support system design. Decision Sciences, Volume 24, pp. 931-952.

Cooper, Robin and Kaplan, Robert S. (1991). Profit priorities from activity-based costing. Harvard Business Review, May-June 1991, pp. 130-135.

Crockett, Fess (1992). Revitalizing executive information systems. Sloan Management Review, Summer 1992, pg 39-47.

Daft, Richard L. and Lengel, Robert H. (1986). Organizational information requirements, media richness and structural design. Management Science, Volume 32, pp. 554-571.

Davenport, Thomas H., Eccles, Robert G., and Prusak, Laurence. (1992). Information politics. Sloan Management Review, Fall 1992, pp. 53-65.

Dearden, John. (1966). Myth of real-time management information. Harvard Business Review, May-June 1966, pp. 123-132.

Denton, Keith D. (1991). What's wrong with these employees? Business Horizons, September-October 1991, pp. 45-49.

Fisher, Marshall L., Hammond, Janice H., Obermeyer, Walter R., and Raman, Anath. (1994). Making supply meet demand in an uncertain world. Harvard Business Review, May-June 1994, pp. 83-93.

Forgionne, Guisseppi A. (1991). Decision technology systems: A step toward complete decision support. Information Systems Management, Fall 1991, pp. 34-43.

Grant, Rebecca A., Higgins, Christopher A., and Irving, Richard H. (1988). Computerized performance monitors: Are they costing you customers? Sloan Management Reivew, Spring 1988, pp. 39-45.

Hapeshi, Kevin, and Jones, Dylan. (1992). Interactive multimedia for instruction: A cognitive analysis of the role of audition and vision. International Journal of Human-Computer Interaction, Volume 4(1), pp. 79-99.

Huseman, Richard C., and Miles, Edward W. (1988). Organizational communication in the information age: Implications of computer-based systems. Journal of Management, 14(2), pp. 181-204.

Jones, Jack William, and McLeod, Raymond Jr. (1986). The structure of executive information systems: An exploratory analysis. Decision Sciences, Volume 17, pp. 220-249.

Kallman, Ernest A. and McKinnon, William P. (1989). Case study: Evaluating IS by using an executive information system. Journal of Information Systems Management, Fall 1989, pp58-68.

Kaplan, Robert S., and Norton, David P. (1992). The balanced scorecard - measures that drive performance. Harvard Business Review, January-February 1992, pp. 71-79.

Kaplan, Robert S., and Norton, David P. (1993). Putting the balanced scorecard to work. Harvard Business Review, September-October 1993, pp. 134-147.

Keen, Peter G.W. (1980). Decision support systems: Translating analytic techniques into useful tools. Sloan Management Review, Spring 1980, pp. 33-44.

Kelly, Floyd (1992a). Program implementation: Measures of success. Working paper, Canadian Ministry of Housing, December 9, 1992.

Kelly, Floyd (1992b). Management information strategy for Rent Control. Working paper, Canadian Ministry of Housing, August 14, 1992.

King, William R., Premkumar, G., and Ramamurthy, K. (1990). An evaluation of the role and performance of a decision support system in business education. Decision Sciences, Volume 21, pp. 642-659.

Knight, Charles F. (1992). Emerson electric: Consistent profits, consistently. Harvard Business Review, January-February 1992, pp. 57-70.

Krachenberg, A. Richard, Henke, John W. Jr., and Lyons, Thomas F. (1993). The isolation of upper management. Business Horizons, July-August 1993, pp. 41-47.

Leventhal, Laura Marie, Teasley, Barbee Mynatt, Instone, Keith, Rohlman, Diane Schertler, and Farhat, John. (1993). Sleuthing in HyperHomes: an evaluation of using hypertext vs. a book to answer questions. Behaviour and Information Technology, Volume 12, pp. 149-164.

McKinnon, Sharon M., and Bruns, William J. Jr. (1992). The Information Mosaic, Harvard Business School Press, Boston MA.

McLeod, Raymond Jr., and Jones, Jack W. (1986). Making executive information systems more effective. Business Horizons, September-October 1986, pp. 29-37.

Meyer, Christopher. (1994). How the right measures help teams excel. Harvard Business Review, May-June 1994, pp. 95-103.

Mintzberg, Henry (1972). The myths of MIS. California Management Review, Fall 1972, pp. 92-97.

O'Shea, Timothy J. (1989). Low-cost approaches to executive information systems. Journal of Information Systems Management, Spring 1989, pp34-41.

Osborne, David, and Gaebler, Ted. (1992). Reinventing Government. Addison-Wesley, Reading MA.

Posner, Bruce G., and Rothstein, Lawrence R. (1994). Reinventing the business of government: An interview with change catalyst David Osborne. Harvard Business Review, May-June 1994, pp. 132-143.

Preece, Jenny, Rogers, Yvonne, Sharp, Helen, Benyon, David, Holland, Simon, and Carey, Tom (1994). Human Computer Interaction. Addison-Wesley, Wokingham, England.

Rockart, John F. (1988). The line takes the leadership - IS management in a wired society. Sloan Management Review, Summer 1988, pp. 57-64.

Rockart, John F. and Crescenzi, Adam D. (1984). Engaging top management in information technology. Sloan Management Review, Summer 1984, pp. 3-16.

Rockart, John F. and DeLong, David W. (1988). Moments of executive enlightenment. Information Strategy: The Executive's Journal (New York: Auerbach Publishers).

Rockart, John F. and DeLong, David W. (1986). Identifying the attributes of successful executive support system implementation. Transactions of the Sixth International Conference on Decision Support Systems.

Rockart, John F. and Treacy, Michael E. (1982). The CEO Goes On-Line. Harvard Business Review, January-February, 1982.

Rodgers, T. J. (1990). No excuses management. Harvard Business Review, July-August 1990, pp. 84-98.

Rodgers T. J., Taylor, William, and Foreman, Rick. (1992). No Excuses Management. Doubleday, New York NY.

Sabourin, Helen (1992). Resource management information. Defence 2000 Achievements Report, pp. 44-47.

Schaffer, Robert H. (1974). Demand better results - and get them. Harvard Business Review, November-December 1974.

Scott, Derek. (1993). Visual search in modern human-computer interfaces. Behaviour and Information Technology, Volume 12, pp. 174-189.

Stevenson, Mark (1994). He sees all, he knows all. Canadian Business, Spring 1994, pp. 30-35.

Tan, Joseph K. H. and Benbasat, Izak. (1993). The effectiveness of graphical presentation for information extraction: A cumulative experimental approach. Decision Sciences, Volume 24, pp. 167-191.

Umanath, Narayan S., Scamell, Richard W., and Das, Sidhartha R. (1990). An examination of two screen/report design variables in an information recall context. Decision Sciences, Volume 21, pp. 216-240.

Vandenbosch, Betty. (1993). Executive Support Systems impact viewed from a learning perspective. Doctoral Dissertation, University of Western Ontario.

Vandenbosch, Betty, and Huff, Sid L. (1992). Executive support systems and management learning. Business Quarterly, Autumn 1992, pp. 33-38.

Watson, Hugh J. and Frolick, Mark (1992). Executive information systems: Determining information requirements. Information Systems Management, Spring 1992, pp. 37-43.

Watson, Hugh J., and Rainer, R. Kelly Jr. (1991). A manager's guide to executive support systems. Business Horizons, March-April 1991, pp. 44-50.

Watson, Hugh J., Rainer, R. Kelly, and Houdeshel, George (1992). Executive Information Systems: Emergence, Development, Impact. (New York: John Wiley & Sons Inc.)

You can contact Floyd Kelly, the author of this paper, by e-mail at This is a classic web review article that "was first placed on the web around 1994". The original citation was Kelly, Floyd, "Executive Information Systems", Nestek Computer Consulting, October 1994 at URL In 1999, Kelly started a website at This paper was featured on that website until it ceased operation in mid-October 2002. This version of the paper was downloaded on October 9, 2002 from

Floyd Kelly provided permission to archive this article and feature it at DSSResources.COM on Tuesday, October 8, 2002. This article was posted at DSSResources.COM on November 7, 2002.