Communications of the ACM
Home/Magazine Archive/August 2003 (Vol. 46, No. 8)/How Cios Manage Their Superiors' Expectations/Full Text
Communications of the ACM
How Cios Manage Their Superiors' Expectations
CIOs must manage their superiors' expectations regarding system delivery and performance if they want to survive [3] and progress as strategic executives and agents of change [1, 2, 4]. This article describes how some of them do it. Insights into CIO expectation management techniques were obtained through content analysis of transcripts of structured interviews with 15 CIOs from the Chicago area.
The sample included CIOs from transportation companies, the credit card industry, metal suppliers, a local hospital, public telecommunications, commercial food services, consumer products, insurance, municipal and nonprofit organizations, international engineering, consulting, national commercial real estate brokerage, and pharmaceutical companies. Over half were Fortune 500 firms. The first of four open-ended questions (all derived from a series of focus group meetings) directly addressed expectation management; the second, the relationship of this management to CIO survival. The third question addressed system performance measurement and expectation, and the final question sought to flesh out the CIO-superior relationship, with a focus on communication and its link to expectation management.
What are your techniques for shaping initial systems expectations, keeping them on course during system development, and reshaping them, if necessary, to fit unanticipated but probable outcomes?
Perspectives and techniques vary depending on the size of the organization and the type of relationship the CIO has with his or her superior. Compared to smaller firms, larger and more mature firms often have more formal venues for structuring communication during strategic formulation and through project management. Steering committees and the like are very useful for establishing initial expectations, as well as for monitoring and adjusting to project progress. Creating teams, distributing project ownership, and using policies that place responsibilities for establishing metrics, outcomes, and accountability with the business unit that sponsors the project are solid project management techniques that factor in extensive communication with stakeholders. Establishing appropriate expectations and keeping them aligned with a project can be further aided by designing some flexibility into the project's management. As one CIO points out, "Good project management can also include preparing options for different points in the development cycle. Initial strategies can be reexamined in light of these or other options that may become viable as things move ahead and come into better focus."
Politically astute CIOs know how to make their boss look good. CEOs should be strongly linked to high-profile, high-return projects, and less so to more modest efforts.
The amount of credibility, access to the superior, and the extant formal procedures for specifying system expectations also bear upon the emphasis CIOs place on creating shared and realistic expectations in the formative stages of a project. Some have bosses who trust them a great deal and have little need or time for updates. These CIOs often have relatively free reign and are usually less susceptible to trouble from unmet expectations of peers or other stakeholders. Those with a more distant relationship or who act in a more consultative role tend to place greater emphasis on surfacing assumptions, clarifying expectations, identifying accountability, and creating joint ownership/buy-in at the beginning of a project.
Which project management techniques and organizational policies and structures will be used also reveals in part the strategic importance of a firm's IT functions. "CIOs manage business projects that may have a heavy IT component, not IT projects per se," remarked a CIO of a Fortune 500 insurance company. In his case, the strategic criticality of IT permeated the entire organization, defining his role as an executive advisor and his department as strategic enablers rather than isolated specialists.
While CIOs in larger organizations rely on informal communication channels as well as these more formal ones, information executives in smaller firms often lack the formal procedures and structures, and place much more emphasis on informal communication. In either case, the frequency and type of communication necessary is in turn dictated to some degree by the relationship the CIO has with his or her superior. This is often a matter of trust, based on length and quality of service, professional reputation, communication style of both parties, and access. Says one CIO of a small firm, "Expectation adjustment is touchy-feely in a small organization, heavily reliant on informal communication with little reliance on formal project management."
Project characteristics also affect expectations. First, some smaller firms may rarely develop their own systems, preferring to purchase off-the-shelf industry standard solutions. In this case, although the CIO and project sponsor still bear responsibility, some of the expectations and accountability shift to the vendor. Second, serious expectations can arise around the development process, not the final system. According to one CIO, "Stakeholders often form inappropriate expectations about the system development process, not the system itself or its performance outcomes. These usually center around time, money, and other resources that the project ultimately requires."
Describe how you think expectation management is linked to CIO tenure or survival.
Our executives definitely agree that expectation management is strongly linked to their career viability. "There is a direct and strong linkage," states a CIO who worked in a variety of companies. "One must manage expectations upward, downward, and laterally, because the boss will eventually hear about unmet expectations from any level." However, the CIOs acknowledged that the sources, accuracy, or effects of expectations cannot always be controlled. "Expectations can be shaped randomly," says one interviewee, alluding to the infamous "seat-pocket dilemma" where an underinformed superior reads about the fantastic benefits of the latest system in an airline magazine. "Outcomes from simple, easy systems may appear more impressive to stakeholders than those from complex, difficult systems," says another. Adds one CIO, "Performing as expected or not, perceptions are often not accurate. Evaluation at the executive level is often vague and impressionistic. Fairness has nothing to do with it." One also remarked that bosses have a very human tendency to remember the rare failure and forget the continuing successes: "That the lights are always on is taken for granted." The impact of unmet expectations is also related to the significance and visibility of the disappointing projecta little invisible setback is obviously less dangerous than a rogue white elephant. The nature of the expectation is also important: "Surprises lead to failuresunanticipated outcomes, scope creep, cost overruns. Dollar surprises are especially bad," offered one respondent.
Trust through track record is indispensable; a relationship of trust with sufficient communication is the first line of defense against unrealistic or unmet expectations. One CIO of a large corporation had worked for four CEOs within one year before being let go. He never had the opportunity to develop a track record with any of these superiors, and the last one brought his own CIO with whom he did have a solid professional history.
Systems outcome measures that address quantifiable efficiency like speed of processing are typically well understood and commonly used, but their usefulness is often limited. Other important measures of systems outcomes that address organizational/strategic effectiveness, value, and popularity may not be well understood, widely used, or not used appropriately or consistently by organizations. How do these various measures, their shortcomings or misuse, affect the systems expectations and your ability to manage them?
Softer measures such as organizational effectiveness and customer or user satisfaction are certainly important, but most of the respondents are not particularly concerned that they are sometimes difficult to measure. Rather, traditional quantitative metrics such as processing speed and power are seen to ultimately translate into how quickly and how well users can do their work. Says one CIO, "Speed, transparency, and flexibility enables work to be done faster, better. This translates onto attitude/satisfaction surveys. Bad technologies tend to show up. Nothing is fast enough."
Trust through track record is indispensable; a relationship of trust with sufficient communication is the first line of defense against unrealistic or unmet expectations.
Good project planning and management include careful metric selection. "Follow the logical path from systems back up to strategic objectives. A balanced score card approach can help show linkages from project to objective," suggests a CIO of a transportation firm. Programming some flexibility in how performance will be measured is also a smart move, according to the CIO of a consumer foods firm: "Asking "what-if" questions is important during planning stages. Huge increases in demand, for example, can obsolete some metrics and make others more appropriate. Build optional metrics and tactics into the project plan and, when necessary, the organizational strategy. Exercise these when needed at different points of the development cycle. The ability to change metrics during system development is often a function of the organizational culture's ability to change as needed (and in a timely fashion)."
Metrics also have their political component, being ignored or selected to make the most favorable case for a new system. They may also be selected because they are what the boss is most comfortable with, rather than their technical appropriateness. "Inappropriate metrics can come with an overzealous "selling" of a proposed system. A lack of clarity can be intentional," claims one executive. Metrics can be ignored once in place, with no follow-up auditing because of time and resource constraints, or politics. Citing his experiences in different companies, one CIO offered, "Some stakeholders will not want a system evaluated, as the system could be a technical success but a business failure (or vice versa). Organizations rarely commit the time and expense to evaluate a running system if it is functioning basically as expected and is delivered on time and on budget. Only large organizations audit systems, and then, only their larger ones. Rarely are underperforming systems disassembled unless they are complete flops."
Measures can also be somewhat myopic, focusing on potential system performance without seeing the larger context of organizational change that may need to occur for the system to be most beneficial. This can be problematic, as one CIO observed, because "Major systems often require changes in organizational structure, including roles, resources, and rewards. When cultural infrastructure needs to be changed during a system development, mismeasurement and/or the selection of inappropriate metrics often occurs." Evaluator orientation is also important. Some metrics should bear upon on what a firm's clients deem important. Finally, performance objectives should be realistically tied to industry standards. Bells, whistles, and aesthetics might be very important for a clothing retailer's Web site, but have little added value in business-to-business e-commerce between, say, suppliers in the bulk fuel industry.
Some experts suggest that CEO involvement in IS management is optimal when it is chiefly symbolic, for example, demonstrating support for the IS agenda. Direct participation, for example through IT department meetings, is less beneficial. If there is a continuum of CEO involvement in IS affairs, with complete detachment on one end and micromanagement on the other, how would you describe the level of involvement you would like to have from your CEO in your work?
The amount and frequency of contact between these two executives is rarely at the discretion of the CIO, who must use his/her communication time wisely. Contact time is useful for educating the boss. "Be proactive, prepared to show them what comes next, rather than react to what grabs their fancy at any given time. Take them deeper than you think you can, and show them that the world is not simple," suggests one of our group. As with the aforementioned seat-pocket dilemma, a little knowledgewhen manifested as unrealistic expectations about a technology or systemcan spell real trouble. Although all our respondents have great respect for their superiors and the work they do, they disclosed it can be wearing when the underinformed boss wants your ear for tech talk. "Close communication is painful, as bosses think they know more than they usually do," remarked one CIO, candidly.
The amount of involvement with IT work needed from a CEO varies with the type of project. Some projects require a great deal of interaction with VPs from other business areas, and that type of contact can be smoothed by an involved (and informed) CEO. Systems that go beyond the bounds of the organization (for example, e-commerce or ERP) can also benefit from the CEO's active sponsorship. According to one CIO: "IT has become a strategic enabler and the CIO a strategic contributor. Many IT-based projects have a greater external focus, and require more CEO involvement. Internally focused projects require less CEO involvement." Politically astute CIOs know how to make their boss look good. CEOs should be strongly linked to high-profile, high-return projects, and less so to more modest efforts. To act accordingly, claims one IT executive, "One must understand how IT affects the CEO's performance and personal bottom line."
Here again, the amount of involvement that exists or is desired is to some degree based on the CIO's track record and history with his or her boss. Those with a long and positive relationship have typically built up a solid degree of trust, so there is less need for any monitoring by the top executive. On the other hand, a CIO who can trust that his or her boss is sufficiently IT-savvy will likely need to devote less time to the education side of expectation management compared to a less fortunate counterpart.
These responses give insight into the complexity of the CIO environment as well as CIOs' strategies for managing expectations. The figure here is a preliminary descriptive model based on this research that groups and summarizes these various factors. The boxes on the left in the figure represent the parts of the CIO's professional environment that bear upon his or her selection of expectation management techniques on the right side of the model.
Organizational characteristics. Important organizational characteristics include size, age, and industry. Typically, the larger and older the firm, the more likely it is to have various formal communication venues (for example, IT steering committees) that the CIO can take advantage of. The more mature organization may have a longer history with IT projects and better-informed senior management. There may also be various methods for educating IT executives. The industry sector plays a role due to variations in IT intensity. Some sectors have leveraged IT strategically for years, and for others, it has been seen as a cost center and not as a source of competitive advantage; this also speaks to the general IT literacy of the organization and its leaders.
Project characteristics. A high-profile system with a high probability of success (and whose virtues can be easily perceived) should be tied to the boss. Low-profile systems or those that need an engineer to determine their merits don't do the superior much political good. The amount of inclusion thus drives how much "upward" expectation management the CIO must perform. Most systems with an external focus (for example, e-commerce) require greater CEO involvement and visibility, and more expectation management than those with an internal focus (for example, inventory). Expectation management-driven communication is also preferable when a project is likely to present some surprises: metric selection and milestone timing are critical at the onset to provide options later. Other techniques such as assumption surfacing and creating a shared vision might be useful only at initial stages.
CIO-CEO relationship. Trust is probably the most important aspect of the CIO's relationship with his or her superior, but it takes time to earn. IT executives, however, may find themselves working for a new boss, where trust is not yet well established. Perhaps worse than having no track record, though, is to work for a familiar but unenlightened boss. Such a person may be more susceptible to inflated claims often made about emerging technologies, and require some proactive education from the IT department or CIO directly. In many firms, access to the chief executive (or other superior) may be extremely limited, placing a premium on effective communication. It may also be the limited access is a product of an organizational culture that does not highly prioritize IT. This increases the likelihood that the CEO is underinformed about IT, and increases the need for education as a form of expectation management, even if only informally and occasionally delivered by the information executive.
Expectation management techniques. Beyond educational dialogue and metric selection, our respondents rely on traditional project management techniques and additional mechanisms woven into these techniques such as milestones that require review of the project's progress, evaluation of objectives, and consideration of alternative objectives or directions for the project. Along with other mechanisms such as a participatory management style, assumption surfacing, and vision sharing, each of these techniques was intended to give the CIO greater avenues for communication as well as legitimate and effective ways of dealing with unforeseen and/or uncontrollable changes in the project's trajectory.
Conclusion
The interviewees intimately understand the need for expectation management and have survived in part through developing this skill. What techniques they use depend a great deal on contextual factors including characteristics of their organization, system project, and their boss. These environmental characteristics are dynamic and sometimes random. Bosses come and go, CIOs change jobs, projects range from the mundane and obscure to the experimental and high profile and may change on the fly, and organizations may be involved in cultural shifts while any of this is going on. Although other peer-level executive managers must deal with satisfying their boss's expectations for new projects, they probably do not have entirely new generations of technology (with new sets of unknowns and inflated claims) facing them on a regular basis. CIOs have a tough time keeping their tough jobs. Survivors, though, work to see that the boss's expectations are realistic, on track, and well satisfied.
References
1. Earl, M.J. A quantity of qualities. CIO Mag., (Sept.15, 1997).
2. Markus, M.L., and Benjamin, R.I. Change agentry the next IS frontier. MIS Q. (Dec. 1996), 385-407.
3. Pastore, R. Survival of the fittest. CIO Mag. (Nov. 1, 1996).
4. Ross, J.W. and Feeny, D.F. The evolving role of the CIO. Framing the Domains of IT Management: Projecting the Future Through the Past. R. Zmud, Ed. Pinnaflex, Cincinnati, OH, 2000.
Footnotes
This research was funded by The Center for Research in Information Management, College of Business Administration, The University of Illinois at Chicago.
©2003 ACM 0002-0782/03/0800 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2003 ACM, Inc.
No entries found