A corporate president put a senior executive in charge of a failing operation. His only directive was “Get it in the black.” Within two years of that injunction, the new executive moved the operation from a deficit position to one that showed a profit of several million. Fresh from his triumph, the executive announced himself as a candidate for a higher-level position, and indicated that he was already receiving offers from other companies.
The corporate president, however, did not share the executive’s positive opinions of his behavior. In fact, the president was not at all pleased with the way the executive had handled things.
Naturally the executive was dismayed, and when he asked what he had done wrong, the corporate president told him that he had indeed accomplished what he had been asked to do, but he had done it single-handedly, by the sheer force of his own personality. Furthermore, the executive was told, he had replaced people whom the company thought to be good employees with those it regarded as compliant. In effect, by demonstrating his own strength, he had made the organization weaker. Until the executive changed his authoritarian manner, his boss said, it was unlikely that he would be promoted further.
Implicit in this vignette is the major fault in performance appraisal and management by objectives—namely, a fundamental misconception of what is to be appraised.
Performance appraisal has three basic functions: (1) to provide adequate feedback to each person on his or her performance; (2) to serve as a basis for modifying or changing behavior toward more effective working habits; and (3) to provide data to managers with which they may judge future job assignments and compensation. The performance appraisal concept is central to effective management. Much hard and imaginative work has gone into developing and refining it. In fact, there is a great deal of evidence to indicate how useful and effective performance appraisal is. Yet present systems of performance appraisal do not serve any of these functions well.
As it is customarily defined and used, performance appraisal focuses not on behavior but on outcomes of behavior. But even though the executive in the example achieved his objective, he was evaluated on how he attained it. Thus, while the system purports to appraise results, in practice, people are really appraised on how they do things—which is not formally described in the setting of objectives, and for which there are rarely data on record.
In my experience, the crucial aspect of any manager’s job and the source of most failures, which is practically never described, is the “how.” As long as managers appraise the ends yet actually give greater weight to the means, employ a static job description base which does not describe the “how,” and do not have support mechanisms for the appraisal process, widespread dissatisfaction with performance appraisal is bound to continue. In fact, one personnel authority speaks of performance appraisal as “the Achilles heel of our profession…”1
Just how these inadequacies affect performance appraisal systems and how they can be corrected to provide managers with realistic bases for making judgments about employees’ performance is the subject of this article.
Inadequacies of appraisal systems
It is widely recognized that there are many things inherently wrong with most of the performance appraisal systems in use. The most obvious drawbacks are:
- No matter how well defined the dimensions for appraising performance on quantitative goals are, judgments on performance are usually subjective and impressionistic.
- Because appraisals provide inadequate information about the subtleties of performance, managers using them to compare employees for the purposes of determining salary increases often make arbitrary judgments.
- Ratings by different managers, and especially those in different units, are usually incomparable. What is excellent work in one unit may be unacceptable in another in the same company.
- When salary increases are allocated on the basis of a curve of normal distribution, which is in turn based on rating of results rather than on behavior, competent employees may not only be denied increases, but may also become demotivated.2
- Trying to base promotion and layoff decisions on appraisal data leaves the decisions open to acrimonious debate. When employees who have been retired early have complained to federal authorities of age discrimination, defendant companies have discovered that there were inadequate data to support the layoff decisions.
- Although managers are urged to give feedback freely and often, there are no built-in mechanisms for ensuring that they do so. Delay in feedback creates both frustration, when good performance is not quickly recognized, and anger, when judgment is rendered for inadequacies long past.
- There are few effective established mechanisms to cope with either the sense of inadequacy managers have about appraising subordinates, or the paralysis and procrastination that result from guilt about playing God.
Some people might argue that these problems are deficiencies of managers, not of the system. But even if that were altogether true, managers are part of that system. Performance appraisal needs to be viewed not as a technique but as a process involving both people and data, and as such the whole process is inadequate.
Recognizing that there are many deficiencies in performance appraisals, managers in many companies do not want to do them. In other companies there is a great reluctance to do them straightforwardly. Personnel specialists attribute these problems to the reluctance of managers to adopt new ways and to the fear of irreparably damaging their subordinates’ self-esteem. In government, performance appraisal is largely a joke, and in both private and public enterprise, merit ratings are hollow.3
One of the main sources of trouble with performance appraisal systems is, as I have already pointed out, that the outcome of behavior rather than the behavior itself is what is evaluated. In fact, most people’s jobs are described in terms that are only quantitatively measurable; the job description itself is the root of the problem.
The static job description
When people write their own job descriptions (or make statements from which others will write them) essentially they define their responsibilities and basic functions. Then on performance appraisal forms, managers comment on these functions by describing what an individual is supposed to accomplish. Forms in use in many companies today have such directions as:
1. “List the major objectives of this person’s job that can be measured qualitatively or quantitatively.”
2. “Define the results expected and the standards of performance—money, quantity, quality, time limits, or completion dates.”
3. “Describe the action planned as a result of this appraisal, the next steps to be taken—reevaluation, strategy, tactics, and so on.”
4. “List the person’s strong points—his assets and accomplishments—and his weak points—areas in which improvement is needed. What are the action plans for improvement?”
In most instances the appraiser is asked to do an overall rating with a five point scale or some similar device. Finally, he is asked to make a statement about the person’s potential for the next step or even for higher-level management.
Nowhere in this set of questions or in any of the performance appraisal systems I have examined is anything asked about how the person is to attain the ends he or she is charged with reaching.
While some may assert that the ideal way of managing is to give a person a charge and leave him or her alone to accomplish it, this principle is oversimplified both in theory and practice. People need to know the topography of the land they are expected to cross, and the routes as perceived by those to whom they report.
Every manager has multiple obligations, not the least of which are certain kinds of relationships with peers, subordinates, and various consumer, financial, government, supplier, and other publics. Some of these are more important than others, and some need to be handled with much greater skill and aplomb than others. In some situations a manager may be expected to take a vigorous and firm stand, as in labor negotiations; in others he may have to be conciliative; in still others he may even have to be passive. Unless these varied modes of expected behavior are laid out, the job description is static. Because static job descriptions define behavior in gross terms, crucially important differentiated aspects of behavior are lost when performance appraisals are made.
For example, in one of the more progressive performance appraisal systems, which is used by an innovative company, a manager working out his own job description prepares a mission or role statement of what he is supposed to do according to the guide which specifically directs him to concentrate on the what and the when, not on the why and the how.4 The guide instructs him to divide his mission into four general areas: (1) innovation, (2) problem solving, (3) ongoing administration, and (4) personal.
In still another company, a manager appraising a subordinate’s performance is asked to describe an employee’s accomplishments, neglected areas, goals, and objectives. The manager is told that he is to recognize good work, suggest improvement, get agreement on top priority elements of the task, clarify responsibility, verify and correct rumors, and talk about personal and long-range goals.
In another company’s outstanding performance appraisal guide, which reflects great detail and careful consideration, the categories are: work, effectiveness with others, problem solving, decision making, goal setting, organizing and planning, developing subordinates, attending to self-development, and finding initiatives. Each of these categories is broken down into example statements such as: “exhibits high level of independence in work”; “identifies problems and deals with them”; “appropriately subordinates departmental interest to overall company goal”; or “gives people genuine responsibility, holds them accountable, and allows them freedom to act.”
Some personnel researchers have advocated role analysis techniques to cope with static job descriptions, and this is a step in the right direction.5
But even these techniques are limited because they lean heavily on what other people—supervisors, subordinates, peers—expect of the manager. These expectations are also generalized; they do not specify behavior.
Nowhere in these examples is an individual told what behavior is expected of him in a range of contexts. Who are the sensitive people with whom certain kinds of relationships have to be maintained? What are the specific problems and barriers? What have been the historic manufacturing blunders or frictions? How should union relationships and union leaders be dealt with? What are the specific integrative problems to be resolved and what are the historical conflicts? These and many more similar pieces of behavior will be the true bases on which a person will be judged, regardless of the questions an appraisal form asks.
Static job descriptions are catastrophic for managers. Job proficiency and goal achievement usually are necessary but not sufficient conditions for advancement; the key elements in whether one makes it in an organization are political. The collective judgments made about a person, which rarely find their way into performance appraisals, become the social web in which he or she must live. Therefore, when a person is placed in a new situation, whether in a different geographical site, at a different level in the hierarchy, or in a new role, he must be apprised of the subtleties of the relationships he will have with those who will influence his role and his career. Furthermore, he must be helped to differentiate the varied kinds of behavior required to succeed.
Some people develop political diagnostic skill very rapidly; often, however, these are people whose social senses enable them to move beyond their technical and managerial competence. And some may be out and out manipulative charlatans who succeed in business without really trying, and whose promotion demoralizes good people. But the great majority of people, those who have concentrated heavily on their professional competence at the expense of acquiring political skill early, will need to have that skill developed, ideally by their own seniors. That development process requires: (1) a dynamic job description, (2) a critical incident process, and (3) a psychological support system.
Dynamic job description
If a static job description is at the root of the inadequacies of performance appraisal systems, what is needed is a different kind of job description. What we are looking for is one that amplifies statements of job responsibility and desired outcome by describing the emotional and behavioral topography of the task to be done by the individual in the job.
Psychologists describe behavior in many ways, each having his or her own preferences. I have found four major features of behavior to be fundamentally important in a wide range of managerial settings. These features have to do with how a person characteristically manages what some psychologists call aggression, affection, dependency, and also the nature of the person’s ego ideal.6
Using his preferred system, one can begin formulating a dynamic job description by describing the characteristic behavior required by a job. This is what these terms mean with respect to job descriptions:
1. How does this job require the incumbent to handle his aggression, his attacking capacity?
Must he or she vanquish customers? Must he hold on to his anger in the face of repeated complaints and attacks from others? Will she be the target of hostility and, if so, from whom? Must he give firm direction to others? Must she attack problems vigorously, but handle some areas with great delicacy and finesse? Which problems are to be attacked with vigor and immediacy and which coolly and analytically?
2. How does this job require the incumbent to manage affection, the need to love and to be loved?
Is the person required to be a socially friendly leader of a close-knit work group? Should the person work closely and supportively with subordinates for task accomplishment? Is the task one in which the person will have to be content with the feeling of a job well done, or is it one which involves more public display and recognition? Will he be obscure and unnoticed, or highly visible? Must she lavish attention on the work, a product, a service, or customers? Must he be cold and distant from others and, if so, from whom?
3. How does this job require the incumbent to manage dependency needs?
Will the individual be able to lean on others who have skill and competencies, or will he have to do things himself? How much will she be on her own and in what areas? How much support will there be from superiors and staff functions? How well defined is the nature of the work? What kind of feedback provisions are there? What are the structural and hierarchical relationships? How solid are they and to whom will the person turn and for what? With which people must he interact in order to accomplish what he needs to accomplish, and in what manner?
4. What ego ideal demands does this job fulfill?
If one does the task well, what are the gratifications to be gained? Will the person make a lot of money? Will he achieve considerable organizational and public recognition? Will she be eligible for promotion? Will he feel good about himself and, if so, in what ways? Why? Will she acquire a significant skill, an important element of reputation, or an organizational constituency? Will he acquire power?
Individuals may be described along the same four dynamic dimensions: How does this person characteristically handle aggression? How does he or she characteristically handle affection? How does he or she characteristically handle dependency needs? What is the nature of his or her ego ideal?
Once the subtleties of the task are defined and individuals described, people may be matched to tasks. I am not advocating a return to evaluation of personality traits. I am arguing for more dynamic conception of the managerial role and a more dynamic assessment of an employee’s characteristics. And only when a person’s behavior is recognized as basic to how he performs his job will performance appraisal systems be realistic.
Critical incident process
Having established a dynamic job description for a person, the next step is to evolve a complementary performance appraisal system that will provide feedback on verifiable behavior, do so in a continuous fashion, and serve coaching-, promotion-, and salary-data needs.
Ideally, a manager and his subordinate will have defined together the objectives to be attained in a certain job, and the criteria by which each will know that those objectives have been attained, including the more qualitative aspects of the job. Then they will have spelled out the subtleties of how various aspects of the job must be performed. They will in this way have elaborated the behavioral requirements of the task.
In order for performance appraisal to be effective for coaching, teaching, and changing those aspects of an employee’s behavior that are amenable to change, an employee needs to know about each piece of behavior that is good, as well as that which for some reason is not acceptable or needs modification. Such incidents will occur randomly and be judged randomly by his manager.
So that there will be useful data, the manager needs to quickly write down what he has said to the subordinate, describing in a paragraph what the subordinate did or did not do, in what setting, under what circumstances, about what problem. This information forms a behavioral record, a critical incident report of which the subordinate already has been informed and which is now in his folder, open to his review. Examples of two incidents can be found in the sidebar, “Examples of Critical Incidents.”
This critical incident technique is not new.7 In the past it has been used largely for case illustrations and, in modified forms, has been suggested as a method for first-level supervisors to evaluate line employees. Supervisors already record negative incidents concerning line employees because warnings and disciplinary steps must be documented. However, efforts to develop scales from critical incidents for rating behavior have not worked well.8 Behavior is too complex to be scaled along a few dimensions and then rated.
But instead of scaling behavior, one might directly record the behavior of those being appraised, and evaluate it at a later date. There are other good reasons for adopting this technique as well. At last, here is a process that provides data to help managers perform the basic functions of performance appraisal systems—namely, provide feedback, coaching, and promotion data. Another plus is that recorded data live longer than the manager recording them.
Here is how behavioral data might be put to use in the critical incident process:
1. Feedback data.
When there is a semiannual or annual review, an employee will have no surprises and the manager will have on paper what he is using as a basis for making his summary feedback and appraisal. Because the data are on record, an employee cannot deny having heard what was said earlier, nor must the manager try to remember all year what have been the bases of his judgments.
Also, as each critical incident is recorded, over time there will be data in an individual’s folder to be referred to when and if there are suits alleging discrimination. Critical incidents of behavior, which illustrate behavior patterns, will be the only hard evidence acceptable to adjudicating bodies.
2. Coaching data.
When employees receive feedback information at the time the incident occurs, they may be able to adapt their behavior more easily. With this technique, the employee will receive indications more often on how he is doing, and will be able to correct small problems before they become large ones. Also, if the employee cannot change his behavior, that fact will become evident to him through the repetitive critical incident notes. If the employee feels unfairly judged or criticized, he may appeal immediately rather than long after the fact. If there are few or no incidents on record, that in itself says something about job behavior, and may be used as a basis for discussion. In any event, both manager and employee will know which behavior is being appraised.
3. Promotion data.
With such an accumulation of critical incidents, a manager or the personnel department is in a position to evaluate repeatedly how the person characteristically manages aggression, affection, and dependency needs, and the nature of his ego ideal. These successive judgments become cumulative data for better job fit.
When a person is provided continuously with verifiable information, including when he has been passed over for promotion and why, he is able to perceive more accurately the nuances of his behavior and his behavioral patterns. Thus, when offered other opportunities, the employee is in a better position to weigh his own behavioral configurations against those required by the prospective job. A person who knows himself in this way will be more easily able to say about a given job, “That’s not for me.” He will see that the next job in the pyramid is not necessarily rightfully his. In recognizing his own behavioral limitations he may save himself much grief as well as avoid painful difficulty for his superiors and the organization.
But the most important reason for having such information is to increase the chances of success of those who are chosen for greater responsibility. In most personnel folders there is practically no information about how a manager is likely to do when placed on his own. Data about dependency are noticeably absent, and many a shining prospect dims when there is no one to support him in a higher-level job. Managements need to know early on who can stand alone, and they cannot know that without behavioral information.
4. Long-term data.
Frequently, new managers do not know their employees and all too often have little information in the folder with which to appraise them. This problem is compounded when managers move quickly from one area to another. For his part, the employee just as frequently has to prove himself to transient bosses who hold his fate in their hands but know nothing of his past performance. With little information, managers feel unqualified to make judgments. With the critical incident process, however, managers can report incidents which can be summarized by someone else.
Some may object to “keeping book” on their people or resist a program of constant reviews and endless reports—both extreme views. Some may argue that supervisors will not follow the method. But if managers cannot get raises for or transfer employees without adequate documentation, they will soon learn the need to follow through. The critical incident process compels superiors to face subordinates, a responsibility too many shirk.
While it might seem difficult to analyze performance in terms of aggression, affection, dependency, the ego ideal, or other psychological concepts, to do so is no different from learning to use economic, financial, or accounting concepts. Many managers already talk about these same issues in other words, for example: “taking charge” versus “being a nice guy”; “needing to be stroked” versus the “self-starter”; “fast track” versus the “shelf-sitter.” A little practice, together with support mechanisms, can go a long way.
Performance appraisal cannot be limited to a yearly downward reward-punishment judgment. Ideally, appraisal should be a part of a continuing process by which both manager and employee may be guided. In addition, it should enhance an effective superior-subordinate relationship.
To accomplish these aims, performance appraisal must be supported by mechanisms that enable the manager to master his inadequacies and to cope with his feelings of guilt; have a record of that part of his work that occurs outside the purview of his own boss (e.g., task force assignments which require someone to appraise a whole group); and modify those aspects of his superior’s behavior which hamper his performance. All of this requires an upward appraisal process.
1. Managing the guilt.
The manager’s guilt about appraising subordinates appears when managers complain about playing God, about destroying people. A great crippler of effective performance appraisal is the feeling of guilt, much of which is irrational, but which most people have when they criticize others.9 Guilt is what leads to the fear of doing appraisals. It is the root of procrastination, of the failure to appraise honestly, and of the overreaction which can demolish subordinates.
Fortunately, there are group methods for relieving guilt and for helping managers and supervisors understand that critical importance, indeed the necessity, of accurate behavioral evaluations. One way is by having people together at the same peer level discuss their problems in appraisal and talk about their feelings in undertaking the appraisal task. In addition, rehearsals of role playing increase a manager’s sense of familiarity and competence and ease his anxiety.
In fact, a five-step process, one step per week for five weeks, can be extremely helpful:
- Week one: Group discussion among peers (no more than 12) about their feelings about appraising subordinates.
- Week two: Group discussions resulting in advice from each other on the specific problems that each anticipates in appraising individuals.
- Week three: Role playing appraisal interviews.
- Week four: Actual appraisals.
- Week five: Group discussion to review the appraisals, problems encountered, both anticipated and unanticipated, lessons learned, and skill needs that may have surfaced.
2. Group appraisal.
By group appraisal, I do not mean peer approval of each other, which usually fails; rather, I mean appraisal of a group’s accomplishment. When people work together in a group, whether reporting to the same person or not, they need to establish criteria by which they and those to whom they report will know how well the task force or the group has done—in terms of behavior as well as results. Group appraisals provide information that is helpful both in establishing criteria as well as in providing each individual with feedback.
At the end of a given task, a group may do a group appraisal or be appraised by the manager to whom they report, and that appraisal may be entered into folders of each of the people who are involved. It will then serve as another basis for managerial- and self-judgment.
3. Upward appraisal.
Finally, there should be upward appraisal. Some beginning voluntary steps in this direction are being taken in the Sun Oil Company, and by individual executives in other companies. Upward appraisal is a very difficult process because most managers do not want to be evaluated by their subordinates. As a matter of fact, however, most managers are evaluated indirectly by their employees, and these evaluations are frequently behavioral.
The employees’ work itself is a kind of evaluation. Their work may be done erratically or irresponsibly. Or they may be poorly motivated. Negative behavior is a form of appraisal, and one from which a manager gains little. A manager cannot be quite sure what precipitated the behavior he sees, let alone be sure what to do about it.
If, however, the manager is getting dynamic behavioral appraisal from his employees, then he, too, may correct his course. But if he asks his subordinates for upward appraisal without warning, he is likely to be greeted with dead silence and great caution. A helpful way to deal with this situation is to ask one’s employees to define the criteria by which they would appraise the manager’s job, not to judge his actual performance.
This process of definition may require a manager to meet with employees weekly for months to define the criteria. By the end of three months, say, the employees should be much more comfortable working with their manager on this issue. And if the manager can be trusted at all, then when he or she finally asks them to evaluate the performance, including specific behaviors, along the dimensions they have worked out together, they are likely to be more willing to do so. Of course, if there is no trust, there is no possibility of upward appraisal. In any event, the upward performance appraisal should go to the manager’s superior so that people do not jeopardize themselves by speaking directly.
Under present performance appraisal systems, it is difficult to compensate managers for developing people because the criteria are elusive. With a developing file of upward appraisals, however, executives can judge how well a manager has done in developing his people. The employees cannot evaluate the whole of their manager’s job, but they can say a great deal about how well he or she has facilitated their work, increased their proficiency, cleared barriers, protected them against political forces, and raised their level of competence—in short, how the manager has met their ministration, maturation, and mastery needs.10 A top executive can then quantify such upward evaluations and use the outcome as a basis for compensating a manager for his effectiveness in developing his employees.
When a group of manager peers experiments with upward appraisal and works it out to their own comfort, as well as to that of their employees, then it might be tried at the next lower level. When several successive levels have worked out their own systems, the process might be formalized throughout the organization. Acceptance of the upward appraisal concept is likely to be greater if it has been tested and modeled by the very people who must use it, and if it has not been imposed on them by the personnel department. With appropriate experience, the managers involved in the process would ultimately evolve suitable appraisal forms.
What about results?
What does adopting the critical incident technique and the dynamic job description mean for judging a person’s ability to obtain results? Does quantitative performance lose its importance?
My answer is an unqualified no. There will always be other issues that managers will have to determine, such as level of compensation or promotability—issues which should be dealt with in other sessions after the basic behavioral performance appraisal.11
Some of the performance appraisal information may be helpful in making such determinations, but neither of these two functions should contaminate the performance appraisal feedback process. There can still be an annual compensation evaluation, based not only on behavior, which is the basis for coaching, but also on outcome. Did an employee make money? Did he reach quantitative goals? Did she resolve problems in the organization that were her responsibility?
No doubt, there will be some overlapping between behavior and outcome, but the two are qualitatively different. One might behave as it was expected he should, but at the same time not do what had to be done to handle the vagaries of the marketplace. He might not have responded with enough speed or flexibility to a problem, even though his behavior corresponded to all that originally was asked of him in the job description and goal-setting process.
Both behavior and outcome are important, and neither should be overlooked. It is most important, however, that they not be confused.
1. Herbert Heneman, “Research Roundup,” The Personnel Administrator, June 1975, p. 61.
2. Paul H. Thompson and Gene W. Dalton, “Performance Appraisal: Managers Beware,” HBR January–February 1970, p. 149.
3. Herbert S. Meyer, “The Pay for Performance Dilemma,” Organizational Dynamics, Winter 1975, p. 39.
4. John B. Lasagna, “Make Your MBO Pragmatic,” HBR November–December 1971, p. 64.
5. Ishwar Dayal, “Role Analysis Techniques in Job Descriptions,” California Management Review, Summer 1969, p. 47.
6. Harry Levinson, The Great Jackass Fallacy (Cambridge: Harvard University Press, 1973), Ch. 3.
7. John C. Flanagan, “The Critical Incident Technique,” Psychological Bulletin, 51:327, 1954, and John C. Flanagan (coauthor Robert K. Burns), “The Employee Performance Record: A New Appraisal and Development Tool,” HBR September–October 1955, p. 95.
8. Donald P. Schwab, Herbert G. Heneman III, and Thomas A. DeCotis, “Behaviorally Anchored Rating Scales: A Review of the Literature,” Personnel Psychology, 28:549, 1975.
9. Harry Levinson, “Management By Whose Objectives,” HBR July–August 1970, p. 125.
10. Harry Levinson, The Exceptional Executive (Cambridge: Harvard University Press, 1968).
11. Herbert H. Meyer, Emanual Kay, and John R.P. French, Jr., “Split Roles in Performance Appraisal,” HBR January–February 1965, p. 123.