Imagine you and three friends started your own business. Each of you has strengths, one friend is good at the financial and accounting functions, another good at marketing, the third knows the product and service inside and out and you consider yourself to be a generalist with excellent people skills. Suppose further, that everything goes well for a couple of years. Each of the friends assumes mastery over his function, and enlists the support of members of his family, wife, children even brothers and sisters to help carry out operations. This informal help eventually becomes formalized and many of these family members are put on payroll.
During the early years, when it remained just the four of you, none of you would have dreamed of bringing up the notion of merit pay, job analysis or performance appraisal. You would meet once or twice a week to discuss how things were going, to plan, and to coordinate upcoming events. Each of you had so much respect for the others that you would not dare raise the matter of “doing better”, even if there were downturns in business. You had started as equals and it did not seem right to change. The business, after all, was not as old as your friendship. In fact, as we said before, one way to define friendship is as two people interacting but allowing the other to retain ownership of him or herself, to more fully develop his or her identity as an individual.
Consider the differences between your small company and a large organization with formal systems. In the case of your business, things are accomplished due to trust, communication, commitment, competence and loyalty. You trust one another, so you do not need to implement a performance appraisal system. You communicate well, so job satisfaction, morale or feelings of being “left out” never get so bad you cannot just talk it out. You are all committed, so there is no need for some sort of merit pay or incentive system. You are all competent so you do not have to worry about training, development or selection systems. And you are all loyal, so the issue of grievances, retention and turnover, never surface because you would not dare let business ruin your friendship.
Would it ever be possible to organize a large number of previously unacquainted people, around the principles of trust, communication, commitment, competence and loyalty; so that there was no need for a performance appraisal system, participation programs, selection, training and development, grievance procedures, retention efforts and the like? It might be, but the challenge would increase as the organization grew in numbers of people and in quantity and variety of subsystems.
Returning to the company of four friends, what would have to transpire to make you need to start formalizing systems? One thing, of course, would be to add employees. Notice that it would be almost impossible for the four of you friends to fully explain to the new people how you operate. They will not be accustomed to such a way of doing business and may take your lack of formality as a sign of disorganization. Or some of them may see the absence of formal systems as a surefire way to get paid for doing whatever they want to do, whether it serves your business or not.
Ultimately, you will probably need to put in place certain checkpoints, boosters and correction devices. Performance appraisal can serve as a way of checking developmental progress, as a boost to an associate’s performance or as a way to correct or readjust the associate’s understanding of what is important to the company. However, come what may, it remains important that the company continue to stress trust, communication, commitment, competence and loyalty. How can this be done?
PA is something of a gamble. If the PA system is poorly designed or implemented, it can serve to undermine the five important variables we are discussing. Merely the idea of putting a PA system in place indicates a doubt in trustworthiness. It might be something like a wife deciding to park in the dark beside the exit from her husband’s workplace and tailing him home every evening. Such behavior on her part suggests at least two possibilities. Either she has found reason to suspect that he is cheating or she has become insanely jealous and cannot be trusted herself, since her judgment has led her to snoop on him with no probable cause.
If he finds out she is doing this, his trust in her will be undermined, if he is not cheating. If he was cheating, he may take a few different courses of action. He may lie and try to cover it up, perhaps evening acting enraged that she would not trust him. He may decide to confess and ask for a divorce. He may not let her know that he knows she is tailing him and figure out stop dropping by and seeing his tryst on the way home from work, if he has previously been doing that. In short, almost no good might come out of snooping on him.
However, if she asks him why it takes him longer than the normal driving time to get home each evening, the confrontation will lead to a lot of the same sorts of outcomes. Not many marriages would last very long if both partners made it a requirement that each tells the other exactly how he or she spent every hour of every day when they were apart. In other words, a performance appraisal system automatically signals distrust, can strain communication, and might undermine commitment and loyalty. Competence or performance of one’s role may even be compromised because of the reduction in commitment.
Of course, the relationship between an employee and the company for which he or she works is not exactly the same as that between two marriage partners. The relationship between two small business partners is closer to that of marriage partners, but in the context of a large number of employees working for the same company, performance appraisal seems more justified.
The formal written job contract is like a prenuptial agreement. You might say the employee-employer relationship is something of a Hollywood marriage, with both partners having been married many times before. (In fact, the employer practices polygamy, right? And so do moonlighting employees. ☺)
What makes PA more acceptable when it involves a relationship between employees and large companies? The employee and the employer haven entered into a limited exchange agreement. The employee agrees to pay the company in time, expertise and effort in exchange for the company’s payment in monetary and other agreed-upon forms of compensation. The company cannot simply agree to pay the person whatever they can afford for a given pay period. Therefore it seems only right that the company expect that the associate not give just the amount of effort he or she has left over after taking care of other obligations, more one week, less another, up and down week to week. The pay stub (or statements about other benefits) is the associate’s opportunity to appraise the contractual performance of the employer, so you might say that a formal performance appraisal system is the employer’s opportunity to appraise the contractual performance of the associate.
What we are looking for in a PA system is an acceptable substitute for intimacy-based trust, loyalty, communication, commitment and competence assurance. The great difference is that for friends who are business partners, the friendship may come first, whereas for bureaucracies and other forms of large organization, friendship is often not possible, since the required level of interaction is precluded by distance, time constraints, or differences in role obligations.
Let us consider the ways that PA can do damage in an organization, prior to looking at how it might be beneficial. (In almost all cases, those who “do well” as judged by the PA system, do so at the expense of those who do not “do well”. This is because most PA processes require that the rater achieve variation in the ratings he or she gives. For the most part, real variance is made secondary to rating variance.)
For every rating situation, there is what might be called a “true score” and then there is the “assigned score”. The true score is never known, but assumed to exist. To varying extents, raters and other consumers of PA results, will imply that the “derived” or “assigned scores” are actually “true scores”. Deep down, everyone involved in the process knows this is not the case.
We all have spent years involved in one of the most elaborate PA processes known to modern humans, namely that which pertains in an academic setting. A professor assigns students positions on either side of an issue relevant to the subject they are studying and asks each of them to lead the in-class dialogue on his or her side of the issue. Consider some of the important factors that help determine the “score” each student receives.
The professor will have had to weigh alternatives of how explicit to be in instructing the students as to what will be considered excellent performance on the assignment. Explicit instructions may take away any possibility of innovation or creativity by the student. Doing this will make the professor’s grading process easier, but it may well turn the entire exercise into something far less meaningful for everyone involved.
Vague instructions may mean that everyone who completes the assignment, does so with different conceptions of how it ought to be done. This makes the professor’s grading job much more difficult since he or she will be comparing “apples to oranges”, so to speak. However, the freedom afforded each dialogue leader may allow for highly creative approaches and thus for a more rewarding experience.
Notice that in work organizations, if performance standards are rigid and detailed, the result may be more reliable criteria by which all those judged will be appraised, however, this benefit to consistency in ratings may come at a sacrifice of interpretive excellence on the part of the individual worker. That is, in the case of explicit criteria, it may be relatively easy to tell those who do an inadequate job of playing their role from those who do a poor job, whereas with more open-ended criteria, the individual performers will each have freedom with which to interpret the role they are playing.
Notice how the job itself will make a difference in this regard. If a person’s job is to remove bushings from a rotating machine, hammer a key in place on one side of the bushing, run it quickly through a liquid that removes any metal shavings from it, and place it in a nearby basket, moving quickly to repeat this process and to do this all day long, it would be possible to simply count the number of correctly processed bushings at the end of each day, and thus have a fairly accurate number approximating a “true score” for that person.
Suppose, on the other hand, the person’s job was to take from and hand to, set-up tools, for each person clocking off one machine and on to another. This person would be stationed at the back of the large room full of machines, possibly in a cage-like area, where employees could not just wander in and out at their leisure. How would this “cage person” be evaluated on his or her performance? Only a few people may occupy the job and the only thing that might be considered measurable may be how many times one of the workers being served complains about how long it takes to turn in their old tools and get new ones. And who knows, these complaints may result from personal disagreements between those served and the “cage person” and not so much how well the cage person actually does the job.
The differences between these two jobs has numerous implications for their respective evaluation possibilities (as well as other interesting implications). Comparisons among workers on the bushing job are made easy by the mechanical output of the process. Notice that when cries of inequity among those doing the job occur, they are more likely to be as a group rather than as individuals, since equity versus inequity will be more readily defined. The entire group of people doing the job may complain together that the production rate is set too high or that the machines malfunction (unless, of course, one person is always stuck with the one machine that breaks down, then the potential for an intra-role equity problem will be quite serious). This means those doing this job are likely to be inclined to raise inter-job equity issues with the outcomes of performance appraisal.
On the other hand, that handful of people who work the cage (across three shifts, possibly two and three at a time), will be more likely to find fault with the PA system as it divides them, giving one an excellent rating and another a poor rating, perhaps with limited data on which to base the ratings. Intra-role equity will be the prevailing complaint in that job.
It is interesting to note here that when most factory jobs were of the type where inter-role equity was likely to come up more often than intra-role equity, unions tended to have more power, and this is likely not a coincidence. A factory full of many people performing a few highly standardized roles would mean that they would be more likely to band together as a group against those not performing the standardized roles, namely management. On the other hand, in a workplace where the jobs are highly various and unstandardized, equity issues will be more vague and people will be less likely to grieve about inequities in a uniform way. In other words, hard feelings will be more likely between individuals than between roles and when you have people fighting against one another, they are much less likely to band together to form a union or argue the same case for justice.
Few people like evaluating others and nobody likes those who do (like evaluating others). Beyond this issue though, W. Edwards Deming pointed out what might well be even bigger problems with evaluating others in the context of organizations.
He contended that the rather American performance appraisal process (others around the world use it too, but we developed and promote it the most) passes the primary responsibility from management (where it belongs) to hourly employees. Deming (1986) said that over 90% of the variance in employee performance is attributable to differences in supplies, procedures and equipment, not under the control of the employee. Once these differences have been removed, then less than 10% of performance variance would remain. He was a world-class statistician, having first become involved in statistical consultation to organizations when he advised the U.S. government on the census.
If most of the variance in employee performance is due to factors other than the individual’s motivation, attention, commitment, and even ability, then the performance appraisal process, which almost everyone agrees, is an undesirable one, is not just distasteful, but it may well be much ado about nothing.
From the perspective of TQM, management’s job is to make sure that processes are in control, reduce common-cause variance and seek out and eliminate special cause variance. It is right for management to enlist employee efforts to control sources of process variance, but is it right for them to blame most of that variation on the employee?
TQM advocates changing processes not workers. It is wishful thinking, Deming would say, for a manager to look at changing the employees who work for him as the key to his or her success. The employees are allies in the fight against the enemy, process variability. They are not the problem, but rather they are an integral part of the solution
(Bowman, 1994).
References
Bhatti, M. Ishaq, Kuldeep Kumar & Michelle Schofield. (1998). On Deming’s principle of human resource management: a statistical perspective. Career Development International, 3(6), 220-232.
Bowman, James S. (1994). At last, an alternative to performance appraisal: Total Quality Management. Public Administration Review, 54(2), 129-137.
Deming, W. Edwards. (1986). 1986. Out of the Crisis. Cambridge, MA: MIT Press.
No comments:
Post a Comment