For example, instead of worrying about the optimal measure or formulas for how well a division performed in absolute, one looks at how well it performed relative to other firms or divisions engaged in similar trade and facing similar environmental uncertainties.
A salesman is compensated not for the absolute level of its sells, but instead for how he sells compare with those of other salesmen in the organization.
As long as the unmeasured exogenous factors influencing performance - such as weather, interest rates, or the odds of a key supplier or customer going bankrupt - affect the individuals or groups in the comparison set in a similar manner, measuring the relative performance of the affected individuals or groups provides a (statistically) better measure of the individuals' or groups' levels or qualities of effort.
But these advantages of comparative performance evaluation can become disadvantages, if the comparison set is not very carefully constructed: 1 - The performance measures that are compared should be similarly affected by uncontrollable environmental variables, so that relative performance reflects to the greatest extent possible things under the control of the individuals involved.
Suppose, for instance, that regional sales managers are evaluated comparatively and that their company manages its advertising nationally, placing ads on national network TV shows and in national magazines and newspapers.
If advertising is important and if sales regions vary in how well those ads reach potential customers, then sales managers in some districts may have a plausible case in claiming that they are being inappropriately compared to sales managers who are fortunate to have been assigned to districts where the company's advertising has better penetration.
In such a case, the company may find itself facing pressure to decentralize its advertising, so that each sales manager can have control over one of the factors that is likely to influence his or her comparative standing with other managers.
Or consider the case of a firm that seeks to measure the performance of its research staff.
There may be good reasons not to evaluate individual members of the research staff relative to the performance of their closest peers.
The firm might therefore choose instead to measure performance relative to members of other research teams within the organization, or even relative to the performance of the research staffs of competitors.
The problem with this is that, in most cases, the more diffuse the comparison set, the less control there is for environmental factors.
If, say, a measure of comparison is the number of patents gained, the specific research agendas of the different research staffs, set to some extent by managerial decree, may become important: Group X is "penalized" because it works in an area that is less likely to lead to many patents.
This may lead to research groups arguing for a greater measure of autonomy: "If our performance is going to be measured on the basis of the patents we gain," will go the argument, "then we should control what we work on.
" 2 - The individuals in the comparison set should be comparable, in the sense that their performances are "on the same scale.
" If salesmen are evaluated based on their relative sales levels and some salesmen are blessed with larger markets than others, those with the smaller markets will properly view the system as unjust.
3 - Individuals in a relatively small comparison set are ''penalized''- usually unjustly, in their view-if the comparison set contains one or more truly exceptional performers.
4 - Comparative evaluation schemes can clash with other important status norms.
For instance, a comparative evaluation scheme that consistently ranks younger employees ahead of those with greater seniority may, in some cultures, give rise to substantial status inconsistency, thereby decreasing perceptions of legitimacy for the evaluation scheme.
Some of these problems can be dealt with by constructing smaller and "tighter" comparison sets-for example, sales personnel might be evaluated only relative to others servicing similar types of customers; professional service workers are compared relative to others in the same experience cohort and practice area; research groups are compared only to other groups that deal with similar scientific problems.
Of course, using smaller groups can exacerbate the third problem listed above.
And constructing smaller and tighter comparison sets often means restricting comparisons to other individuals (or groups) inside the specific firm or specific business unit of the firm; we'll see some potential problems with this below.
In theory, these difficulties can also be addressed by handicapping the relative performance measures: To compensate for their relative disadvantages, we might give "extra credit" to: the sales manager working in an area relatively less well covered by national advertising; the research group working on a subject less prone to developing patents; the salesperson with the smaller market area; and individuals stuck in a group with an exceptional star performer.
In practice, however, handicapping is a fragile and contentious process.
The basis for the handicapping scheme may be disputed: The salesmen assigned to the larger and potentially more fruitful territory will point out that sales take time and effort, and she has to work harder.
Or she may argue that the size of her territory is partially of her own making, as she has built up its clients.
And even if the basis for the handicapping scheme is agreed upon, the size of the handicap can be contentious.
Almost every handicapping process is rife with the potential for politicking and feelings of procedural injustice by those who face the handicaps, and even by those who are advantaged by handicaps but not (in their own view) advantaged enough.
In some instances, handicaps can be constructed reasonably objectively, using statistical techniques.