Human resource Management
danrid
05ch_weathington_compensation.pdf
5 Competitive Compensation: Measuring Performance and Considering the Market
DigitalVision/Thinkstock
Learning Objectives
After reading this chapter, you will be able to:
1. Understand and define performance in relation to the workplace.
2. Describe the common errors and mistakes made in performance measurement.
3. Distinguish between trait-based, comparison-driven, goal-driven, and behavior-based performance mea- surement systems.
4. Discuss what is meant by a market competitive pay system.
5. Understand the role of compensation surveys in a well-designed compensation system.
6. Understand the purposes of pay bands in relation to external and market pay considerations.
7. Discuss what adjustments a small company might have to make in measuring performance.
abc82339_05_c05_107-138.indd 107 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Introduction As we have discussed in previous chapters, properly designed compensation systems tie orga- nizational incentives to desired employee performance. Whereas job analysis and job evalua- tion (discussed in Chapter 4) focus on the components of the job, performance measurement focuses on the behavior of the employee. Attracting and retaining the best performers is a key function of compensation and benefits systems, especially given that top-performing employ- ees can outperform typical or low-performing employees by an order of magnitude—3 to 12 times as much, by some estimates (Campbell, Gasser, & Oswald, 1996). A well-designed and -executed compensation strategy can result in a competitive advantage for an organization.
Designing a compensation and benefits strategy, however, requires more than just a consider- ation of performance inside the organization. Companies compete in a business environment, not just for market share but also for employees. Attention must be paid to the compensation practices of competing organizations in order to attract, select, and retain the best employees.
In this chapter we will begin by discussing performance measurement and then consider how compensation surveys can be used to help gain a competitive advantage. We will then discuss how pay bands can be used to build a structured, yet flexible, compensation system. Lastly, we will discuss how organizational size and resources impact performance measurement, surveys, and setting pay bands.
5.1 Performance Measurement At first glance, performance appears to be a topic that needs no definition. After all, to per- form is to do something. In fact, the Merriam-Webster dictionary defines performance as “the act of doing a job, an activity, etc.” (“Performance,” n.d.). Thus, performance is behavior.
Upon further reflection, however, it becomes clear that beyond simply looking at behavior, we must also consider the context of that behavior.
• Why did it occur? • Whom did it impact? • What will happen in the future? • How will one individual’s behavior influence, either positively or negatively, other
people? • Is the behavior furthering or hampering organizational goals?
All of these questions, and many more, need to be considered when designing, implementing, and managing compensation and benefits. So, instead of asking what performance is, the bet- ter option may be to consider what causes performance.
One way to conceptualize performance is to utilize the equation below:
Performance = Ability × Motivation + Error
abc82339_05_c05_107-138.indd 108 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Of particular importance in this equation is the multiplicative nature of the relationship between ability and motivation. An individual’s ability to perform may be poor or high. Poor performance does not result in accomplished goals, while good performance does. Motivation can actually be focused in a negative or counterproductive way. As we discussed in Chapter 3, the behavior that is performed is that which is incentivized (Kerr, 1975). Thus, a person of high ability can actually produce quality performance in the wrong direction (i.e., doing a really good job but not for the company). This can result in counterproductive behaviors such as theft, poor quality, lost time, and missed deadlines as well as the employee’s actions not moving the company forward.
Of course, performance is also impacted by the support, resources, and so on that an employee receives. Accordingly, error, in this equation, represents everything else that has the poten- tial to impact performance but is outside the control of the individual. Error can be positive or negative as well as systematic or random (Weathington, Cunningham, & Pittenger, 2012). Positive error represents something that makes performance seem better than it actually is. Negative error makes performance seem worse than it actually is. Systematic error is in a set direction (i.e., positive or negative) and for a set amount. Random error is variable and can be either positive or negative and of varying magnitude. For example, the performance of a salesperson working a geographic territory is in part determined by the quality of that terri- tory as well as broader economic considerations. The quality of the territory would represent positive or negative error when trying to utilize number of sales as a measure of performance. Fluctuating economic conditions would then add random error into this equation, complicat- ing the measurement of performance even further. We will cover error as it relates to per- formance measurement in more depth after we discuss the basic issues and legal aspects of measuring performance.
Despite the difficulties that arise, there is a need to actually assess how employees are doing on the job. The most common term used for this is performance management. In practice this term is often used interchangeably with performance appraisal; however, performance man- agement is a more comprehensive term because it encompasses all aspects of performance, whereas performance appraisal focuses more on administrative concerns. Appraising per- formance is only part of a comprehensive performance management process. All too often, performance appraisals are viewed by managers as a bureaucratic requirement that takes time and effort away from “real” work. This perception is unfortunate since performance can- not be improved if it is not first assessed and then tied to organizational goals.
Performance management focuses on iden- tifying, evaluating, and developing employee performance. All three components are equally important and build upon each other. In order to be effective, a performance man- agement system is based on the foundation of a job analysis that identifies the behaviors, tasks, KSAOs, and competencies needed to fulfill the organizational function for which the job exists. Of course, we must also remember that we are evaluating people and morale, and future performance can be impacted by the accuracy and perceived fairness of evaluations.
Critical Thinking
What is performance? How is it different from motivation?
abc82339_05_c05_107-138.indd 109 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Key Concepts in Performance Measurement
As we discussed at the beginning of the chapter, performance is best understood in the con- text that it is being measured. This encompasses knowing who we are measuring (individual, team, division, or organization as a whole), what we are measuring (behaviors, traits, etc.), and why we are measuring the performance (promotion purposes, employee feedback, com- pensation decisions, etc.).
From a compensation standpoint, performance can be assessed at the individual, group or team, or organizational level (Austin & Villanova, 1992). If the goal is to promote teamwork, then care must be taken to link rewards with team goals and objectives. Alternately, the com- pensation for individual contributors needs to be tied to those results over which they have the most control. When designing a compensation system, the contingency between behav- ior and reward is the key determinant of behavioral incentives, as we learned in Chapter 3. Therefore, typically, organizational-level rewards are used for executives who can actually impact the bottom line of organizational performance, while individual rewards are used for rank-and-file employees. Rewards based on business-unit performance can be implemented as long as there is a direct connection to the results, such as with managers.
We must also consider what we are actually attempting to assess. Performance is multidimen- sional and is not simply proficiency in accomplishing job-related tasks (Austin & Villanova, 1992; Campbell et al., 1996). Not all jobs serve the same purpose, and a framework is useful when considering what is actually being done on the job. Just because individuals have the same job title doesn’t mean that they are doing the same thing. In organizations, performance is defined and organized into categories through both formal and informal processes. Some- times this process is structured and scientific, while other times it is organic and occurs with- out much conscious thought. Either way, the results should be tied to organizational strategy and functioning. A comprehensive and useful framework was proposed by Campbell et al. (1996) that categorizes performance based on the following eight general factors:
1. Job-specific task proficiency—core job tasks (i.e., specific activities that differentiate this job from other jobs)
2. Non-job-specific task proficiency—tasks that are nonspecific to certain jobs but apply across jobs
3. Written and oral communication task proficiency—different jobs will have different levels of each
4. Demonstration of effort—motivation, commitment, and persistence 5. Maintenance of personal discipline—degree to which negative behaviors are
avoided (e.g., drug use, tardiness) 6. Facilitation of team performance—includes peer leadership and enablement of
interdependent tasks 7. Supervision/leadership—influence subordinates directly and set goals for
subordinates 8. Management/administration—set unit goals, attain resources, problem solve, and so on
Not all jobs contain all of these aspects of performance, and a job analysis, not simply a job description, is the best source of information regarding what constitutes performance for a given job. Once the components of performance for a given job have been identified, they must then be assessed.
abc82339_05_c05_107-138.indd 110 1/4/16 3:08 PM
DeficiencyContamination Relevance
Predictor Criterion
Section 5.1 Performance Measurement
The measurement of these components is based on a representative sample of tasks that encompass the eight general factors discussed above. This measurement will take the form of a predictor (i.e., the actual device used) that is intended to measure a criterion, or sample of performance, that is considered to be necessary for the successful accomplishment of job objectives. The parts of the predictor that are not related to successful job performance are a source of error and can be thought of as contamination that clouds our ability to actually assess what we want to assess (i.e., the criterion). Conversely, deficiency is the aspects of performance that we do not assess at all. These are the things we would like to measure but for one reason or another are not able to do so. As an example, for a car salesperson we might be interested in sales ability and how much revenue the individual could bring in to the company. It seems obvious that we could measure this by looking at the number of cars sold during a specific time period. However, what about factors such as customer satisfaction, repeat customers, online reviews of the sales experience, regional differences in style preferences and number of customers, and fuel costs? All of these factors represent contamination and deficiency (i.e., error). In addition, some are more controllable by the employee than others. Therefore, even a seemingly objective measure of performance, like number of sales, is not perfect. If we then incentivize salespeople based purely on this measure, what are we actually incentivizing— long-term profitability or short-term gain? The part that we get right, the overlap between the predictor and criterion, is relevance. This is where our measure is accurate or valid.
This relationship is demonstrated graphically in Figure 5.1.
Figure 5.1: Predictor-criterion relationship
Taking the time to assess the measurement of job performance is important in order to ensure the results are truly accurate.
Based on Schmitt, N. & Klimoski, R. J. (1991). Research methods in human resource management. Cincinnati: South-Western.
DeficiencyContamination Relevance
Predictor Criterion
abc82339_05_c05_107-138.indd 111 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Contamination and deficiency occur because we are trying to predict performance in a real- world environment where all variables cannot be controlled. Performance evaluations con- sist of three main elements—the raters/those being rated, the rating instrument, and the rating context (Landy & Farr, 1980). All three of these are potential sources of both accurate and inaccurate information. More on sources of error will be discussed later in this chapter.
One last consideration is the purpose or reason that performance is being evaluated. After all, one of the main purposes of measurement is to obtain information that can then be used. This textbook is focused on compensation and benefits, but it is rare that performance is evaluated for only one purpose. Among the many possible purposes (see Cascio, 1991) for performance measurement are
• personnel decisions/administrative (e.g., promote, train, fire), • research (e.g., test validation), • setting objectives for training, • coaching or providing feedback to employees, and • facilitating organizational diagnosis and development.
Compensation is just one possible use of performance measurement, so care should be taken that conflicting goals and objectives do not result in poor evaluations of performance. For example, performance measurement done for purely administrative reasons rarely provides the level of detail needed for coaching. Performance measurement, while one of the most common activities in human resource management, is also one of the most commonly mis- used and neglected practices.
Measures of performance can take three basic forms: observations of behavior, objective measures of output, and ratings of nontangible factors. Each of these can be measured using multiple techniques, and there are pros and cons to each. This will be discussed in more detail later in the chapter.
Performance Measurement and the Law
Before getting into the specific methods for measuring performance, we must consider the legal environment. Care must be taken to avoid discrimination based on protected categories such as age, race, sex, religion, color, national origin, and disability (see Chapter 2 for a more detailed discussion of legal issues). Increasingly, performance measurement is becoming part of litigation involving layoffs, promotions, discharge, and merit pay (Martin, Bartol, & Kehoe, 2000). Performance measurement used to make employment decisions is subject to the same scrutiny and standards applied to employee selection.
Werner and Bolino (1997) examined hundreds of court cases involving performance mea- surement and found that the key elements for an effective performance appraisal system include a valid job analysis, clearly communicated and understood written instructions, two- way communication about appraisal results, and, if multiple raters are involved, the amount of agreement between them. Companies need to make sure they have competent employees developing the performance measurement system to help avoid legal issues as well as good communication and training about the system for all parties involved in the process.
abc82339_05_c05_107-138.indd 112 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Sources of Error in Measuring Performance
There are several sources of error that may contribute to a poor or inef- fective performance measurement process. Each should be considered when implementing and evaluating performance. An awareness and con- sideration of possible errors can help us avoid them when designing mea- surement systems and when interpret- ing results. Many can be alleviated by proper training and effective design of the performance measurement system.
Halo Effect The halo effect occurs when a rating on one or more job dimensions or cat- egories affects the ratings on others, either positively or negatively. Sometimes, negative effects are called horns, while the term halo is reserved for the positive impact. For our purposes, halo is used to refer to both.
For example, a supervisor may rate an individual high on several traits and continue rating him or her in the same fashion on subsequent traits. Training and norming evaluator ratings can help alleviate halo effects, but evaluators tend to drift back to their strict or lenient rating patterns over time. In order to maintain employee trust and perceptions of fairness, retrain- ing and renorming of evaluations should occur on a fairly regular basis.
Bias Bias involves the tendency to, consciously or unconsciously, allow differences to impact the evaluation of performance. The bias can be in favor of the employee or against the employee. Often, bias based on race, gender, age, sexual orientation, and so on is encountered, even though bias based on these factors is illegal, albeit sometimes hard to prove.
We often make judgments about other people very quickly. https://www.youtube.com/watch ?v=84OT0NLlqfM&feature=youtu.be
These judgments are influenced by our awareness of negative or positive stereotypes. For example, when an applicant interviews for a job, the interviewer will make judgments about the person’s competence based on physical characteristics such as height, dress, and attrac- tiveness—even though these may not be related to job performance (for more on this, see Malcolm Gladwell’s discussion of the Warren Harding error in his book Blink: The Power of Thinking Without Thinking [2007]). In compensation decisions, this tendency could result in an employee being paid less for a job simply because he or she looks young or because the
iStock/Thinkstock
Look out for sources of errors when measuring performance and work to alleviate them through proper training and design.
abc82339_05_c05_107-138.indd 113 1/4/16 3:08 PM
Section 5.1 Performance Measurement
applicant is a female applying for a stereotypically male job or vice versa. Uncontrolled bias can lead to both ethical issues and costly legal problems.
Generalization Errors Halo effects occur when a supervisor rates an employee positively or negatively on one par- ticular rating dimension (e.g., timeliness) and then subsequently generalizes that same rat- ing to other dimensions. The effect also may occur if the supervisor receives information from several sources and mentally generalizes to include most, if not all, feedback on that appraisal. For example, because an employee is likable and everyone, including the manager, enjoys working with her, then the tendency is to give good evaluations even when dealing with aspects of job performance that have nothing to do with likability (e.g., technological knowledge).
A similarity effect may impact the appraisal to the extent that the supervisor believes the employee has similar beliefs, attitudes, or personality characteristics and lets that impact the rating given to the employee (see O’Leary, Durham, Weathington, Cothran, & Cunningham, 2009). Sometimes this is based on observable, surface-level characteristics such as race and gender. Other times this is based on shared interests (e.g., liking the same sports team). We tend to like and favor those we view as like ourselves. Therefore, care must be taken to rate actual performance and not just perceived similarity.
Contrast Errors Contrast errors occur if a rater evaluates one employee’s performance against another’s. This is particularly a concern when the first employee is rated either extremely high or very low. There is a common tendency to exaggerate the differences between those being rated, particularly if the appraisal is strongly or entirely subjective (i.e., the next rate is rated higher because of the poor scores of the first rate or vice versa), so care must be taken to avoid this tendency.
Regression to the Mean and Central Tendency In statistics, regression to the mean deals with the circumstance of more extreme scores becoming less extreme over time (i.e., high scores are not as high and low scores are not as low). The implication of this for ratings of performance is that those individuals rated highest one time are more likely to receive a lower score at another time. Correspondingly, those who receive the lowest ratings are more likely to receive a higher rating in the future. Over time, individual ratings have a tendency to move closer to the arithmetic mean of the distribution of all ratings. In other words, they are rated as more typical or average in terms of performance (they move closer to the center of the distribution).
From a pragmatic perspective, this same trend occurs in performance measurement when raters are required to provide written explanations of extreme scores (high or low), which requires more time and effort. Under such conditions, the supervisor is more likely to rate everyone near the middle of the scale because that makes the process easier for the supervisor.
abc82339_05_c05_107-138.indd 114 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Strictness and Leniency With respect to strictness and leniency, some people simply have a tendency to judge or eval- uate others harshly, while others tend to rate too easily. If a supervisor is generally strict with his or her employees when conducting performance appraisals, the ratings are likely to reflect such a belief with low ratings, whereas another supervisor evaluating the same group of employees may be more lenient, which would result in higher scores.
When administering a performance appraisal, a supervisor who tends toward strictness will rate an employee low and in essence will ignore the top part of the rating scale. Leniency is the opposite and refers to supervisors who rate employees higher than would be expected if objective criteria were used. Lenient raters ignore the bottom part of the rating scale. In practice, this results in a lack of variability and an inability to distinguish between employees. Instead, the entire performance scale should be used to accurately evaluate performance.
Maximal Versus Typical Performance When individuals know they are being assessed, they tend to put forth their maximum amount of effort. This results in their best performance possible (i.e., their maximal performance). On the job, however, individuals tend to underperform their maximum potential, resulting in typ- ical performance. There are many reasons for this difference in performance levels beyond people being lazy.
For one, it is almost impossible to perform at a maximum level for long. Maximum perfor- mance strains our physical and cognitive resources. Sustained performance at maximum level exhausts these resources, which causes errors, injuries, and burnout. Consider a sports example. Many states have recently passed laws that limit the amount and level of difficulty for high school football practices. The reason for this is that most injuries occur in practice and not during actual games. By limiting maximal performance, the goal is to reduce overall injury levels.
From a compensation perspective, we must be aware of what kind of performance is being assessed so that we can reward the performance appropriately. Is the performance we are assessing maximal or typical? The one we observe or even want in a given situation depends on the specific context of the performance.
Multiple Errors It should be noted that the types of errors discussed are not mutually exclusive. That is, multiple errors may impact performance measurement at the same time. For example, a supervisor who is particularly lenient in her or his evaluation also may tend to rate employees highly, but not highly enough that the appraisal requires the rater to provide detailed, written justification for a particularly high rating. The more we can eliminate the errors, the better the performance measurement system and the better the results. Making a concerted effort to design the system to minimize errors in measuring performance will help eliminate the impact of multiple errors.
Critical Thinking
What are the common errors that lead to inaccurate performance measurement?
abc82339_05_c05_107-138.indd 115 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Methods of Measuring Performance
The purpose of performance measurement is twofold—to reduce ambiguity and to provide standardization (Weathington et al., 2012). This purpose can be accomplished through the use of both objective and subjective evaluations. Objective measurement is based on easily quantifiable factors, whereas subjective measurement depends upon human judgment.
Objective measures typically include production data (e.g., dollar value of sales, number of widgets produced) and personnel data (e.g., accidents, absenteeism). These kinds of mea- sures are intuitively appealing because they do not rely on possibly biased interpretations and conjectures; how- ever, there are problems with objective measures. First, some performance metrics are outside the control of the employee (e.g., price, demand). Sec- ond, the focus is on the outcomes of behaviors and not on the actual behav- iors that occur. Therefore, no distinc- tion may be made between legitimate and questionable, or even illegal, busi- ness practices. The “why” may be as
important as the “how much” when it comes to performance. Last, measures of objective per- formance are not always available or may be very difficult to obtain.
Subjective measures of performance attempt to fill this gap by making inferences about the “why.” Unfortunately, behavioral and cognitive restrictions often limit the accuracy of sub- jective measures. As applications of agency theory (discussed in Chapter 3) would suggest, these often take the form of interpersonal barriers and political barriers that blur the relationships between organizational intentions and manager/employee actions. Interper- sonal barriers are when raters and those being rated have different expectations about the method or objective of performance measurement. Political barriers deal with managers not utilizing measurement systems properly in order to avoid conflict or to advance their own agendas.
To consider interpersonal barriers, consider the multiple roles that a manager must play when interacting with subordinates. Part of a manager’s job is to coach and develop employ- ees. At the same time, managers are also supposed to make formal appraisals of subordinate performance that impact raises and promotions. These two roles could be in conflict with each other. In order to develop and mentor employees, the manager must have accurate infor- mation about performance—both good and bad. However, it is in the employee’s best interest to only present good information to the supervisor. In short, raters and those being rated may not use the same metric for making evaluations of performance, and even when they do, not all information is shared equally.
Science Photo Library/Superstock
The purposes of measurement are to standardize and reduce ambiguity.
abc82339_05_c05_107-138.indd 116 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Consider the following scenario. A long-term employee working for a telecommunications company has worked his or her way up to a middle management–level position where he or she oversees the resolution of customer complaints. This employee excels in the job and com- plaint resolutions are up, along with customer satisfaction and retention. Instead of reward- ing this employee, and potentially setting him or her up for a promotion out of the business unit, the manager of the unit takes personal credit for the performance of the work group and rewards everyone in the group equally—even though the manager knows the results are primarily due to the work of one employee. Do you think the employee and manager have equivalent opinions about this outcome? What political barriers do you think are causing these differences?
Political barriers occur when a manager’s goals do not match those of the organization or do not focus on the best interests of the employee. In the example above, the manager takes credit for the employee’s performance, both to enhance his or her own reputation and to pre- vent the source of that recognition (i.e., the employee) from leaving. As an alternative exam- ple, a manager could give a poor employee a good rating or give everyone the same rating in order to avoid interpersonal conflict and to maintain the status quo.
In the world of work, two compensation-related elements must come together to ensure that the best employees are attracted to and remain with a firm: (1) a viable performance- measurement process and (2) the ability of supervisors to conduct effective performance appraisals. Given the above discussion, it is not a surprise that there is a low relationship or correlation between objective and subjective measures (Cascio, 1991). Barriers such as this are a major concern since research suggests that often the majority of good performance is attributable to a small group of elite performers (O’Boyle & Aguinis, 2012).
So what do we do? How do we turn error-prone evaluations of performance into tools that can aid both the worker and the organization? The key is standardization, clear communica- tion, and considering the perceptions of both raters and those being rated.
The most common measures of performance are based on one or more of the following: traits, comparisons, goals, and behaviors. While behaviors are typically the most easily quantifiable measures, all techniques have strengths and weaknesses. The following sections will discuss each measure in greater detail.
Trait-Based Systems Traits are characteristics or aspects of an individual related to job duties and results. A trait- based system uses these characteristics to judge how an employee is doing in his or her job. For example, employees may be rated on their diligence, communication skills, ability to solve problems, ability to be a team player, and so on, depending on the specifics of the job. While providing some basis for awarding pay increases, promotions, and the like, problems with this approach may arise. For example, if a different supervisor were to evaluate the same employee over the same time frame, would the employee receive the same scores? This is an issue of interjudge reliability (consistency of ratings) across different raters that may cre- ate problems in feelings of fair treatment by the employees as well as impacting pay equity, promotions, and the like. Evaluations of traits, however, can provide insight into underlying
abc82339_05_c05_107-138.indd 117 1/4/16 3:08 PM
For each of the following traits, indicate which score best represents the employee being assessed.
1 The employee
does not possess this trait.
2 The employee
is below average on this trait.
Problem Solving
Communication
Adaptability
Trait Score
3 The employee is average or
typical in this trait.
4 The employee
is above average in this trait.
5 The employee is among the best employees seen
in expressing this trait.
Section 5.1 Performance Measurement
characteristics that are very useful for employee development and strategic planning of career trajectories.
Another consideration when evaluating traits is if the factor being evaluated is truly a long- term, stable characteristic of the individual or if it is temporary and will change based on the specific situation and individual’s feelings or emotions. In the empirical literature, these short-term, malleable characteristics are termed states (Chaplin, John, & Goldberg, 1988). States can be changed through training and coaching, while traits are more enduring and harder to modify. The distinction between states and traits, however, is not always clear (Allen & Potkay, 1981), and in practice the focus should be on behaviors that can be coached or trained. As we covered in Chapter 3, compensation can be used to motivate both short- and long-term change, but effective change is dependent upon feedback and identifying what traits (and corresponding behaviors) need to be changed.
Trait evaluations are fairly easy to develop. One method of trait assessment is to ask raters to evaluate how much a specific trait represents the employee by utilizing a Likert-type rating scale (see Figure 5.2).
Figure 5.2: Sample trait rating scale
While compensation can be used to motivate change, providing feedback to employees is how effective change happens.
For each of the following traits, indicate which score best represents the employee being assessed.
1 The employee
does not possess this trait.
2 The employee
is below average on this trait.
Problem Solving
Communication
Adaptability
Trait Score
3 The employee is average or
typical in this trait.
4 The employee
is above average in this trait.
5 The employee is among the best employees seen
in expressing this trait.
Another option is to do a graphic rating scale, which ties desired traits to observable employee behavior (typically through a job analysis). This is usually done to address the more subjec- tive nature of trait evaluation and aid in quantifying measurements of performance. With this method, an employee’s performance is rated on specific traits related to job duties that are broken down and measured by clear behaviors. An example of a trait-oriented graphic rating scale, intended to be completed while observing an employee, is shown in Figure 5.3. It is also
abc82339_05_c05_107-138.indd 118 1/4/16 3:08 PM
Employee’s Name: Date:
Rate the employee on each of the following by marking the appropriate rating on the line next to each item.
1 Always
2 Often*
1. 2. 3. 4.
behavior prevents others from maintaining a clean workspace does not leave time at end of production cycle for cleanup leaves debris in common work area cleans without supervisor present
at least four times for period observed two or three times for period observed only one time for period observed
* **
***
Trait Being Assessed: Diligence in Maintaining Workstation
3 Occasionally**
4 Seldom***
5 Never
Section 5.1 Performance Measurement
useful to rate both positive and negative behaviors (as is done in Figure 5.3). This way, raters must think about each rating being made and are not tempted to mark all fives or all threes for an employee without considering the specific item being assessed.
Figure 5.3: Sample graphic rating scale
Asking a rater to assess both positive and negative traits forces the rater to spend time thinking about each trait and lowers the chance of giving all extremely high or extremely low scores.
Employee’s Name: Date:
Rate the employee on each of the following by marking the appropriate rating on the line next to each item.
1 Always
2 Often*
1. 2. 3. 4.
behavior prevents others from maintaining a clean workspace does not leave time at end of production cycle for cleanup leaves debris in common work area cleans without supervisor present
at least four times for period observed two or three times for period observed only one time for period observed
* **
***
Trait Being Assessed: Diligence in Maintaining Workstation
3 Occasionally**
4 Seldom***
5 Never
Advantages of a trait-based graphic rating scale include the fact that it is the most widely used system in the world and is easy to develop and administer. This type of rating scale also ties well with information typically produced in a job analysis. Additionally, results enable com- parison across groups of employees.
Disadvantages for using trait-based graphic rating scales include that it is prone to the halo effect and that the rating may be based on a false impression since the employee knows he or she is being observed and may perform better than is typical. Overall, trait rating approaches, especially when not tied to actual behavior, tend to be more subjective, less reliable, and less legally defensible than other systems.
Comparison-driven Systems A second approach to evaluating performance involves comparing the performance of one employee to another. For objective measures of performance, this is as simple as rank- ordering employees based on the assessed metrics (e.g., number of cars sold, accounts won). Remember, however, that many other contextual factors, such as sales territory or economic conditions, have the potential to impact seemingly objective measures. For subjective mea- sures, this approach involves making judgments and cognitively evaluating each employee against other employees. The most common approaches are alternation ratings, paired com- parisons, and forced distributions. Each of these will be discussed in turn.
abc82339_05_c05_107-138.indd 119 1/4/16 3:08 PM
Customer Service
Employee A
Employee B
Employee C
Employee D
Employee E
Employee F
Employee G
All Employees
Employee F
Employee C
Employee E
Employee A
Employee D
Employee G
Employee B
1.
2.
3.
4.
5.
6.
7.
Ranked Employees
Think about all the aspects of customer service you want your employees to demostrate. Consider each employee and how he or she has performed regarding customer service. Now list all of your employees on the left-hand side of the form. Next, put the highest-performing employee on the top line on the right side of the form. Then put your lowest-performing employee on the bottom line. Continue this process until all employees have been ranked.
Section 5.1 Performance Measurement
Alternation Ratings
With alternation ratings, employees are sorted by extremes. Using this approach, the supervi- sor would select a given job-related task, behavior, or trait—such as customer service—and rate the highest employee on the trait, then the lowest employee, then the next highest, then the next lowest, and so forth. An example of an alternation rating scale is illustrated in Figure 5.4.
Figure 5.4: Sample alternation rating scale
After completing an alternation rating scale, the rater will have a rank-ordered list of employees from top to bottom based on their performance related to a certain task, trait, or behavior.
Customer Service
Employee A
Employee B
Employee C
Employee D
Employee E
Employee F
Employee G
All Employees
Employee F
Employee C
Employee E
Employee A
Employee D
Employee G
Employee B
1.
2.
3.
4.
5.
6.
7.
Ranked Employees
Think about all the aspects of customer service you want your employees to demostrate. Consider each employee and how he or she has performed regarding customer service. Now list all of your employees on the left-hand side of the form. Next, put the highest-performing employee on the top line on the right side of the form. Then put your lowest-performing employee on the bottom line. Continue this process until all employees have been ranked.
The advantage of an alternation rating scale is that it results in a succinct ranking of employ- ees from highest to lowest. Focusing on extremes (highest, then lowest, and so on) helps rat- ers make more accurate evaluations; however, the ratings in the middle of the distribution are not as accurate as those at the extremes. This ranking does not allow employees to be rated the same, even if in actuality they are performing at the same level. Additionally, there are not equal distances between rankings. In other words, the difference in performance between the highest-ranked employee and next-highest-ranked may not be the same as the difference between any other employees who are ranked next to each other.
Paired Comparisons
As an alternative to ranking all employees at once as to who is best, worst, and in between, a paired comparisons approach, as shown in Figure 5.5, may be used. With the paired com- parisons approach, the supervisor would compare each employee to every other employee
abc82339_05_c05_107-138.indd 120 1/4/16 3:08 PM
Customer Satisfaction
Score 3 1 2 2 2
As Compared
to: A B C D E
+ + – – –
– + + + +
+ – + + –
+ – – + +
+ – + – +
A
B
C
D
E
Section 5.1 Performance Measurement
being evaluated. So, in this example, employee B was ranked higher than employee A but lower than all other employees. To determine relative ranking, we would add up the totals of pluses (+) for each employee to generate a numerical score for each given trait.
Paired comparisons results in a more nuanced ranking of employees than the alternation approach. Instead of a list, we end up with a grid that compares all employees to all other employees. This added complexity, however, sometimes makes the information more difficult to interpret. Consider Figure 5.5. Employee A rated higher than all of the other employees except B. Employee B, how- ever, was ranked below all of the other employees except A. How do you inter- pret this result since it doesn’t look logi- cal (if A is better than all other employees than B, why isn’t B ranked higher than all employees)? This method requires more than just examining the numbers. It requires actually talking with raters about why they made the ratings the way they did.
Forced Distributions
Forced distributions require the evalu- ator to place a percentage of people in groups such that a few employees are considered to be above average or star employees, most employees are performing ade- quately, and a few employees are below average or poor performers. Many different numbers have been proposed and used to determine how many employees should be in each groups, such as a 10%-80%-10% split or a 10%-70%-20% split. The feature box on the next page discusses how GE used the forced-distribution method of performance measurement.
An advantage of this approach is that it enables an overall ranking of employees in a given department or unit. Additionally, it aids the employee with respect to his or her career by providing a list of strengths and developmental areas. Differentiation of performance across employees is important and allows the identification of not only strong performers but also those who would benefit from coaching and training as well as alternative forms of moti- vation. For differentiation to truly work and be widely successful, it must be endorsed and championed from the top, as in GE’s case with Jack Welch.
Figure 5.5: Sample paired comparison
scale
Paired comparisons offer an alternative to a simple list of best to worst employees, but the complexity of the results may make it more difficult to decipher.
Customer Satisfaction
Score 3 1 2 2 2
As Compared
to: A B C D E
+ + – – –
– + + + +
+ – + + –
+ – – + +
+ – + – +
A
B
C
D
E
abc82339_05_c05_107-138.indd 121 1/4/16 3:08 PM
Section 5.1 Performance Measurement
Disadvantages of this approach are negative perceptions about the process and the encour- agement of nonproductive competition between employees. The halo effect, strictness or leniency, and recent positive or negative interactions with a given employee also may impact the rankings. Additionally, the supervisor must ensure that the employee believes the trait is valuable. For example, if an employee is told his creativity is his best trait but the supervisor, department, or organization does not value such a trait highly, he may not perceive his high score as helpful to his job, pay increases, or ability to be promoted. This could cause him to neglect or lessen the use of that trait in favor of emphasizing ones that are more readily rec- ognized by the company. The company then loses the benefits of unique talents, which could have been competitive advantages.
One proposed reason for forced distributions is that performance is normally distributed and forms a normal or bell-shaped curve. While the normal curve is, in fact, an accurate represen- tation of nonrestricted performance, most collections of employees were selected based on a
Compensation and Benefits in the Real World: General Electric (GE)
Perhaps the most widely publicized forced-distribution model of performance measurement was popularized by Jack Welch, former CEO of GE. Beginning in the 1980s, Welch restructured GE to be a flatter organization, with the goals being to speed decision cycles, move information throughout the organization rapidly, utilize quick and effective feedback, and reward managers on such attributes as leadership, candor/openness, simplicity, and integrity. He believed the move to global economic growth and competition demanded a different approach and set of values than the ones upon which GE was built (see Slater [1998] for a more complete discussion of GE’s strategy under Jack Welch’s leadership). As part of this transformation, Welch put in place a forced-distribution system.
Part of the forced-distribution performance management system he created involved what was known as a “Work Out” process in which the business leaders from each of the departments were put in front of their people to let the leaders hear what the employees thought about the company, its business plans, and its leaders. The “Work Out” was supported by a performance appraisal system referred to in the media at the time as “Rank- and-Yank.” Welch intended it strictly as a differentiation system, although it ended up being a forced-distribution methodology that placed employees in the top 20%, middle 70%, or lowest 10% in the company. In his mind, great companies built great people and teams through communication, consistency, transparency, and candor. As Welch said in an article in the Wall Street Journal in 2013, “It’s about aligning performance with the organization’s mission and values. It’s about making sure that all employees know where they stand. Differentiation is nuanced, humane, and occasionally complex, and it has been used successfully by companies for decades” (Welch, 2013).
Welch’s “humane” point is particularly important since negative morale and impressions can result from systems viewed as “rank and yank.” Microsoft experienced the detrimental effects of a “rank and yank” system before changing course to a system that better supports a collaborative, innovative environment. For more information on Microsoft’s experience, see http://www.businessinsider.com/microsoft-was-destroyed-by-its-stack-review-process -according-to-new-vanity-fair-expose-2012-7 and http://www.businessinsider.com/ microsoft-just-killed-its-controversial-stack-ranking-employee-review-system-2013-11.
abc82339_05_c05_107-138.indd 122 1/5/16 2:09 PM
Normal Distribution Negatively Skewed Distribution
Section 5.1 Performance Measurement
system that was designed to weed out potential poor performers. This results in a restriction of range for current employees being evaluated utilizing a forced-distribution system. In other words, there should be a greater number of higher performers than would be expected with a normal curve because there are fewer poor performers in the company since they weren’t hired in the first place (see Figure 5.6).
Figure 5.6: Normal versus skewed distribution
Potential poor performers should be filtered out during the hiring process, so being forced to give adequate employees a lower ranking can cause morale problems and unproductive competition among workers.
Normal Distribution Negatively Skewed Distribution
Therefore, instead of fitting employees into an accurate distribution of performance, a forced- distribution system artificially imposes a distribution of performance that does not fit real- ity. The system is measuring relative performance (performance in relation to others) rather than absolute performance (a true depiction of the employee’s results). In GE’s case, it forced the ranking of employees at the bottom even if they were performing well or on par with expectations, resulting in the dismissal of solid-performing employees as well as a fearful and overly competitive work environment.
Goal-Driven Systems The best example of a goal-driven performance measurement system is management by objectives (MBO). MBO was first coined by Peter Drucker (1954) and then popularized by George Odiorne (1965). It was widely disseminated by manufacturing companies such as the R. G. Barry Corporation (see Chapter 1) and aims to provide mutually agreed-upon standards of work performance and results.
Creating and using an MBO system effectively requires clearly communicated and mutu- ally agreed-upon behaviors that produce measurable results that support the firm’s busi- ness strategies. The objectives must be specific, measurable, attainable, realistic, and timely
abc82339_05_c05_107-138.indd 123 1/4/16 3:08 PM
Section 5.1 Performance Measurement
(SMART goals; see discussion in Chapter 3) and, most important, they must be defined and agreed upon by both the employee and supervisor. While some claim that MBO is most useful for professional and managerial employees, the process works well for any level of employee. For example, R. G. Barry was extremely successful using an MBO-styled approach with its hourly, managerial, and professional employees. The success of the program was based on following the MBO guidelines, embracing the philosophy by the company’s management, and having a supportive organizational culture.
Goal-driven systems are based on goal-setting theory, which, as we discussed in Chapter 3, is one of the most heavily researched and supported theories of motivation. The focus on objectives, however, does not always lend itself to understanding what characteristics of the person performing the job resulted in success or failure. Also, objectives must be accepted by both employees and managers. Without agreement and “buy in” on both sides, goals do not direct behavior.
Behavior-Based Systems A behavior-based system rates employees on the objective actions performed. Because this approach is objective and based on observation, it is less susceptible to many of the errors that plague other, more subjective measures of performance. There are three main forms of behavior-based performance appraisal systems: critical-incident technique, behaviorally anchored rating scales (BARS), and behavioral observation scales (BOS).
Critical-Incident Technique
The use of critical incidents in job analysis was discussed in Chapter 4. This technique can also be used to evaluate performance. Critical incidents require the supervisor to keep an ongoing log of events, interactions, behaviors observed, and results for each employee. Super- visors and subordinates work together to identify critical incidents that distinguish between successful and poor performance. Periodically, the supervisor and employee meet to discuss the information so that the employee is not surprised when performance appraisal time comes along.
As an example, consider fire safety in a manufacturing environment. One safety guideline is to always clear away any potentially flammable materials before welding. Failure to do so could result in a fire that could spread very quickly in some manufacturing environments and result in both injury and monetary cost. This is a critical incident, and failure to clear away flammable materials properly would represent a major failure to perform on the part of the employee.
The critical-incident technique has both pros and cons associated with it. A pro for this type of appraisal technique is that it includes regular, ongoing communication about performance between the supervisor and the employee. The employee has a chance to interact with the supervisor and adjust expectations and goals or correct possible misunderstandings of supervisory observation. Additionally, the halo effect is mitigated and the use of facts and data collected throughout the year minimizes the impact of recent events unduly influencing the measurement and evaluation of performance.
abc82339_05_c05_107-138.indd 124 1/4/16 3:08 PM
Demonstrates effort and completes work in a timely and accurate manner
21 3
• works slowly, poor effort
• poor work quality, multiple errors, missed deadlines
• not proactive, avoids or resists development opportunities
• puts forth sufficient effort to accomplish work tasks
• acceptable work, not many errors, usually meets deadlines
• takes on extra tasks when assigned, engages in development opportunities when directed
• superior effort, little wasted effort
• high-quality work, errors rare, exceeds deadlines
• shows initiative, proactive, seeks out development opportunities
Needs Improvement Satisfactory
4 5 6 7
Superior
Section 5.1 Performance Measurement
Possible cons include that the supervisor must be fully committed to this approach and well trained in one-on-one facilitation, an expectation that should also exist for all forms of per- formance feedback but is often lacking. Additionally, the supervisor must make the time to maintain a file with detailed notes on each employee. The greater the number of employees reporting to a given supervisor, the less likely he or she will be inclined to keep detailed, ongo- ing records.
Behaviorally Anchored Rating Scales
Behaviorally anchored rating scales (BARS) represent a mechanism for measuring perfor- mance whereby supervisors connect rating scales with detailed examples of performance (see Figure 5.7).
Figure 5.7: Sample behaviorally anchored rating scale
Though using the BARS approach is time consuming, it provides meaningful examples, ongoing feedback, and clearly communicated standards.
Demonstrates effort and completes work in a timely and accurate manner
21 3
• works slowly, poor effort
• poor work quality, multiple errors, missed deadlines
• not proactive, avoids or resists development opportunities
• puts forth sufficient effort to accomplish work tasks
• acceptable work, not many errors, usually meets deadlines
• takes on extra tasks when assigned, engages in development opportunities when directed
• superior effort, little wasted effort
• high-quality work, errors rare, exceeds deadlines
• shows initiative, proactive, seeks out development opportunities
Needs Improvement Satisfactory
4 5 6 7
Superior
While the steps are time consuming, the result is specific and detailed and allows compari- son to other employees doing similar jobs at any organizational level. The steps include the following:
1. Maintain a file of critical incidents (discussed in the preceding section and in Chapter 4) by asking supervisors and others who know the job well to provide detailed information regarding what are successful and unsuccessful performance parameters.
2. Ask subject matter experts to group the critical incidents into a small number of dimensions or competencies.
3. Verify the groupings by asking a second team of individuals who know the job well to reallocate the critical incidents into the same or different clusters, thus validating or modifying the original groupings.
abc82339_05_c05_107-138.indd 125 1/4/16 3:08 PM
Section 5.1 Performance Measurement
4. The second group determines how effectively or ineffectively the behaviors described in the incident reflect actual performance on the job and make adjust- ments as needed.
5. The final step involves selecting a small number of the incidents (perhaps six) to represent the behavioral anchors for a given job, which are then used to rate the employee’s performance.
Apply It: Using BARS to Get Results
As an applied example, an entertainment services company was having problems with internal communication and job performance within its in-house tech services group. A consultant was hired to fix these issues, and after meeting with executives and internal customers, he decided to focus on improving performance measurement within the organization in order to tie it to measurable objectives and further corporate objectives. He developed and used a modified version of BARS that involved people knowledgeable about a given job who determined the correct behaviors and critical incidents with which the job could be anchored. The “rules” of the first group meeting required that they achieve true consensus about the incidents and required behaviors to ensure widespread acceptance and commitment to the new approach. Initially, however, the group attempted to use a majority-rule approach, which could have created a less-than-ideal result. While the first meetings were extraordinarily long with few results occurring, subsequent meetings and results were rapid and commitment was widespread. The second, or validating, group consisted of managers, supervisors, and executives, thus ensuring acceptance, support, and comprehension of the elements of the BARS program throughout all levels within the organization. Initially, the program started with the information technology (IT) department as a “pilot” program but was later used throughout the company.
The BARS approach provides detailed information with true numerical measurements based on realistic data from employees who know the job well. In other words, the BARS approach provides meaningful examples and includes on-the-job behaviors as a guide for evaluations. As a result, acceptance and commitment will be high and results may be tracked accurately and thoroughly. Feedback is ongoing, standards are clear and widely communicated, and there is less likelihood a rater will rate someone high on all dimensions simply because the employee was highly rated on another dimension (the halo effect).
Disadvantages with this approach include the length of time and effort involved by many people. Some managers, supervisors, and executives may perceive the effort as being nones- sential to productivity. To help change that perception, the program requires a senior-level executive champion to support and nurture the effort and a well-respected HR department to overcome resistance.
Behavioral Observation Scales
Behavioral observation scales (BOS) are similar to BARS in that the technique involves a process of identifying the key tasks for a particular job. The difference is that employees are evaluated based on how often they perform the required behaviors needed for successful job
abc82339_05_c05_107-138.indd 126 1/4/16 3:08 PM
Section 5.1 Performance Measurement
performance, and only those that result in a positive outcome are assessed. The scores for each of these observed behaviors are then added together to produce an overall performance score. Aspects of behavior are weighted to reflect the relative importance of the measure to the overall job.
A BOS is created using the same basic technique as a BARS and shares the same strengths and weaknesses. A BOS is time con- suming to maintain, and it is hard to ensure accuracy in a dynamic workplace. For this reason, they are less commonly used than either critical incidents or BARS.
Acceptance of Performance-Measurement Systems
As with goal-setting theory, which we discussed in Chapter 3, performance measurement must be accepted in order to be accurate. The goal of measurement must be understood, and the rater and the person being rated must trust both the reason why performance is being evaluated and how the results of that evaluation will be used. It is also usually a good idea to collect evaluations using multiple techniques, such as 360-degree feedback, that gather information from not just supervisors but also subordinates, colleagues, customers, and oth- ers who are knowledgeable of different aspects of the employee’s performance. Assessing performance via means of a variety of sources can result in different yet equally valid evalu- ations of the employee, depending on the assessor’s understanding of the job, dealings with the employee, and so on. These divergent ratings can create some challenges due to the dis- parate results, but they provide meaningful information that should be incorporated into the overall evaluation.
Employees are more likely to accept the results and believe they were rated fairly if eval- uations are made frequently, ratings are made by supervisors familiar with their work, employees have the opportunity to express themselves, and new goals were set based on the appraisal (see Cascio, 1991). On the rater side of the equation, evaluators must accept that performance measurement and feedback are important organizational goals and must accept the system as a means of reaching those goals. To achieve this, a well-designed performance- measurement system must be created and used. Proper communication and training must be given to explain the process, and continual evaluation of the system must be done to make necessary adjustments in light of changes in the legal, economic, and business environments.
In practice, employee objectives are often set at the beginning of the year, or at the end of the prior year, and then nothing is done until the mad scramble at the end of the year to get it completed before the deadline. This can result in just a cursory treatment of the performance measurement system. To help combat this issue, some companies are trying to move to a more fluid process where discussions, both formal and informal, are conducted throughout the year instead of only one time at the end of the year before raises are given. With ongo- ing discussions, employees should not be surprised when they get their formal performance appraisal because they’ve been kept informed of their performance throughout the year.
Critical Thinking
What is the difference between trait- based and behavioral approaches to measuring performance?
abc82339_05_c05_107-138.indd 127 1/4/16 3:08 PM
Section 5.2 Market Competition and Compensation Systems
5.2 Market Competition and Compensation Systems Knowing how employees perform is just one part of the equation. The other part of the equa- tion is how to compensate those employees for their performance. To help determine this, a comprehensive compensation system must be developed that accounts for the company’s culture, goals, and so on and that takes into account the market environment in which the company operates.
Compensation systems, as we mention numerous times in this book, can be a source of either a competitive advantage or disadvantage for employers. As the saying goes, “The people make the place.” Compensation and benefits are significant factors in not just the attraction and retention of employees but also the attraction and retention of the right employees.
However, what is the right compensation level or strategy for a given company? Paying too much or too little has implications for corporate strategy, competition, and profitability. In general, companies will establish a compensation strategy that leads, matches, or lags that of competitors. Paying at market means your midpoint will match the mean or arithmetic aver- age salary for that position, lagging the market will set a midpoint below the mean salary for that position, and leading the market will set a midpoint above the mean salary for that position.
Strategic Considerations
Both factors that are external and internal to the company should be considered when setting compensa- tion levels. Paying above market, even if your company can afford to do so, is not always the best approach. High pay can undermine low-cost strategies and tie up funds that could be invested in other areas such as research and development or expansion into other markets. Low pay levels, however, may make jobs less attractive to potential employees and cause current employ- ees to leave the company for better- paying positions elsewhere. Be sure to consider benefits (discussed in Chap- ter 7) and not just salary levels when developing corporate compensation strategies.
It is also possible to combine approaches. For example, some companies use a market lead strategy for highly skilled employees and those in key positions through the organization (managers, specialists, research and development, etc.) but use a market lag approach for less specialized positions where there are plenty of potential workers. With a combined approach,
Caia Images/Superstock
An effective compensation and benefits plan depends on a strategic consideration of the big pic- ture that incorporates both internal and external factors.
abc82339_05_c05_107-138.indd 128 1/4/16 3:08 PM
Section 5.3 Compensation Surveys
however, care must be taken to ensure that the differing policies do not result in friction across categories of employees.
Legal Issues—Antitrust Law
In collecting survey data, care must be taken to avoid any appearance of collusion across orga- nizations. Antitrust laws are in place to protect the free pricing of labor, and companies must avoid any suspicion of conspiring with other companies to set or fix wages (Fischer, 1985). More information, including recent court decisions, can be found on the U.S. Department of Justice Antitrust Division website (www.justice.gov/atr).
It is always recommended that you consult legal counsel if you feel you are in danger of violating any law. However, there are some basic steps that can be taken to help avoid the appearance of collusion. First, do not imply (or explicitly state) that future compensation practices are dependent upon similar behav- ior by other companies—in other words, that your company will match what your competi- tors are doing and nothing else. Second, de- identify data such that responses can’t be tied back to specific individuals. Third, when possi- ble, utilize outside consultants or vendors not directly affiliated with your organization. Last, avoid the appearance of “closed door” meetings with competitors or others in the industry.
With these safeguards in place, survey data can be a valid and useful tool in setting and modi- fying compensation strategy.
5.3 Compensation Surveys A compensation survey is a tool used to determine the typical incentives that are provided to employees across a specific job, industry, geographic region, and so on. Data is collected from multiple employers and is analyzed to develop an understanding of the overall amount of compensation paid.
The purpose of surveys is typically to gather information about competitors. In this context, competitors are those who recruit the same type (in terms of occupation, skills, experience, etc.) of employees as your company. This could be those who offer the same products or ser- vices as you or it could be companies in different industries. You are looking at competitors in the labor market, not necessarily competitors of your products and services.
Some surveys are conducted by individual companies, while others are commissioned by professional organizations, consulting firms, and/or governments. Due to changing economic and business issues, the data from compensation surveys may become out of date quickly. Most surveys conducted by professional organizations and government agencies are updated on a regular basis and clearly identify the date of data collection.
Critical Thinking
What legal considerations impact both performance measurement and compensation surveys?
abc82339_05_c05_107-138.indd 129 1/4/16 3:08 PM
Section 5.3 Compensation Surveys
Deciding What Information to Collect
The data collected in a compensation survey should be relevant for how the information is intended to be used. Therefore, the first step is deciding why survey data is being collected. Questions such as the following should be answered:
• Is the data being used for job redesign? • Is this for an existing or new position? • What is the company’s compensation strategy? • Who are the company’s competitors? • What is the target industry?
Once the context is understood, then it can be decided what information to collect, such as the following:
• Base salaries • Recent changes in salary levels (i.e., raises, bonuses, etc.) • Merit increases • Salary ranges • Starting salary • Incentives/bonuses • Educational requirements • Geographic location • Source of hire (internal/external) • Working conditions • Benefits provided (i.e., medical insurance, retirement plan, etc.)
One of the biggest decisions to be made is if a new survey needs to be conducted (either using in-house expertise or by hiring consultants). Depending on the need, it is possible that existing salary surveys are already available that can answer the desired questions. Caution should be exercised in collecting market data solely based on job titles since these can vary widely and encompass a huge spectrum of duties, responsibilities, and qualifications.
Custom Surveys
The advantage of custom surveys is that they can be tailored to fit specific organizational goals. In practice, however, it is often difficult to collect the data needed with in-house-developed surveys. One of the most difficult problems is that those firms you most want to collect data from understandably do not want to provide data to a competitor. Therefore, custom surveys are most commonly sourced out to consultants or consulting firms. This may be costly, and it is usually necessary to share the results of the survey with all of the companies that participated.
Existing Surveys
There are many existing sources for survey data. In the United States, the Bureau of Labor Statistics (BLS; www.bls.gov) is one of the most useful. The BLS has been collecting data on labor practices since the 1890s and updates the National Compensation Survey (NCS; www. bls.gov/ncs/) on a regular basis. The NCS provides data on quarterly changes in employer
abc82339_05_c05_107-138.indd 130 1/4/16 3:08 PM
Section 5.4 Integrating External Pay Rates (Pay Bands)
costs through the Employment Cost Index (ECI), quarterly employer cost levels through the Employer Costs for Employee Compensation (ECEC), and incidence and provisions of employee benefits.
Other sources of survey data are consulting firms such as William H. Mercer (www.mercer.com), Towers Watson (www.towerswatson.com), and the Hay Group (www.haygroup.com). Profes- sional organizations such as WorldatWork (www.worldatwork.com) and the Society for Human Resource Management (www.shrm.org) also provide broad surveys. Additionally, specialized organizations such as the Society for Industrial and Organizational Psychology (www.siop.org) and the American Society of Mechanical Engineers (www.asme.org) often prepare specialized surveys that are useful to the membership of the organization. Online resources such as Salary. com (www.salary.com) and O*Net (www.onetonline.org) also provide useful information.
Using Survey Data
When evaluating survey data it is important to make sure the compensation survey was con- ducted in as objective, reliable, and valid a manner as possible. The survey data can only be generalized to the extent that it was collected from a representative sample. For example, if data on a specific job was collected from only two companies, then calculating a mean and range for compensation levels would not be useful.
Survey results should also be reliable. Reliability is consistency. This means that the results were collected in the same manner across locations and would provide you with similar results if administered again. Be sure to examine how the information was collected and if the questions asked accurately represent the interpretations of information that are being provided. The report that presents the results should define terms and provide the reader with enough information to understand the data. Definitions should be provided for technical terms, jargon, and buzzwords. Adjustments might be needed to account for inflation, cost-of-living differences, and so on.
Finally, be sure the survey reports the infor- mation you actually need and in the detail required. Information on the total compensa- tion package, including bonuses and benefits, is typically more useful than just salary data.
5.4 Integrating External Pay Rates (Pay Bands) Once you have established what to pay based on your market, you then establish pay bands, also known as pay grades, as a method of maintaining flexibility while recognizing external market information. Pay bands are established informally any time more than one level of compensation is paid to employees with the same job title. A much more effective approach is to establish bands based on what the typical person in a job should be paid (i.e., the midpoint of the pay range) as well as the minimum and maximum levels that should be paid for a given job. These levels will be influenced by the company’s compensation strategy of leading, lag- ging, or matching the compensation amount of competitors.
Critical Thinking
How do compensation surveys aid in establishing competitive advantage?
abc82339_05_c05_107-138.indd 131 1/4/16 3:08 PM
Section 5.4 Integrating External Pay Rates (Pay Bands)
As an example, consider Table 5.1, which presents the pay scale utilized by the U.S. federal government (http://www.opm.gov/policy-data-oversight/pay-leave/salaries-wages/2014/ general-schedule/).
T a
b le
5 .1
: G S
ra ti
n g
sc a
le
A n
n u
a l
R a
te s
b y
G ra
d e
a n
d S
te p
G ra
d e
St e
p 1
St e
p 2
St e
p 3
St e
p 4
St e
p 5
St e
p 6
St e
p 7
St e
p 8
St e
p 9
St e
p 1
0
W IT
H IN
G
R A
D E
A
M O
U N
T
1 $1
7, 98
1 $1
8, 58
2 $1
9, 18
0 $1
9, 77
5 $2
0, 37
3 $2
0, 72
4 $2
1, 31
5 $2
1, 91
1 $2
1, 93
4 $2
2, 49
4 VA
R IE
S
2 20
,2 17
20 ,6
98 21
,3 67
21 ,9
34 22
,1 79
22 ,8
31 23
,4 83
24 ,1
35 24
,7 87
25 ,4
39 VA
R IE
S
3 22
,0 58
22 ,7
93 23
,5 28
24 ,2
63 24
.9 98
25 ,7
33 26
,4 68
27 ,2
03 27
,9 38
28 ,6
73 73
5
4 24
,7 63
25 ,5
88 26
,4 13
27 ,2
38 28
,0 63
28 ,8
88 29
,7 13
30 ,5
38 31
,3 63
32 ,1
88 82
5
5 27
,7 05
28 ,6
29 29
,5 53
30 ,4
77 31
,4 01
32 ,3
25 33
,2 49
34 ,1
73 35
,0 97
36 ,0
21 92
4
6 30
,8 83
31 ,9
12 32
,4 91
33 ,9
70 34
,9 99
36 ,0
28 37
,0 57
38 ,0
86 39
,1 15
40 ,1
44 1,
02 9
7 34
,3 19
35 ,4
63 36
,6 07
37 ,7
51 38
,8 95
40 ,0
39 41
,1 83
42 ,3
27 43
,4 71
44 ,6
15 1,
14 4
8 38
,0 07
39 ,2
74 40
,5 41
41 ,8
08 43
,0 75
44 ,3
42 45
,6 09
46 ,8
76 48
,1 43
49 ,4
10 1,
26 7
9 41
,9 79
43 ,3
78 44
,7 77
46 ,1
76 47
,5 75
48 ,9
74 50
,3 73
51 ,7
72 53
,1 71
54 ,5
70 1,
39 9
10 46
,2 29
47 ,7
70 49
,3 11
50 ,8
52 52
,3 93
53 ,9
34 55
,4 75
57 ,0
16 58
,5 57
60 ,0
98 1,
54 1
11 50
,7 90
52 ,4
83 54
,1 76
55 ,8
69 57
,5 62
59 ,2
55 60
,9 48
62 ,6
41 64
,3 34
66 ,0
27 1,
69 3
12 60
,8 77
62 ,9
06 64
,9 35
66 ,9
64 68
,9 93
71 ,0
22 73
,0 51
75 ,0
80 77
,1 09
79 ,1
38 2,
02 9
13 72
,3 91
74 ,8
04 77
,2 17
79 ,6
30 82
,0 43
84 ,4
56 86
,8 69
89 ,2
82 91
,6 95
94 ,1
08 2,
41 3
14 85
,5 44
88 ,3
95 91
,2 46
94 ,0
97 96
,9 48
99 ,7
99 10
2, 65
0 10
5, 50
1 10
8, 35
2 11
1, 20
3 2,
85 1
15 10
0, 62
4 10
3, 97
8 10
7, 33
2 11
0, 68
6 11
4, 04
0 11
7, 39
4 12
0, 74
8 12
4, 10
2 12
7, 45
6 13
0, 81
0 3,
35 4
So u
rc e:
U n
it ed
S ta
te s
Fe de
ra l G
ov er
n m
en t,
O ff
ic e
of P
er so
n al
G ov
er n
m en
t. h
tt p:
// w
w w
.o p
m .g
ov /p
ol ic
y- d
a ta
-o ve
rs ig
h t/
p a
y- le
a ve
/s a
la ri
es -w
a g
es /2
01 4/
g en
er a
l- sc
h ed
u le
/
abc82339_05_c05_107-138.indd 132 1/4/16 3:08 PM
Section 5.4 Integrating External Pay Rates (Pay Bands)
Consideration must be paid to the number and width of the bands as well as the overall cor- porate strategy regarding compensation. The next section provides steps that can be followed to establish bands.
Steps in Establishing Pay Bands
The first stage in establishing pay bands is to evaluate jobs using a job evaluation or similar process (see Chapter 4). Jobs that are similar are grouped together, and the midpoint of pay is established by considering which jobs in the band are considered to be key or benchmark positions. All jobs are then placed within the established bands. To maintain flexibility, there is typically overlap in pay levels across adjacent bands.
Many techniques have been proposed on the best way to determine the width of each pay band. The easiest of these is to just go 20% above and below the midpoint for determining minimum and maximum amounts. The key is to be consistent and clear about the methodol- ogy used.
There are no established standards, but there are some guidelines that will help establish consistency. First, drive the program from the top. Executives and managers should serve as subject matter experts in sorting jobs into bands as well as establishing the midpoint, mini- mum, and maximum pay levels for each band. Also, consider strategic considerations: Does the company want to lead, match, or lag the market when it comes to compensation level? Establish in advance how often bands will be evaluated and adjusted. Do not just put the bands in place and then never look at them again.
Following are the steps that outline the process of establishing pay bands.
Step 1: Determine Overall Pay Range Determine overall minimum and maximum pay across the organization. The minimum will be for the lowest grade and the maximum will be for the highest grade. Use a ranking of all company positions or job groups and current salary survey data relative to those positions to set these numbers. Be sure to incorporate the company’s compensation strategy to lead, lag, or pay at market.
A pay range will generally spread plus or minus about 20% of the midpoint, but this number is simply a guideline and not a rule. Business concerns and organizational strategy should be used to establish the range. For example, consider if the lowest-paid position is a sales clerk and you want to pay at market, and survey data for that position shows an average sal- ary of $20,000. The minimum salary would by $16,000, the midpoint would be $20,000, and the maximum would be $24,000. This would mean that $16,000 was the absolute minimum salary that would be paid by the organization. You would then repeat this process for the highest-paid position to determine the maximum salary the company would pay. It is also not necessary to establish the same range across pay bands.
abc82339_05_c05_107-138.indd 133 1/4/16 3:08 PM
Section 5.5 Considerations Due to Company Size
Step 2: Determine Number of Bands Determining the number of bands to be used is both an art and a science. When possible, base this number off of the results of a job evaluation. Larger companies will typically end up with more bands than smaller companies. Other factors, such as the number of geographic loca- tions and the number of job titles, will also determine the number of bands used; the more locations or titles an organization has, the more bands it will need. There is no best number of bands; however, simplicity is usually preferred over complexity.
Step 3: Decide on Range per Band Next, set a minimum and maximum for each grade. The maximum of one grade should usu- ally overlap the minimum of the next-higher grade. Midpoints for grades can be determined either using survey results or based on benchmark jobs that are considered to be paid accu- rately. Be sure to maintain flexibility by making the bands large enough to deal with differ- ences in performance.
Step 4: Create a Pay Grade Chart Using the sum of the minimum and the maximum, calculate the midpoint by dividing that sum by two ([Max + Min] / 2 = Midpoint). Another way is to determine the midpoint for each band and then calculate the minimum and maximum for each band based on the spread used, such as 20%. Refer to Table 5.1 for an example of a completed chart.
Step 5: Review and Revise As we mentioned above, do not just set bands and then never evaluate them. Set a policy as to how often the pay bands will be evaluated and who will be in charge of making sure it is done. Salary increases, inflation, and cost-of- living adjustments to pay will necessitate the updating of pay bands to keep them accurate. As a rule of thumb, evaluate bands at least every three years.
5.5 Considerations Due to Company Size Performance measurement serves the same purpose in organizations regardless of the size of the company. What changes are the resources available. Smaller companies may have a more difficult time identifying appropriate individuals to serve as evaluators of performance. Addi- tionally, training and budgetary constraints may limit the complexity of the methods that can be used. Whereas large companies can afford to hire consultants to develop customized mea- sures, smaller companies typically have to depend on the use of off-the-shelf measures. Care must be taken to remember that the context in which performance is measured determines how the results of that measurement can be interpreted.
Critical Thinking
What are pay bands and how are they useful?
abc82339_05_c05_107-138.indd 134 1/4/16 3:08 PM
Summary & Resources
Much of the same situation exists when considering compensation surveys. Larger compa- nies have more resources and often have more readily identifiable competitors with dedi- cated compensation departments that are willing and able to share information on pay levels and compensation practices in order to gain similar information in return. Smaller companies are more likely to rely upon surveys conducted by professional or national organizations.
The practice of setting pay bands is often neglected by smaller organizations due to the perceived complexity in implementing and maintaining the bands. Also, small compa- nies often do not think they need pay bands since they just have a small number of employees and very few overlapping job titles. Pay bands, however, are valid tools for small companies as well as midsized and large companies.
Critical Thinking
Does a systematic approach to performance measurement and pay bands apply to small companies? Explain your answer.
Case Study A Moat for Your Castle Inc.: Get in Sync
A Moat for Your Castle Inc. has continued to grow and, based on the insight gained from Chapter 4, has managed to stem the hemorrhaging of long-time employees and convince them to remain with the company. In addition, job analyses were used to identify the key KSAOs and competencies needed for successful job performance. The company is continuing to make a profit, but with increased overhead, the margin is not as high as before. The company needs to find a way to rein in costs, the largest being payroll, without impacting the morale of employees.
How could a formal performance-measurement process help the company make sure employees are being compensated appropriately? How can knowledge of individual employee performance help with this process?
How can the job analysis and job evaluation techniques covered in Chapter 4 be combined with a consideration of market factors to establish a pay structure? Would pay bands be useful for the company?
Summary & Resources
Summary Performance is a multidimensional concept that cannot be interpreted without considering the context or environment in which it is being measured. Measures of performance can be either subjective or objective and can assess traits, on-the-job behavior, or the accomplish- ment of organizational objectives. A number of approaches, such as focusing on traits, com- parisons, goals, or behavior, were discussed.
abc82339_05_c05_107-138.indd 135 1/4/16 3:08 PM
Summary & Resources
All attempts to measure performance are prone to errors such as halo effect, bias, contrast effects, regression to the mean, strictness, and leniency. In order to address these errors, many different techniques have been developed to measure performance. There is no one best technique; instead, the method used needs to be based on the reason why performance is being measured and upon environmental and organizational constraints.
Compensation surveys allow companies to gain information about competitor practices and set internal compensation accordingly based on strategic decisions to lead, meet, or lag the market in terms of compensation level. Pay bands represent a method of establishing inter- nally consistent compensation levels that complement organizational compensation strategy while still leaving the flexibility to reward all employees accordingly.
Small and midsized companies without a dedicated compensation department often neglect to establish formal compensation strategies due to the perceived difficulty of the process. However, the advent of the Internet and availability of information has brought competitive compensation within reach of most organizations. Compensation does not have to be viewed as just the largest expense for most organizations. It can also be a competitive advantage.
Key Terms
alternation ratings Method of perfor- mance comparison where the highest- ranking employee is identified, followed by the lowest; this process is then followed until all employees have been evaluated.
behavior-based system Evaluation of per- formance based on observable actions and the completion of tasks.
behavioral observation scales (BOS) A technique for measuring performance in which employees are evaluated based on how often they perform the required behav- iors needed for successful job performance and only positive behaviors are assessed; similar to BARS in that it involves a process of identifying the key tasks for a particular job.
behaviorally anchored rating scales (BARS) A mechanism for measuring per- formance whereby supervisors connect rating scales with detailed examples of performance.
bias The tendency to allow differences such as race, gender, age, sexual orienta- tion, and so on to impact the evaluation of performance.
compensation survey Tool used to collect information about compensation practices across jobs, employers, industries, geograph- ical regions, or other aspects of interest.
contrast errors Occur if a rater evalu- ates one employee’s performance against another’s.
criterion A sample of performance that represents the successful attainment of job goals.
critical incidents A method of perfor- mance measurement that requires the supervisor to keep an ongoing log of events, interactions, behaviors observed, and results for each employee.
abc82339_05_c05_107-138.indd 136 1/4/16 3:08 PM
Summary & Resources
error Anything that distorts the accuracy of perceptions; it can be positive or negative as well as systematic or random.
forced distributions Method of perfor- mance comparison that requires the evalua- tor to place a percentage of people in groups such that a few employees are considered to be above average or star employees, most employees are performing adequately, and a few employees are below average or poor performers.
halo effect Ratings on one or more job dimensions affect the ratings on others, either positively or negatively.
interpersonal barriers Raters and those being rated may have different ideas of stan- dards being used to rate performance.
management by objectives (MBO) Pro- cess of defining objectives such that both management and employees accept them and understand what needs to be done in order to achieve them.
objective measurement Measures of performance that are easily quantifiable and externally verifiable.
paired comparisons An evaluation of per- formance in which every employee is com- pared against every other employee being evaluated.
pay bands Strategy designed to promote flexibility by establishing pay ranges for dif- ferent classifications of jobs; also called pay grades.
performance appraisal The process of evaluating employee performance.
performance management The process of identifying, evaluating, and developing the performance of employees.
political barriers Managers not utilizing measurement systems properly in order to avoid conflict or to advance their own agendas.
predictor Measure intended to forecast successful job performance.
similarity effect A source of error that may impact the performance appraisal to the extent that the supervisor believes the employee has similar beliefs, attitudes, or personality characteristics and lets that impact the rating given to the employee.
subjective measurement Measures of per- formance that are dependent upon human judgments and opinion.
traits Characteristics or aspects of the indi- vidual related to job duties and results.
abc82339_05_c05_107-138.indd 137 1/4/16 3:08 PM
abc82339_05_c05_107-138.indd 138 1/4/16 3:08 PM
