Page 194 - Bruce Ellig - The Complete Guide to Executive Compensation (2007)

180 The Complete Guide to Executive Compensation

Selecting the Jobs
Unfortunately, not all jobs can be surveyed (due to a combination of magnitude of effort and
lack of comparability). However, a careful attempt should be made to ensure that the survey
has a good balance of horizontal and vertical representation. Horizontal representation
requires the inclusion of the various organization functions (e.g., research, production, mar-
keting, sales, finance, legal, and employee resources). The vertical requirement focuses on
promotional career paths within a particular function (e.g., plant manager to general pro-
duction manager to vice president of production). The number of jobs to be surveyed is a
function of availability (and extent of cooperation of the surveyed companies) as well as the
method of evaluation. Point factor plans require surveying only a few key jobs representing
the various levels, whereas market pricing requires surveying as many jobs as possible,
perhaps 25 percent or more.

Job Comparability
Before obtaining any data, it is important to determine the manner in which the surveyed
companies will be reporting on comparable jobs. Essentially, there are three methods: job
matching, job evaluation, and multiple regression.
Job Matching. Job matching is the most common method of ensuring comparability of
data. It requires matching jobs to ensure that similar responsibilities exist. This is important
because job content can vary significantly between two organizations, even though the jobs
may have the same title. For this reason, organization charts showing reporting relationships
along with job descriptions are included to maximize good job matches.
Nonetheless, it becomes increasingly difficult to find companies with similar positions
at senior levels. In many instances, the differences in job content for the same job title in
different companies are as great as, if not greater than, those for jobs with dissimilar titles.
Further, the implicit assumption is that like responsibilities will be valued comparably by
other organizations. Unfortunately, there are too many instances where this is simply not
true. Be it a formal job-evaluation plan or a simple ranking program, different companies
come up with different perceptions of the importance of different disciplines. Thus, even if
two jobs in two different companies were identical, the pay practices of the companies would
differ if they placed different values on the position.
Another difficulty many companies have is matching up to a company with either fewer
or more management levels. For example, assume a survey has data on the top functional
heads at the corporate, group, and division levels. How does a company with four levels—
corporate, business, multidivision, and division—relate to this data? The answer: by match-
ing first those most comparable (namely, corporate and division) and then interpolating (e.g.,
between corporate and group for business, and between group and division for multidivi-
sion). Obviously, less confidence should be placed in data that has to be massaged in this
manner than in data that matches well.
Another technique that may be employed in job matching is adjusting for degree of devi-
ation. For example, the data received from each company could be weighted based on degree
of closeness of fit with the survey job. Therefore, as shown in Table 5-8, a company that had
a job either noticeably heavier or lighter would be given a weighting of 0.5, whereas those that
were only slightly different would be given a weighting of 0.8. This technique is typically
called survey leveling.