Salary Surveys and Understanding Market Data
July 2008
Workforce Management Group
Determining appropriate pay levels for employees is one of the biggest challenges
managers and human resource professionals face. Paying too little may cause an
employee to leave for a better paying offer elsewhere, or it may cause an
employee to feel undervalued for the work they perform. Paying too much causes
a higher operating cost for the organization which impacts the bottom line. The
balance is found through using salary surveys which provide compensation data of
comparable positions in organizations. This report will review the types of
compensation surveys, the data components, and how to use the data when
comparing internal positions to keep pay levels competitive.
Types of Surveys
There are several types of survey data that should be considered when making
that comparison of organizational positions to external jobs, better known as
“market pricing” jobs. Formal compensation surveys usually administered by
compensation consulting companies (Watson Wyatt, Mercer, Hewitt, etc.) are the
most common. These survey sources follow a rigorous procedure for collecting,
verifying and reporting data. Compensation professionals generally provide the
data submission and follow the standards for reporting data. Respondents are
careful to accurately match jobs only when 70% or more of the job responsibilities
are equal. These surveys tend to be more expensive and often take more time
between data submission and data availability.
Many trade associations and professional organizations do a compensation survey
of their members. These are useful because they are usually strong job and
industry matches. However, they sometimes are less thorough with data
verification and reporting accuracy. Many times incumbents submit their own data
in these surveys; so survey results should be viewed with that understanding.
There are also websites that report salary information. These have historically
provided anecdotal information with some questions regarding accuracy and
completeness of the information. The biggest drawback has been the data is self-
reported, similar to some trade/professional organizations as noted above.
Visitors to the websites are asked to enter there own data as a requirement to
receive information. Some of these sites have improved their data management
and reporting capabilities over the years, or they have other options besides
reporting only self-reported data.
Regardless of the survey source, World at Work in a September 2007 article
“Understanding and Managing Compensation Survey Data Conflicts” reported
these points to “Sizing Up Compensation Surveys”:
Is there substantial participant input in the design of the survey through a steering
committee, advisory board or similar participant group?
How robust are the job descriptions and leveling charts?
Are job descriptions and leveling charts periodically reviewed and updated?
Do survey participants regularly engage in peer-to-peer job-matching discussions?
Does the survey cover necessary compensation elements (base, bonus, sales
incentives and long-term incentives)?
Does the survey have a robust method for collecting non-base compensation
elements?
Do all firms participating in the survey provide data on all compensation elements?
Do survey managers routinely provide direct feedback on submission quality?
Is feedback regularly solicited from participants?
Does the survey manager provide a designated contact for participant questions?
Can you actually speak to someone when you need assistance?
Are the data elements included in the survey reports clearly explained and
documented?
Are all compensation computations clearly documented and available to participants?
Does the reported date have an acceptable degree of face validity? (For example, is
the 50th percentile similar to the average?)
Are reweighting rules incorporated into the survey report to ensure no single firm
represents more than a specified percentage of the survey results?
Are minimum reporting guidelines implemented in accordance with safe-harbor rules
(e.g., no data is reported when fewer than five firms match)?
Are custom peer-group reports readily available (e.g., within hours)?
Are reweighting rules applied to custom peer-group reports?
Are there useful and user-friendly online survey tools?
Are survey reports available in both PDF and worksheet formats?
Is the survey produced on a timely basis?
One additional word of caution in using any source of compensation information
that is self-reported data. Self-reported data is compensation data that comes
from the incumbent. Some data sources ask for salary information from trade
association members or visitors to their website. The veracity of self-reported data
should be questioned. Did the respondent match to the right job? Do they
understand leveling? Will they tend to exaggerate compensation information?
Some sources that rely on self-reported data follow a rigorous data cleaning and
verification process others report data that may be less reliable and accurate.
What is reported in Salary Surveys?
The basic data reported in salary surveys is:
Base salary
Variable compensation
o Annual/short-term
o Long-term
Total Cash
Number of incumbents
Number of organizations
Few salary surveys report benefit data
The data is commonly reported using these statistics:
Average or Mean – e.g. the sum of the reported salaries divided by the
number of companies reporting.
Weighted Average – e.g. the sum of reported salaries divided by the total
number of incumbents.
Median – the middle value in a list of values sorted from high to low
Percentiles – 25th, 50th, 75th – e.g. the 25th percentile means 25% of the
reported values are less
The weighted average is used most commonly since this statistic weighs each of
the incumbents equally. Some salary surveys only report the average or mean
instead of the weighted average, in which case that value is utilized. The next
most common is the median and is useful because extremely high or low values
don’t influence the value of the median. The percentiles can be useful if there is
an intentional philosophy of paying higher or lower than average.
How the data is sliced
Since the value of jobs can differ for various reasons such geographic area
(region, sub-region, state, metropolitan area, and etc.), industry and employer
(asset, number of employees) size, most surveys report salary data by these
different cuts or scopes. Depending on the number of incumbents and companies
reporting data, it can be possible to find job data for employers in a specific
geographic region, industry, and company size all in one.
Most jobs are reported by job level. These levels are from entry to senior and from
individual contributor to management. This is particularly helpful when reviewing
job families.
Aging Survey Data
Survey data is usually collected as of an effective date. The effective date can
vary from survey to survey, but is still captured as one point in time for each
survey. It is generally accepted that the value of jobs continuously grows
throughout the year. So, in order to keep the reported values of survey jobs with
different effective dates consistent, the data should be aged to a common date.
There is an assortment of ways to select that common date. Picking a date in the
middle of the year (July 1) provides an average estimate of the value of the jobs
through out the year. Using the current date gives an estimate of today’s value of
the job. Additionally, using the same date the salary ranges are effective keeps
the market data in synch with the salary structure. Regardless of which reason
used to select a common date to age the survey data, be consistent so that all
market pricing is as of the same date.
Market Pricing a Job
Market pricing a job is as much an art as a science. There is always a balance
between the number of incumbents/companies reported in the survey and the
match in terms of scope (geography, industry, company size and job level). In
order to be comfortable with the number of incumbents reported, the industry or
geographic region may not be as specific as one would like – a bit of a give and
take to get a large enough incumbent group.
In order to do the best job of matching survey data to the internal job, it’s
imperative to know the organization’s jobs both in terms of content and level (entry
to senior level and individual contributor to manager). Don’t rely on job titles; use
the job description and talk to the manager to determine the job content. Jobs
should be priced based on job content alone.
The association’s internal equity is also a consideration. Know how the
association values different sets of responsibility in relation to the market.
Consider what skills the association values differently than the market. For
example, Farm Credit entities typically value credit leadership responsibilities
higher than the market. This makes sense because Farm Credit’s main product is
extending credit to customers and thus the impact of that position is greater than
some comparable jobs in the market.
Decisions will need to be made throughout a market pricing project as to what kind
of match to use. It’s important to be consistent with these decisions. Questions
may arise such as:
Should I use as many matches as I can find?
Should I rely on one or two of the most solid matches?
Should I use only one match from any source?
Can I use both geographic and industry scopes?
As much as possible, be consistent with the scopes used. For example, if a sub-
region doesn’t have as many useful matches as the region, then use the region for
all matches; particularly those within the same job family and/or job level. Note
that a broader industry scope may need to be selected so that the matches used
have adequate incumbents.
There will invariably be some jobs being market priced that have responsibilities in
different job families. These are jobs that are a blend of responsibilities and
require using different job family matches. If the job is half accounting and half
credit analysis, then weight the accounting matches 50% and the credit analysis
matches 50% to get an accurate picture of the true compensation level.
What do we learn from market pricing a job?
The goal of market pricing a job is to determine its Estimated Market Value (EMV).
The EMV can then be used to determine the target salary or to place the job in the
correct grade (or both). The EMV can be either in terms of the base salary or the
total compensation (includes variable pay) for the job.
Summary
Compensation data is readily available to organizations, and it is equally as
available to employees. Different surveys sources gather data differently, ranging
from a thorough analysis to self-reported data. Knowing and applying sound
practices regarding the usage of salary survey information will result in a
consistent application throughout the organization. Being comfortable with the
data used to may pay decisions is critical; if managers don’t fully understand this
aspect it will be difficult for them to discuss it with employees.
Pay is a critical aspect of the total rewards package for employees, and it’s also a
reason why employees leave the organization. Instilling sound practices to review
pay levels and make decisions is what best practice organizations do to reward
and retain employees. According to Capital H Group, organizations see talent as
their last competitive business advantage and are willing to invest to get the talent
they need to position themselves for the future. Compensating competitively, and
sometimes creatively, will help Farm Credit organizations attract and retain that
talent.
Additional questions or assistance with salary surveys and market pricing jobs?
Please contact either compensation consultants from the Foundations team:
Greg Law
Senior Workforce Management Consultant
greg.law@farmcreditfoundations.com
1-800-322-6554
Resources:
Cardinal, Kenneth R. (2007) Understanding and Managing Compensation Survey
Data Conflicts, Workspan Magazine, September 2007
Capital H Group (2007) Challenging Conventional Wisdom: A Caution About
Compensation Survey Data, News You Can Use, January 2007
Fox Lawson & Associates, LLC (2006) Making Sense of Market Data, Newsletter,
Volume 10, Number 2
Grigson, D., Delaney, J., & Jones, R. (2004) Market Pricing 101: The Science and
the Art, Workspan Magazine, October 2004
WorldatWork C17: Market Pricing – Conducting a Competitive Pay Analysis,
Course Materials, 2006