The Anatomy of a Compensation Survey (And Why It Matters)

The Anatomy of a Compensation Survey (And Why It Matters)

The Anatomy of a Compensation Survey

Whether your organization is looking to market price a brand new position for a job posting, make an offer to a potential new employee, overhaul your entire salary range structure, check to see if your wages are keeping pace in your industry, or simply trying to stay true to your compensation philosophy, there is one tool that is indispensible—the compensation survey.

Comp surveys come in all shapes and sizes and can vary in terms of how and what data is reported depending on your needs. Admittedly, sometimes the complexity and shear volume of the data tables in a comp survey can be overwhelming and it can be tempting to cut corners. With a quick Google search you can almost instantly, and cheaply (or freely), get to the single dollar figure that you want. Of course while this search engine methodology sounds a lot less painless than thumbing through hundreds of pages of compensation data, there are significant issues with validity, reliability, and just plain being wrong that could end up causing your organization more time and resources in the long run than you saved in the first place.

To help keep the headaches to a minimum and get you the best data possible when it really matters, here is a breakdown of some of the key components of compensation surveys that you are likely to encounter the next time a comp project lands on your desk.

Effective date

Defined

The date or range of dates during which the individual employees were being paid the reported rates.

Why it matters

In short, the effective date tells you how up-to-date the data is that you are using. This is different from the data collection period, which only tells you how long the survey was open for participation. Analyzing thousands of individual data points takes time, so there is likely to be some lag between the effective date and when the report is published.

The key is that the time frame should be standardized so that all of the data collected is from the same period.

Demographic breakouts

Defined

The specific breakouts can vary from survey to survey, but typically include options such as employer size, industry, company revenue, non-profit status, union status, and geographic location.

Some reports also provide breakouts related to the number of years the individual employees have been with their organizations.

Why it matters

Most surveys provide overall demographic information for categories that match up to the breakouts reported within each position. Looking at these demographics can help you decide whether or not the report is going to be a good fit for your organization. Is the data dominated by large manufacturing organizations with massive revenue figures, but you are a small non-profit? The survey may or may not be the best fit, depending on the other resources available to you.

If you are similar geographically and have common office or administrative type positions, the reality is you probably need to know what other employers in your area are paying regardless of industry, etc.

Even if you can’t keep up with the reported dollar figures in the report, you can at least be prepared to demonstrate what other non-compensation type benefits your organization has to offer in order to stay competitive in the market.

Sample size

Defined

For the full survey report this is the number of employers that provided data during the participation period. Within each position reported, this can be both the number of employers reporting data as well as the number of individual employees for which data was reported.

Why it matters

Within each position, employer and employee count is critical for two major reasons.

  1. First, survey publishers are required by the federal government to exclude any jobs/breakouts for which fewer than five employers provided data.
  2. Second, the more data, the more consistent and accurate the results will be.

Positions or breakouts with only single digit employer or employee counts are more prone to be influenced up or down by variations in the data both within the position for the current survey, as well as if you are comparing data year over year from the same survey.

While this doesn’t mean the data is “no good”, if the stats seem way off from what you expected, checking the sample size can be one way to explain the variation.

Job titles, descriptions, numbers, and families

Defined

These terms are probably the most familiar to HR folks working with compensation survey reports. Job title is the name of the position, the number is a number assigned to the job, and the description is a summary of the positions major job duties/functions.

The job family is a grouping of positions that typically fall within the same department or functional area at most employers.

Why they matter

The job numbers are usually assigned base on job family and both are really more for purposes of organizing the data within the survey than interpreting the data itself. The job title and description should be used in combination to determine whether or not the job you are trying to price out at your organization is a good “match” with the title/description in the survey.

Even though a job title from a survey may be the same as what your organization uses on the surface, making sure the description itself matches at least 70% of your organization’s description will give you the best data comparison.

Median

Defined

This is the value at which 50% of the employers pay at and below, or at and above. The median is the middle number and is derived when data points are arrayed from lowest to highest. If there is an even number of data points, the median is derived by averaging the middle two data points.

Why it matters

Because of the way it is calculated the median, as a general rule, tends to be less influenced by possible outliers than other measures of central tendency and therefore is the key statistic you will want to glean from a compensation survey for each position.

Furthermore, the median is also how you arrive at the gold standard for compensation a.k.a. the “market rate”. This figure is calculated by averaging the medians for the same position from three different sources.

Other Percentiles

Defined

These percentiles are calculated much like the median, but using different a percent of the employers paying at and below (10% and 25%), or at and above (75% and 90%) each number.

Why they matter

Picking the “right” percentile for your needs will depend largely on your organizations compensation philosophy and/or what your comp project is trying to achieve.

Does your organization aim to pay well above “market” in order to attract and retain the best and brightest? Then the 75th percentile might have the dollar figure your need.

Or do you have a robust benefits package and company culture that allows you to still maintain an excellent workforce, but maybe don’t need to pay out quite as much in cash compensation? Then perhaps you can look at the 25th percentile for guidance.

Employer and weighted averages

Defined

Average per employer is determined by adding each employer’s rate of pay for the reported position and dividing by the total number of employers reporting data for that position. Weighted average is determined by adding the rates of pay for each employee in a position and dividing by the total number of employees in that position.

Why they matter

These averages are still measures of central tendency for the “going rate”, but are not as highly upheld in the eyes of most compensation experts as the median.

When a report provides both the average per employer and the weighted average, comparing the two figures can help illustrate whether or not certain employers in the sample tend to pay more or less than others.

Types of pay reported

Defined

This could include anything from just straight base pay to bonuses to commissions to stock options. When these later types of pay (apart from base pay) are included in calculations it will likely be indicated by using terms such as “variable pay” and “total compensation”.

The same statistics, median, percentiles, averages, can all be reported out within these larger categories.

The individual survey should specify exactly what kinds of pay were included in each category near the beginning of the report.

Why they matter

Again, this is a matter of making sure you are paying competitively and comparing apples to apples. Particularly for higher-level positions or those positions that may receive a large chunk of their total income based on commission, knowing what types of compensation are included in the number in front of you is key.

Potential employees at the executive level or in sales will likely expect you to spell out much of this additional compensation information when reviewing an offer of employment and not addressing these items may cost you a talented employee.

Conversely, if all of the salaries seem excessively high to you within a particular survey, you might want to check what kinds of pay participating employers were asked to include in their submissions.

Compensation & Benefit Survey Data

ERC publishes compensation and benefit survey data covering local, regional, and national samples across a wide variety of industries.

Access Survey Data