We live in an era when data is king. Now more than ever, everything is counted, analyzed, and reviewed in almost every facet of our lives. Over the last several decades, most organizations have significantly increased the amount data available, but are struggling now with how to analyze it in a meaningful way that actually benefits their organization. Labor market analysis is no different. Strong advances have been made in data collection and now we more data than ever before. However, how we analyze the data, what it actually means, and how we use it to manage compensation are concerns most of us deal with on a routine basis. A simple illustration of a single job that I have used with clients demonstrates the importance of answering these questions.
If we collect data on a job that pays $25,000 at the midpoint from three peer organizations. In most markets, we would want to collect more than three peers, but three are utilized in the illustration to keep it simple. Figure 1 summarizes the results of each peer. The matches vary in staffing, midpoint pay, and match quality. In this case, the largest employer pays the most at $25,500 while the smallest pays the least at $23,000. However, the smallest possibly due to duties assigned, span of control, or differences in organization of work is a less than ideal match. Figure 2 presents four common methods of analyzing the results in Figure 1.
Figure 2 reports some rather large differences in results. Moreover, the conclusions with these results would be quite different. A simple average would indicate that the employer is paying approximately three percent greater than the market. The weighted average taking into account the size of the respondents finds that the employer is ten percent behind the market. The matched average or average that includes weights for the quality of the match shows a less than one percent differential. The refined match which weights after dropping the weakest matches places the employer about one percent behind market. It should be evident that we could draw very different perspectives on the market placement of this job depending on the approach we select. Furthermore, this issue is exacerbated by the fact that these differentials are magnified by the volume of jobs included in the analysis.
So, is this job ahead, behind, or at market? Should we take some type of action? The only way to answer these questions is to have more information before, during, and after the analysis process.
The biggest thing is to really take the time to understand the meaning of the data as well as the analysis. There are three major steps we can follow that will minimize the chance we have these quandaries:
- Summary of the analyzed classifications
- Analysis of results
- Overview of significance
Summary of Analyzed Classifications
The first component of the analysis as well as reporting process should “set the stage” for the analysis results. The minimum that should be presented is the name of the classification, a short description of the class, current pay range, and distributional characteristics of the associated range. In addition, relevant survey data should be added to the summary that includes number of respondents, type of respondents, quality of each match on some rating scale, and any distributional concerns related to the respondents. Now that the reader can understand the source, relevance, and scope of the data, the results can be presented in a more meaningful way.
Analysis of Results
There is some debate over the analysis and presentation of results in a labor market survey. Typically, results are presented in a number of ways to capture the relative position of the market vis-à-vis the employer conducting the analysis:
- Ranges values(minimum, midpoint, and maximum)
- Range spread (percentage)
- Actual salary values
There are two differing points of view related to the best method of comparative analysis. The first camp favors range comparison as the best method of assessing market positioning. The merits of this approach include ensuring that the breadth of options are included in the analysis, potential earning power is taken into account, and individual actions have less effect on overall results. The opposing camp focuses on comparing actual salaries. This approach centers on the differentials that exist in actual individual pay and provide a snapshot of actual pay practices. However, it suffers from comparability weaknesses related to mitigating factors that all impact compensation levels, such as time in class, longevity, pay increases, and general pay plan mobility.
Once the unit of analysis is selected, the results should be summarized with the following:
- Standard deviation
Overview of the Significance
Once the data is collected and analyzed, it must be interpreted. What does it really mean? A good set of questions to ask to assess the significance for each job pertains to the type of respondent: talent competitor, similar industry, or regional leader. A talent competitor takes employees from your organization as well as you recruit from them. Data collected from a similar industry would be relevant since the products and services produced by the organization most resemble that of the employer conducting the study. A regional leader is typically a large employer that possesses the size and influence to define and adjust the resources and perceptions in the labor marketplace.