Data in its raw form, a string or table of numbers, is of little to no value in compensation. Data has to be summarized, described, and presented with particular analysis so that a decision can be made. The key to understanding compensation market data is to report data in a way that is easy to understand.
In the WageWatch Market Competitive Survey, subscribers select their competitors which create a custom set of data from peer organizations. The PeerMark Report™ contains useful statistics to measure the location of data with the set. The report shows where the central part of the data is in the form average and percentiles.
In prior blogs, we have presented central tendency in part as a discussion of average (simple mean) and median. Since then we have also introduced weighted average to the report which is another measure of central tendency. These points are provided on the report because all are needed to answer the question, ‘Which statistic to use for benchmarking?’ The answer is ‘It depends’.
The weighted average is used when the number of incumbents in the competitive set needs to be taken into account such as when comparing data sets that contain both very small and very large companies. In an extreme example, if we had a city with only two hotel companies one with 10 housekeepers and the other with 1,000 housekeepers – then the one with 1,000 housekeepers dominates the marketplace and would determine the prevailing rate for this job. Here, weighted average would best represent the market rate.
The simple mean, also called average, is an unweighted calculation. Weighing competitors equally is a hedge against the unknown and is the recommended calculation when there are other competitors in the marketplace that have not participated in the survey. We do not know if the missing companies are big or small, pay low or high. This is why the simple mean is considered representative of market values.
There is an advanced metric called the Combination Mean which is the average of the simple mean and the weighted average. This gives equal importance to the number of incumbents and number of companies in the set.
We introduced percentiles in Part 1 of this blog topic. The percentiles and averages together describe what the shape of the data distribution of the competitive set “looks” like. The bell shaped distribution with a peak in the middle and tails on both ends, is called a normal distribution. While a true normal distribution is purely theoretical, there are many examples in the data are approximately normal.
The closer the market median and the market average wages are to each other, the more normal the distribution. In most reports, the median and averages rates will be within a dollar of each other. If the average is much greater than the median, the hump of the curve is skewed, leaning to the right creating a long tail on the left. This indicates a “hot job”, where the trend is to pay above market median. If the average is less than the median, the skew is to the left creating a long tail to the right indicating the job is cooling down.
WageWatch reports provide data tools and report statistics for analysts of all experience levels. Please contact WageWatch if you need assistance with interpretation the statistics reported, help building custom reports, or have a need for our wide range of consulting services. For more information on our services and surveys please call WageWatch at 888-330-9243 or contact us online .