What are differentials?
Differentials (or data / benchmark differentials) compare compensation between locations, for a selection of jobs, levels, and industries. They are presented as a percentage difference relative to an anchor (reference) location a user selects. For example the entire US can be compared to New York for all Engineering jobs.
How are differentials calculated?
The following methodology is applied:
- Apply job, level, industry, and location filters.
- Calculate the median salary (benchmark) for the anchor (reference) location, for each of the levels selected.
- Calculate the median salary (benchmark) for each selected location and for each of the levels selected.
- Divide each level + location benchmark by the equivalent benchmark for the anchor location, where there is data available for a level, for a location and an anchor.
- Take the average of all matching combinations.
Note - where a selected or anchor location does not have data for a selected level, this will be excluded from the calculation. If a user hovers over a location in the differentials chart, the tool will highlight and levels which have been excluded from the calculation.
Example tooltip showing levels excluded from the differential calculation in the Philippines.
Are differentials based on mean or median salaries?
Differentials are based on the median salary for matching incumbents in each selection.
Are differentials adjusted/normalized to reflect the Employee Count in each location?
Differentials are not adjusted for Employee Count in each location.
Differentials are normalized for level, where there is sufficient data for a level in a location.
While Differentials are simple to understand, they are sensitive to the distribution of data available at each location selection, including the spread of different jobs in different locations.
For a normalized differential, consider Smart Differentials. These provide a model estimated differential, which is normalized across selected jobs and levels.
What if there is no data or very limited coverage for a job in a location I am interested in?
If there is no data for a job in a location you have selected, then this job will not be included in the differential. The differential will be based on the remaining jobs which do have coverage in that location.
In these circumstances, consider Smart Differentials. These provide a model estimated differential, which reflects all jobs and locations even if there is no survey data available for a job in a location. This is helpful for hiring in areas like new or emerging markets where there is little existing data.
Are Data Dominance and Suppression applied to differentials?
Yes - Data Dominance and Suppression rules are applied to data differentials.