Predictive Analytics For Initial Placement

PGN HR hands image for predictive analytics

As a strategy for actually placing a job into a well established salary (pay) structure, one of the methods used by the experts at PGN HR is predictive analytics. More specifically, linear algebra, regression, and classification.

A very common issue that comes up in all compensation studies and job classification interventions is the question of initial placements. Questions surface usually in the middle of the project and right before the implementation stage where roadmaps for implementation and planned roll-out is about to be issued.

This happens because turnovers and natural attritions can never be stopped. Just because you are doing a compensation and classification study does not mean you can discount or ignore turnovers and attritions. Organizations will lose employees and hire replacements right in the middle of the project, and leaders will invariably ask "where should I place this new candidate? — I know that you are in the middle of updating the salary structure and in the middle of reclassifying the job!"

So, how do we deal with this very specific issue? Continuous Predictive Analytics!

Predictive Analytics for Compensation and Job Classification

PGN HR uses linear algebra, geometrically progressive matrix and the following statistical tests: T-tests, Chi-square tests, Kruskal -Wallis H-Test, Kendall's Tau, Spearman's Rho and multi-variate regression for analytics. We use the results of these tests and analyses to recommend changes to policies and practices affecting initial placement in the structure, which in turn affect internal equity.

  • Define and establish parameters based on compensation strategy for initial hire
  • Establish employee cohorts (categorical) and numerical variables
  • Using the geometric salary matrix, we map the “applicants’ data” degree and Years of Experience onto the matrix, and obtained the corresponding salaries
  • We perform a T-test. Both two-tailed and one-tailed, with unpaired data sets to show the validity of the process
  • We then perform a multiple regression analysis and will use thee following equation:
    • IP = ß1 + ß2 * Degree + ß3 * Outside Experience (Rounded)
    • IP being Initial Placement
Predictive Analytics for Promotion and Lateral Progression

Similar to our initial placement process, PGN HR will use the results of statistical tests to recommend changes to policies and practices affecting initial placement in the structure, which also impacts internal equity. Our procedure for predicting promotion and lateral progression include:

  • Utilization of Years in Position measure instead of Years in Service
  • Map the newly promoted incumbents’ salaries to the salary structure using a geometric matrix.
  • Conduct internal equity analysis update

To learn more about our method and approach, please reach out to one of our experts.

- Elena Mason, CCP, SPRH, Prudential Global Network

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: