Sign up for our newsletter and get the latest big data news and analysis.

Artificial Intelligence & Compensation Benchmarking

Currently, compensation benchmarking compares employees across and within firms on the basis of titles. In many cases, titles mean different things between companies or even within a company. Therefore, solely title-based compensation benchmarking can be very imprecise. An entire compensation consulting industry has arisen to help companies calibrate and benchmark their employees’ compensation through a manual process beginning with title-level survey data and resulting in a consultant report after considerable manual adjustments.

This somewhat imprecise, expensive and manual process can be made highly precise, inexpensive and automated through artificial intelligence. The process begins with gathering employee-level responsibility data and then running machine learning models over this highly granular data set to provide accurate comparisons customized to each employee. This sophisticated software can digest and classify large amounts of data are able to look beyond an employee’s department and title to identify the most similar employees.

Title-based compensation assessment is thought to compare the most similar employees. This may have been true when industrial jobs represented the vast majority of the workforce and a large percentage of each employer’s staff performed few identical tasks.

In services jobs (which of course now dominate the US economy), employees with the same title might have different responsibilities to match employer needs. Consider hypothetical servers working at tip-free casual dining restaurants Alpha and Bet’s. Although servers at both Alpha and Bet’s have the same title and the common responsibility of taking food orders from guests, servers at Bet’s are also expected to run food from the kitchen to tables and serve cocktails in the bar area.

A solely titles-based benchmarking of these employees would not assess the servers’ significant day-to-day differences. Bet’s might experience higher employee turnover because servers can go to Alpha and make the same salary while doing less work. A responsibilities-based benchmarking system would reveal that servers at Bet’s have more demanding responsibilities. Bet’s could solve their turnover problem by increasing compensation for their servers or delegating these responsibilities to other employees.

This highly simplified example illustrates how titles only tell part of the story. Gathering employee responsibilities and identifying overlap is important to properly benchmark employees. But at large employers and employers with highly skilled employees, this manual task becomes a time-consuming and error-prone exercise. Machine learning applications solve these issues of manpower by efficiently and accurately organizing responsibilities and identifying the most similar employees.

As artificial intelligence applications become more sophisticated and are fed with more data of different types, these systems can reveal the valid and most important bases for comparing employees. Returning to our Alpha and Bet’s hypothetical, assume Bet’s employees are paid less. An artificial intelligence system could, for instance, reveal that servers at Alpha are better screened in interviews and forget orders less frequently.

At the end of the day, artificial intelligence is a human augmentation tool. Artificial intelligence applications give humans wide berth to draw intricate conclusions from massive sets of data. In the context of compensation, humans can decide for themselves what variables are most important and what HR decisions should be made in the context of the business’s goals.

About the Author

Adam Zoia is the CEO of CompIQ, founder of Glocap, and co-founder of Stella.ai and has been in the compensation and recruiting business for 20+ years. Adam has a B.S. in Economics from Wharton, a B.A. in History from UPenn, an Honours B.A. in Politics, Philosophy & Economics from Oxford University, and a JD from Harvard Law School.

 

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: