ThinkWhy applies AI to better determine compensation

Workforce analytics company ThinkWhy announced today it is adding a feature that leverages the AI capabilities of its software-as-a-service (SaaS) application platform to analyze compensation trends across more than 20,000 classes of jobs.

That capability will enable human resources (HR) teams to better refine an offer for each job title based on variable rates of compensations derived from the latest salaries being offered in a specific geographic region, said ThinkWhy CTO David Kramer.

The compensation feature is based on a proprietary machine learning algorithm that synthesizes research in every metropolitan area in the U.S. by tracking 18 trillion data points collected from more than 8.6 million paystubs, Kramer added. The platform then verifies that data and surfaces salary recommendations and talent supply forecasts by region.

ThinkWhy aggregates compensation data, including bonuses and incentives, for each job title and regularly updates it. That eliminates the need for employers to pull together multiple sources or crowd-source data gleaned from other HR professionals, noted Kramer.

Making salaries make sense

Collecting that data across multiple geographies is becoming more pressing because in the wake of the COVID-19 pandemic, many organizations have become more inclined to hire the best talent wherever they can find it. However, it doesn’t always necessarily follow that they want to pay someone residing in Alabama the same level of compensation they would someone residing in San Francisco or New York. Another critical factor when an organization is evaluating a potential applicant is the level of education each job applicant has attained.

The challenge, of course, is as the overall unemployment rate remains low, companies often need to balance their priorities against market realities. “You can’t hire what isn’t there,” Kramer said.

There’s also a natural tendency to miscalculate salary for any given role when they are based on human intuition. Offer too much and total costs rise. Arguably worse, offer too little and an employee it took six months to train may jump ship for a competitor.

Finally, a data-driven approach can reduce bias in salary decisions, protecting companies from being either being sued or fined by a regulatory agency once it becomes apparent salary decisions were not made in a consistent fashion across genders and races. Organizations often need to conclusively prove that it was an increase in the overall supply of job candidates that resulted in a measurable decrease in salaries, not a new hire’s race, creed, color, gender, or sexual orientation.

Finding the optimum compensation

It’s not clear to what degree an AI approach to hiring might result in better overall talent management within an organization. However, a data-driven approach should result in more decisions being based on science rather than the “gut feeling” that may have as much to do with indigestion as it does with insight. Of course, no AI platform can guarantee any job applicant will ultimately succeed in their chosen field. Nevertheless, the turnover rate within organizations might be reduced thanks in part to data science. It’s not uncommon for HR directors to find themselves trying to find candidates to fill the same positions over again year after year simply because the salary being offered is too low.

Compensation can be one of the thorniest issues any organization has to navigate. Not every manager or business owner is willing to put their faith in AI just yet to help resolve those issues. However, as more data becomes accessible, AI platforms will tend to get closer to the mark more often than their human counterparts.