Follow Datanami:
October 5, 2016

ADP Leverages Huge Database to Gauge Worker Attrition

A database of more than 30 million employee records is the basis of a new cloud-based analytical tool designed to lower worker turnover by identifying in advance those most likely to quit.

Human capital management specialist ADP said its “Turnover Probability” tool running on its data cloud uses a predictive model to help employers identify top performers most likely to give notice. The tool also is touted as helping organizations identify likely “hotspots of attrition” while mitigating the risks of losing workers in key positions, departments and locations.

Citing tight a labor market that ironically includes the very data scientists who would analyze employee data on the ADP platform, the company said its turnover tool could help develop retention policies in departments with rising turnover rates. “Employers need to know where the highest levels of turnover occur within their organization, how their likelihood of turnover compares to other companies in their industry and what factors might be causing employees to leave,” Marc Rind, ADP’s chief data scientist and vice president of product development, noted in a statement.

ADP said its turnover tool forecasts the likelihood of losing employees within job types, locations and teams as well as focusing on individual workers. The tool analyzes factors such as compensation, job descriptions, organizational structure and “dynamics” as well as employee demographics.

It also allows users to compare their turnover risk against industry benchmarks based on ADP’s database of more than 30 million employee records. In one example, human resources personnel at a hypothetical company with about 2,300 employees were benchmarked with a 14.3 percent turnover probability within their industry sector while the company’s actual turnover rate was more than 1 percent lower than the industry benchmark.

“You need industry benchmarks to compare yourself to peer organizations when assessing the relevant factors,” Rind explained. “We were able to go back in time and score tens of thousands of employees in our unique and expansive dataset and give them a low, medium or high turnover probability. We compared that risk assessment with whether they actually left their jobs one year later and found that our model was highly accurate in identifying who voluntarily left the organization.”

ADP’s massive employee database stems from the fact that it issues pay checks for one in every six U.S. workers. The turnover probability tool leverages that dataset by with a predictive analytics capability running on the company’s DataCloud, which was rolled out in 2015.

So-called workforce analytics tools have mostly focused on the recruitment and hiring process. Lately, new analytics tools have emerged to gauge job performance along with worker retention frameworks like ADP’s. Another emerging approach dubbed “organizational analytics” looks for ways to boost worker productivity by, for example, gauging how much time each day workers focus on “core” tasks.

Recent items:

Microsoft Delves Deeper Into Worker Analytics

Can Workforce Analytics Help Us Find Work-Life Balance?

Datanami