Predictive Analytics in Human Resources

sherrie-suski-analytics

Unemployment rates across the nation are at 3.9%, the lowest we have seen since 1969. Unemployment rates in the college degreed 25+ age group are even lower hovering around 2.1%.  In this kind of an economic environment, it becomes critically important that we can retain our good employees and attract the kind of talent that can ensure we stay ahead of the competition.  The question is HOW??

Enter the world of Predictive Analytics. It is, at its core, a technology that learns from existing data and uses this to forecast individual behavior. This means that predictions are very specific.  Instead of predicting turnover as an aggregate number for the end of the year, using PA, we can stat to predict which employees have a greater chance of turning over. Predictive analytics involves using a set of various statistical (data mining) techniques used to predict uncertain outcomes.

People analytics today brings together HR and business data from different parts of the business and is now addressing a wide range of challenges: analyzing flight risk, selecting high-performing job applicants, identifying characteristics of high-performing sales and service teams, predicting compliance risks, analyzing engagement and culture, and identifying high-value career paths and leadership candidates.

To put theory into practice, we first need to gather clean data.   This may prove more difficult than you originally thought but stick with it.  You need to identify as many variables as possible. Think of basic pieces of information like homes address, to calculate driving time and distance to work, gender and pay grade and comparatio.  Think also about collecting more obscure data like number of vacation hours or sick time used, personality or behavioral assessments and whether or not they elect health insurance through your company.  You are looking for any data where you would find a stronger correlation between that variable, or set of variables, and employees who terminate versus employees whom you retain. Once you have designed the algorithm, it will be time to put it to the test.  Give yourself at least 6 months to collect data and test the validity of the algorithm.

Although 79% of organizations consider people analytics to be an important trend, according to Deloitte’s Global Human Capital Trends (2016) report, only 8 % of the organizations had this capability in 2015.  It is one of the best tools we have for bringing HR out of the emotional, gut feel realm and into the world of  data based decision making, yielding quantifiable and sustainable results.