Data Analytics: Using HR Analytics to Support Managerial Decisions
HR analytics is receiving increasingly greater formal attention because of its direct link to organizational effectiveness and profitability. HR analytics facilitates workforce optimization, derives metrics for workforce performance improvement, enables a more strategic contribution from HR and have greater impact on organizational decisions.
What is HR analytics? It is the systematic identification and quantification of the people drivers of business outcomes. The global HR analytics market size was valued at $2.25 billion in 2019 and is expected to grow at a compound annual growth rate of 14.2% from 2020 to 2027.
As a case in point, a major manufacturer posed four questions related to the potential value in HR analytics. Each question and the derived insights are provided below.
Question 1: “How can we improve identification of employees who have a high risk of leaving?”
Through analysis of existing employee performance and incentive data, the following insights were derived:
- Employees who had been transferred or promoted were more likely to leave – especially those with lower merit payouts.
- The window of employment with the highest probability of departure was around three to six years – this was the timeframe that the employer needed to invest the most time and energy in retention.
- Where there was a mismatch between experience, tenure, and job bands (position and personal), the employee had a high likelihood of voluntary attrition. This was true regardless whether the position job band was lower or higher than would be expected given tenure and experience.
- Employees who left voluntarily were consistently performing higher with regular merit increases or were performing lower with regular merit decreases than employees who did not leave.
- Discrepancies with incentives were always associated with attrition (e.g., higher merit scores but lower actual vs. target payouts).
Question 2: “Does the company do well retaining high-performers?”
While generally the answer was found to be “Yes“, HR analytics generated unexpected findings.
- While all the high performers, across all regions in this company had a lower voluntary attrition rate than average and low performers, the top performers in the European region had higher voluntary attrition rates than did the average performers. This led to different incentive policies in that region.
- The profiles of high performers who leave voluntarily tend to be younger, have higher actual vs. target pay, higher merit scores, have two to three years of tenure, and were promoted or transferred recently.
Question 3: “Does the previous year’s merit and payout drive higher performance for the following year?”
In general, high merit was found to be positively related to high performance, but not for actual vs. target payouts. But this finding varied by region. For example, in Asia, increased merit and actual vs. target payout contributed to higher performance, however in South America, increased merit and actual vs. target payouts lead to decreased performance in the following year. Again, this finding informed incentive policy by region.
Question 4: “How do we analyze engagement survey data, including both structured (i.e. categorical responses) and unstructured data (i.e. free form text)?”
This question of integration of traditional structured survey data with unstructured text data is increasingly driving conversations in HR Analytics – until the last several years, text-based data was almost impossible to analyze. With advancements in techniques like natural language processing, the ability to extract meaning and sentiment from large text fields and mine the results for insights is commonplace.
With this specific manufacturer, the structured employee survey data was primarily numerical scores of each survey question on a scale from 1-5. The unstructured data was collected from the employees’ free form responses to two survey questions “What things excite you about working in the company?” and “What things don’t excite you about working in the company?”
For the structured data, the survey questions were grouped into five categories: growth, positive leadership, autonomy, competence, and pride. The structured data was then linked to the survey results related to departure type (e.g., voluntary or involuntary) to determine relationship with termination. Results revealed that higher engagement scores in growth, positive leadership, autonomy, were associated with higher voluntary attrition while higher competence and pride scores were associated with lower voluntary attrition. These results were then combined with an entity sentiment analysis on the free form text.
Entity sentiment analysis combines both entity analysis and sentiment analysis and attempts to determine the sentiment (positive or negative) expressed about entities within the text. For example, if the employee’s answer was “I don’t like the HR system”, the algorithm would identify the entity at issue as “HR system” and give “HR system” a negative sentiment score. Using this integrated approach, results demonstrated that employees had the most positive attitudes (sentiment) towards development, growth and opportunity, environment and atmosphere, teamwork, and group relationships. Employees had the most negative attitudes (sentiment) towards leadership, manager, training, development and opportunity, work and life balance.
Answering these questions enabled this manufacturer to understand its employee more clearly, predict the employees’ turnover more accurately, make managerial decisions more effectively, and reduce the operational cost more significantly.
For more details, please check our paper at https://dl.acm.org/doi/pdf/10.1145/3374135.3385281
 Grand View Research. Global HR analytics market SIZE: Industry REPORT, 2020-2027, from https://www.grandviewresearch.com/industry-analysis/hr-analytics-market