Staff turnover: How to predict it with accuracy

Staff turnover: How to predict it with accuracy


Employee turnover is one of the most significant problems an organization can encounter throughout its lifecycle, as it is difficult to predict and often introduces noticeable voids in an organization’s skilled workforce. Service firms recognize that the timely delivery of their services can become compromised, overall firm productivity can decrease significantly and, consequently, customer loyalty can decline when employees leave unexpectedly. As a result, it is imperative that organizations formulate proper recruitment, acquisition and retention strategies and implement effective mechanisms to prevent and diminish employee turnover, while understanding its underlying, root causes. In this article we will explore the implications of employee turnover and how to predict it with accuracy.

 

 

Employee turnover has drawn management researchers’ and practitioners’ attention for decades because turnover cost affects an organization’s operational capabilities and budget. Employee turnover is both costly and disruptive to the organizational function (Kacmar et al., 2006; Mueller and Price, 1989; Staw, 1980); and both private firms and governments spend billions of dollars every year managing this issue (Leonard, 2001). Turnover costs involve recruiting, selecting, and training. According to the U.S. Department of Labour, turnover costs a company one-third of a new hire’s annual salary to replace an employee, which is about $500 to $1500 per person for the fast-food industry and $3000 to $5000 per person for the trucking industry.

 

Advertisment

 

Furthermore, turnover disrupts social and communication structures and causes productivity loss (Mobley, 1982). Turnover also demoralizes the remaining employees and leads to additional turnover (Staw, 1980). Sagie et al. (2002) found that a high-tech firm lost 2.8 million US dollars or 16.5% of before-tax annual income because of employee turnover.

 

 

These researchers also found that turnover reduced profits, increased the organization’s total risk, and triggered more turnover among the organization’s other employees. Therefore, understanding and predicting turnover at the firm and departmental levels is essential for reducing it, as well as for effectively planning, budgeting, and recruiting in the human resource field.

 

 

Previous studies indicating that turnover reduces organizations’ team or work performance and financial performance and significantly changes an organization’s direction when a top executive leaves. Glebbeek and Bax (2004) found that excessive employee turnover harms firm performance (profits). According to Hancock et al. (2013)’s study, turnover has a strong negative relationship with organizational performance in the manufacturing and transportation industries. According to Kacmar et al. (2006)’s study, employee turnover increases customer waiting time and reduces profits. Turnover also reduces restaurants’ profitability and customer satisfaction because of declining productivity.

 

 

Overall, the inability to predict employee turnover and to replace that individual reduces organizational performance and profits and disrupts the organizational structure.

 

 

The value of being able to predict turnover is directly related to the time it takes an organization to hire and on-board employees as compared to their needs. This time is not only dependent on the organization, but possibly the type of employees that are being hired. Governmental organizations that are the basis of this analysis require a lengthy period to hire and on-board employees, because of clearance and security issues. Predicting turnover early allows HR to proactively plan for the possible turnover and to be prepared to launch the search, shortening the time-to-hire and on-board employees.

 

 

Therefore, Predicting employee turnover helps reduce the hiring lead time, thus eliminating some turnover costs. The lead time of employee turnover and replacement involves five stages: providing a leaving notice, advertising the job opening, interviewing, doing a background check, and completing on-board training. An employee usually provides a leaving notice at least two weeks before the actual leaving date. Then the employee’s manager must prepare for advertisement and hiring committees, requiring a week to complete. The hiring committees need two to eight weeks, sometimes even longer, to release the hiring advertisement and select interview candidates. During the interview period, hiring committees need one to two weeks to make a final decision. One week to six months is needed to complete the finalist’s background security check. Finally, the employee’s on-board training takes at least one week.

 

 

 

 

Thus, an organization usually spends at least two months replacing a new employee. Prediction models or systems provide the predicted turnover number a year in advance. Based on a combination of this number, the organisational plan, and the budget, the HR department determines a final demand number. Ignoring employees’ notice periods, HR can either advertise job openings and start hiring immediately based on the final predicted employee demand number (thus reducing the training period) or wait until an employee provides the leaving notice. Therefore, predicting turnover at the firm and departmental levels reduces hiring lead time (Kacmar et al., 2006). The other benefit is to identify unusual patterns in turnover and possibly investigate root causes.

 

 

Statistical methods are used to develop employee-turnover prediction models. These methods are time series, survival analysis, logistic regression, and data mining. Time series methods capture employee turnover’s seasonal and cyclical patterns and predict an aggregated turnover number, in terms of headcount, by using a historical turnover number.

 

 

Survival analysis identifies significant internal and external turnover factors and builds a Cox PH model to predict turnover at the individual, departmental, and entity-wide levels. Logistic regression method determines whether significant factors identified in the literature are also significant in the organization studied. A set of decision rules is created based on employees’ tenure using a decision tree method.

 

 

For one to produce accurate prediction results, the sources of data should be well organised and be trusted sources. HR data is usually fragmented depending on the systems that the organisation uses. All data from different sources need to be consolidated into one dataset.

 

 

Once the data has been consolidated, the process of cleaning it will start. This is one of the most important steps if we want accurate results. If wrong information is fed to the models, they will return wrong prediction. After cleaning the data is modelled using different types of machine learning models. The best model will then be chosen for future use.

 

 

Implementing an employee-turnover prediction model is also a key part of workforce planning in a lean management system. The model is inserted into a software program using a user-friendly interface. The human resource department installs this program, imports the employee’s information into the program periodically, computes prediction information, and exports the results. Based on the results, HR either modifies the employee retention and promotion strategies for the targeted employees with high turnover probabilities to reduce the employee turnover rate or prepares to hire new employees to prevent reduced productivity.

 

 

Benjamin Sombi is a Data Scientist, Entrepreneur, & Business Analytics Manager at Industrial Psychology Consultants (Pvt) Ltd a management and human resources consulting firm.

 

 

Reference

  1. Kacmar, K. M., Andrews, M. C., Van Rooy, D. L., Steilberg, R. C., and Cerrone, S. (2006). Sure everyone can be replaced but at what cost? turnover as a predictor of unit-level performance. Academy of Management Journal, 49(1):133{144}.
  2. Mueller, C. W. and Price, J. L. (1989). Some consequences of turnover: A work unit analysis.
  3. Human Relations, 42(5):389{402}.
  4. Staw, B. M. (1980). The consequences of turnover. Journal of Occupational Behaviour, pages 253{273}.
  5. Leonard, B. (2001). Turnover at the top.
  6. Mobley, W. H. (1982). Some unanswered questions in turnover and withdrawal research.
  7. Academy of Management Review, 7(1):111{116}.
  8. Sagie, A., Birati, A., and Tziner, A. (2002). Assessing the costs of behavioral and psychological withdrawal: A new model and an empirical illustration. Applied Psychology, 51(1):67{89}.
  9. Glebbeek, A. C. and Bax, E. H. (2004). Is high employee turnover really harmful? An empirical test using company records. Academy of Management Journal, 47(2):277 {286}.

Benjamin Sombi
Consultant
This article was written by Benjamin a Consultant at Industrial Psychology Consultants (Pvt) Ltd

Related Articles





Notifications

Sign up now to get updated on latest posts and relevant career opportunities