How HR can leverage people analytics in new world of work
As induced by the pandemic, organisations are going through rapid digital transformation, and the skill search has offset the demand-supply equilibrium, as more and more organisations move from legacy to cutting-edge tech platforms.
Companies across the globe are looking to curate a best-in-class talent retention strategy that builds a vibrant and innovative work culture to harness employee stickiness and retention. There is a huge impetus for organic talent development. Flexible working principles and practices, progressive wellness and benefits, new-age self-learning, and long-term incentives are on a rise.
Internal career mobility, both local as well as global, customised career and learning paths, employee engagement as well as enhanced communication find their place at the heart and centre of the overall talent retention strategy.
Retention efforts that are based on data-driven approaches have proven to have a higher success rate. Reactive data analysis focused on identifying and representing the number of people who are leaving is as important as understanding the reasons behind employee separation. A game-changer could be where data is leveraged to predict employees that are at the highest risk of attrition so that a targeted retention effort could be made to ring-fence key players in the organisation.
Data science based on predictive algorithms that explain the classification or clustering of employee profiles is one of the highly effective ways to identify at-risk employees. Especially since these are based on data attributes like age, experience, past ratings, position to market compensation, last promotion, location, number of leaves taken, supervisor, training investments, job family, skill, and career velocity among others.
Under normal or stable circumstances, employee attrition is predictable to a large extent, since it is dependent on certain predictable parameters like retirement age, low performers planned for the exit, planned automation and outsourcing to name a few. Employee attrition also depends on multiple unforeseeable variables such as external funding, company performance, organisational redesign, and market dynamics as it pertains to talent and company culture to name a few. It is therefore important to leverage both data intelligence and market insights to be able to combat attrition.
There are multiple options when it comes to choosing a machine learning algorithm or statistical design to curate an organisational fit model to predict employee attrition. For example, a “logistic model’ based on predicted ‘attrition risk’ parameters that produce employee scorecards. Whereas the “Classification model” calculates the risk of an employee leaving the organisation, and segregates employees into broad parameters like high-risk/ low-risk or more likely to quit/less likely to quit. ‘Non-linear regression’ model has a direct correlation to the ‘probability of attrition’ when the outcomes are bifurcated.
Likewise, the ‘decision trees’ model works on factors like information gain and variation reduction to estimate employee loss. However, when the model involves multiple parameters, the decision trees tend to become very large and complex. In these situations, the ‘‘Random Forest method comes handy as it combines multiple decision trees using several algorithms to classify and predict.
The seamless implementation depends on organisational context, the requirements laid down by the decision-makers, data availability, budget, and computational power of the team executing this.
The rich insights that are gathered in the process with time data-learns-from-data, can also help the organisation across multiple areas.
There are multiple ways in which HR can leverage people analytics insights to better understand and manage work. Data insights could be leveraged to ascertain the possible reasons behind attrition and can help in taking appropriate measures to prevent it. Data can also be used to curate a predictive attrition model that not only helps take preventive attrition measures but can also help in enhancing hiring decisions. Deriving logical trends in candidates’ past performance to predict future trends can further help curate a well-balanced onboarding as well as training.
Predictive attrition trends when coupled with the right compensation and market movement data can help to ring-fence key talent that is critical for success. A step further, classifying “Employee Persona” can help design policies and communication in a way that results in culture building and employee stickiness.