Today, predictive analysis is no longer considered a technology of the future. Many sectors are already using it to optimise management processes, manage risks and fraud, innovate products and services, and for marketing purposes. What about HR? More and more companies are adopting an approach based on data analysis and human interactions. Studies also show that HR departments have a positive attitude towards HR analytics.
What is predictive analysis?
How useful is this for HR departments?
Predictive analyses can be used to identify both threats and opportunities, and to respond appropriately: identify employees likely to resign, assess medium- and long-term skills shortages, determine the drivers of employee engagement, improve training processes, etc.
For example, predictive recruitment is now within reach. With the methods available today, it is possible to accurately identify the interindividual characteristics that largely explain performance in a given organisation, role or profession. How? By analysing the links between performance data and skills-related data for employees working in the roles and organisation in question. It is advisable to use as many relevant indicators as possible to characterise both performance (objective criteria, comparative performance, subjective criteria) and skills (personality, motivations, beliefs, abilities).
These contextualised methods can be used to map the unique characteristics of internal workforces. If used properly, they produce highly accurate and reliable results. “Ready-to-use”, automated solutions are also available, which provide predictive skills models based on input criteria (job, sector, role, etc.) and do not require contextual analyses. Note, however, that when using solutions such as these, the quality and quantity of the data must be ascertained!
“Algorithms are powerless without contextual analysis, perspective and human intelligence”
Challenges and best practices
The challenges are less about science than about the ability of human beings to unify and interpret
Predictive analysis can only be carried out on the basis of structured information such as that found in a database. Automated HR processes – such as payroll, talent, absenteeism and performance management – combined with survey, assessment and training results, etc. are now producing a large volume of mostly scattered data. However, data analysis is effective only if systems are interconnected and generate consistent data sets!
Best practice: It is important to unify data to produce a consistent set. Unfortunately, machines do not do this kind of preparatory work, which is essential to the widespread use of predictive
The challenge is to analyse all this information and use it to produce reliable models. Besides statistical skills, HR analytics requires an understanding of organisational and individual operating models, and an awareness of legislation on the use of personal data. Today, however, it is still a relatively new or experimental area.
Best practice: It is possible to develop these new skills within HR departments. Another solution is to use psychometric test publishers, which have thousands of assessment data and strong data analysis expertise.
Besides upstream data processing (aggregation of HR systems) and analysis skills (which are readily found), one other core competence is essential to avoid post-analysis failure: an ability to understand and share data.
Data is never really intelligent in itself; all the more so because
Finally, and this is probably the most important aspect of this approach, the data must tell a story; a meaningful and shared story, for a genuine purpose.
Therefore, a seamless man-machine alliance is required to ensure that HR analytics is more effective than intuition.
Subscribe to our Newsletter!
Enjoyed This Article? You will also like…
The world is changing. So, too, are the skills we expect to find in a manager. An update on the changes in the pipeline with Alexandra Didry, R&D Manager at PerformanSe... I myself find the various studies published by the World Economic Forum in 2015, 2016...