Home > Blogs > Top Concern For Talent Analytics
Several organizations believe that talent is better served ‘customized’ through human resource department. Each person can be served through a series of rules (defined by organizational policies) which are flexible to selected individuals.
During hiring, the candidate conversion rate of 1% is not atypical in certain sectors. The metric is so small and goal is so weak that the motivation to achieve it is diminished.
Therefore, talent analytics has to be implemented holistically. Selected implementation leads to half knowledge, which is worse than ignorance. The cycle should encapsulate use cases from recruitment, talent management, compensation and retention.
Recruitment Analytics should start with a talent gap analysis to validate if the talent is indeed required for the organization and whether this skill is required in that team. Hierarchical model of team governance in an organization is highly conducive for repetitive hiring of similar talents across several teams.
Candidate sourcing, screening and interviewing should use advanced natural language technologies for profile match, resume shortlist, keyword tag matches, and LinkedIn skills evaluation.
Actionable insights can be generated from employee feedback forms. Data can be scraped from emails, yammer and several other places to find clues of behavior inconsistent with company’s’ culture like bullying, harassment etc. Such issues are potentially very damaging to company’s image and can potentially lead to huge financial and legal cost.
Data can validate if any inequality persists after statistically accounting for skills, location, grade and other such parameters e.g. gender inequality.
Retention prediction of individual employees can be done by building predictive models that use data from employee, company, and external economic factors. Such models have to be back tested aggressively and regularized strictly to avoid over fitting.
One possible way to evaluate the model is to not provide intervention for some time after prediction, meanwhile evaluating the actual turnover against the predictions. The drawback is that while the model is validated but lots of model parameters dynamically change with time (like economic situation etc.) and the model may not remain stable.
Alternatively, model can be tested though AB testing. Turnover for two populations is compared and some get the intervention while others don’t. The relative change is measured and significance is established.
Holistic application of talent analytics can help organization source talent differently. It can be used as a tool to manage the talent effectively while making sure those activities inconstant with company’s vision is limited. Unconscious bias against gender equality in compensation can be managed effectively. Effective retention strategies can be built and implemented.
Manish Gupta heads sales for Banking and Data Science CEE for Manipal Global Education. An engineer by qualification, Manish has done his MBA from Symbiosis Centre for Management and HRD and is currently pursuing his second PG in Data Science from Manipal University.
Nishant Chandra, Ph.D.
Dr. Nishant Chandra is the center head of AIG Science India and data science leader. He develops natural language, text mining, and machine learning models for the insurance industry. Prior to AIG, Dr. Chandra has driven innovation in BFSI, e-commerce, R&D, and mobile telecom industries in USA and India. He was recently acknowledged as one of the top 10 data scientists in India.