Abstract
Despite widespread adoption of Human Capital Management (HCM) systems, organizations struggle to extract systematic workforce intelligence from their accumulated HR data. Using rich contextual information derived from four years of HCM data from a leading financial services firm and textual data in job postings, we systematically measure how well employees’ current skills match the requirements of new job openings. These skill-fit scores are validated using actual hiring decisions from 1,034 internal applications. The measure robustly predicts hiring success: applicants with one standard deviation higher skill fit are 2.25 times more likely to be selected when competing for identical positions. This demonstrates that job postings and HCM data contain rich, decision-relevant skill signals previously untapped by organizations. These validated skill-fit measures also reveal important workforce patterns – a) 14.8% of employees face limited skill-aligned opportunities, and b) 54.9% of unsuccessful job applicants are overlooking suitable alternatives. Our empirical approach provides a replicable framework for developing validated skill-fit measures from readily available HCM data without additional data collection or validation effort, creating a foundation for deeper workforce analytics.
Citation
George, Nikhil, Ramayya Krishnan, and Rahul Telang. 2025. “From Posting to Prediction: Building Validated Workforce Analytics.” Working Paper.
@techreport{george2025prediction,
author = {George, Nikhil and Krishnan, Ramayya and Telang, Rahul},
title = {From Posting to Prediction: Building Validated Workforce Analytics},
year = {2025},
month = {jun},
institution = {SSRN},
type = {Working Paper},
url = {https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4906323},
note = {Last revised: June 2025}
}