Measuring employer attractiveness in diverse talent markets

Lun Li, Theodoros Lappas, Rong Liu

Research output: Contribution to journalArticlepeer-review

Abstract

Previous work has measured employer attractiveness via proxies, such as the number of new hires or employee sentiment mined from surveys and reviews. The first drawback of such metrics is that they treat all talent as one homogeneous population with uniform recruitability patterns. Second, they ignore that a firm's attractiveness is heavily influenced by that of its competitors, as they all vie for the same talent. Third, they assume attractiveness remains constant across time. We address all three drawbacks via a methodology that can cluster job titles to identify prevalent talent types, as well as compute and monitor employer attractiveness for each type in a competitive market. Our clustering approach groups job titles based on their lexical similarity, their co-occurrence in professional resumes, and their overlap in terms of the skills that they require. Our work directly informs the decisions of employers trying to recruit in competitive talent markets, as well as of workers trying to identify the best destination for their career. We compare our approach to competitive baselines and demonstrate how it can improve the results of key applications.

Original languageEnglish
Article number114079
JournalDecision Support Systems
Volume177
DOIs
StatePublished - Feb 2024

Keywords

  • Employer attractiveness
  • Job title clustering
  • Workforce analysis

Fingerprint

Dive into the research topics of 'Measuring employer attractiveness in diverse talent markets'. Together they form a unique fingerprint.

Cite this