Analytics-Led Talent Acquisition for Improving Efficiency and Effectiveness
Analytics-Led Talent Acquisition for Improving Efficiency and Effectiveness
复制标题
以分析为主导的人才招聘，提高效率和效益
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Mahek Shah
中科院分区：
文献类型：
--
作者：
Girish Keshav Palshikar;Rajiv Srivastava;Sachin Sharad Pawar;Swapnil Hingmire;Ankit Jain;Saheb Chourasia;Mahek Shah
大型IT组织每年通过多个招聘渠道雇佣数万名员工，以实现他们的增长和人才补充。假设对每名员工至少审查和评估10个潜在的配置文件，人才获取(TA)人员最终处理50万个具有多种技术和领域技能的应聘者配置文件。由于业务的规模和紧迫的时间表，由于对候选人简介的误解或技术评估不足，可能会出现次优的人才选拔。这种因人工、有偏见和主观评价而实施的招聘流程可能会导致工作和组织匹配度较低，从而导致人才素质较差。随着越来越多地采用数据和文本挖掘技术，招聘过程也被重新设想为有效和高效。主要信息来源，即候选人简介、职务描述(JDS)和助教流程任务结果，都在eHRM系统中捕获。作者提出了一套为提高招聘过程的效率和效力而构建的关键功能组件。通过在一家大型跨国IT公司进行的多个真实案例研究，这些组件的有效性得到了验证。本文阐述的一些重要组件包括简历信息提取工具、工作匹配引擎、技能相似度计算方法，以及用于验证和完成JD for Quality职位规格说明的JD完成模块。使用简历和JD以及求职搜索引擎的文本提取模块的大数据集进行的测试显示出高性能。
Large IT organizations every year hire tens of thousands of employees through multiple sourcing channels for their growth and talent replenishment. Assuming that for each hire at least ten potential profiles are scrutinized and evaluated, the Talent Acquisition (TA) personnel ends up processing half a million-candidate profiles having multiple technical and domain skills. The scale and tight timelines of operations lead to possibility of suboptimal talent selection due to misinterpretation or inadequate technical evaluation of candidate profiles. Such recruitment process implementation due to manual, biased, and subjective evaluation may result in a lower job and organizational fit leading to poor talent quality. With the increased adoption of data and text mining technologies, the recruitment processes are also being reimagined to be effective and efficient. The major information sources, viz., candidate profiles, the Job Descriptions (JDs), and TA process task outcomes, are captured in the eHRM systems. The authors present a set of critical functional components built for improving efficiency and effectiveness in recruitment process. Through multiple real-life case studies conducted in a large multinational IT company, these components have been verified for effectiveness. Some of the important components elaborated in this paper are a resume information extraction tool, a job matching engine, a method for skill similarity computation, and a JD completion module for verifying and completing a JD for quality job specification. The tests performed using large datasets of the text extraction modules for resume and JD as well as job search engine show high performance.