People Analytics at Work. Intensive Case Studies of Pioneering Companies in Germany and Switzerland

人员分析是依赖数据的更普遍的管理转变的一部分（Davenport 2006）。它专注于实时获取、分析和预测员工绩效和敬业度、工作和协作模式。基本假设是，人类在组织决策中的经验和直觉应该在很大程度上被数据所取代，从而形成一种“基于证据”的人力资源管理形式和数据驱动的文化（BI Trend Monitor 2020; Kremser/Brunauer 2019）。通过收集和连接各种各样的数据，人员分析旨在建立预测、评估和控制个人和组织行为的新方法（Loi 2020； Sousa等人2019；Manuti/de Palma 2018； Leicht-Deobald等人2019；Goodell King 2016; Marler/Boudreau 2017），从而在工作场所产生新的可见性和算法强加的等级制度。人力分析与未来工作的高度相关性与人力资源领域缺乏关于基于算法的决策系统的实践和影响的经验知识形成鲜明对比。人员分析的实现不能仅仅从技术角度来理解。恰恰相反：以“企业社交图谱”的形式（Höller/Wedde 2018），它将社交网络的社会既定逻辑转移到工作和组织领域（Staab/Geschke 2019）。它依赖于将工作流程和组织行为持续转化为数据，并以多种方式持续连接这些数据，从而为分析人员和工作提供新的角度（“数据化”）。这种基于算法的社会秩序逻辑（cf. Yeung 2017）如何在组织上实施，是该项目的一个主要焦点。为了分析新可见性的产生，它借鉴了量化社会学（Diaz-Bone/Didier 2016）和“估值研究”（Lamont 2012）等领域的见解。从这个角度来看，人员分析不仅仅是以客观的方式代表组织的现实。相反，他们的特点是他们的表演性和反应性（Espeland/Sauder 2007），以各种方式通知员工的社会行为，自我表现和身份。通过对德国和瑞士领先企业的案例对比研究，该项目分析了人员分析的实施动态。它调查了一种基于数据、计算和（至少部分）基于人工智能/机器学习的新控制制度实际上是如何谈判的。在案例研究层面，它侧重于主体化和算法控制的工作，从而揭示了人物分析的微观基础。从比较的角度分析了公司特有的数据文化和外部软件供应商的影响，以及国家立法对这些过程的影响。

People Analytics is part of a more general shift in management relying on data (Davenport 2006). It is focusing on the real-time acquisition, analysis, and prediction of employee performance and engagement, work and collaboration patterns. The fundamental assumption is that human experience and intuition in organizational decision-making should, to a substantial degree, be replaced by data leading to an ’evidence-based’ form of human resource management and a data-driven culture (BI Trend Monitor 2020; Kremser/Brunauer 2019). By collecting and connecting a great variety of data, People Analytics is designed to establish new ways to predict, evaluate, and control individual and organizational behavior (Loi 2020; Sousa et al. 2019; Manuti/de Palma 2018; Leicht-Deobald et al. 2019; Goodell King 2016; Marler/Boudreau 2017), leading to new visibilities and algorithmically imposed hierarchies at the workplace. The high relevance of People Analytics for the future of work stands in stark contrast to a lacking empirical knowledge about the practices and effects of algorithm-based decision-making systems in the area of Human Resources.The implementation of People Analytics cannot solely be understood from a technical perspective. Quite to the contrary: in the form of a “corporate social graph” (Höller/Wedde 2018), it transfers the socially established logics of social networking to the sphere of work and organization (Staab/Geschke 2019). It relies on the ongoing transformation of work processes and organizational behavior into data, and on continuously connecting this data in manifold ways allowing ever new angles for the analysis of people and work (‘datafication’). How such an algorithm-based logic of social ordering (cf. Yeung 2017) is organizationally implemented, is one main focus of the project. To analyze the production of new visibilities, it draws on insights from the fields of the sociology of quantification (Diaz-Bone/Didier 2016) and ‘valuation studies’ (Lamont 2012) alike. From this perspective, People Analytics are not simply representing organizational reality in an objective manner. Rather, they are characterized by their performativity and reactivity (Espeland/Sauder 2007) informing the employees’ social actions, self-presentations, and identities in various ways. Based on comparative case studies of pioneering companies in Germany and Switzerland, the project analyzes the implementation dynamics of People Analytics. It investigates how a new regime of control based on data, calculation, and – at least in part – on Artificial Intelligence/Machine Learning is actually being negotiated. On the case study level, it focuses on work which is both subjectified and algorithmically controlled, hereby revealing the micro-foundations of People Analytics. In a comparative perspective, it analyzes the influence of company-specific data cultures and external software suppliers as well as the impact of national legislation on these processes.