People Analytics: Hype, Fear and Real Opportunities

Data analyses are gaining importance in many areas of work. But what does it mean when such analyses increasingly include the data of an organisation’s employees? In this article, we discuss who may benefit from these people analytics, for whom they pose risks, and what organisations should consider when using such data. 

For many organisations, the analysis of various types of data is part of everyday business. Increasingly, organisations are also turning to the data of their own employees. Under the term people analytics, data about employees is collected, often combined with external data, and statistically analysed. As a result, HR decisions are augmented or certain HR activities automated. The public debates on people analytics are often highly polarised. 

Between hype and fear 

In 2021, a team from Bayerischer Rundfunk (Bavarian Broadcasting Service) tested a recruiting software that was supposed to generate a behaviour-based personality profile of applicants based on voice, language, gestures, and facial expressions. The result: The system evaluated applicants as less conscientious if they wore glasses, and more open if they wore a headscarf. Amazon’s attempt to develop a hiring algorithm that identifies ideal applicants based on profiles of successful hires from the past 10 years failed as well. It recommended almost exclusively hiring male applicants. Finally, a “productivity score” designed to capture how, when and for how long users in an organisation used Microsoft 365 services caused public criticism. The score was intended to offer administrators insights into the use of the internal IT infrastructure. Initially, however, the data was reported on an individual basis. Today, hardly any article on the risks of people analytics can do without one of these three scandalous examples. 

On the other hand, the vendors of people analytics software themselves shape the debates with promises such as “measure soft skills […] validly and objectively”, “[eliminate] unconscious biases with AI-supported video interviews”, or “[understand] employees of your company with ease”*. These promises raise high and barely attainable customer expectations and promote hype around people analytics, while scandals reported in the media trigger fears. Both these aspects shape our understanding and use of people analytics.

*These quotes are from people analytics vendor websites and are used here for illustration.

For whom does people analytics pose risks? 

The use of people analytics is associated with risks, which are evident in the examples presented above. Bayerischer Rundfunk’s research on AI-based personality analyses shows that complex analyses are often only explainable to a limited extent and can lead to self-fulfilling prophecies, for example, when recruiters learn before a job interview that an applicant has been classified as narcissistic by a software programme. The much-cited example of Amazon’s hiring algorithm clearly shows that algorithms are not neutral and the use of historical data can inhibit change. Finally, when huge amounts of personal data are collected and analysed, questions of data protection and employee privacy arise, as the example of Microsoft’s productivity score illustrates. People analytics can also pose a risk to management, for example through high costs associated with its implementation, while the financial benefits are difficult to assess. 

Who benefits from people analytics?

People analytics can be a profitable business involving many different actors inside and outside organisations. IBM, for example, claims to have saved more than 100 million US dollars within one year through the use of artificial intelligence (AI) in human resources. Several start-ups have also specialised in the business of people analytics and are benefiting from the hype: the AI-based talent management platform eightfold, for example, recently raised 220 million US dollars and is now valued at over 2 trillion US dollars. 

At the same time, the polarised debates between fear and hype often make it difficult to address the real potential of people analytics – for organisations and employees. Beyond distant, hypothetical scenarios of the future, people analytics can already support solving urgent HR management issues today, as our interviews with people analytics managers have shown. The shortage of skilled workers, for example, is a major challenge for many HR departments today. Here, analytics can support HR planning by combining internal data on skills, roles, and needs, with external market data to assess where future shortages may arise. Retention of existing employees also becomes relevant here: Automated, anonymous pulse surveys are used to evaluate job satisfaction in real-time or to determine which factors are particularly important to employees, such as work-from-home policies. In addition, many organisations now set their own diversity and equality goals. In this context, algorithms are used, for example, to analyse salary data to determine how large the gender pay gap is in such organisations. Overall, our interviews have shown that people analytics offers real opportunities to organisations, some of which differ greatly from the examples presented in the media. 

Needs orientation, participation, contextualisation and data protection as key foundations

Protecting employees is important to realise the potential of people analytics. When developing use cases, the real needs of employees and organisations should be put front and centre. Therefore, those responsible should not construct use cases solely according to feasibility and data basis. It is also important to involve employees and their representatives in the development process right from the beginning. Works councils are often involved too late, even though they know the workplace very well and can thus assess feasibility. When interpreting the analyses, users should always consider the context of the data, as people analytics will never fully capture the complexity of humans. At the same time, data protection must be at the centre of the development of use cases, for example, by already technically limiting data misuse or by only analysing data on a team or organisational level instead of on an individual level. 

Focus on real opportunities instead of hypothetical scenarios of the future

The use and analysis of employee data will continue to play an important role in the world of work. It can well support some activities in HR management. However, when it is used to evaluate or rank people it becomes problematic. When discussing people analytics, it is therefore important to look at the real potentials that the technology can already offer today instead of focusing on hypothetical scenarios of the future. 

This article is based on a presentation by the authors at re:publica 2023. 

This article includes findings from a project for which the authors received funding from the Hans Böckler Foundation.

Image credits: Philipp Schmitt / Better Images of AI / Data flock (digits) / CC-BY 4.0