The introduction of people analytics in the HR sector offers enormous opportunities for companies. The process, also referred to as “Human Resources Analytics” or “Workforce Analytics,” describes a phenomenon in digitizing the working world that is becoming increasingly important in Germany as well. It concerns the use of special software to improve the quality of HR decisions through targeted data analysis and use. The data-supported decisions are intended to improve recruiting decisions that were previously made on the basis of “personal experience” or “gut feelings.” The process allows to avoid costly wrong decisions, save time, and even strengthen employee loyalty to the company through targeted measures. There are a number of legal hurdles to its deployment potential, however. After all, the method focuses on analyzing personal data from an employment context, so that in some instances it will be of highly sensitive origin. Although the use of people analytics in Germany is still in its infancy, it is not an entirely new phenomenon. Google, for example, already started using people analytics software to optimize internal staffing decisions in 2009.
How it works
The core element of people analytics software is the linking of corporate, HR, and other data from generally accessible sources, such as career networks. This data is filtered, linked, and statistically evaluated using complex algorithms. The software shows users correlations that can be used to make predictive forecasts for the future as part of “Predictive Analytics.” Still, the right conclusions must be drawn from the relationships shown, however. The use of appropriate software can thus complement decision-making processes, but cannot replace people. The limit, both legally and socio-psychologically, is always the “automated recruiting decision.”
Areas of application
The potential field of application of such software is extremely diverse. It may be used for HR decisions of various kinds. In the area of recruitment, decision-making can be influenced in the application process, for example by
- analyzing the chances of success of individual applicants in the specific position
- assessing an applicant’s general character suitability for a position on the basis of a behavioral prediction.
In monitoring, the evaluation of employee satisfaction, such as based on the analysis of email communication, also plays a key role. In addition, it is possible to determine which factors have a lasting influence on the well-being of the workforce. It is even possible to predict when and how likely it is that employees will quit. This may be a cost-saving instrument, especially for the resignation of highly qualified employees, if it can be counteracted in advance.Discrimination
Although individual software manufacturers advertise with slogans such as “business beyond bias,” there is a risk of “digital discrimination,” which in certain cases may also become relevant under the General Equal
On the one hand, algorithms may already be programmed in a discriminatory way if people include their evaluations and prejudices in programming. On the other hand, “accidental discrimination” based on an absurd correlation may also occur.
Depending on the method on the basis of which the data is collected, there may still be concrete disadvantages for people of certain ethnic origins or people with disabilities. The use of language analysis software, which creates personality profiles of applicants on the basis of automated telephone interviews serves as an illustrative example. It must be ensured that the software takes into account the linguistic peculiarities of non-native speakers and people with disabilities when analyzing emphasis, language rhythm, etc.Data protection
Especially with respect to the GDPR and the new German Federal Data Protection Act, there are legal problems associated with the use of people analytics software. Anonymizing data, which would exclude the application of the data protection provisions, is hardly an option in the employment relationship. After all, the point is to make recruitment decisions. First of all, therefore, a legal basis for data processing is required. Article 6(1)(f) GDPR and Section 26(1) Federal Data Protection Act new version can serve as such, according to which in particular data processing is permitted for purposes of the employment relationship. Comprehensive personality analyses may only be carried out, however, on the basis of consent if at all. Whether voluntary consent in the employment relationship is possible at all is, however, largely negated due to dependency. In addition, the prohibition of automated decisions as stipulated in Article 22(1) GDPR is of great relevance in the context of data protection law. When people analytics software is used, the final decision must still be taken by a human being. People analytics can therefore not reduce the decision-making process, only support it.
A particular risk is the possibility of violating the general right to privacy (Article 1(1), Article 2(1) German Constitution). To prevent such violations, the boundary to “personality screening” of employees or applicants by the software users must be maintained. Conclusion
The enormous potential of use of people analytics software can lead to significant progress in HR. The legal hurdles cannot be ignored, however. To comply with the current legal framework, it is necessary to regularly check the software used for discriminatory elements and the facts of justification in terms of data protection up to obtaining declarations of consent where necessary. The final HR decision should always be made by a human being. Thus, not only the prohibition of automated decisions under Article 22(1) GDPR is complied with, but also the socio-psychological component of the decision.