Finally, the HR world has discovered how to utilize Big Data. HR managers agree that “data are the oil of 21st century”. There are big expectations placed on these bits and bytes that sleep softly in numerous databases. Could it be that the expectations are too high?

What is Big Data?

As the term implies, Big Data refers to a collection of data sets that are too large and complex to analyze with standard statistical applications and data base systems. Unfortunately, there is no BMI to tell data when they are getting big or when they become obese. Suppliers of standard applications have already started to enhance their products with new features so they are capable of handling larger data sets. This just makes the definition more vague. It has been reliably estimated that the amount of data doubles every second year. HR departments contribute to this since IT–systems have replaced paper personnel files with a wide range of tools from application tracking systems to computer-generated letters of resignation.

What is the potential benefit of Big Data?

Talent analytics, talent management, feedback, e-learning and many other systems’ databases accumulate enormous sets of usable data, offering great potential to data driven HR processes. In many organizations, relevant data about the company’s workforce is stored over the entire HR life cycle. Data analyses can be used for a variety of tasks, such as determining the efficiency of a recruiting system. Nevertheless, data collection is a long way from producing beneficial information:

  1. Defining the purpose: Particular information must be identified (filtered down): the purpose, relevant and useable data, KPIs (not everything you can measure is automatically a KPI), and the potential independent variables.
  2. Integration: Data from several sources have to be integrated before the analyses can begin. Issues here are aggregation, matching, and data mining. This makes a tight collaboration between the HR and IT department necessary.
  3. Analysis: For an adequate data analysis two things are necessary: 1) technical (statistical) know-how on adequate handling of big data sets and 2) expertise in organizational research with a strong HR background.
  4. Interpretation and Reporting: The final goal of all visualizations and interpretations is to recommend what to do in practice. Decision makers need a reporting system that is understandable and not more complicated than necessary. Otherwise, the analysis is a pretty headstone in the data graveyard.
Be aware: bigger, is not necessarily better, nor smarter

Even if the brave new data world offers a lot of potential; data are not a substitute for using our brains. Furthermore, the “end of theory” suggested by some big data enthusiast will have to wait. Relationships in the HR field and in organizational research are often quite complex (mediation- and moderations effects rather than linear relationships). Nevertheless, many ad-hoc analyses reveal nonsense or trivial relationships and therefore result in wrong conclusions. HR KPIs like job performance or turnover rate are driven by multiple factors (for people with deeper interest read the Journal of Applied Psychology).

Organizations and their contextual factors vary. Even within an organization factors such as personal characteristics (demographic as well as psychological ones), leadership- and organizational climate vary between departments. One of the main weaknesses of the big data approach is that it ignores micro organizational processes. Other analyses fail because of a simple lack of basic statistical knowledge: probabilistic relationships are used for deterministic predictions, cause-effect relationships are derived out of correlations, third variables are included and predictions made on past data (regression) might be useless when bordering conditions are changing. It is just a question of time, before the next statistical super brain will get to the conclusion that the hiring of employees is one of the main risk factors for fluctuation (a mean comparison revealed that rejected candidates showed a significant lower turnover rate, 0.0%, to be exact). Another emerging topic due to big data in HR is data security and privacy concern. Data related to employees are legally more protected that other ones.

Further reading:

Big Data, Trying to Build Better Workers (New York Times, 2013)