The Good and Bad of People Analytics
Decisions tend to be bad, judges, for example, decide on a hungry stomach, and low sugar level, purchasing managers and regular managers think they know it better because mostly they decide with System 1, the ancient brain system, which is full of biases (Overview of biases on Visual Capitalist). It is quick, but not satisfying. Studies showed that data-driven models are much better in decision making [Ross J. (2019), Designed for Digital]. Here, people analytics jumps in. People analytics is designed to provide additional tangible information in order to make hypothesis-driven and data-supported decisions concerning employees, collaboration and communication within the company [Wikipedia, People Analytics].
As current examples show, algorithms are not free of biases either. Amazon’s recruitment tool, which was later abandoned, developed a bias against women. Dr Sandra Wachter said, “Garbage In Garbage Out” [Business Insider, Why it is totally unsurprising that Amazon’s recruitment AI was biased against women]. Algorithms learn from the past, and if the historical data reveals that successful candidates were mostly male, the algorithm cannot know it better and will develop a bias against women.
Another issue which people analytics tools have is the data they need. To learn and get deep insights into an employee, they need vast amounts of data from various sources, which is problematic because the employer or third party vendor will get insights in all conversations an employee has in all possible digital channels the company offers. Can the company guarantee that no data will be leaked and thus personal and company data get available on the internet?
As in the Business Daily Podcast “Being watched at work” noted, more problems arise if the employer implements such analytics, but not informing employees which data is collected for which purpose. Another one is the later misuse of data which was purposed for another reason. It was analysed how employees move in the buildings, to plan future offices. However, later on, they decided to analyse when employees arrive and leave the building to let employees go which are coming late and leaving early [BBC, Business Daily, 2020, Being watched at work].
Not all pitfalls are foreseeable, as the like button on Facebook. The inventor Justin Rosenstein never intended that teens and adults commit suicide when not getting enough likes [The Social Dilemma, 2020]. However, it is of high importance to act fast and fix the issue!
It is, therefore, of the utmost urgency that companies that want to use tools such as people analytics first consider whether they really need such tools. If the answer is yes, then companies must deal with ethical issues. Before a company starts collecting and evaluating data, it should be clear what should be achieved and why. After clarity about the purpose, enter into conversations with the staff, explaining what exactly will be done and for what purpose. Furthermore, the employees must always be able to view and download the data collected about themselves. Moreover, if harm is identified, fix it quickly. Harness the power of humans and technology together to truly operate as a social enterprise [Deloitte Global Human Capital Trends, 2020, P 108].
As Daniel Pink (2011) showed in his book Drive, when the issue of the money is gone, the three intrinsic elements of motivation are Purpose, Mastery and Autonomy. Douglas McGregor (1960) describes similar views with his theory y.
Use this knowledge to handle employees’ data responsibly and ask critically if there are not alternatives instead of monitoring every step of the employees.