The ever-increasing generation of data of all business aspects and departments has revolutionized business decision making. In the era of big data, real-time dashboards and data analytics, companies need to loosen their reliance on gut instinct and manager experience in favor of more evidence-based decision making (Weill & Woerner, 2015) as over the years is proven that decision making based on data can improve business performance substantially (Provost & Fawcett, 2013). To enable the full potential of data-drivenness; business elements, processes, and procedures need to be quantified. However, one essential element of the business is hard to quantify: human performance.
Until now, the changes that have been made due to the implementation of data-driven decision making have only improved human performance. It enabled them to focus on their core tasks and therefore to be more efficient and effective (Brynjolfsson & McAfee, 2017). However, if we keep increasing the importance of standardization this might extend to decreasing returns. Human resources and their performance are not rigid, but variable. Their output can vary due to qualitative, psychological, elements like motivation and support. These psychological elements are directly influenced by, but not limited to, their tasks and targets (Heath & Heath, 2010) resulting in two reasons why data-drivenness of human performance is hard. (1) Humans are social creatures and need emotion and stimulation in order to perform to their best capacity (Heath & Heath, 2010). Standardizing, quantifying and data-driven their tasks and targets can, therefore, decrease their performance. (2) The psychological elements that impact their performance can also vary due to other external factors inside and outside of the workplace. It is, therefore, nearly impossible to quantify or standardize human performance accurately.
This will set a limit to in which extent data-drivenness can be implemented, as human performance will always be an essential element of business performance. Even though due to recent developments in Artificial Intelligence, machines can outperform humans in certain activities, they will only complement and not replace human activities (Brynjolfsson & McAfee, 2017). Where do you draw the line in data-drivenness to prevent that the decrease of human performance will overshadow the surplus of evidence-based decision making?
Bibliography
Brynjolfsson, E., & McAfee, A. (2017, July). The business of artificial intelligence. Harvard Business Review.
Heath, C., & Heath, D. (2010). Switch: How to change things when change is hard.
Provost, F., & Fawcett, T. (2013). Data science and its relationship to big data and data-driven decision making. Big Data, 1(1), 51-59.
Weill, P., & Woerner, S. L. (2015). Thriving in an Increasingly Digital Ecosystem. MIT Sloan Management Review, 56(4), 27-34.
Hi Jeannot,
You point out that there is an alarming increase in the use of data to quantify human performance. What I would argue is that the situation is not as bad as it seems. This is for two reasons:
(1) Data driven decision making regarding human performance does not necessarily mean that the person under evaluation needs to know the metrics that measure the performance. If the metric remains in the background and is used as a support instead of the lead-driver in performance evaluation, is the situation still that alarming?
(2) It’s not in the best interest of managers to act upon data that does not increase performance. If such a measure is put into use and it appears to work counterproductive, it will not last long until the metric used will be abandoned.
I’m curious about your view on these two points.
Hi Lars,
Thank you for your comment.
(1) I agree, communication (what to communicate and what not) will be fundamentally important to make a distinction in the quantification of human performance and the influence on their performance. However, as argued in the comment it will be hard to set up such accurate metrics to measure human performance due to its inconsistencies.
(2) Yes, that’s true up until a certain point. If for instance, human performance is the bottleneck in an operation, the company will have to find a way to improve it. The decision will have to be made if the task is still relevant to be done by humans, or if it can be oursourced to machines.
Thank you for bringing up a very interesting topic. Accidentally, i used to have a similar question while working at the previous company. Especially in a multinational company where there are numerous layers to manage, leaders always require quantifiable KPIs to evaluate overall an employee’s performance. It’s extremely difficult to get a reasonable quantifiable target due to the job nature (FMCG data analysis per client in my case). Therefore, dissatisfaction and inequality emerges, discouraging many employees to fully dedicate for the works. Although there are many performance metrics currently in use (e.g., Subjective appraisal by manager periodically, product defects, number of errors, net promoter score, 360-degree feedback, etc.), i think how reasonable it is to employees and how well-developed the supportive system is are the main factors to consider. However, advanced technologies make it easier to provide correct data, thus support employers to make the right metrics for evaluating human performances. Immersive technologies AR/VR can be effectively used for enhance training and assessment platform to extract intensive data and unprecedented insights for improvements.
There are many research on how to use and analyze the big data effectively in evaluating human performances. I’m currently interested in a Consortium on Data-Driven: Human Performance and Personalized Interventions (HAPPI) from the University of Southern California. The research involves computer scientists, engineers, and biologists to build the area of Human Performance Analytics and Informatics, reaching the goals of “efficiently collect and analyze large-scale multi-modal sensor data across a multitude of users in highly noisy and dynamic environments” and “how to exploit such data to advance human performance through personalized interventions”. Additionally, the applications of People Analytics are increasingly implemented in HRM or HR analytics to enhance recruitment, learning & development, and of course performance management. Conclusively, despite the variable nature of human performance, data can be extracted and effectively used in various ways to gain insightful analysis results, resulting in a more reasonable and precise management tools boosting employees’ satisfaction, enthusiasm, and dedication. So, the first mission is probably to fully understand your employees and HR patterns.
https://www.exoinsights.com/post/2018-the-year-human-performance-met-data-analytics
https://www.analyticsinhr.com/blog/what-is-hr-analytics/#fourth
https://dornsife.usc.edu/labs/consortium-for-human-performance-information
Laker, P. A. v. d., 2018. Data-Driven Human Resource Management, Netherlands: Ridderprint BV, Ridderkerk.
Interesting insights An, thank you for the contribution!
I think that an underlying idea of this question is the ability of computers to completely take over human activities. However, machines will never be able to fully replace humans as you mention in your post, especially because of the psychological elements. Therefore, machines won’t be threat to the existence of human activities. Neither will they cause standardization of people’s work as this is impossible given the vast majority of complex, human-specific tasks they have to do. Rather, machines will continue supporting people in their work and give employees more time to focus on the non-standardized tasks. Therefore, it will become more important for companies to separate the standardized taks from the human-specific tasks in order to keep data suitable to support employees in their daily work.