Modeling and Validation of Key Performance Indicators for Business: An Entropy-Based Approach

Abstract

A company management system based on key performance indicators (KPIs) allows optimizing the company’s ecosystem by helping managers effectively manage available resources in key functions based on information about the company’s economic and operational activities.

The article focuses on developing KPIs in a company and understanding the necessary criteria to be taken into account to improve the efficiency of the new system implementation at the micro and macro levels. This will help not only to achieve short-term goals but also in the strategic aspect will allow us to remain at the appropriate level of competitiveness, adapt to changes, and develop despite external challenges.

The article actualizes that the effectiveness of KPIs directly depends on how well the indicators are structured, and how correctly they are interpreted and updated. In this study, we propose utilizing information theory tools to develop complexity-informed decision-making systems through precursor indicators. Specifically, classical information, approximation, fuzzy, permutation, and distribution entropies was examined. To test the effectiveness of these indicators, stock returns of the four most and least developed and globalized companies were analyzed. The findings reveal that the determinism or chaotic behavior of these companies is variable, as indicated by entropy metrics calculated using a sliding window algorithm. The performance of companies fluctuates with market conditions, showing increased efficiency and demand, reflected by rising trends and minimal entropy during certain periods, and maximum entropy during periods of disturbance. Empirical results demonstrate that approximation-based entropies are particularly resilient across varying market conditions and effectively distinguish between developed and chaotic companies.