How to predict performance management systems success? Utilization of the critical check points

Purpose The prediction of the success of Performance Management Systems (PMS) is difficult and may lead to contradictory results because there are many potential factors locating at different hierarchical levels of an organization and usually only reflecting a part of success. Therefore, in this study we make use of the PMS chain theory developed by Kadak and Laitinen (2016). This theory is based on fifteen check points (CP) of compatible key factors (KF) which form a logical and comprehensive chain for PMS success. The purpose of this research paper is firstly to compare the relationship between PMS success and CP variables in original (2015) and newer (2022) samples (RQ1). Secondly, the purpose is to assess different statistical methods to predict the success of PMSs using information from these CPs (RQ2). Design/methodology/approach This approach is based on the chain theory of PMS success introduced by Kadak and Laitinen (2016). For the empirical research, data have been collected from 73 Estonian companies that use PMS. This survey data contains information on 15 CPs and assessments of the success of PMS. The research compares five different statistical methods (sum of variables, linear regression, logistic regression, ridge regression, and a numerical Solver solution) to predict the success of PMS with the help of CP variables information. The success of PMS is in this study measured by the self-assessed impact on organizational performance reported by the management. Findings The results of the study show that in the sample all 15 CP variables are highly correlated both with PMS success and with each other. The main contingencies are similar in the older and newer sample. This supports the chain theory and also RQ1 on the similarity of the two samples. If we evaluate the performance of statistical methods in explaining PMS success with the help of ROC curve, the best results were produced by the logistic regression analysis and the numerical Solver optimization method (RQ2). In addition, we found three CP variables that produce incremental information about PMS and affect therefore on the PMS success more than other CPs. Practical implications The research produces models that can be practically used to design and to predict the success of PMS. The simplest model is to calculate the sum of the CP variables, while the most effective model is based on the logit of these variables or the weighted sum of the three most important CP variables solved by a numerical optimization method. In addition of predicting, the results can be used also in constructing an efficient predictor of PMS success of companies in time. Originality/value The study adds to the literature a detailed and comprehensive, and multi-faceted, methodology for assessing and predicting the success of PMS. Used methodology is more detailed than the general consensus in literature (about which elements critically form a PMS) and shows measurable PMS components under each PMS element.