Designing a performance eva‎luation model based on balanced scorecard an‎d machine learning - a case study

12/1/2025 12:00:00 AM

Performance eva‎luation of organizations, particularly in dynamic an‎d data-driven environments, is a critical component of strategic decision-making. Models that can analytically an‎d accurately monitor organizational performance based on data play a key role in enhancing productivity an‎d achieving long-term objectives. This study was conducted with the aim of designing a data-driven model grounded in the Balanced Scorecard (BSC) framework an‎d machine learning techniques. A case study was carried out in an online cosmetics an‎d personal care store. Key performance indicators (KPIs) were identified in collaboration with the strategy team across the four perspectives of the BSC model, an‎d data analysis was performed according to the stan‎dard CRISP-DM methodology. To design the initial conceptual model, correlation analysis an‎d expert opinions were utilized. Subsequently, using the K-Means algorithm, three performance clusters were identified, an‎d causal relationships among the indicators were examined separately within each cluster. The results indicated that only in the second cluster did the regression model demonstrate an adequate statistical fit; therefore, the meaningful paths of this cluster formed the basis of the final conceptual model. Based on the significant relationships in the second cluster, two improvement scenarios were developed for the organization. In the first scenario, the focus was placed on enhancing human capital skills. The analysis showed that a 5% increase in the quality of training programs could, through improvements in support processes an‎d customer complaint management, lead to an approximate 2.5% growth in sales. In the second scenario, a 5% increase in employee satisfaction with benefits was found to reduce operational errors, increases 1.5% by improving overall organizational performance. This research demonstrated that integrating managerial frameworks with data-driven analytics can lead to the design of precise, practical, an‎d effective models for performance eva‎luation an‎d strategic planning.

كارت امتيازي متوازن , ارزيابي عملكرد , يادگيري ماشين , رگرسيون , چارچوب CRISP-DM , تحليل همبستگي , خوشه بندي K-Means

Balanced Scorecard , Performance eva‎luation , Machine Learning , Regression , CRISP-DM Framework , Correlation Analysis , K-Means Clustering

Amir Abyari

Mohammad Mahdavi Mazdeh