Conceptual Modeling for Indicator Selection

Abstract

Indicator-based management enables decision makers to make decisions based on quantitative measures. This approach has been successfully applied in multiple domains beyond traditional business-related ones, including Education, Healthcare, and Smart Cities, among others. Yet, it remains a difficult and errorprone task to find suitable Key Performance Indicators (KPIs) that are aligned with business goals. Indeed, there is a general lack of adequate conceptualizations and formal models of indicators, that captures the subtle yet important differences between performance and result indicators. Moreover, there is a lack of approaches interleaving business modeling techniques with data analysis in an iterative process. In order to tackle these deficiencies, we propose a methodology for eliciting, selecting and assessing explicitly KPIs and Key Result Indicators (KRIs). Our methodology is comprised of (i) a novel modeling language that exploits the essential elements of indicators, covering KPIs, KRIs and measures, ii) a data mining-based analysis technique for providing domain experts with data-driven information about the elements in their model and their relationships, thereby enabling them to validate the KPIs selected, and iii) an iterative process that guides the discovery and definition of indicators. Finally, we apply our approach to a water management case study to show its benefits.