Developing the KPI Tree Model with Prediction Procedures

Abstract

The development of controlling over the past few years through Big Data, artificial intelligence, the development of various mathematical statistical methodologies, and digitalization has created models for management that have allowed for more efficient decision-making and planning. Various prediction and causal controlling models greatly increased the information content needed to make decisions. Through innovation processes, a number of new operational measurement opportunities and data acquisition sources have emerged. However, these data alone carry very little information content. On the other hand, in a structured system or model, this data can become a set of information that can even analyze the causal relationships of a whole organization's operating model and support various management functions from design to control.
The KPI tree is a controlling model used and developed in most cases by multinational organizations operating in the industry. The KPI tree organizes KPIs, formulated in different ways, into groups of objectives and correlations with a logical structure built on one another. KPIs measured in different ways are determined by both the literature and corporate practice. The data generated by Big Data and Industry 4.0, on the other hand, allows for the creation of new KPIs and the collection of existing KPIs up to the minute. These new KPIs provide an opportunity to both compare daily plan- fact and objective based daily analysis based on it, as well as to measure various lean management and other business management methods.
By utilizing and implementing data from the KPI tree model, various prediction models can effectively predict future performance and deviation from the target. In addition, various mathematical statistical methods and advanced Big Data parsing algorithms are able to predict operational anomalies that can be extracted from KPI tree data.