Luisa Naumann Messung und Überwachung der Qualität abrechnungsrelevanter Daten in der Fahrzeuglogistik

About this book

In the course of digitalization, companies are driving forward the automation of their business processes by using a wide variety of systems. High data quality must be ensured across all systems in order to guarantee a smooth process flow and achieve optimal business performance.

In order to identify and improve data-quality problems within the process chain in good time, it is essential to measure and monitor the current data-quality situation across all processes. In the field of quality management, there are both subjective and objective measurement methods for the quality of a data set, a data table or a database within a system. In multi-stage processes, however, several systems are used and data is transferred to system interfaces.

The area of supply chain management and logistics in particular is characterized by extensive system landscapes. The literature from this area is based on subjective assessments of data quality, which do not provide a suitable method for continuous and automated measurement of data quality.

The aim of this dissertation is to develop an automatable model for measuring and monitoring data quality within process chains in which multiple systems are used. For this purpose, a literature review was first conducted in relation to supply chain management and logistics in order to analyze the measurement methods of data quality in this area and to identify relevant data-quality dimensions that can be measured automatically.

As part of a case study, the billing-relevant data of the vehicle logistics process within Volkswagen Konzernlogistik GmbH & Co. OHG was examined. The billing-relevant data runs through the entire process chain and is suitable as a basis for developing a cross-process measurement method. As part of a survey, employees were asked to assess the data-quality situation on the basis of ten automatable data-quality dimensions and their satisfaction with this situation. Regression analyses helped to find out which data-quality dimensions have a significant influence on employee satisfaction. In the application case, these were consistency and accuracy.

Variance analyses were used to identify differences in opinion between the departments. For the measurement model, metrics were formulated for all ten data-quality dimensions at various levels of detail in order to summarize them together with the regression coefficients as one metric for the overall process. The applicability of the measurement model was verified with the help of a consistency measurement for the use case. A monitoring tool was visualized in the form of a possible dashboard. Furthermore, it was explained how improvement measures can be derived from the dashboard and how the most cost-efficient improvement measure can be identified using an algorithm.

This dissertation offers a cross-process measurement method for data quality that can be transferred to other processes and serves as a basis for improving data quality.