Service Blueprinting using Process Mining Techniques

Abstract

Service blueprinting is one of the visualization techniques to comprehend and analyze a service. It provides a holistic view of a service environment by visualizing an entire service process and components. However, it is difficult to draw a service blueprint because of service intangibility. Kotler, P., Armstrong, G., Franke, G., and Bunn, M. D. (1990) said that “Service intangibility means that services cannot be seen, tasted, felt, heard or smelled”. That is why it depends on the experience of people who create a service blueprint. Thus, the generated service blueprints might not be the same because people could have different experiences and thought each other. That is, the inaccurate service blueprint might be drawn and it might lead to inefficient service improvement. This thesis proposes a data-driven service blueprinting method using process mining techniques such as activity clustering, process discovery, performance analysis, and so on. The proposed method consists of five steps. First, the similarity scores between activities are calculated in the process and resource perspectives. After that, highly correlated activities are clustered based on the scores. The reason for the clustering is that service blueprints describe service processes at a high level. Secondly, the extracted clusters are adjusted considering time distance because Günther, C. W., & van der Aalst, W. M. P. (2006) said that activities performed together in a short period of time can be considered highly correlated in terms of proximity. Afterward, several fail points such as duplicated works, bottlenecks, and waiting points are extracted by performance analysis. Lastly, a process is extracted by inductive mining from the event log. And then, the extracted process model would be simplified by merging activities which exist in the clusters. In this step, the relations between activities are considered. With the extracted fail points and the simplified process model, the more objective and accurate service blueprint can be created. In this thesis, the prototype system is implemented to evaluate the proposed method. Two case studies were conducted for the evaluation. In the first case study, it shows that a high-level process suitable for service blueprints can be derived by using the method. In addition, it is possible to set parameter values from users, so that the users can perform the analysis from a desired point of view and create a service blueprint. Also, in the second case study, we can see that the method can be applied to a complex process and produce meaningful results. Previously, there was a limit to generate a service blueprint for a complex process. However, the components of service blueprints such as fail points, a high-level process can be derived quickly and accurately if we use the proposed method. These evaluation results show that the method can create service blueprints by extracting more accurate fail points, process models and the additional information such as the improvement direction of the service, low-level activities which make up high-level activities.