During the development of POIL, we conducted both empirical and non-empirical re- search activities in the research phases as introduced in Section 1.4; i.e., (1) problem analysis, (2) requirements analysis, (3) solution design, and (4) solution validation.
Design Science Research Framework. We mainly used the research principles of de- sign science as introduced in Section 1.4, since the goal of this thesis is to provide a solution (i.e., the POIL framework) for a real-world problem (i.e., the gap between process information and business processes) by creating innovative artifacts. We used several methods such as case studies, experiments, prototypes, and use cases in order to evaluate the POIL framework. We provided contributions in the area of the design artifact. In addition, we applied rigorous techniques such as graph theory or linear al- gebra. Moreover, we studied prototypes that instantiate posed or newly learned design
prescriptions from our research. We provided comprehensive information to technical as well as managerial audience. Finally, we introduced technical implementations but also risks and benefits when applying the POIL framework.
Research Phase Empirical Activity Non-Empirical Activity
Problem Analysis Requirements Analysis Solution Design Solution Validation Online Survey Exploratory Case Study 3 Performance Tests Survey
Step 2 Step 3 Step 4
Step 1 Exploratory Case Study 1 Exploratory Case Study 2 Literature Survey Controlled Experiment Research Phases of this Thesis Research Activities of this Thesis
Design Science Research Framework
Research Framework of this Thesis
Research Framework
Figure 11.2: Research activities along research phases.
Step 1: Problem Analysis. During the problem analysis, the gap between process information and business processes was investigated. In practice, process information is not only stored in large, distributed and heterogeneous data sources, but also managed separately from business processes. Hence, in practice, process information and business processes are only linked manually, statically and partially, e.g., in enterprise portals connecting specific process information with business processes and associated process tasks. To cope with this problem, the POIL framework automatically aligns process information with business processes and their tasks in a context-aware manner. In this case, the problem analysis was supported through the following research activities:
• Exploratory Case Study 1 (cf. Section 3.3): We conducted an exploratory case study in the automotive domain. We analyzed business processes such as the review of product requirements or the identification of system specifications. We performed eight interviews and received other questionnaires with additional data. In particular, this case study allowed us to identify problems related to the gap that exists between process information and business processes.
• Exploratory Case Study 2 (cf. Section 3.4): Moreover, we conducted an exploratory case study in the clinical domain. Again, we analyzed business pro- cesses such as the admission of patients to a surgical clinic. We performed eight interviews and received questionnaires with further data. The clinical case study allowed us to compare results with the automotive case study and to gain further insights helping us to generalize our problem investigation.
Step 2: Requirements Analysis. During this step, requirements enabling POIL were elicited. The requirements reflect wishes and needs of process participants such as knowl- edge workers and decision makers. They further concern technical issues enabling the delivery of relevant process information to process participants. In this case, the require- ments analysis was supported through the following research activities:
• Exploratory Case Studies 1 and 2 (cf. Sections 3.3 and 3.4): From the aforementioned case studies, we also identified requirements. These are mainly driven from a practical perspective since the POIL requirements were derived from the interviews and filled-out questionnaires.
• Online Survey (cf. Section 3.5): We conducted an online survey with 219 employees from more than 100 enterprises. It allowed us to increase the validity of our case study results.
• Literature Survey (cf. Section 3.6): We conducted a comprehensive literature survey. Its goal was to approve the requirements identified during the empirical studies. Additionally, the literature survey allowed us to derive further POIL requirements.
Step 3: Solution Design. During the solution design, we created the POIL framework to meet the identified requirements. POIL aligns process information with business processes, both at the process schema and instance level [53]. In turn, this enables a process-oriented, context-aware delivery of process information to process participants. The main idea of POIL is to split up business processes into their constituent process elements and to integrate the latter with comprehensive process information [7]. As mentioned in Chapter 4, enabling the POIL framework requires four architectural layers: data layer, semantic layer, context layer, and application layer (cf. Figure 11.3).
Application Layer
Context Layer
Semantic Layer
Data Layer
Handling of Process Information and Business Processes
Set of Data Sources
Handling of Context Information Delivery of Process Information
Figure 11.3: POIL architecture levels.
Step 4: Solution Validation. During this step, the POIL framework was thoroughly validated. Unlike the other research activities, the solution validation investigated the concepts of the POIL framework. The solution validation analyzed whether the POIL framework works as desired and fulfills the expectations and requirements. For this purpose, prototypes, statistical methods, and test were used. The goal of the solution validation was to demonstrate the applicability and feasibility of the POIL framework. Moreover, lessons learned from the development and application of the prototypes were taken into account as well. In this case, the solution validation was supported through the following research activities:
• Survey (cf. Section 9.2): We conducted a survey in the automotive domain to prove that the SIN LP and SIN RP algorithm (cf. Sections 5.5.1 and 5.5.2) ac- tually support process participants when performing knowledge-intensive business processes. We applied the algorithms to a real-world use case from the automotive domain and then compared their outcome with the results of a survey among expe- rienced automotive engineers. Overall, 20 experts from an automotive OEM par- ticipated. Particularly, results have shown that the algorithms can indeed replace the costly and time-intensive human determination of relevant process information. • Exploratory Case Study 3 (cf. Section 9.3): We conducted a case study in the automotive domain. Unlike the first two case studies, this case study proved the validation of the POIL framework. More specifically, we proved that the SIN maintenance algorithms presented in Section 5.4.3 are able to maintain SINs. Parts of this empirical validation were interviews with employees regarding the usefulness of SIN maintenance. Overall, eleven employees from an automotive manufacturer were interviewed. Results confirm that the algorithms can indeed replace the costly and time-intensive human maintenance of SINs (cf. Section 9.3.3).
• Controlled Experiment (cf. Section 9.4): We conducted a controlled single factor experiment in the agricultural domain. Its main goal was to investigate the benefits of a process information portal implementing POIL compared to a con- ventional enterprise portal implementing hard-wired IL. Overall, twelve employees participated. They mainly stemmed from the sales department, but also from the department responsible for project management. During the development of the process information portal, we gained a lot of know-how and experience impor- tant for the design of POIL. Specifically, we implemented the architectural layers of the POIL framework (except for the context layer) in a process information portal called iProcess (cf. Section 9.4.1). We then compared the experimental re- sults of POIL with the ones of a hard-wired IL solution, which has been used by the involved enterprise (cf. Section 9.4.2). The results show that the POIL frame- work can indeed close the gap between process information and business processes. The framework can decrease the time needed for handling and managing process information during business process execution (cf. Section 9.4.3).
• Performance Tests (cf. Section 10.2): We conducted 22 performance tests to determine potential weaknesses of the POIL framework as well as options for its improvement. In particular, we investigated the core component of POIL; i.e., the SIN. We measured the time needed for performing the SIN creation phases as introduced in Section 5.3.2.