Organizations are complex entities that consist of units and people that work
together in order to satisfy the needs of customers. Many organizations are inclined
towards vertical-thinking i.e. placing the focus on functional and hierarchical
structures. However, as organizations today require flexibility and ability to address
emerging business challenges, they often shift towards horizontal-thinking through
better understanding of their business processes
The number of organizations adopting BPM is quickly increasing. By this means,
so is the number of business process models as result of BPM initiatives. Within a
single organization this number often ranges from hundreds to even thousands of
business process models. In order to handle such large amount of business process
models, organizations structure them in terms of a process architecture. A process
architecture defines how the set of business process models of one organization can
be systematically organized
The purpose of a process map is to show a visual and holistic view of all business
processes of one organization. A process map provides an overview of how an
organization operates as a whole without necessarily going into the process details
Despite their importance, the design of process maps is still more art than science,
essentially because there is no standardized modeling language available for process
map design. Whereas managing process model collections has recently been a focal
point of research
This doctoral thesis focuses on the development of a language for designing process maps. In particular, the thesis provides the following contributions.
Reference BPM Framework: First, we present a holistic reference BPM
framework. The framework is a consolidation of procedural frameworks introduced
by prominent BPM researchers
Method for assessing cognitive effectiveness of process maps used in practice:
We found that it is important for models to be designed appropriately, because the
appeal of a model has an effect on the user using the model
Quality requirements for modeling languages and models: Modeling languages are typically used by both individuals and organizations, all having diverse backgrounds and stemming from various industries. Therefore, during the development of modeling languages this diversity needs to be taken into consideration, such that languages would serve relevant users in the most effective and efficient ways. One way of ensuring this is to develop the language for designing process maps according to well-established quality requirements for modeling languages. Thus, a third contribution of this thesis is the extensive list of quality requirements for modeling languages and models we have derived by conducting a systematic literature review on quality of modeling languages and models. Quality requirements are used as things according to which other things are developed, judged or measured. We found 79 relevant papers from which we retrieved 319 quality requirements. We use a subset of these quality requirements as basis for developing the language for designing process maps.
A language for designing process maps: Fourth, we define the intra-language parts (abstract syntax, semantics, concrete syntax) of the language for process maps. We follow an explorative method; hence we rely on empirical data for the language development. We primarily relied on empirical data in order for the language to reflect actual usage in practice. The empirical data we used for the language development was a collection of process maps and a qualitative in-depth study with one of our industry partners. We used the process map collection to define the process map meta-model which includes all concepts and concept relations organizations include in their process maps. The qualitative in-depth study enabled us to ensure the process map meta-model is saturated. Accordingly, we reuse symbols in our language which have already been used in practice as part of process maps. We follow this approach in order to ensure the language will consist of elements already familiar to organizations. During the language development, the quality requirements we collected by means of the systematic literature review are also followed. As a result, we ensure the language for designing process maps is also grounded in literature.
Method for checking language suitability for a particular purpose: BPMN has
become the de-facto standard for business process modeling. However, despite its
wide user acceptance, even organizations that use BPMN for modeling their business
processes do not use this language for designing their process maps. Thus, our fifth
contribution is a method for checking language suitability for a particular purpose.
We use this method to assess the suitability of BPMN for designing process maps,
since the purpose of a process map is different than the one of a business process
model. The method includes using the four semantic mapping relationships of
equivalence, intersection, subsumption and disjointness
Experimental evaluation of the language for designing process maps: Last, we evaluate the developed language for designing process maps by means of an experiment. In order to be able to evaluate the language for process maps, we need to assess the effectiveness of each concept the language offers. To test the effectiveness and efficiency of a modeling language, a counterpart is required against which we compare the language. However, since, to the best of our knowledge, there exists no other dedicated modeling language for designing process maps, we cannot assess process maps created by two different modeling languages. We have already evidenced that a process map designed using BPMN would be a subject of potential misinterpretation. Therefore, we decided to use textual representation as a benchmark against which we compare the visual language for designing process maps. We consider text as the best alternative, because it will induce the least bias from the participants. Also, using text will also help us represent the correct semantics of all process map concepts, which, for instance, is not the case with BPMN. Hence, we assess the effectiveness and efficiency of process maps designed using elements from our language against process maps that have not been designed using our language, instead the concepts have been textually represented. We found that, a process map designed using the elements of the newly developed language leads to users performing better in terms of both accuracy and time, than a process map not using the full range of the elements from the newly developed language for designing process maps.
The research we have conducted for this thesis and the consequent results have strong implications for BPM research and for practice. In terms of practice, we showed that up until now process map design has been more an issue of a craftsmanship rather than science. Thus, the language we have defined will support practitioners when designing their process maps. It will release them from the burden of, first relying on their own creativity when designing process maps, and second choosing among the numerous elements existing process modeling languages offer in order to be able to capture their requirements. Most importantly, having a language which offers appropriate elements for all concepts shown in the process map would assist practitioners in ensuring they depict a correct overview of their company’s operations, accordingly decreasing threats of potential misinterpretation.
Furthermore, a process map could be used as a tool to abstract from the details of many process models to a single model which captures the essence of their performance. Thus, research about the design of process maps complements prior research on managing process model collections. However, compared to the existing literature which has a more technical focus, process maps could be used to understand process model collections from a more strategic perspective. In this way, managing process model collections could take the turn of adopting a top-down approach, rather than the bottom-up one followed until now. Beyond the language for designing process maps we present in this thesis, the approach we follow to develop the language is one that, to the best of our knowledge, has not been applied before when undertaking this task. Therefore, this approach could be followed by all who wish to develop a modeling language which is grounded in both practice and literature.
The research concerned with process maps presented in this thesis is only a
starting point from where a stream of potential research arises. First, the language we
propose does not yet include the full range of visual variables introduced by
Moreover, the relations between the processes shown on the process map and the
process details stored in the lower levels of the corresponding process architecture
have not been fully identified. We are aware of the typical hierarchical decomposition
of processes