Designing Controls for Network Organisations: A Value-Based Approach

ion of procedural details of controls. The second purpose of the value viewpoint for procedural controls goes back to the complexity of networks. The design of inter-organizational controls can be very complex, because, for instance, different perspectives of the multiple stakeholders are involved. A way to deal with this multi-perspective problem is to use a multiviewpoint approach and in this thesis we have focussed on two viewpoints: the economic value viewpoint and the process viewpoint. By introducing a value viewpoint, we provide a high-level model of controls, without focusing on procedural details. As can be concluded from the case studies, an economic value analysis of control guides the design of the underlying process or information system solution. The latter can be seen as a solution that ensures the occurrence of all value transfers of the corresponding value model. From Key research outputs 299 this point of view, the preliminary value modelling is important, because different economic value solutions require different operational models to implement it. For instance, in the music case study, a business model for paid radio would require a different information system than the business model for free radio considered in our case. The econtrol methodology allows us to focus on pure value-related aspects of control mechanisms, and provides additional functionality compared to the process perspective on controls. In the health care case, we found the economic value perspective important in the design of a new controls service, the Social Chart, which was the most important result for the health care domain experts of the FRUX project [Droes et al., 2005]. Modelling the Social Chart from the economic value perspective yields such issues as who will fund the Social Chart, what service does it provide and who will operate the Social Chart. A further evalue analysis enables us to answer these questions, which is not provided by procedural modelling. The analysis of procedural controls along with their value aspects is also important because the choices made in value models should not contradict the control principles. The issue of who will operate the Chart is important not only from the value perspective, but also from the control perspective. The operator of the Social Chart should be independent from the care providers or insurance companies, who would like to advertise the service. Similar issues were encountered in the Beer Export case. In the Beer Export case, where the parties who deliver the TREC service should be impartial and not not be prejudiced in favour of a reduction in excise payments. To summarize these arguments, we conclude that the combination of process and value perspectives provides a valuable tool to analyze both control and economic value issues related to the introduction of these controls. The analysis performed by econtrol is more powerful than that provided from a process perspective alone. 14.1.3 The e3control ontology To design opportunistic behaviour, we have suggested the econtrol ontology, which is based on the existing evalue ontology, introduced by [Gordijn and Akkermans, 2003]. The evalue ontology represents concepts to describe a network organization as a set of actors that exchange objects of economic value with each other, however it does not incorporate control-related aspects, such as opportunistic behaviour. Therefore, we have extended the evalue ontology to specifically include the concepts required to model the effects of opportunistic behaviour on value transfers. The concepts we have introduced in the econtrol ontology include, among other things, subideal value object and sub-ideal value transfer to show that the result of sub-ideal behaviour is an incorrect exchange of a value object or an exchange of an inappropriate value object. The transfers of sub-ideal value objects are marked in graphical models with dashed lines. In addition, we have introduced a liability token, a concept that allows us to identify the party who behaves sub-ideally. As was discovered in the case studies, all the control problems, which we have elicited from domain experts or from literature reviews, could be modelled with the concepts of the econtrol 300 Conclusions and future research ontology. In all cases, the economic motivation behind the control problems could be identified and modelled with the concepts of econtrol . However, econtrol concepts were not sufficient when we wished to design more processor information systems-related aspects of controls. For example, econtrol cannot be used to express the sequential order of activities. This means that when procedural or other non-value details become important, an analysis should use the appropriate tools. In our work, some procedural aspects of control are described in the control patterns using UML activity diagrams. 14.1.4 The e3control patterns In addition to the econtrol ontology, the designer also needs to have knowledge about possible control problems and mechanisms that are possible and how they should be implemented in a proper way. This knowledge was structured in econtrol patterns, which are defined as descriptions of generic and re-usable control mechanisms for a recurring control problem. In this thesis, we have described four procedural control patterns and two contractual control patterns. The procedural patterns include: Partner Screening, Proper Contracting, Execution Monitoring and Execution Confirmation. The Partner Screening and Execution Monitoring patterns describe monitoring-based mechanisms to combat a counter actor who executes his contracting obligations sub-ideally. The difference between the two patterns is that Partner Screening suggests an ex-ante monitoring control, and Execution Monitoring suggests an ex-post monitoring control. Further, the Proper Contracting pattern describes how the contracting process should be arranged to avoid misunderstandings about the contracting commitments the actors make with each other. Finally, the Execution Confirmation pattern suggests how the final process of a transaction should be arranged to avoid misunderstandings about the performed activities. The control mechanisms described in the procedural patterns affect only the process viewpoint, unless a control requires some activity to be delegated. Delegation creates a service, modelled at the value level as a value transaction. In such a case, the control changes the initial value model. Delegation is modelled by combining procedural patterns with special delegation patterns. Furthermore, there are two contractual patterns: Penalty and Incentive. Both patterns describe an actor’s choice of behaving ideally or sub-ideally. The Penalty pattern describes a mechanism of punishing sub-ideal behaviour. The Incentive pattern describes a mechanism of rewarding ideal behaviour. In these patterns the control mechanism affects both the value and the process viewpoint. The patterns have been demonstrated in detail in two case studies: the ROC case and the Beer Export case. In the ROC case we describe the Execution Monitoring, the Incentive and the Penalty patterns. The Beer Export case includes the Execution Monitoring and Partner Screening patterns. Other patterns, although not demonstrated, could also be found in the case studies. The patterns can also be observed in the other two case studies. For instance, in the music case, the rights societies monitor the number of listeners and tracks played by the radio station. This ex-post monitoring process can be described by the Execution Monitoring pattern. The witnessing of the number of listeners and tracks is done by the listeners, who report this information to the rights societies. The information is then sent by the listener to the rights societies in a seamless model so that the listener does not interact with the process. Reflection on experience with the econtrol approach 301 In the health care case, several patterns can be found as well. For example, the government agency CVZ monitors the performance of care providers to ensure that patients get good quality care. This corresponds to the controls described in the Execution Monitoring pattern. Another scenario is when a care provider may claim to have no commitment to provide care for a given budget, and hence refuses to continue to provide care. In accordance with the Proper Contracting pattern, the government only commits a specific budget, if a care provider promises to provide the agreed care for that budget. The patterns in the health care case are analyzed in more detail in [Kartseva et al., 2007b]. We have not invented the controls described in the patterns. We have structured existing knowledge of controls by employing the patterns approach. The structuring of the knowledge of patterns is a contribution in itself, since, in our view, the controls in the internal control theory are not structured in an efficient way. Typical descriptions of internal controls are very domain-specific. Firstly, separate risks and controls are described for every specific activity of each transaction cycle. Secondly, they assume the presence of some specific documents or parts of IS (e.g. picking tickets, packing slips, bar codes, and credit card processing systems). As a result, a control designer must be acquainted with a huge number of control descriptions. For example, in the controls described in [Romney and Steinbart, 2006], a control expert, who wants to design controls for the sales and procurement processes inside an organization should have thorough knowledge of about one hundred controls. In the patterns, we describe the controls at a higher level of abstraction, without a specific focus on a business process or an industry. This is possible, since we start by modelling the control problems with domain-independent value models and initially describe a control problem as a sub-ideal value transfer. Thus, the controls