تعیین توالی ماژول پشتیبانی از تصمیم در فرایند پیاده سازی ERP ــ استفاده از روش ANP

The paper addresses the alignment between business processes and information technology in enterprise resource planning (ERP) implementation. More specifically, we concentrate on one of the key decisions at the tactical alignment level: the decision on the implementation sequence of the ERP modules. Since the module sequencing problem involves a myriad of organizational and technical issues, connected to each other in networked manner, the analytic network process (ANP) methodology is applied. As a result of the study, we present first a general level conceptual framework to sequence ERP module implementations and expand the model to a more detailed level in a case study. The priorities for the implementation sequence of the ERP modules are determined in the case study.

Contemporary organizations operate in a global environment that is characterized by constant change and different cultural settings. To avoid the fragmentation of the operational data in different business units, companies implement enterprise wide information systems, such as ERP (enterprise resource planning) systems that can increase control, improve coordination and communication, and create the picture about the corporate functions on the aggregate level. Typically, ERP-systems support financials, human resources, operations and logistics, and sales and marketing functions (Davenport, 1998). The objective of ERP systems is to conduct the business processes more efficiently and effectively, in an integrated manner. They are a way to control functions of the organization and to make all the units perform in a more uniform way. Dramatic operational improvements are possible, through integration and redesigning of processes. It is assumed that the standard software package more or less fits all organizations, and all units inside one organization, which creates risks for ERP investments. In theory, business processes are modified to fit the systems, since customizing the system is considered too expensive and too risky. But in practice both business processes and ERP systems are suspects to be changed during the implementation process. The technical nature of ERP software is already quite well described and known (see, for example, Klaus et al., 2000), but because of the ERP implementation related context-sensitive social diversity is huge, the perceived results and outcomes of ERP implementation vary a lot (Ross and Vitale, 2002 and Scott and Vessey, 2000). There are a lot of problems reported in the implementation of these systems, and even the objectives of the system investments may evolve during the process (Glover et al., 1999, Nandhakumar et al., 2003 and Themistocleous et al., 2001). Technology and the organization are in a continuous interaction with each other during technology adoption and use, according to Orlikowski (1992). The alignment between business and IT is a dynamic process with bi-directional interaction. The decisions affecting the alignment of the business and IT are made on strategic, tactical and operational levels (Hendrickx, 2002). The strategic decisions concerning ERP projects include, for example, which modules are implemented and how much business process re-engineering is conducted (Parr and Shanks, 2000 and Mabert et al., 2003). The decisions on issues, such as, using internal or external resources for implementation and in which order the modules are implemented can be considered tactical in nature. Finally, the operational decisions are those concrete decisions affecting daily workflows. The implementation of ERP software packages necessitates disruptive organizational change (Soh et al., 2000) and that's why there are considerable risks involved. When the problems surface they require countermeasures that require resources and time, such as providing more personnel training when the ERP implementation introduces significant changes in workflows. The risks and misfits between the organization and technology should be identified early for planning the change management issues (Soh et al., 2000 and Soh et al., 2003). The scope of decisions that a company can make becomes smaller when the implementation process approaches actual operational implementation. The implementation can be conducted either as “big bang” or as phased implementation by site or by module (Mabert et al., 2000). This paper studies the alignment of business and IT in a case where it was decided to phase the implementation by the ERP modules. A survey investigating companies in the United States reported that about 17% of organizations in their sample phased the ERP implementation by modules (Mabert et al., 2000), indicating that phasing the implementation by modules is rather common. Thus, the module sequencing decision can be considered as a relevant research problem. The aim of this paper is to provide the analytical means to analyze the investment parameters, including the risks, before the actual implementation starts, and based on the analysis to decide on the implementation sequence of the ERP modules. The implementation of some modules may need to be postponed, because the identified risks require time consuming countermeasures, whereas other modules may need to be implemented as soon as possible, because, for example, the legacy system is not working properly and has to be replaced. For some monolith and rigid ERP packages this module sequencing decision may be driven by technical imperatives, but more flexible technologies and service-oriented architectures will call for social negotiation to decide optimal sequence for available ERP modules and services. Since the decision problem of the module sequence involves a myriad of organizational and technical issues, which are interconnected in networked manner, we propose the analytic network process (ANP) method (Saaty, 2001 and Saaty and Özdemir, 2005) to be applied. The ANP methodology includes defining the decision-making criteria and their interrelationships as well as the decision alternatives. ANP methodology supports complex, networked decision-making with various intangible criteria. It improves the visibility of the decision-making process and generates the priorities between the decision alternatives. ANP has been applied to a variety of decision problems, including, for example, allocating proper service concepts to the different IT market segments (Partovi, 2001), investment evaluation (Kengpol and O’Brien, 2000), evaluating componentized Enterprise Information Technologies (Sarkis and Sundarraj, 2003), ERP systems evaluation (Shyur, 2003), and R&D project selection (Meade and Presley, 2002). The application of the ANP methodology requires, first, developing the general level conceptual framework and, second, developing the framework at the detailed level for applying it in an actual case. In the next two sections we review the literature on ERP implementation and business and IT alignment to provide the conceptual foundations for the paper. Section 3 also describes the principles of the ANP method and defines the focus of analysis in the paper. Section 4 presents the results of the case study. Finally, section five discusses the results and conclusions of this paper.

We believe that the process presented in the paper represents a relevant model for improving business and IT alignment for ERP implementation. The ANP methodology itself seems to be applicable for this kind of decision-making, although its application is somewhat laborious and also requires theoretical knowledge on the method. However, the ANP approach enforces the analytical comparison between the alternatives as well as improves the decision visibility, for example, for external auditors and boards investing huge amounts of money in the ERP software. The analysis conducted in the present paper brings insights to the ERP module sequencing problem. We developed an ANP model for deciding on the implementation sequence of ERP modules, using the alignment of business and IT and the IT investment perspective as the theoretical foundation for the model. We described the decision-making process for ERP module sequencing and the application of the ANP method in detail to enhance the practical usability of the method. The method can be considered rather easy to use and very informative in terms of providing clear results. Our model is technology oriented focusing on software module sequencing, meanwhile social construction elements are embedded into business requirements, risks and solution constraints. However, the method has its limitations as well, such as the model becoming easily rather complicated, which may limit its use in some situations. The benefits from utilizing this ANP model for ERP module sequencing do not only come from the explicit priorities between the alternative modules. Major value can be derived from the knowledge transfer during the ANP data collection process, because answering to pairwise comparisons requires an enhanced dialog between the ERP vendor, the IS department and the business units. ERP implementation process is more knowledge, resource, requirements and change management challenge than technical IT system deployment. Thus, issues regarding IT and business alignment, as well as IT governance might surface during the data collection process. Key business requirements may be biased, current understanding about solution constraints may be limited, and priorities may be unintentionally sub-optimized because of bounded rationality and past experiences from non-integrated operations and systems with limited data models. The reason for choosing the ANP method was that it is suitable for networked decision problems. Similar analysis could be conducted, for example, by applying a system dynamics model in combination with the AHP model. An example of this approach can be found in Kivijärvi and Tuominen (1991). The main strength of the ANP method is that it is goal oriented and it also allows the feedback loops in the model (which is not possible in AHP, for example). The most obvious limitation of the applicability of the ANP method is the very high number or pairwise comparisons that need to be conducted. However, there are several different ways in which the data input can be conducted, such as graphic, verbal, matrix and questionnaire formats that can be used to input the data, which can make the data collection easier. As AHP, ANP also has the rank reversal problem, meaning that adding an indifferent alternative or criterion to the model may actually reverse the whole solution (Dyer, 1990 and Pérez et al., 2006). Furthermore, the results of the sensitivity analyses in ANP are rather difficult to interpret and may thus have very limited value. Moreover, a certain amount of theoretical knowledge is needed to choose the method for calculating the limitmatrix. There are many different ways to calculate the results, which may be confusing to a non-expert user. This may limit the practical use of the method. In the present paper we used the “calculus type” for calculating the limit matrix, which is recommended by, for example, Adams (2001). Finally, there are also limitations related to the software used in the analysis, for example, the Superdecisions software does not allow multiple users to work on the model simultaneously, which is allowed, for example, when using AHP and the Expert Choice software. Furthermore, the overall consistency of the model is more difficult to calculate in ANP than in AHP. ERP module sequencing is not a simple task: it requires patience and careful analysis of the key business requirements, the solution constraints, the risks and the costs as well as the alternative ERP modules. Since the decision involves a myriad of issues, the results of the analysis must be addressed with caution and they may not be considered an absolute recommendation. For example, there are other options than implementing only the top prioritized ERP modules or implementing all the relevant ERP modules from 1 to 9 at the same time. As described in the previous section, one possible option might be the implementation in module groups. For practical usability, the abstraction level of the ANP model should be carefully considered and the analysis should focus on a rather limited decision-making domain to control the complexity of the analysis. The ANP process itself, however, is valuable for enhanced knowledge transfer and improved communications regarding the ERP module sequencing decisions, and simultaneously improving business and IT alignment. The future research could include combining the ANP method with other decision support methods to provide a more comprehensive methodology for the ERP investment process. The ANP model developed could also be used in ERP software selection phase: comparing alternative modules and solution constraints from various ERP packages could improve understanding of the differences between different technologies.