Modelling in the Large

LP0560417

Chief Investigators: Prof Michael Rosemann, Prof. Peter Green (UQ), Prof. Graeme Shanks (Monash) and Ms Wasana Sedera (QUT)

Partner Investigators: Assist. Prof Michael zur Muehlen (Stevens Institute of Technology) and Dr. Wasim Sadiq (SAP Research)

Funding: ARC linkage
2005: $89,819
2003: $91,885
2004: $48,296
plus contributions (cash and in-kind) from the industry partner

Category: 2801 - INFORMATION SYSTEMS

Administering Institution: Queensland University of Technology

Industry Partner: SAP Australia (SAP Research)

This research project has been initiated by two presentations from Michael Rosemann and Graeme Shanks, and Michael zur Muehlen and Michael Rosemann at the SAP Innovation Congress in Orlando, Florida, in February 2004.

Summary

Business modelling supports a variety of approaches, which are currently of significant interest for Australian organisations. This includes business process management, risk management or Enterprise Architecture design. The trend to centralised and integrated business modelling leads to projects with a high number of models, modellers, users, and modelling purposes. This phenomenon is called modelling in the large. The proposed research project explores the major issues within these projects and develops improved modelling techniques and project management approaches for large modelling initiatives. The outcomes will streamline the design, integration, maintenance and communication of models with a focus on Enterprise Systems.

Overview

Conceptual models are widely used for the design of complexity-reduced visualisations of relevant parts of the real world. Over the last 10 years, significant changes in the actual usage of conceptual modelling in practice can be observed.

First, a focus on data models has been replaced by a focus on business process models. The rather static data view provided valuable insights into the structure of master and transaction data, but lacked a clear impact on business decisions. Business Process Modelling became popular with the uptake of Business Process Management and is now widely applied. A recent study supported by the Australian Computer Society indicated that four out of the top six modelling purposes address the understanding and improvement of business processes.

Second, business processes are closer to the interest of managers, business representatives and end users than the less intuitive data models. Thus, an overall shift from models that specify system requirements to models that capture business requirements can be observed. The related research community is traditionally based in the field of Computer Science, and does typically not conduct comprehensive research in the area of business modelling.

Third, modelling takes place more and more in an environment of off-the-shelf-solutions such as the SAP solutions. Unlike classical requirements analysis that starts literally each time from scratch, Enterprise Systems modelling is typically based on reference models. Reference models have to be designed, communicated, maintained and individualised. This situation implies a design-for-reuse / design-by-reuse paradigm and demands modelling techniques that can cater for the needs of model configuration. The majority of books on systems analysis and design, however, does not even mention the term "reference model".

Fourth, the general uptake of modelling leads to an increased number of modelling purposes within one organisation. This situation requires integrated modelling techniques that interrelate, for example, data models, process models, organisational models and application models. Current research, however, is often focused on isolated modelling for a particular purpose (e.g. Enterprise Systems, process re-design, risk management or workflow management).

Fifth, the popularity of modelling for many different purposes motivates many large organisations to centralise all modelling initiatives. The aim is to provide an integrated repository that supports all involved modellers and stores a large number of models often in excess of 500. In contrast to this trend, it can be observed that the core of research on modelling is on rather small modelling examples that do not reflect the complexity as it can be seen in practical projects,

Sixth, in correspondence with the increased demand for a more mature modelling environment, sophisticated modelling solutions such as ARIS, Proforma, Casewise or SystemArchitect have been developed and applied by Australian organisations. These tools are labelled as Upper CASE (Computer Added Software Engineering) tools as they are not able to generate software from the models. However, these tools have received little focus in academic research.

Overall, the related research on conceptual modelling does not sufficiently acknowledge the current trends in business modelling. The focus is often on the development and evaluation of modelling techniques and tools. Empirical work on conceptual modelling in general and on the management of very complex business process models is the exception. Thus, practitioners face significant challenges in comprehensive modelling projects.

The main aim of this project is to improve the quality of enterprise-wide business modelling initiatives with a focus on projects driven by Business Process Management. The project concentrates on issues related to "modelling in the large", in particular, the representation of large-scale models and the design, communication, maintenance and use of a number of interrelated models by a team of modellers. It includes both conceptual models that are developed within organisations and the reference models that describe enterprise systems such as SAP. We will adopt available research outcomes as well as develop new approaches to deal with the existing complexity. These methods will make modelling more intuitive and facilitate the tracking of modelling intentions.

The project includes three subsidiary aims represented by the following research questions:

1. How can large scale process models be represented for better understanding and communication?
2. How are multiple process models managed in large scale modelling projects?
3. What are the main issues faced by Australian organisations undertaking "modelling in the large" and how are they being addressed? How does the situation in Australia compare with current practice in the USA and Germany?

Significance and Innovation

Improving the quality of enterprise-wide business process modelling initiatives is significant for the following reasons:

Many Australian organisations are currently undertaking enterprise-wide business process modelling projects involving hundreds of modellers, and exploring enterprise-wide standardisation and centralised management of their modelling initiatives. There are currently few guidelines for this type of project. It is important that best-practice is understood and synthesised into relevant and appropriate issues and guidelines.

Working with the industry partner, SAP, and our international colleagues will enable a global perspective on modelling in the large to be developed. In particular the comparison of issues and means of addressing those issues in Australia, Germany and the USA will ensure that the guidelines developed will be useful internationally.

The combination of field and laboratory-based research approaches will ensure the outcomes of the study have strong validity at both at the detailed individual and broader organisational levels.

Outcomes of the project will be of great importance and relevance not only to organisations undertaking modelling in the large projects and to our industry partner, SAP, but also to a large number of Australian consulting companies that function as SAP implementation partners.

The deep understanding of modelling in the large issues will provide SAP with rich insight into how to represent and manage reference models for their enterprise software packages, and how reference models can be used more effectively in large scale business process modelling projects.

There are four major outcomes from this research project:

• A set of issues that Australian organisations are facing when undertaking large-scale process modelling projects and a corresponding set of guidelines for addressing these issues.
• A set of guidelines for improved representation of large-scale business process models.
• A set of guidelines for the management of multiple models in large-scale business process modelling projects.
• A comparison of large-scale business process modelling issues and means of addressing these issues in Australia with Germany and the USA.

Research Team

Professor Michael Rosemann
Professor Peter Green (University of Queensland)
Professor Graeme Shanks (Monash University)
Ms Wasana Sedera
Assist. Professor Michael zur Muehlen (Stevens Institute of Technology, USA)
Sonia Lippe, SAP Research
Hui Min Tan, QUT
Corina Raduescu, University of Queensland
Malini Jayaganesh, Monash University
Dr. Wasim Sadiq (SAP Research)

Documentation

Raduescu, C., Tan, H. M., Jayaganesh, M., Bandara, W., zur Muehlen, M., Lippe, S. (2006). A Framework of Issues in Large Process Modeling Projects. (PDF, 126Kb) Paper presented at the 14th European Conference on Information Systems, Goetesburg , Sweden , 2006.