Situational Method Engineering in Practice: A Case Study in a Small Enterprise

Abstract

In this paper we report on our experience in the application of the assembly-based situational method engineering approach in practice. In particular, we demonstrate the construction of a situation-speci c method for a small size company operating in the domain of e-commerce. The aim of this work was to help the company to specify its business model and activities in order to gain in common understanding and to provide means for further business evolution and innovation.

Keywords

• Jolita Ralyte

-

application, while in section 4 we discuss advantages and di culties identi ed during this case study. 2

Situation-Speci c Method Design

The discipline of Situational Method Engineering (SME) [ 2 ] promotes situationspeci c method construction on the y by reusing existing method components. The aim of such method construction is to t the situation and method requirements of a particular organization or a project. In our study, we followed the assembly-based SME approach [ 4 ][ 6 ], which uses the notion of method chunk as the main building bloc. A method chunk is de ned as an autonomous and coherent part of a method, which combines method product and process perspectives in the same component (as opposed to a method fragment which separates them into product fragments and process fragments [ 2 ]). The assembly-based SME process consists of three main steps: method requirements speci cation, method chunks selection and assembly of selected method chunks (see Fig. 1). As shown in Fig. 1, there are two strategies for specifying situation-speci c method requirements: intention-driven and process-driven. The intention-driven strategy is especially suitable for an existing method adaptation by adding new intentions and/or strategies, while the process-driven strategy is relevant in the case of a brand new method construction. The aim of our study was to construct a project-speci c method dedicated to specify the company's business situation with a set of models and to identify potential business innovations. Construction of a new method was necessary because the company did not use any particular method and we could not identify any existing method fully satisfying the situation of our project. This strategy requires: rst (1) the assessment of the enterprise/project situation, then (2) the identi cation of a set of engineering intentions that are required to be ful lled by the new method, and nally (3) the identi cation of potential strategies to achieve the intentions and ordering of these intention and strategies by using the Map formalism [ 7 ]. The result of this step is a generic process model, named requirements map, represented as a graph where nodes are process intentions and arcs represent di erent strategies to achieve them. In our case, we have evaluated the situation of the company with a set of criteria, some of them are shown in Table 1. Very small enterprise (7 employees) with an aim to grow.

Founded in 2010 with one specialized e-commerce web store. In 2011 the company grows to 9 web stores. The objective for 2012 { 20 web stores.

Market

Niche market based on the "market to demand" model.

Need for High: the company has to permanently look for new niche products innovation and to increase their o er { to create new web stores. evolution Need for High: the company has to watch technology and business innovations strategic watch to stay competitive in the market.

Management of By project: The creation of each new web store is managed as a growth project.

High IS/business No formal or semi-formal documentation available concerning busidocumentation ness activities, information system and applications.

High interest and enthusiasm but no experience in modeling and describing business activities.

Criteria Size of the company Maturity Impact of a new project Skills in modeling

The characterization in Table 1 demonstrates that, in order to stay competitive in the market, this company has to constantly increase its o erings and to innovate its business strategy. However, it does not have any well-structured documentation concerning its business model, activities and data. In a few years, this company will not be a start-up anymore and having an appropriate documentation will therefore be key for the evolution of the enterprise. Therefore, we have identi ed two main objectives that should be satis ed by the new method: (1) to document the enterprise business situation and (2) to discover potential business evolution options based on the analysis of the produced models. In particular, it was decided to use three modeling perspectives: business, business process and information (data) in both method phases with potentially di erent strategies to manipulate these models. During the rst phase of the method application, these models should serve to specify the As-Is situation, while in the second phase they should allow to discover and evaluated the potential To-Be situations. The requirements map for the new method construction is illustrated in Fig. 2. It says that the method will be based on two main intentions (Document enterprise business situation and Discover business evolutions) and identi es the types of approaches/techniques that should be used to achieve these intentions in terms of generic strategies (e.g. business modeling techniques, process modeling techniques, etc.). In the next step, we use this requirements map to select appropriate method chunks, i.e. modeling and exploration techniques. Once the method requirements have been speci ed, the selection of the method chunks matching these requirements can start. The Requirements-driven strategy (see Fig. 1) helps to formulate method chunks selection queries by giving values to the attributes used to specify method chunks descriptors and interfaces (e.g. type of engineering activity, situation or source information/product, intention to achieve, etc). For example, Select method chunks where Engineering activity = "Design" AND Technique = "Process modeling" AND Situation = ("Business activity expertise" OR "Business activity description") AND Intention = (Verb = "Construct" AND Target = "Process model").

The detailed method chunk metamodel can be nd in [ 4 ]. At least one method chunk has to be selected for each requirements map section (i.e. <source intention, target intention, strategy>). Evaluation, Decomposition, Aggregation and Re nement strategies (see Fig. 1) can be used to re ne the candidate chunk selection by re ning the selection query and analyzing more in depth if the chunk matches the requirements. In our case, we did not have a fully operational repository and the selection of method chunks was mainly based on the literature review and the author's (plying the method engineer role in the project) Business modeling techniques Process modeling techniques Model consistency check personal method knowledge. For example, for business model construction and innovation, we have selected two method chunks: e3value [ 1 ] and Business Model Canvas (BMC) [ 5 ]. These two business modeling techniques allow capturing complementary business model perspectives. While e3value focuses on the collaborations with business partners and value exchanges, BMC puts forward the business value propositions and describes how an organization creates, delivers and captures value. Business model patterns (BM patterns) proposed in [ 5 ] facilitate business model evolution and can be used together with BMC for discovering how an enterprise business model could evolve { modeling possible To-Be situations. Furthermore, in [ 5 ] the authors propose a set of design approaches such as Customer insights, Ideation, Visual thinking, Storytelling, etc. that are considered as creativity and innovation techniques and can be combined with business model canvas for creating new business models. In particular, we have selected and tested the techniques named Empathy map, What-if questions, Scenarios and Brainstorming. The Empathy map technique helps to identify di erent categories of enterprise customers, to better understand their environment, behavior, concerns and aspirations, and to design better value propositions and more appropriate customer relationships. The What-if questions help team members to break free of constraints imposed by current models, while Scenarios allow the imagining of new ways of realizing business activities. We recommend completing the application of these techniques with concluding Brainstorming sessions. Table 2 lists the method chunks selected for each requirements map strategy. e3value [ 1 ], Business model canvas (BMC) [ 5 ].

BPMN was selected for modeling enterprise activities in terms of process models. In particular, we recommended using guidelines provided in [ 8 ].

Data modeling Any type of class/entity diagram ts very well for producing conceptechniques tual domain models. We have used simpli ed class diagrams based on two types of relationships: existential dependency and specialization. We did not nd any method chunk allowing to check consistency between the three types of models (business, process and data). We have de ned a set of consistency validation rules to satisfy this method requirement.

Innovation and Business modeling patterns (BM patterns) [ 5 ]. Techniques extracted exploration form [ 5 ]: Empathy map, What-if questions, Scenarios, Brainstorming. techniques End of use

The application of the method stops when the enterprise decides to stop the use of the documentation obtained by applying this method.

Method Chunks Assembly As shown Fig. 1, the assembly-based SME approach identi es two strategies, named association and integration, to assemble selected method chunks into a new method. The integration strategy has to be applied if selected method chunks have similar engineering goals, their process and/or product models overlap (i.e. contain same or similar elements) and they are used to produce the same deliverable (e.g. the same model). Otherwise, the association strategy is used to position the method chunks in the new method and provide guidelines for their execution. In our case, the method chunks representing two business modeling techniques { e3value and BMC { have the same engineering goal { to construct a business model, but they produce di erent and complementary business models. Therefore, the association strategy is su cient to indicate that these two method chunks can be applied in parallel without any particular ordering. In contrast, BMC and BM patterns are overlapping method chunks because they use the same canvas model. However, the integration of these two method chunks already exists because they are extracted from the same approach [ 5 ]. Besides, in our method these two method chunks are used for di erent purposes: BMC for business model description and BM patterns for business model innovation. Other selected method chunks deal with complementary engineering goals and simple association is su cient to combine them into the desired method. Fig. 3 illustrates the process model of the assembled method.

Examples of Method Application

Several collaboration and modeling sessions have been organized together with the employees of the company in order to specify the initial business documentation. In particular, we have developed e3value and BMC models to represent the company's business model. Enterprise activities were formalized with business process models by using BPMN and the data models with class diagrams. This project permitted the production of semi-formal documentation of enterprise business and uni ed employees' awareness of their company business structure and activities. For illustration purpose a few models are shown in Fig. 4.

In order to test the second phase of our method we have experimented the Empathy map technique followed by a brainstorming session. Participants were divided in two groups and explored two di erent user pro les. Then, the brainstorming on these two pro les permitted the identi cation of new ways to improve customers loyalty and even to transform them into purchasing advisors. During this study, we have frequently observed that the employees of the company (who are also its co-creators and business partners) were not necessarily speaking the same language and therefore not always having the same understanding of things. Each of them was having his/her own interpretation of enterprise business model, activities and his/her own roles and responsibilities. This work has assisted the company to clarify and unify their understanding and to produce a semi-formal documentation of its business structure, processes and domain concepts. In addition, this study has motivated the managers of the company to think about some improvements, innovations and changes to be implemented in a near future.

Various discussions and collaborative sessions with the company employees demonstrated that the method was well adapted to the project situation and requirements. Overall, they understood models that we have developed and techniques to produce them, and found them quite intuitive. They are determined to use them as a support for enterprise evolution management in the coming years.

This case study demonstrates that not only big companies need methods and modeling techniques to describe their business, activities, information systems and enterprise architecture in order to manage their complexity and evolution. Small and medium size enterprises can also bene t from model-based documentation to establish a common understanding and agreement on enterprise activities and to facilitate future development strategies. However, a small company can be easily lost in the jungle of modeling approaches and techniques proposed in the literature and di erent Internet sources. It needs help in selecting and combining appropriate method chunks, and it needs training in applying them in practice, at least at the beginning of the documentation process.

References

1. Gordijn , J. , Akkermans , H.: E3-value: Design and Evaluation of e-Business Models . IEEE Intelligent Systems 16 ( 4 ), 11 - 17 ( 2001 )

2. Henderson-Sellers , B. and Ralyt , J.: Situational Method Engineering: State-of-theArt Review . Journal of Universal Computer Science , Vol. 16 ( 3 ), 424 - 478 ( 2010 )

3. Johannesson , P. : The Role of Business Models in Enterprise Modelling . In: Krogsie, J. , Opdahl , A.L. , Brinkkemper , S. (eds), Conceptual Modelling in Information Systems Engineering , pp. 123 - 140 , Springer ( 2007 )

4. Mirbel , I. and Ralyte J. : Situational Method Engineering: Combining AssemblyBased and Roadmap-Driven Approaches . Requirements Engineering 11 ( 1 ), 5878 ( 2006 )

5. Osterwalders , A. , Pigneur , Y. : Business Model Generation: A Handbook for Visionaries, Game Changers and Challengers . John Wiley & Sons, Inc. ( 2010 )

6. Ralyte , J. , Deneckere , R. , Rolland , C. : Towards a Generic Model for Situational Method Engineering . In: Proceedings of CAiSE 2003 , pp. 95 - 110 , LNCS vol. 2681 , Springer ( 2003 )

7. Rolland , C. , Prakash , N. , Benjamen , A. : A Multi-Model View of Process Modelling . Requirements Engineering , 4 ( 4 ), 169 - 187 ( 1999 )

8. Silver , B.: BMPN Method & Style. Cody-Cassidy Press ( 2009 )