Outsourcing in software development projects has become a widely adopted utility model for dealing with competitive issues such as cost pressure and the lack of skilled human resources. Established outsourcing relationships in terms of captive o shoring centers or clientvendor partnerships are prone to fail due to the lack of quali ed outsourcing decisions for the allocation of software components to either in-house or external/o shored development. Therefore, this design-oriented research endeavor aims at developing a methodology and a corresponding tool to inform and improve the outsourcing decision in global software development projects on the basis of software components. For that reason, the results of explorative case studies researching outsourcing decision making as well as concepts of collaborative software development are combined and form the proposed methodology for supporting project leads in decision making and in avoiding of instinctive and spontaneous decisions on the allocation of software components.
From a national and global perspective, the sourcing of application systems has
signi cantly matured and been widely adopted over the past years. A signi
cant amount of software companies nowadays engage in software development
outsourcing (SDO) as they expect bene ts such as cost reduction,
round-theclock-development, or additional skilled workforce [
From a global software development (GSD) perspective rms have to deal
with collaboration of team members that are separated in di erent locations.
Since separation in terms of location or time challenge GSD projects in regard to
their success, collaboration platforms and smart concepts for communication and
collaboration have been established. Amongst others, traceability has become a
widespread methodology in GSD for controlling, tracing, and decision making
in globally distributed software projects [
As stated above, the sourcing decision of software components has only been scarcely researched while the development setups of companies have shifed to a global perspective. Hence, GSD provides approaches and concepts that fertilize the decision making in SDO. For that, the research questions of this doctoral thesis are as follows: 1) Which in uencing factors have an impact on the sourcing decision of software components? 2) How can a sourcing decision for software components based on the in uencing factors be made? These research questions lead to this theory-informed and design-oriented research endeavor including the following research objectives: 1. Developing a methodology to inform the outsourcing decision of SDO projects on the allocation of software components to either in-house or external/o shored teams and 2. to design and to implement an information system for supporting and informing the component-based outsourcing decision.
The remainder of this research proposal is structured as follows: First, basic information about SDO and GSD as well as outsourcing decisions as foundation of this research endeavor are described in the next section. Following, related research in the area of outsourcing decisions is presented while selected theories providing determinants for the component-based outsourcing decision are introduced. In the consecutive subsection, the proposed approach and its deduction, the used research methodology, and the preliminary ndings are described. Finally, this proposal is conclude in the last section. 2 2.1
Development of software products in a globally distributed setting is the
preferred software engineering mode for an increasing number of companies [
Software development with virtual teams is highly interconnected with
outsourcing in software development projects which "is a multifaceted and complex
activity in which clients and vendors interact in many di erent ways to produce
and deliver the software services required" [10, p. 960]. This results in di
erent settings of client-vendor-relationships which are considered as outsourcing of
software activities to an external provider (classical outsourcing), as outsourcing
to an external provider in an o shore/farshore country (o shoring), or as internal
provision at an o shore location (captive o shoring) whereof each outsourcing
type is challenging di erently but is also strongly supported by concepts of GSD
[
The variety and availability of di erent outsourcing types are not only means
for large software companies. Even small and medium-sized enterprises make use
of the described sourcing models as they want to participate in the advantages
of GSD and SDO as well, which are for them primarily the access to skilled
workforce, the exibility to allocate and to clear resources, or to save costs [
As a consequence, this research endeavor aims at analyzing the concepts of
GSD, especially the concept of traceability and rationale management (TRM),
with regard to failing factors of SDO projects. As traceability and rationale
information in distributed software development projects represent a network
consisting of software development artifacts, connections between these artifacts,
and annotations to these connections, the stringent maintenance of such data
provides detailed information about planned software components [
Decision making as a part of SDO is a vital element for minimizing the faultproneness of an outsourcing relationship. It comprises the analysis of multifaceted in uencers regarding their implications on the sourcing strategy of a company. Hence, a range of outsourcing decision making approaches has been proposed in past research.
[
However, current outsourcing decisions are generally strategic in nature [
Decision Making in Software Development Outsourcing
The aim of this dissertation work is to analyze determinants and contingency
factors that in uence the allocation decision for software components in SDO
(cf. research question 1). Additionally, this research endeavor seeks to develop a
decision support approach for evaluating software components according to their
outsourcing potential (cf. research question 2). Therefore, the wheel has not to be
reinvented, since there already exist collections and literature synopses of
decision determinants as shown by [
Support of the decision making process in SDO has experienced little
attention in IS research only. Similar to determinants and contingency factors in the
paragraph above, decision making processes have been developed on a strategic
level thoroughly but have been unattended on project level, where the question
of component allocation arises. Thus, the dynamic model of o shore software
development by [
In the course of a theory informed design science research approach, relevant
theories from IS literature that are used to inform the search for decision
determinants have been selected and are explained in the following: The Resource-based
View (RBV) considers a rm as a set of productive resources [
The Knowledge-based View (KBV) has its origin in the RBV of the rm and is
a theoretical perspective that regards knowledge as the most important resource
of the rm [
Additonally, Transaction Cost Economics (TCE) consider the market participation as costly while identifying conditions under which market governance is more advantageous than internal governance of a rm [36, 37]. Hence, transaction costs have to be considered when participating in the market and the assumptions of bounded rationality and opportunistic behavior of other participants have to be taken into account. For protecting against opportunistic behavior of another market participant transaction costs occur.
Finally, the Systems Theory (ST) [
In order to meet the intended research goals, the structure of this dissertation
endeavor is design-oriented in nature [
In a rst step, case studies with eight small and medium-sized enterprises (SME) that engage in GSD with virtual teams have been conducted. The study describes the software development phases that are suitable for outsourcing, organizational and operational aspects of outsourcing decision, and characteristics of software components that qualify for outsourcing (the publication is under review at the moment). The study reveals the need for decision support in the role Step Results / Planned Activity Status Case studies Identify need for a decision support methodology Done in SDO projects on the basis of software components.
Deduction of require- Deduction of requirements from RBV and KBV Done ments from theory of the rm, from ST, and from software engineering principles. Development of a t model including determinants of the allocation decision of SDO projects that a ect software components.
Complement determi- Extend the set of determinants with exploratory In nants with input from case studies to prove the relevance for practice. progress practice Development of a deci- Development of an outsourcing decision support In sion support methodol- methodology for the allocation of software com- progress ogy ponents (in-house development versus external or o shored development).
Development of a deci- Implementing the requirements of the decision Planned sion support tool support methodology as mobile web-application by enriching the determinants and contingencies with available data from collaboration platforms of GSD teams.
Evaluation Evaluation of the the developed methodology for Planned component-based outsourcing decisions as an empirical study in an organizational setting. of a project manager and shows that software development outsourcing is mainly a ecting the implementation phase of a software project as software components are allocated either for in-house or outsourced development. Therefore, this research endeavor intends to develop a methodology and tool support for project managers in order to make informed outsourcing decisions for components.
In a second step, requirements for the intended decision support methodology were deduced from theories described in the previous subsection. From the RBV, the KBV, and the ST the following requirements have been derived: { Software components that are highly interacting with other components of the software system should not be developed over company boundaries (ST). { The priorities of connected requirements or requirements sets of a component describe the importance of the software component and, thus, increase the competitive advantage of a rm. Hence, the subsystem should not be outsourced (KBV). { In the case of intense communication between software developers for the planning/implementation of a single component, the development should not be outsourced (RBV). { Depending on the required knowledge for the software component to be developed, the component is either developed in-house or outsourced, depending on the skill availability. The required knowledge has to be di erentiated between functional knowledge, business process speci c knowledge, and technical knowledge (KBV). { The development of a component should not be outsourced if a frequent exchange of information is required (TCE).
Additional requirements were derived from outsourcing literature and software engineering practices in GSD: { Traceability and rationale information resulting from collaboration platforms should be used to improve the sourcing decision making process (e.g. use traceability details such as requirements and their connectivity to software components in order to determine the importance of a component by the priority of connected requirements). { The solution should be optimized for project leads, since they are primarily in charge of outsourcing decision making. Thus, the software tool should be lightweight and should be provided as mobile solution but must not be lacking the interfaces to latest data stored in the rm's collaboration platform.
As described in Table 1, in a third step, the requirements that were deduced from existing theories are enriched by multiple case studies of an explorative research design. The case studies consist of interviewing experts and reviewing as well as analyzing documents and data of software companies. For this purpose, semi-structured interviews according to a questionnaire consisting of 42 open questions are conducted [38]. The case studies aim at complementing the in uencing determinants and contingency factors of component-based outsourcing decision making. Thus, the target groups of the interviews are project leads or deciders of SDO projects in small, medium-sized, or large software engineering enterprises. Currently, six companies have been analyzed and at least six more cases are planned. The preliminary results disclose further determinants and contingencies that in uence the sourcing decision. Hence, the SDO decision and its determinants can be modeled by the t model as presented in Figure 1.
According to the proposed conceptual t model, the development of a
component-based decision support methodology for SDO projects is in the focus
of the next step. By incorporating the explored determinants of the
outsourcing decision, the approach makes use of the TRM approach of GSD. For each
determinant the possibility of using TRM data is analyzed and conceptually
described. Hence, this research endeavor aims at informing the sourcing decision
to a high degree with structured data from used collaboration platforms and to
increase the utility of the proposed methodology. An initial instantiation of the
component-based decision support methodology has already been published [
Finally, in step ve and six, the implementation and evaluation of the
conceptual methodology complete the theory-informed artifact building from case
studies [
Researching component-based outsourcing decision making in SDO combined with the development of tool support by implementing concepts of GSD is highly demanded by companies in corresponding outsourcing relationships. However, not only the relevance for practice but also the request in academic research for theoretical contribution to outsourcing decision support on a component level lay the foundation for this research endeavor that has the goal to facilitate such sourcing decisions by structured and well informed analyses of the component speci cs. At the same time this research work extends the knowledge base of decision making theory in SDO.
The next steps of this research endeavor include the completion of the exploratory phase by conducting the remaining case studies as well as the nal definition of the described decision support methodology. Afterwards, the intended artifact in form of a software tool is implemented and used for the evaluation and testing of the underlying methodology.
This doctoral research is being supervised by Prof. Armin Heinzl (University of Mannheim, heinzl@uni-mannheim.de) and is supported by the German state of Baden-Wuerttemberg as part of the research project "GlobaliSE". [36] Williamson, O.E.: Markets and Hierarchies: Analysis and Antitrust Implications.
Free Press, New York (1975) [37] Williamson, O.E.: The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. Free Press, New York, USA (1985) [38] Yin, R.K.: Case Study Research: Design and Methods. Sage Publications, Los Angeles, California, USA (2010)