Designing Organizational Models for Service Integration in a Multi-sourcing Context

Size: px
Start display at page:

Download "Designing Organizational Models for Service Integration in a Multi-sourcing Context"

Transcription

1 Designing Organizational Models for Service Integration in a Multi-sourcing Context Marius Goldberg Karlsruhe Institute of Technology marius.goldberg@kit.edu Full Paper Gerhard Satzger Karlsruhe Institute of Technology gerhard.satzger@kit.edu Abstract With the persistent shift towards multi-sourcing, the complexity of service delivery is continuously increasing. This presents new challenges for clients who now have to integrate interdependent services from multiple providers. As other functions, service integration is subject to make-or-buy decisions: clients can either build the required capabilities themselves or delegate service integration to external functions. To define detailed organizational models, one requires understanding of specific tasks and how to allocate them. Based on a qualitative and quantitative expert study, we analyze generic organizational models, and identify key service integration tasks. The allocation of these tasks to clients or their providers generates a set of granular organizational structures. We analyze drivers for delegating these tasks, and develop typical allocations in practice. Our work contributes to expanding the theoretical foundations of service integration. Moreover, our findings will assist clients to design their service integration organization, and to build more effective multi-sourcing solutions. Keywords Service Integration and Management, Service Integration, Multi-sourcing, IT Service Management, IT Outsourcing Introduction With the constant shift towards multi-sourcing, the complexity of IT service delivery is continuously increasing (Bapna, Barua, Mani, & Mehra, 2010; Wiener & Saunders, 2014). This introduces new challenges for clients that have to manage and integrate services from a multitude of providers into endto-end services (Jin, Kotlarsky, & Oshri, 2014; Plugge & Janssen, 2014). In accordance, academic research and practice are increasingly focusing on service integration (Bapna et al., 2010; Wiener & Saunders, 2014). This concept is concerned with orchestrating services delivered by various internal and external providers into seamless end-to-end services (Anderson and Parker 2013; Bapna et al. 2010; Davy 2014; Goldberg et al. 2014a; Schermann et al. 2006). In this paper, we develop and evaluate a set of key service integration tasks. They enable the definition of granular organizational models by allocating individual tasks to internal or external providers. In doing so, we aim to provide a structural basis for future research regarding service integration organizations. As various providers increasingly deliver interdependent services, the service landscapes of clients are becoming more complex (Arora et al., 2013). In accordance, Gartner reports that only very few companies are ready to perform service integration (Longwood and Heiden, 2012). Goldberg et al. (2015) find that organizational structures need to be adapted to the changed requirements to enable successful service integration. To deal with the increasing complexity, practitioners and researchers are looking for models to transfer service integration responsibilities to external entities (Bapna et al., 2010; Davy, 2014). Academic research identifies a set of typologies for these service integration organizations (Goldberg et al. 2014a). Extant research, however, does not give enough guidance on how to build these. There is no clear understanding of how to distribute service integration tasks within the complex multi-sourcing Twenty-first Americas Conference on Information Systems, Puerto Rico,

2 environments that is, which tasks to keep in-house and which to outsource in order to achieve effective service integration (Anderson and Parker 2013; Goldberg et al. 2014a; Urbach and Würz 2012). We intend to address this research gap by investigating the relevant organizational models proposed by academic research. We will complement them with integration tasks, and examine which of these tasks are typically outsourced and what the reasons are for doing so. Hence, our paper addresses the following research questions to expand the current knowledge of service integration: Which relevant basic organizational models do exist for service integration? Which drivers influence the client s decision to outsource individual service integration tasks? How are service integration tasks typically allocated between client and providers? To answer the research questions, we perform an explorative, multi-stage research study. Building on the results of a structured literature analysis, we carry out multiple expert interviews. Based on expert knowledge, we then identify the three most relevant organizational models, discuss their advantages, and derive important service integration tasks. Subsequently, we perform an empirical, quantitative study. It assesses which factors will influence the decision to outsource individual tasks, and determines typical resulting allocations of these tasks between clients and providers. The findings of our work allow refining existing theoretical models. From a practical perspective, our contribution should help outsourcing clients to better understand service integration, assess their current situation, and design effective and mature client organizations. The paper at hand is structured as follows. First, we provide an overview about related research in the following section, before outlining our research method. In the fourth section, we identify and describe the most relevant service integration models from a client perspective including a discussion of their advantages. The subsequent section then outlines service integration tasks, relevant decision drivers and typical task allocations. Finally, we conclude with a summary, a discussion of limitations and future research possibilities, as well as managerial implications. Related Work Service integration is an emerging topic both in practice and theory. In the following, we will present and discuss related literature work. While much research has been published on single- and multi-sourcing (Lacity, Khan, & Willcocks, 2009), Bapna et al. (2010) argue that multi-sourcing with interdependent services is fundamentally different and more complex. That is, services are characterized by a high degree of interdependence and not delivered as silos. As a consequence, existing concepts and findings cannot simply be transferred. To deal with the increasing complexity of multi-sourcing, practitioners and researchers are looking for ways to outsource service integration activities. This leads to a new shift of management responsibilities from the client towards external entities (Goldberg et al. 2014a). The respective organizational structures and governance aspects for service integration require additional research (Anderson and Parker, 2013; Bapna et al., 2010; Goldberg et al., 2014a; Urbach and Würz, 2012). In general, integration can be defined as the level of collaboration between organizational units jointly delivering interdependent tasks (Anderson & Parker, 2013; Gulati, Lawrence, & Puranam, 2005; Lawrence & Lorsch, 1967). Service integration often also denoted as service integration and management (e.g. Goldberg et al. 2014a) refers to various internal and external service providers delivering interdependent IT services. It aims at integrating and managing their various services and service elements into end-to-end services (Anderson and Parker 2013; Davy 2014; Goldberg et al. 2014a). The level of service interdependence thereby determines the need for service integration (Bapna et al., 2010). Plugge and Janssen (2014) provide a first overview of governance requirements for multi-sourcing with interdependent services. They investigate how resources are coordinated in multi-sourcing arrangements and how different organizations interact. Their research does not cover different organizational configurations but assumes a generic structure. In contrast, our work validates strategic models for structuring the multi-sourcing organization. Twenty-first Americas Conference on Information Systems, Puerto Rico,

3 Outsourcing strategic service integration tasks is a means for risk sharing in complex multi-sourcing. Satzger and Kieninger (2011) discuss risk sharing in service contracts and service systems. The focus of the paper is mainly on decision criteria regarding selection and development of these contracts. In contrast, our work is focused on organizational models. An important basis for our paper is the work of Goldberg et al. (2014a). The authors present and discuss five organizational models for structuring service integration. They provide a theoretical overview based on typologies. Their work, however, stays mostly generic. In contrast, our work provides detailed insights into selected models regarding decision drivers and allocations of tasks. Goldberg et al. (2014a) differentiate a strategic and operational service integration function. They argue that the clients retained organizations should always perform the strategic function. Retained organizations are generally defined as the parts of the former IT organization that remain with the client. They form an interfacing organization that steers the sourcing environment (Gewald and Helbig 2006; Goldberg et al. 2014b). Hence, the models differ with respect to the operational service integration function resulting in five basic organizational typologies. We will introduce these in the fourth section (see section Evaluating Basic Organizational Models for Service Integration), where we also discuss their advantages and relevancy for practice. An important aspect in multi-sourcing service integration is IT service management (Schermann et al., 2006). Goldberg et al. (2015), however, argue that traditional IT service management needs to evolve into end-to-end service management to meet the requirements of service integration. A vast body of practitioner-oriented literature is available on IT service management and the ITIL framework. Work specifically addressing service integration is still relatively sparse, though. Holland (2015) provides an initial overview of ITIL in the light of service integration. While IT service management and ITIL are mainly focused on required IT service management processes, our work investigates the organizational structures required for service integration. Research Method To investigate and contribute to detailed organizational models and task allocations, we performed an explorative research study. Building on a literature review, we carried out a series of qualitative expert interviews before collecting quantitative results based on a questionnaire (see Figure 1). To reduce subjectivity and mistakes, two researchers performed the steps outlined below independently. Literature review Literature Review Expert Interviews Questionnaire Figure 1. Overview of the research method To capture the state of the art of research, we performed a literature review following the methodology outlined by Webster and Watson (2002). In Google Scholar, we identified 167 articles and papers in an initial keyword-based 1 full-text search. In an abstract based perusal, we selected a total of 18 papers relevant to our research questions. We selected peer-reviewed journal and conference papers that deal with interdependent multi-sourcing or, more specifically, service integration. Originating from these, we identified 8 additional papers in a systematic forward and backward search following the same selection criteria. The set of 26 papers is the basis for framing our research and foundations. We selected a subset of eleven papers to identify service integration tasks and decision drivers. The selected papers discuss service 1 Two independent searches were based on combinations and variations (synonyms) of the following two sets of keywords: (1) multi-sourcing and service integration, and (2) multi-sourcing, IT outsourcing, Twenty-first Americas Conference on Information Systems, Puerto Rico,

4 integration and multi-sourcing with interdependent services. We analyzed each paper in a qualitative content analysis (see subsection Qualitative Content Analysis). 2 Qualitative study based on expert interviews Based on the initial insights from literature, we conducted a series of twelve expert interviews. They involved client representatives, consultants and service providers (see Table 1). Upfront, we defined a profile to select interviewees. It required that interviewees had at least three years of experience with multi-sourcing or service integration. They were required to work in a service integration function, consulting clients in this matter, or working for a provider at the interface with a multi-sourcing client. # Interviewee Position Company #1 Management Consultant Global Management and IT Consultancy #2 IT Service Integration and Management Consultant Global Management and IT Consultancy #3 IT Management Consultant Global Management and IT Consultancy #4 IT Sourcing Manager Service Integrator for Global Industrial Company #5 Sourcing Architect Service Integrator for Global Financial Services Company #6 CIO International Insurance Company #7 Sourcing Manager International Insurance Company #8 Project Manager, Global SIAM Project Global Pharmaceutical Company #9 Service Engineer Global Outsourcing Provider #10 IT Service Management Consultant Global Outsourcing Provider #11 Lead Client Sourcing Architect Global Outsourcing Provider #12 Project Executive (SIAM Project) Global Outsourcing Provider Table 1. Overview of the interview participants The interviews followed the method proposed by Gläser and Laudel (2010). They lasted between 45 and 90 minutes and were based on an interview schedule. In the mostly open-ended questions, interviewees described their experiences with multi-sourcing particularly focusing on organizational aspects, challenges and success factors. At the end of each interview, we discussed the five organizational models extracted from literature. The interviewees rated the models relevancy, pointed out advantages and disadvantages, and outlined relevant decision drivers (e.g. service integration capabilities, complexity of integration). We digitally recorded, transcribed, and then carefully analyzed each interview (see subsection Qualitative Content Analysis). Qualitative Content Analysis In a qualitative content analysis (Mayring, 2008), we analyzed both identified literature and interview transcripts. Qualitative content analysis is a widely applied technique to systematically derive intersubjectively verifiable insights from qualitative texts. It is particularly helpful for extracting categories and items. From the above sources, we extract advantages and disadvantages, service integration tasks and decision drivers. 2 A list of the 26 papers and the eleven papers analyzed in the qualitative content analysis is provided in the appendix. Twenty-first Americas Conference on Information Systems, Puerto Rico,

5 We iteratively analyzed each paper or transcript extracting paraphrases into a master table. We continuously clustered and combined paraphrases until we reached a theoretical saturation. That is, we were unable to further combine items without losing relevant information. To achieve inter-subjectiveverifiability and to ensure consistency, two researchers performed the steps outlined independently. Quantitative study based on questionnaire To further develop our findings and provide quantifiable results, we performed a quantitative study based on a questionnaire. Based on the questionnaire results, we rank decision drivers and develop typical service integration task allocations. In total, we received 18 completed questionnaires, which corresponds to a response rate of 31% based on a sample population of 59. Participants were client representatives, IT consultants and employees of external service integrators or service providers (see Table 2). Most data was collected with closed, prompted and pre-coded questions. In additional open-ended questions, the participants were able to provide further details and information (Oppenheim, 1992). Amount Relative Amount Type 4 22% Client-side representatives 7 39% IT consultants 4 <22% Provider-side representatives Table 2. Overview of the questionnaire participants The questionnaire consisted of two sections. First, the participants rated decision drivers identified in our qualitative study that determine which service integration tasks a client is likely to outsource. Based on these, we determine the importance of the decision drivers. In the second section, participants assigned typical service integration tasks to the client and a potential service integrator. We asked our participants whether activities should be best retained or be delegated to a guardian vendor or an independent service integrator, respectively. On average, their assignments reflect typical allocations of tasks. We asked for definite answers. Participants, however, were able to select both the retained organization and the external entity in case the responsibility for a task should be shared (multiple selections possible). Evaluating Basic Organizational Models for Service Integration In this section, we discuss relevant organizational models for service integration in multi-sourcing from the client perspective. Based on insights from our qualitative study, we identify the three most relevant models and discuss their key advantages and disadvantages. Academic research describes five theoretical typologies for organizing service integration (Goldberg 2014a)(also see Related Work). They are presented in Figure 2. The typologies differentiate a strategic and an operational service integration function. While clients retained organizations should always perform the strategic function, the models differ with respect to the operational function. Roles 1 CIO 2 Sourcing Managers 1 SIAM Project Manager 2 Management Consultants 4 IT Management Consultants 1 IT Service Management Consultant 2 Sourcing Architects 2 Project Executives 2 IT Sourcing Managers 3 17% Service integrator representatives 1 Sourcing Architect Twenty-first Americas Conference on Information Systems, Puerto Rico,

6 2.1 Retained Organization Retained IT Organization Service Integration Internal Services External Provider External Provider Guardian Vendor Retained IT Organization Internal Services External Provider External Provider... Service Integrator 2.3 Independent Service Integrator 2.4 Prime Provider Retained IT Organization Service Integration Retained IT Organization Internal Services External Provider External Provider... External Provider Prime Provider Service Integrator External Provider 2.5 Shared Service Integration Retained IT Organization Service Integrator Internal Services External Provider External Provider... Service Integrator Service Integrator Service Integrator Figure 2. Organizational models for service integration (based on Goldberg et al. 2014a) First of all, retained organizations can perform operational service integration themselves (Figure 2.1). For most clients, this is the first step into service integration for continuously growing multi-sourcing portfolios (Bhattacharya et al. 2012; Goldberg et al. 2014a). Alternatively, service integration can be partly delegated to an external entity. In the guardian vendor model, one of the existing outsourcing providers is taking over integration in addition to service delivery (e.g. Bapna et al., 2010; Goldfayn, 2006) (Figure 2.2). As a third option, a specialized independent service integrator is performing service integration tasks without assuming additional delivery responsibilities (e.g. Longwood & Heiden, 2012; Longwood, 2010)(Figure 2.3). Or, service integration can be performed by a prime provider (e.g. Plugge & Janssen, 2009)(Figure 2.4). From the client perspective, this model behaves similar to single-sourcing, as the prime provider is responsible for contracting and integrating subcontractors. Last, service integration can be shared among clients and their providers (e.g. Bhattacharya et al., 2012)(Figure 2.5). Providers are made responsible for directly negotiating delivery agreements with one another. Organizational Model Interviewee # #1 #2 #3 #4 #5 #6 #7 #8 Table 3. Relevancy of organizational models for service integration To identify the most relevant models from for clients, we discussed them with service integration experts. To evaluate the relevancy, we focused on the interviews with client representatives and consultants. We found that three models are relevant from a client perspective in a multi-sourcing context (see Table 3). Total Retained Organization Model 8 100% Guardian Vendor Model 5 63% Independent Service Integrator Model 4 50% Prime Provider Model 2 25% Shared Service Integration Model 1 13% Twenty-first Americas Conference on Information Systems, Puerto Rico,

7 First, the client can choose to keep service integration with the retained organization which all experts consider relevant. Or, they can chose to outsource service integration to a guardian vendor or to an independent service integrator 63% or 50% confirmed their relevancy. On the contrary, the prime provider model does resemble single-sourcing from the client perspective and is considered as outdated. The shared service integration model is perceived as too complex and associated with unclear governance structures. Accordingly, most experts consider both models irrelevant for clients. Based on these insights, we decided to focus on the three aforementioned models and excluded the latter two from our evaluations. Retained Organization Independent Service Integrator Guardian Vendor Advantages Clear distribution of responsibilities Clear governance and accountability Full contractual authority All control with client Low overhead costs for small provider portfolios Existing tools customized to client environment Objectivity and neutrality (low opportunism) Specialized skills and tools Predefined tools and processes Economies of scale and scope Risk sharing Part of delivery increases endto-end understanding Good SI knowledge and skills based on experience Existing relationship and trust Economies of scale and scope Risk sharing Knows client environment and culture Disadvantages Missing specialized know how No economies of scale / scope Continuous investment in tool maintenance Poor access to leading edge tools and technology No contractual authority Additional loss of client competencies Unclear distribution of responsibilities Difficulties in managing client internal providers New to client environment and culture Not part of delivery reduces end-to-end understanding Increased complexity and overhead costs No contractual authority Additional loss of client competencies Unclear distribution of responsibilities Difficulties in managing client internal providers Potential hold-up and opportunistic behavior Potential resentment of other providers Table 4. Advantages and Disadvantages of the Proposed Models Each of the three selected models has specific advantages that we extracted from the interviews based on qualitative content analysis (see Table 4). Clients that decide to keep service integration in-house stay in full control of the integration process. In addition, the model provides a simple governance structure with clearly distributed accountabilities and responsibilities. Particularly for small provider portfolios, the model is associated with lower costs because of less coordination and communication overhead in comparison to the other two models. On the other hand, many clients lack required skills and technology for performing service integration. For large provider portfolios, costs may be higher as clients cannot realize economies of scale and scope. Handing over service integration tasks to an external party will often make specialized skills and technology available. Costs may be lower if economies of scale and scope can be realized. In addition, clients are able to partly share integration risks. On the other hand, clients are losing control over the integration process, the overall integration complexity is increased, and the governance model is less clear. Particularly, interviewees often mentioned the lack of formal contractual authority of the external party as an issue. Key advantage of an independent service integrator in comparison to a guardian vendor is their neutrality while the guardian vendor may opportunistically leverage their strategic position to the disadvantage of other providers. This neutrality comes at the cost of added complexity and a lack of endto-end delivery understanding. Twenty-first Americas Conference on Information Systems, Puerto Rico,

8 We expect that the client s individual evaluation of the advantages outlined above will influence the basic strategic direction for selecting a model. We will analyze the three models and the selection criteria in more detail in the next section based on service integration tasks. Developing Detailed Organizational Models for Service Integration In literature, the three relevant organizational models for service integration are defined as ideal types which are abstract models that represent theoretical extremes. Likely, they do not exist in these pure forms in reality but rather in hybrid forms (Doty and Glick 1994; Goldberg et al. 2014a). In this section, we introduce a set of service integration tasks derived from interviews to detail the organizational models. They enable us to represent and analyze more fine granular, realistic hybrid models and subtypes. After discussing drivers that influence clients to retain or outsource tasks, we present and discuss typical allocations based on our questionnaire results. Organizational service integration tasks The service integration organization consists of a strategic and an operational function. While the strategic function should remain with the client, operational tasks can be retained or outsourced to an external entity (Goldberg et al. 2014a). In the latter case, service integration tasks need to be allocated between retained organization and external service integrator. Which tasks are delegated depends on the client s individual situation and the desired underlying model. To enable more detailed model definitions, we identify important service integration tasks. They are extracted from literature and our qualitative study based on a qualitative content analysis (see section Research Method). In total, we identify and distinguish 15 basic service integration tasks. An overview of these tasks including descriptions, and the number of supporting literature as well as interviews are provided Table 5. For example, multi-sourcing services need be selected and actively managed as a service portfolio (Service Portfolio Management). We found evidence for this task in three papers and coded paraphrases in 75% of the expert interviews 3. Or, providers need to be selected (Provider Portfolio Management) and managed according to their contracts (Contract Management) which is mentioned by 67% or 42% of the interviewees, respectively. An important aspect in our interviews was the development, management and measurement of end-to-end service levels. Approaches to define complex service levels are also discussed in literature (e.g. Kieninger et al and Kieninger et al. 2013). According to our interviewees, the tasks outlined above are required to achieve integrated end-to-end services. Drivers influencing the decision to outsource service integrations tasks The last subsection outlined specific service integration tasks that clients can perform themselves or delegate to an external entity. Clients will outsource service integration in different degrees (Goldberg et al. 2014a). The question at hand is which tasks clients will likely retain or outsource. The results of our qualitative study show that they are assessing advantages and disadvantages of the service integration models differently depending on the individual circumstances (for example if they have the required in-house capabilities). In the following, we elaborate on the rationale for outsourcing tasks and, thus, for selecting specific models. We will present and discuss important decision drivers influencing this decision. The drivers are extracted based on qualitative content analysis from both selected literature and interviews (see section Research Method). The drivers influence whether a specific task is retained or outsourced. Consequently, they also influence the choice of one model over another. 3 A mapping of literature and experts to specific service integration tasks is provided in the appendix. Twenty-first Americas Conference on Information Systems, Puerto Rico,

9 Service Integration Tasks Service Portfolio Management Description Determine the service portfolio in alignment to business requirements. Table 5. List of Service Integration Tasks Literature paraphrases 3 Interview paraphrases (75%) Demand Management Elicit and understand business requirements. 4 8 (67%) Financial Management Manage financials, costs and charge back. 4 6 (50%) Service Catalog Management Define, structure and publish the IT service catalog. 2 6 (50%) Transformation and Organizational Change Provider Portfolio Management Contract and Provider Management Relationship Management SLA Management Operational IT Service Management OLA Management Contract-Monitoring Manage the overall IT transformation as well as organizational and cultural changes. Evaluate and select providers that fit the service requirements. Negotiate and harmonize outsourcing contracts an manage providers, accordingly. Establish and foster strong relationships with the business and across providers. Define, negotiate and manage end-to-end service levels across all providers. Implement and perform (operational) ITSM focusing on cross-provider processes. Define and manage agreements between the internal and external IT service providers. Monitor performance of providers based on contracts, SLAs and OLAs. 3 8 (67%) 3 8 (67%) 8 10 (83%) 9 12 (100%) 4 10 (83%) 4 12 (100%) 4 4 (33%) 3 5 (42%) Service Measurement Measure end-to-end performance of services. 3 9 (75%) Transition Management Service Management Tool Operation Manage service transitions and handover to new providers. Operate, maintain and improve the service management tool landscape. 3 6 (50%) 3 10 (83%) We assessed the importance of decision drivers as part of our questionnaire (see section Research Method). Participants selected which drivers typically determine the clients decision (see Table 6). The three most influential drivers are service integration capabilities, expected cost, and tool capabilities. That is, clients with strong in-house capabilities in a service integration task will most likely keep it within their retained organization; or, with the prospect of increased overall integration capabilities, choose an external party. The second driver regards cost of service integration. Often, the primary goal of outsourcing is the reduction of operating costs (Lacity et al., 2009). The same seems to be true for the service integration decision and clients may choose to outsource inefficient areas. Third, the availability of an appropriate tool basis seems to influence the decision to retain or outsource service integration activities. Clients lacking specialized tools and tool knowledge for service integration are likely to outsource relevant activities. For example, this could be the case for tool operation and service measurement. Other factors are perceived as less important. Yet, they may be essential in specific situations. For example, interviewees mentioned that legal requirements might force clients to keep control over the integration process in-house in specific industries. In addition, the decision drivers may be influenced by overall contingency factors that we did not consider in our study. For example, larger company sizes could increase integration complexity and make larger companies more likely to outsource service integration. Twenty-first Americas Conference on Information Systems, Puerto Rico,

10 Decision Driver Table 6. Key Decision Drivers Influencing the Decision (N=18) In conclusion, service integration models are perceived more or less appropriate depending on specific factors. Clients need to evaluate whether it is more feasible to retain or outsource specific service integration tasks. The ranking provides a starting point for understanding the selection process. To better understand which tasks clients are typically delegating, we provide an overview of different allocations in the next subsection. Typical allocations of service integration tasks Frequency 1 Service Integration Capabilities (in performing service integration task) 13 2 Cost (to perform service integration task) 11 3 Tool Capabilities (required to perform service integration task) 7 4 Strategic Importance of IT (in client company) 4 5 Service Integration Quality (aspiration to increase quality) 4 6 Level of control (over integration process) 3 7 Complexity (of integration and service landscape) 3 8 Risk (aim to share risks / risk adversity) 2 9 Legal Requirements (require to retain service integration task) 2 Clients can contract an independent service integrator or a guardian vendor to take over service integration responsibilities. If they decide to outsource tasks, clients need to decide how to allocate them. Based on questionnaire results, we determine two typical allocations in the following. We asked our participants whether tasks should be best retained or be delegated to a guardian vendor or an independent service integrator, respectively. We asked for definite answers wherever possible. Participants, however, were able to select both the retained organization and the external entity in case the responsibility for a task should be shared (multiple selections possible) (see also section Research Method). In general, most integration tasks are mainly seen under the responsibility of the client (see Table 7). In a model with an independent service integrator, eleven tasks are typically assigned to the retained organization, including strategic and business-facing topics like demand management or financial management. The majority of our participants assigned six tasks to an independent service integrator. For example, contract monitoring and tool operation are typically delegated. Interestingly, service portfolio management is the only task that all participants would retain without external involvement. All other tasks are at least partly outsourced to or shared with an external entity. There is mostly unison regarding the distribution of responsibilities in the independent service integrator and the guardian vendor model. For example, portfolio management, demand management and financial management are ranked almost identically. However, there are also significant differences between the two models. Obviously, most participants would delegate contract monitoring and transition management to an independent service integrator but not to a guardian vendor. We see the same difference for provider portfolio management: 56% would involve an independent service integrator while only 11% would delegate it to a guardian vendor. Similar differences exist for Service Catalog Management and Contract Monitoring. Based on interview comments, we expect that the neutrality of independent service integrators is the reason for these differences. Clients seem to be less willing to hand over more strategic tasks to a provider that could use them opportunistically. Twenty-first Americas Conference on Information Systems, Puerto Rico,

11 Service Integration Task Table 7. Typical Task Allocations Between Retained Organization and External Service Integrator (N=18, multiple choices possible) In conclusion, the distribution of responsibilities not only depends on the specific contingency factors but also on the selected service integration model. While few tasks should always be kept in-house, many tasks can be shared with or handed over to an external service integrator. Due to potential opportunistic behavior, clients typically delegate fewer responsibilities to a guardian vendor than to an independent service integrator. Conclusion Independent Service Integrator Retained Organization Ind. Service Integrator Retained Organization Guardian Vendor Guardian Vendor # % # % # % # % Service Portfolio Management % 0 0% % 0 0% Demand Management % 2 11% % 1 6% Financial Management % 3 17% % 2 11% Service Catalog Management 17 94% 4 22% 17 94% 5 28% Transformation and Org. Change 17 94% 11 61% 17 94% 10 56% Provider Portfolio Management 16 89% <-> 10 56% 16 89% <-> 2 11% Contract and Provider Management 15 83% 10 56% 16 89% 4 22% Relationship Management 15 88% 11 65% % 5 29% SLA Management 12 67% 10 56% 13 72% 9 50% (Operational) Service Management 11 61% 14 78% 10 56% 14 78% OLA Management 9 50% 13 72% 8 44% 14 78% Contract-Monitoring 8 44% 15 83% 16 89% 9 50% Service Measurement 8 44% 15 83% 8 44% 15 83% Transition Management 7 39% 15 83% 16 89% 5 28% Service Management Tool operation 3 17% 17 94% 4 22% % In this paper, we present the findings of our explorative study on multi-sourcing service integration. We performed a literature analysis, a series of twelve expert interviews and a quantitative study based on 18 completed questionnaires. Our findings provide a basis for future research in this important area. Additionally, clients can use our insights to better understand service integration, to guide their organizational design decisions, and to improve their service integration solutions. Our results show that the decisions regarding the service integration organization are highly dependent on the specific circumstances. Based on study insights, we identify three organizational models that are most relevant in multi-sourcing from the client perspective. The models differ with regards to the operational service integration function while a strategic function always remains with the client. That is, clients may choose to retain all service integration responsibilities, or outsource them partly to a guardian vendor or an independent service integrator. To enable a detailed analysis, we introduce a set of 15 important service integration tasks. The decision to retain or externalize a task is influenced by various contingency factors and the underlying service integration model. The most important factors are the clients service integration capabilities, costs and tool capabilities. Many strategic service integration activities should stay with the client, while other activities can be shared or completely handed over. Clients seem to be less willing, however, to hand over more strategic tasks to a guardian vendor than to an independent service integrator. Twenty-first Americas Conference on Information Systems, Puerto Rico,

12 With regard to our work, a number of limitations need to be considered. First, a limitation of our study is that the number of 18 completed questionnaires is relatively small. Caution must be applied, as results may not be completely generalizable. The research topic, however, is new and we aim to develop fundamental insights. Together with the data from twelve expert interviews, the number is sufficient to derive a basis for future research and insights for practice. Another source for potential criticism is the selection of study participants. We investigate service integration from the client perspective. Although we involved client side representatives, many interview partners and questionnaire participants are either consultants, provider side representatives or service integrator employees. We expect, however, that the combination of these different sources provides particularly meaningful insights. It involves experience from all relevant parties involved in service integration. Nevertheless, we encourage additional studies to re-evaluate our findings with additional clients. Service integration is under-researched and we see several possibilities for future research. In this paper, we analyze three organizational models and basic task allocations. Future research should validate and confirm our findings. Particularly case study research and quantitative studies should be conducted. Although we identified important decision drivers, they are only loosely coupled with tasks and their allocations. It would be particularly interesting to analyze in depth which factors influence clients to retain or outsource a specific task. We expect that a combination of case studies and large quantitative studies could provide meaningful insights. Additionally, our overviews of task allocations provide a basis for detailed governance models. Future studies should try to combine organizational models and our insights to develop more detailed governance structures. Last, parallels are evident between our service integration tasks and IT service management processes. Interviewees frequently referred to ITIL processes, which led to a similar wording. Literature already outliend that IT service management needs to evolve into an end-to-end service management (Goldberg and Satzger 2015). Based on the service integration tasks presented in this work, researchers should start to investigate required adaptions to IT service management processes. The insights regarding model selection and task allocations provide clients with a starting point for designing their service organization model and organization. If clients choose to delegate activities to an external function, the exact responsibilities need to be defined. Our discussions of the advantages and decision drivers help them to select the model that fits their requirements best. During negotiations with the prospect service integrator, the list of tasks provided should enable a guided discussion. Clients should focus on the activities that are typically externalized. They should then evaluate which additional activities could be handed over and which activities should be kept in the authority of the retained organization. Particularly strategic tasks should not be externalized. Consciously selecting and systematically defining the service integration function should enable clients to better realize benefits from multi-sourcing and to build effective outsourcing solutions. Service integration is an important contributor to multi-sourcing success. REFERENCES Anderson, E. G., and Parker, G. G Integration of Global Knowledge Networks, Production and Operations Management (22:2), pp Arora, J., Sengupta, C., and Joshi, Y Service Integration and Management: An Idea Whose Time Has Come, Everest Group Research. Bapna, R., Barua, A., Mani, D., and Mehra, A Research Commentary-Cooperation, Coordination, and Governance in Multisourcing: An Agenda for Analytical and Empirical Research, Information Systems Research (21:4), pp Bhattacharya, S., Gupta, A., and Hasija, S Single Sourcing versus Multisourcing: The Role of Effort Interdependence, Metric-Outcome Misalignment, and Incentive Design, INSEAD Working Paper. Davy, A Sourcing of IT services - Industry Trends, Information Systems in Management (3:2), pp Doty, D. H., and Glick, W. H Typologies as a unique form of theory building: Toward improved understanding and modeling, Academy of Management Review (19:2), Academy of Management, pp Twenty-first Americas Conference on Information Systems, Puerto Rico,

13 Gewald, H., and Helbig, K A Governance Model for Managing Outsourcing Partnerships: A View from Practice, in Proceedings of the 39th Annual Hawaii International Conference on System Sciences, pp Gläser, J., and Laudel, G Experteninterviews und qualitative Inhaltsanalyse als Instrumente rekonstruierender Untersuchungen (4 th ed.), Wiesbaden: VS, Verlag für Sozialwissenschaften. Goldberg, M., Kieninger, A., and Fromm, H Organizational Models for the Multi-sourcing Service Integration and Management Function, in Proceedings of the 16th Conference on Business Informatics, Geneva, pp Goldberg, M., Kieninger, A., Satzger, G., and Fromm, H Transition and Delivery Challenges of Retained Organizations in IT Outsourcing, in Proceedings of 5th International Conference on Exploring Service Science, Geneva, pp Goldberg, M., Satzger, G., A Capability Framework for IT Service Integration and Management in Multi-Sourcing, in Proceedings of the 23rd European Conference on Information Systems, Münster. Goldfayn, E Organization of R&D With Two Agents and Principal, Bonn: University of Bonn, pp Gulati, R., Lawrence, P. R., and Puranam, P Adaptation in vertical relationships: Beyond incentive conflict, Strategic Management Journal (26:5), pp Holland, K "An Introduction to Service Integration and Management and ITIL," White Paper Axelos. Jin,., Kotlarsky, J., and Oshri, I Towards Understanding Knowledge Integration in Multi- Sourcing Engagements, in Information Systems Outsourcing, R. Hirschheim, A. Heinzl, and J. Dibbern (eds.), Berlin: Springer, pp Kieninger, A., Westernhagen, J., and Satzger, G The economics of service level engineering, in Proceedings of the 44th Annual Hawaii International Conference on System Sciences, Kauai, HI, pp Kieninger, A., Kimbrough, S., Satzger, G., Schmitz, B., Straeten, D "Leveraging service incident analytics to determine cost-optimal service offers," in Proceedings of the 11th International Conference on Wirtschaftsinformatik, Leipzig, pp Lacity, M. C., Khan, S. A., and Willcocks, L. P A review of the IT outsourcing literature: insights for practice, The Journal of Strategic Information Systems (18:3), pp Lawrence, P. R., and Lorsch, J. W Differentiation and Integration in Complex Organizations, Administrative science quarterly (12:1), pp Longwood, J The Role of the Multisourcing Service Integrator in Delivering End-to-End Outsourced Services, Gartner, Inc. Longwood, J., and Heiden, G Should You Insource or Outsource the Multisourcing Service Integrator Role?, Gartner, Inc. Mayring, P Qualitative Inhaltsanalyse. Grundlagen und Techniken, Weinheim: Beltz. Oppenheim, A. N Questionnaire design, interviewing, and attitude measurement, London: Continuum. Plugge, A., and Janssen, M Managing change in IT outsourcing arrangements: an offshore service provider perspective on adaptability, Strategic Outsourcing: An International Journal (2:3), Emerald Group Publishing, pp Plugge, A., and Janssen, M Governance of Multivendor Outsourcing Arrangements: A Coordination and Resource Dependency View, in Governing Sourcing Relationships, J. Kotlarsky, I. Oshri, and L. P. Willcocks (eds.), Berlin: Springer, pp Satzger, G., and Kieninger, A Risk-reward sharing in IT service contracts - A service system view, in Proceedings of the 44th Annual Hawaii International Conference on System Sciences, Kauai, HI, pp Schermann, M., Böhmann, T., and Krcmar, H Integration of IT services: Towards a pattern-based approach for eliciting service integration requirements, in Proceedings of the 12th Americas Conference on Information Systems, Acapulco, p Urbach, N., and Würz, T How to Steer the IT Outsourcing Provider, Business & Information Systems Engineering (4:5), SP Gabler, pp Webster, J., and Watson, R. T Analyzing the past to prepare the future, MIS Quarterly (26:2), pp. xiii xxiii. Wiener, M., and Saunders, C. S Forced coopetition in IT multi-sourcing, Journal of Strategic Information Systems (23:3), pp Twenty-first Americas Conference on Information Systems, Puerto Rico,

14 Appendix Mapping of literature to service integration tasks Service Integration Task Literature Bapna et al. (2010) Beulen and Tiwari (2010) Goldberg et al. (2014) Goldberg et al. (2014b) Ilmo and Nahar (2010) Jin et al. (2014) Plugge and Janssen (2014) Plugge et al. (2013) Rajamaki and Vuorinen (2013) Schermann et al (2006) Wiener and Saunders (2014) Service Portfolio Management Demand Management Financial Management Service Catalog Management Transformation and Org. Change Provider Portfolio Management Contract and Provider Management Relationship Management SLA Management Operational IT Service Management OLA Management Contract-Monitoring Service Measurement Transition Management Service Management Tool Operation Appendix 1.1. Mapping of service integration tasks to literature Twenty-first Americas Conference on Information Systems, Puerto Rico,

15 Mapping of interview experts to service integration tasks The following table shows whether paraphrase indicating a service integration task was coded for a specific interviewee. Interviewee # Service Integration Task #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 Total Service Portfolio Management 9 75% Demand Management 8 67% Financial Management 6 50% Service Catalog Management 6 50% Transformation and Org. Change 8 67% Provider Portfolio Management 8 67% Contract and Provider Management 10 Relationship Management % SLA Management 10 83% Operational IT Service Management 12 OLA Management 4 33% Contract-Monitoring 5 42% Service Measurement 9 75% Transition Management 6 50% Service Management Tool Operation 10 Appendix 1.2. Mapping of interviewee paraphrases to service integration tasks 83% 100% 83% Twenty-first Americas Conference on Information Systems, Puerto Rico,