Finding the next edge in service innovation: a complex network analysis

Zhen Zhu and Dmitry Zinoviev

Introduction

Network theory has been recognized as a valuable tool, used in research across the natural and social sciences. It offers explanations for social phenomena in a myriad of disciplines, ranging from psycholog)' to economics, communication to political science, and other fields (Borgatti, Mehra, Brass, & Labianca, 2009). One particular type of network — a semantic network that represents relationships among concepts within a knowledge framework - has emerged in recent years in business and management fields. As a potent tool for literature review and knowledge categorization, it often generates novel insights for understanding the trajectories of past knowledge development and reveals directions for future research (Kovacs, 2010).

In the services field, the topic of service innovation (SI) attracts substantial interest among both service researchers and industrial practitioners. After more than four decades of conceptual evolution, SI has entered a multidimensional phase characterized by rapidly expanding components and scope (Carlborg, Kindstrom, & Kowalkowski, 2014). Meanwhile, the conceptualization of SI, including its domains and its involvement with other concepts in services, remains relatively understudied compared to goods innovation (Ostrom, Parasuraman, Bowen, Patricio, & Voss, 2015). Following the call for a systematic review of the existing knowledge base (Gallouj & Savona, 2009) and inspired by other recent review works on service innovation (e.g., Antons & Breidbach, 2018; Witell, Snyder, Gustafsson, Fombelle, & Kristensson,

2016), this chapter achieves three research purposes. First, it adopts the network analysis method to map the structural status of the SI mindscape in order to detect its inner structure and subdomains. Second, it aims to clarify understanding of the constructs of each subdomain by highlighting critical theoretical concepts, commonly employed research methods, and highly related industrial sectors. Last, this chapter identifies the connections and ‘structural holes’ among the subdomains (where one node is connected with two other disconnected nodes in the network), revealing potential combinative frontiers for future research in SI.

Methodologically, this chapter follows the guide of network theory (Borgatti & Halgin, 2011; Newman, 2010) in adopting the term ‘vector models’ from automated network analysis to plot business scholars’ interests and foci in SI studies. An output of a semantic analysis is a scale-free complex network that represents the concepts ‘communities’ (also termed clusters or domains) of the SI field. This is expected to elaborate and clarify the meaning and scope of SI.

The impetus of this chapter is to identify' and explicate the differences and distances among SI subdomains, to bring awareness to the need for understanding other subdomains beyond one’s own in terms of theoretical concepts, research methodologies, and industrial contexts. This is anticipated to set a foundation for more combinative creations in the future among researchers from different subdomains in the SI field. As a secondary contribution, this chapter illustrates the network analysis procedure and outcomes for conceptual mapping. Thus, the contribution is readily transferrable to other research topics.

This chapter is organized as follows: the network theory is introduced first (especially assumptions regarding use of structural holes for idea generation). This is followed by the details of the network analysis method, which is explained for exploring the structure of the SI field. Next, findings on the detected subdomains, including their features and interconnections, are delineated. This chapter closes with a conceptual framework on SI, which is proposed based on the findings and the implications for future research and contributions of this study.

Networks and network theory

A network consists of a set of actors or nodes that are directly or indirectly connected through a set of ties of a specified type (such as friendship or similarity) (Borgatti & Halgin,

2011). The pattern of ties in a network results in a particular structure, and nodes occupy positions within this network. Much of the theoretical relevance of network analysis stems from characterizing network structures (e.g., inner communities) and node positions (e.g., centrality) and relating these to group and node outcomes. A complex network featuring a heavy or flat tail in the distribution curve for the number of edges per node (i.e., the frequency of ties connecting to each node decreases quickly), a high clustering coefficient, and a clear community structure can reflect real-world complexity, such as in social networks.

A fundamental premise of network theory is its emphasis on structure, including structure shape, position, and structural environment of the nodes in the network. Whereas traditional social studies investigate characteristics of an actor, be it a member of a social network or a theoretical concept in a certain literature, as a function of other characteristics of the same actor (e.g., income as a result of education and gender), network research focuses on one’s influence (from and to) and similarity with other actors, in other words, one’s position in the network. Common outcomes of network analysis include graph-theoretic features of a network (such as modularity) and overall distribution of ties (such as density and weak ties).

Structural properties of a network can be analyzed by various network measures and metrics, among which centrality, ‘betweenness,’ and clusters are the most basic and also typical in network research (Newman, 2010, p. 9). Therefore, these are explained here. Centrality quantifies how important each node is in a networked system. A simple but useful measure of centrality is degree, indicating the number of edges attached to a node. Nodes with unusually high centrality or degree can be marked as hubs, which can exercise disproportionate effect on others despite being few in number. In conceptual maps, such hubs play central roles, often forming significant themes in concept clusters. Another special measure of centrality is betweenness; this term refers to the extent to which a node lies on paths between other nodes. Nodes of high betweenness may have considerable influence within an overall network, owing to their unique position in bridging otherwise disparate nodes or groups. Without such bridging nodes, structural holes appear instead. Clusters or communities are tightly knit groups within a larger, looser network. For semantic networks, the way a network breaks down into clusters can reveal levels and concepts of a knowledge field that are otherwise hard to conjecture.

Recently, network analysis in the social sciences has developed into a coherent and generalized research paradigm organized around four core features — a focus on the ties that link individual actors (rather than the on attributes of the actors per se), and the heavy use of systematic empirical data, graphic imagery, and mathematical and/or computational models (Freeman, 2004). This chapter engages all four of these features in later sections.

One prominent network theory is concerned with missing ties or structural holes within a network. By definition, a structural hole exists when a focal actor links to a pair of disconnected actors. According to structural hole theory, networks are rift with structural holes, which expose an actor to novel communities, diverse experiences, unique resources, varying preferences, and multiple thought worlds, providing superior opportunities (Burt, 2004). The association between novel ideas and structural holes is crucial to the social capital of brokerage, and the actors who possess the feature of structural holes are considered to be in critical locations to identify future innovations. Meanwhile, a bridge built to mend a structural hole often starts weak, lacking the shared viewpoints and methods to realize or germinate novel ideas. Following these guidelines, this chapter explores major structural holes in the existing network of concepts related to SI and suggests potential combinative ideas for future service research.

In contrast with structural holes, dense networks where members of a community are tightly connected are recognized to have the advantage of carrying out collaborative actions through frequent communications and strong ties that often entail trust, norms of cooperation, and effective exchange of complex knowledge. Such dense clusters are defining features of networks where ties are denser within than between the clusters. Close bonding among members sets the boundary or scope of a community, differentiating the in-group members from the outsiders (Borgatti et ah, 2009). For a conceptual network, nodes within the same cluster, be it opinions, keywords, or information, are more homogeneous, forming converging themes. On the negative side, dense networks may not offer novel concepts or ideas, because members have already been widely assimilated and information fully shared (Obstfeld, 2005). This chapter detects dense conceptual clusters and identifies primary themes of each cluster within the SI domain.

Semantic network analysis method

For conceptual mapping, such as in linguistics literature, two major approaches to semantic analysis are identified as broadly defined concepts: ontologies and term vector models (Manning, Raghavan, & Schiitze, 2008). An ontology is a form of knowledge representation, usually within a certain domain, that uses a shared vocabulary (Gruber, 1993). An overview of ontology construction techniques is given by Perez-Corona et al. (2012). All of these techniques (e.g., free list, bounded list, and double entrance matrix) require extensive human involvement. On the contrary, term vector models, also known as “vector space models” (Salton, Wong, & Yang, 1975) rely on the availability of previously defined keywords or subject tags that can be automatically harvested and compared for similarity by customized software and programming. An output of a semantic analysis is a small-world semantic network (Steyvers & Tenenbaum, 2005; Zhang, Luo, Xuan, Chen, & Xu, 2014) that represents the term associated with the original concept (in this chapter, ‘service innovation’) and therefore elaborates and clarifies its meaning. Semantic network analysis (SNA) has also been used to examine concepts, such as differences in national cultures or the evolution of the communication discipline.

Comparing SNA and citation analysis

It is noteworthy that SNA differs from citation analysis in that the latter features the connections among authors (agents) or articles (artifacts), but not associations among the subject tags (the terms). (Please see Figure 9.1.) Although citation analysis has been widely used to track knowledge evolution in research fields, it bears the limitation of potentially overestimating the impact of certain popular or methodological citations that have remote or no conceptual connections to the concept of interest. In addition, because academic journal articles often include dozens of citations each, the network among articles can be unnecessarily dense, reducing the clarity of key association paths. In contrast, each academic article has a limited number of subject terms assigned by the database or journal editors, so the subject tag network has the potential to offer a crisper map of interconnectedness among the concepts. Furthermore, because each subject term is expected to convey the most important ideas in the article, a subject tag network is likely to reveal links among the most substantive meanings around the central concept, enhancing their clarity.

Comparing SNA and content analysis

SNA and content analysis differ both in terms of human labor involvement and processing approach. A typical meta-analysis involves an intensive and laborious content analysis requiring multiple reviewers to independently and manually comb through original verbatims and sort them into often predetermined categories. In contrast, automated SNA does not always require a priori definitions of the categories to be used. Instead, this method allows the use of the natural language of the participants to determine their shared meaning. As a result, threats to reliability or validity due to the coders or categories are not an issue in SNA (Rice & Danowski, 1993).

Another critical difference between SNA and content analysis lies in the object of analysis. First, consider content analysis. Meta-analyses often focus on identifying and classifying the meaning and scope of a particular theoretical concept per se. For instance, Witell et al. (2016)

Figure 9.1 Data structure.

identified 84 definitions of SI in 1,301 articles published between 1979 and 2014. Carlborg et al. (2014) reviewed the progression of the SI concept between 1986 and 2010, based on 128 articles published in leading marketing and innovation journals. In contrast, the small-world semantic network represents a conceptual ecosystem in which the central concept resides (Witell et al., 2016). Instead of investigating “what is SI?” directly, SNA investigates “what are the concepts that are studied together with SI within the literature?” or “what concepts are similar to SI?” In this sense, SNA is similar to social network analysis because both are able to provide structure to a network. Semantic networks are analyzed based on shared meaning, whereas social networks are based on communication partners; connections in SNA are formed by the use of overlapping concepts, whereas instances of social interaction constitute social links (Doerfel & Barnett, 1999).

This chapter employs the vector space modeling approach in forming the conceptual network of SI. The SNA approach is expected to augment or complement traditional literature review techniques by providing an alternative but holistic network view.

SNA Of SI

To explore the mental space or the mindscape related to ‘service innovation,’ this chapter applied semantic decomposition (Kovacs, 2010; Salton et al., 1975; Zinoviev, Stefanescu, Swenson, & Fireman, 2013) consisting of two phases: (1) construction of a semantic network associated with the concept in question on the collected corpus or body of information, and (2) extraction of semantic domains to describe the concept. Figure 9.2 depicts the framework of the analytical process used in the chapter. The analysis starts by acquiring Si-related data from academic journal articles; this is followed by selecting analysis terms, calculating similarities between the terms, and clustering the terms based on their similarity into service domains. Last, the topicality is extracted, including key concepts, methods, and related service industries for each domain of SI.

Data sources and acquisition

The importance of SI has been embraced by a broad spectrum of business disciplines, including financial services, management, information systems, hospitality and tourism, public administration, etc. To gather data, the Elton B. Stephens Co. (EBSCO) database Business Source Complete was selected as appropriate, owing to its extensive content and audience coverage that matches with the broad SI theme. For instance, EBSCO offers more than 375 full-text and secondary research databases and provides access to audiences at academic institutions, schools, public libraries, hospitals and medical institutions, corporations, associations, and government institutions. In addition, EBSCO has been previously used as a search engine in social network literature (e.g., Gondal, 2011), which increases confidence in using this data source for SNA.

Figure 9.2 Framework of analytical steps.

Search term selection

The EBSCO database facilitates searches of articles by either keywords or subject terms, instead of only by author names, article titles, or full text (so it is less limited than the Thomson Reuters Web of Science). The keywords are suggested by authors, whereas the subject terms are created and applied by editors, professional lexicographers, and subject specialists to fit the scope of the database. This research hypothesizes that subject terms are more objective and authoritative by nature than are keywords, because the latter are likely to be employed as a personal statement of research interest or a marketing tactic to show association to trendy topics and thereby attract readership. Thus, the present analysis uses the subject terms from the selected articles to form semantic networks of SI.

The dataset is built from articles published between 2003 and 2018, retrievable by searching the subject terms “service innovation” or “service” + “innovation.” The corpus includes original terms from 2,341 Si-related articles published in 786 scholarly journals, resulting in 4,318 unique subject terms. The number of subject terms per article averages seven (standard deviation = one). Among the included scholarly journals, Health Affairs published more than 60 articles related to SI; it ranked the highest on the ‘Top 25 Journals’ list, followed by Research Policy, and Journal of Business Research (see Figure 9.3a). At the concept level, the top five most frequently used subject terms — technological innovations, innovations in business, service establishment/merchant wholesalers, customer services, and service industries — each appeared more than 250 times in the overall dataset (see ‘Top 25 Subject Tenns’ in Figure 9.3b). To highlight the major clusters

Figure 9.3a Top 25 Journals Publishing in Service Innovation.

Figure 9.3b Top 25 Subject Terms Related to Service Innovation.

of the network and reduce unwanted noise, 668 tenns were retained, each of which appeared at least four times in the original tenn corpus and was significantly similar to at least one other term (with generalized similarity >.70) in the SNA.

Similarity analysis

Defining similarity among concepts of a semantic network is a critical concern in a large sample literature review. How and why concepts are perceived to be similar (or not) and how they are then categorized or classified has far-reaching impact on cognitive structures. In this context, a research approach to defining similarity among Si-related concepts shapes the conceptualization of the domains in the mindscape of SI. Previous works have primarily depicted similarity by frequency among terms that co-occur in the same articles (e.g., Antons & Breidbach, 2018). This chapter, however, adopts the recursive definition of generalized similarities proposed by Ко vacs (2010): two terms (such as subject tenns) are similar if they are associated with similar artifacts (such as articles); two artifacts are similar if they are associated with similar terms. This critical departure from previous works is obtained by relaxing a point connection (i.e., same article) to a generalized group or class connection (i.e., any article in a group of similar articles). This approach has been adopted in several recent works in both political science and medical research for SNA (Mu, Goulennas, Korkontzelos, & Ananiadou, 2016; Zhang & Zinoviev, 2018).

Figure 9.4 Visualized difference between Pearson correlation and generalized similarity.

The main advantage of the generalized similarity over other similarity measures (such as Pearson correlation or cosine distance) is that it allows comparison of tenns that are not directly associated with the same artifacts (and vice versa). As visualized in Figure 9.4, the Pearson correlation of two terms (T | and T?) is zero when they are associated only one-to-one with the artifacts (A| and A2), but the generalized similarity is possibly nonzero if A| and A2 are associated through other tenns (T3 and T4). Along the same line, on the artifact side, A3 and A_t are not correlated (using Pearson correlation) but they are likely to be similar when the similarity between T₃ and T₄ through their links with A, and A₂ are considered. Through a simulation and two empirical examples, Kovacs (2010) shows that in two-mode data or a bipartite network, the generalized similarity measure magnifies within-group similarities and between-group dissimilarities, illustrating the effectiveness of this approach in detecting otherwise less salient community or domain patterns, particular those muted by sparsity of data. This feature is particularly valuable in detecting emerging or early-stage trends within the overall semantic network.

The software Python is used to process the adjacency matrix of 2,341 artifacts (research articles) and 668 tenns (subject tenns), from which a two-mode dataset or a bipartite non- directional network is formed. The values of the generalized similarity measures in the matrix range from —1 to 1, where 1 denotes perfect similarity and —1 denotes perfect dissimilarity. Values around 0 indicate independence (or neutrality) between the actors. The detailed description and mathematical illustration of the analytical procedure can be found in Kovacs (2010).

Extracting term clusters

Graphic illustrations of the SNA in this chapter were produced using Gephi - a powerful software package for network visualization and analysis (Bastian, Heymann, & Jacomy,

2009). Overall network size, average degree centrality, density, number of clusters, average clustering coefficient, and modularity coefficient are calculated for the network (see Table 9.1). The outline of the semantic network is plotted in Figure 9.5. Subject terms in the SI field are found to be densely interconnected, with average degree centrality (a count of how many connected neighbors a node has) at 63.7 and average clustering coefficient of the network at a level of .634. Meanwhile, six term communities (clusters in different colors in Figure 9.5) identified from the network graph are considered conceptual domains of the SI

Table 9.1 Network measures of the SI domains

Figure 9.5 Interconnectedness among SI domains.

field (Ottenheimer, 2006). Overall modularity (which measures the ratio of edge densities within clusters to edge densities between clusters) reaches .62, indicating a community structure that shows relatively stronger within-domain connection while allowing certain conceptual similarity between domains of the same field.

Figure 9.6a Domain of innovation in business and service strategy.

Six SI domains

To label the clusters that emerged, the top 25 most frequent subject terms in each cluster were first listed to identify the common theme among them. The common theme of each cluster was then used to name the conceptual domain. This analysis reveals subject terms reflecting different aspects as compared to the published articles. Most domains (with the exception of Domain 6) consist of three types of subject terms - theoretical concepts, research and analytical methods, and industrial contexts. Thus, unique connections are found between SI concepts and the industrial backdrops in which they are most prominent and therefore also most frequently investigated. To manifest the theory—industry connection, several domain names were labeled with a combination of one critical theoretical theme and one salient practical context. In particular, the six conceptual domains of SI are labeled as: (1) Innovation in Business and Service Strategy, (2) Technolog)' Innovation and Customer Interfaces, (3) Digital Innovation and Financial Sendees, (4) Trans- fonnative Services and Public Services, (5) Service Ecosystem and Knowledge-Based Services, and (6) Agency Services in Cultural and Entertainment Industries. (Graphs of SI domains are provided in Figures 9.6a through 9.6f.) Conceptual properties of each domain are summarized in Table 9.2. Example works in each SI domain are provided in Table 9.3. In the next section, conceptual meanings, featured research contexts, and frequently used research methods of each domain are further elaborated to provide more nuanced understanding of the SI field.

Domain 1: innovation in business and service strategy

The largest conceptual domain in terms of unique subject terms (226), total count of subject terms (5299), and ties between terms (7791) within the SI network relates to both the innovation practices in business as well as service strategies. The first sub-theme in the domain includes ‘innovation in business,’ ‘innovation management,’ ‘new product development,’ and ‘R&D.’

Figure 9.6b Domain of technology innovation and customer interfaces.

1 77

Figure 9.6c Domain of digital innovations and financial services.

Figure 9.6d Domain of transformative services and public services.

Figure 9.6e Domain of service ecosystem and knowledge-based services.

Figure 9.6f Domain of agency services in cultural and entertainment industries.

Table 9.2 Key concepts, research methods, and industries within each SI domain

Table 9.3 Example works in each SI domain

(Continued)

Table 9.3 (Cont.)

These sub-themes address the processes and practices of goods and services design and development, mostly at the innovation project level. From these sub-themes, calls for service-specific innovation measures for service sectors emerged. The second sub-theme investigates strategic- level, service-related business model innovation, enlisting subject terms such as ‘business model,’ ‘strategic planning,’ ‘CRM,’ and ‘knowledge management’ (e.g., Drejer, 2004; Van Wijk et al., 2008). The third sub-theme consists of terms like “customer service,” which is seemingly unrelated to innovation in business but detected nevertheless in the domain. One possible explanation for this finding can be traced to the argument of Hipp and Grupp (2005) that human elements, including an employee’s individual-level experience and service skills, contribute to organizational knowledge and non-technological components in SI processes. Structurally, the customer service sub-theme bridges Domains 1 and 2 (Technology Innovation and Customer Interfaces), whereas the business model innovation sub-theme links to Domain 5 (Service Ecosystem and Knowledge-Based Services).

Although the themes of ‘innovation in business’ and ‘business model innovation’ are widely investigated across many service industries, several service sectors — consulting, busi- ness-to-business services, tourism, high tech, and logistics — are among the most typical contexts for empirical examinations in the domain. Meanwhile, industrial research (e.g., Drejer, 2004; Hipp & Grupp, 2005) and meta-analysis (e.g., Van Wijk et al., 2008) are among the frequently used research methods for this domain. In addition, regression analysis and conjoint analysis are commonly employed in new SI design, converging with their traditional roots in the new product development literature.

Domain 2: technology innovation and customer interfaces

The second conceptual domain in SI addresses technology innovations, primarily in traditional consumer-facing service encounters. This domain contains 114 unique subject terms that appear 2,937 times in the corpus and form 1,902 links among the nodes. Possibly because of the broad inclusion of service sectors in the domain, the overall density (.295) of the network is among the lowest across the six conceptual domains. Three themes coexist in the domain. The first centers on technology and innovation adoption concepts, such as technology innovation, innovation adoption, self-service, and automation. Empirical works in the past 15 years have extended further into new technology-based services such as self- service technology, automation, mobile services, and smart interactive services (Meuter et al., 2005; Wunderlich et al., 2012; Zhu et al., 2007). These are based on earlier works, such as the technology readiness concept (Parasuraman, 2000), the technology acceptance model (TAM) model (Davis, Bagozzi, & Warshaw, 1989), and the flow and control theories offered by Hoffman and Novak (1996). The second theme of this domain spans over innovation in an array of high-touch, non-technology-based customer service interfaces as well as some well-documented service concepts, including customer satisfaction, quality of service, customer relationship, service failure and recovery, omni-channel user experience, and relationship management in retailing Sis. Evidence of overlap between the two themes can be found in studies that feature technological failure and recovery' in the user experience of computer-mediated Sis (Dong et al., 2008; Zhu et al., 2013).

Technology innovation and customer interface research is found to be connected with digital innovation and financial services (Domain 3) through shared interests in digital services (e.g., e-commerce, mobile technology, and internet marketing) and the TAM. This domain is also linked to the Innovation in Business and Service Strategies Domain through classic concepts such as ‘customer service’ and ‘quality of service.’ In addition, a weak link to the Transformative and Public Services Domain is detectable via the ‘innovation adoption’ concept and other shared analytical methods. One salient feature of the domain is the presence of the super node — ‘technology' innovation’ — which implies an overwhelming emphasis on technology-based SI in this domain. Meanwhile, innovation with regard to human elements (such as employee creativity) or non-technological process innovation is potentially understudied in the consumer service experience literature.

The most commonly used industrial contexts in this research domain include e-commerce, retailing, hotel and food services, construction, automobile services, and internet marketing. Methodologically, the Delphi method, Radio Frequency Identification Systems (RFIS), statistical sampling, structural equation modeling, and logistic regression analysis are often employed for data collection or analysis.

Domain 3: digital innovation and financial services

The third most developed domain addresses innovations, especially' those based on information technology and adoption of technology-mediated interfaces in an array of financial services, from consumer lending, saving and checking, investment advising, stock exchange, insurance, to accounting services, to name just a few. This domain contains 117 unique subject terms which appeared 1,718 times in the corpus. Among the subject terms, 3,367 edges were established to form a dense conceptual community, reaching a density level of .496. A wide range of studies and disciplines joined the discourse regarding the adoption of financial innovation and its effect on consumers, financial institutions, and society at large. This domain centers on corporate strategies in developing, adopting, and deploying digital innovations of any kind. For earlier digital technologies, Alvarez and Lippi (2009) employed econometrics methods to model branch distribution and deployment of ATMs to meet consumer money or cash demand; Forman (2005) investigated the corporate digital divide in financial services in early internet adoption decisions. In recent years, more disruptive analytical and operational technologies, such as artificial intelligence (AI), machine learning, and blockchain are embraced under the banner of Fintech to further digitize business and consumer financial solutions (Gomber et ah, 2018; Qi & Xiao, 2018). Organizational learning and adaptive capacities appear to be the common theoretical underpinnings of the related studies (e.g., Dlazevic & Lievens, 2004; Weigelt & Sarkar, 2012).

Apart from corporate innovation practices and strategies, the second theme, individual consumer adoption attitudes toward and behaviors regarding innovative financial solutions (such as e-commerce payment systems and mobile banking services) are also investigated (e.g., Xue et al., 2011; Zhu et al., 2007). The third theme in this domain looks further into the far-reaching effects of digital innovation in financial services on broader society, often from the economic development and legislation viewpoints. Unique issues such as information transparency strategy, financial inclusion, federal safety nets, and patent protection have attracted attention from researchers in recent publications (e.g., Lumsden, 2018; Oney, 2018). Structurally, the technolog)' and society theme bridges to Domain 4 (Transformative Services and Public Sendees) and Domain 6 (Agency Sendees in Cultural and Entertainment Industries), whereas the consumer adoption of digital innovation theme ties to Domain 2 (Technology Innovation and Customer Interfaces).

Publications in this domain expand across multiple disciplines, including business strategy, finance, marketing and consumer behavior, accounting and auditing, and business law. Thus, the research contexts include financial services, related legal services, and news publication industries. It is noteworthy that most of the cutting-edge digital innovations in financial services have been investigated outside of the United States, indicating that European countries, and some emerging countries, are leading digital innovations in financial services thus far. Methodologically, data mining, data envelopment analysis (DEA), and AI algorithms are among the frequently used research methods for this domain. Interestingly, grounded theory qualitative research has also been employed to explore the unknowns of the radical innovations rising in the related sendee sectors.

Domain 4: transformative services and public services

The fourth SI domain concerns the innovation and design of public services and the transfonna- tive value brought about by various service practices. This domain gathers 158 unique subject terms that appear 1,856 times in the dataset and establish 5,460 ties within the domain. A number of public services, such as civil services, education, and insurance provision, have been set as research contexts for studying innovations in the transformative and public services. In terms of methodological approaches, conceptual, analytical (i.e., mathematical modeling), and empirical methods (i.e., surveys and comparative studies) are all found in this service domain.

The first theme addresses a variety of innovation design and adoption issues in government and public services. For example, concepts such as e-government, public service leadership and social innovation within public administration and civil service, information storage and retrieval, and government policy are frequently cited as subject terms for related publications (Alves, 2013; Osborne & Brown, 2011; Scheirer & Dearing, 2011; West, 2004). The second theme of the domain highlights the public, transformative value of sendees beyond the typical public sectors (Anderson & Ostrom, 2015; Baron et al., 2014; Finstenvalder et al., 2017). For example, Blocker and Barrios (2015) drew from service-dominant logic and structuration theory to conceptualize the transformative value of service design and innovation as a social dimension of value creation that illuminates uplifting changes among individuals and collectives in the marketplace, which differs from the routine or habitual values offered in such services. Researchers in the domain have investigated the impact of holistic value propositions and communal practices on consumer and employee wellbeing (Rosenbaum, 2015), inclusion and accessibility to vulnerable consumers or members of society (Dickson, Darcy, Johns, & Pentifallo, 2016; Rosenbaum et al., 2017), and mitigating inequalities due to poverty (Martin & Hill, 2015).

The network density of this domain reaches .44, indicating a good balance between internal congruency and external connection. Transformative services and public services are found to be conceptually linked to Digital Innovations and Financial Services (Domain 3) through shared concerns about health insurance, government regulation, legal services, and communication of technical information. Meanwhile, transformative services are also loosely connected with Technology Innovation and Customer Interface (Domain 2) through concepts such as work environment and employee well-being. Interestingly, the connection to the Sendee Ecosystem and Knowledge-Based Services (Domain 5) is mostly driven by shared system thinking and system-level qualitative research methods. In addition, transformative services innovation ties loosely to Agency Services in Cultural and Entertainment Industries (Domain 6). Thus, the Transformative Services and Public Services Domain arises as the most broadly connected domain within the overall SI mindscape, which also reflects the intersectional and cross-disciplinary nature of many works in the domain (Corns & Saatcioglu, 2015).

Domain 5: service ecosystem and knowledge-based services

Two emerging SI domains were identified in the conceptual network. These present significantly fewer nodes, and links among nodes, as compared to other more fully developed domains. The first new domain — namely, Service Ecosystems and Knowledge-Based Services — entails 33 unique subject terms that appear 547 times in the corpus and form 231 edges among each other, reaching a network density of .438. It appears to be spun out of the border area between Transformative Services and Public Services (Domain 4) and Innovation in Business and Service Strategy (Domain 1). Knowledge-Based Services include tasks that require use of detailed processes or technical knowledge. Professional services, such as consulting, healthcare, and software engineering, are typical knowledge-based services.

Despite the small number of subject terms in the domain, the current pool offers a rich set of concepts for theoretical development and practical examination along two themes. One theme focuses on innovations related to system-level infrastructure and resources, such as those in resource management, management information systems, and information architecture (Voss & Hsuan, 2009). Such concepts are especially relevant to value networks in platform-based economies, such as those that many societies are moving toward (Agarvval & Selen, 2009; Lusch & Nambisan, 2015). The other theme of the domain points to human (and hence managerial) issues, such as corporate culture, knowledge management, organizational change, stakeholders, and employee empowerment in service ecosystems. As Blocker and Barrios (2015) point out, because all social and economic actors are resource integrators, the locus of value creation can be broadened beyond a provider—customer dyad and toward a view of service ecosystems. Other stakeholders such as customers, employees, suppliers, and complementers are embraced into the value creation of the service ecosystem. One notable issue, the customer’s journey to interact with service deliver)' network (Tax, McCutcheon, & Wilkinson, 2013), has attracted recent attention in the SI field. The presence of the two-theme structure resembles the social/technological subsystems or information technology/organization pair, as highlighted conceptually by Leonardi and Barley (2008). Structurally, the system-level infrastructure theme closely bridges the domain to Transformative Services and Public Services (Domain 4). Meanwhile, the managerial theme spans toward Innovation in Business and Service Strategy (Domain 1).

Industries such as ecosystem services, logistics, support services, and office administrative services are found in existing publications. Methodologically, more qualitative methods (e.g., best practices and case studies) than quantitative research methods are used in these studies, possibly owing to the difficulty of accessing system-level data in empirical research on service ecosystems, which will continue to be a challenge in future research.

Domain 6: Agency Services in Cultural and Entertainment Industries

Another nascent domain, Agency Services in Cultural and Entertainment Industries, emerges from the present research; this may be the first time this industry appears in a literature review on SI. Among the 20 nodes in the domain, agency decisions and behaviors of agencies, managers, publishers, or promoters of various art and cultural designs and offerings take shape into an identifiable theme (e.g., Scott, 2006; Sunley et al., 2008). Research works on agency emphasize that design emerges from interfaces that synthesize and recombine diverse knowledge so as to produce emergent effects and new designs.

Modular architecture and noncontractual innovation (or open innovation) are found to proliferate in industries, such as visual art and gaming sectors, where uncertainty of pertinent problems and their solution is high and the relevant knowledge is widely dispersed (e.g., Christensen, 2006; Olleros, 2007). This industry may eschew centralized platforms or leader- dominant ecosystems, choosing instead distributed and self-organized innovation processes driven by the desire of many independent agents to exploit some platform potentialities, bringing positive or negative surprises to platform leaders (Cusumano & Gawer, 2002; Iansiti & Levien, 2004). Thus, noncontractual innovation processes not only add conceptual completeness to the understanding of a platform economy, as suggested by Olleros (2007), but also enrich the innovation frameworks for service offerings.

Domain 6 features a high-density structure. Innovation phenomena from motion picture, performing arts, radio network, theater, television broadcasting, and publishing industries are closely interconnected within the domain, reaching a density level of .847. This high similarity (or convergence) among the different creative fields is impressive given the fact that, although they often share a common tendency to focus on ideas and to hold a strong aesthetic component, they differ starkly in relationships with markets, distribution channels, and intellectual property rights. Meanwhile, the high density leaves little room for other concepts outside the domain to set bridges or links. Only two weak outward ties were detected for Domain 6, one connecting to Domain 3 (Digital Innovation and Financial Services) through news publishing topics and the other to Domain 4 (Transformative Sendees and Public Sendees) through ‘metropolitan areas’ and ‘transportation’ topics. Although no method subject tenn has been detected in this domain, possibly because it is still in early development, inductive approaches such as case studies and interviews are commonly used in this domain. Meanwhile, its methodological void indicates a promising research opportunity for future studies in the creative service domain.

Recent shifts among domains in SI literature

Over the last 15 years, the SI literature has experienced rapid growth. As shown in Figure 9.8, the annual publication rate increased from below 100 prior to 2007, to more than 150 afterward, with a peak of more than 200 in 2015. Meanwhile, the distribution of publications among the six SI domains shows an uneven and dynamic pattern (see Figure 9.7). The Digital Innovation and Financial Services domain was the most prominent in the early 2000s, appearing in 35% of the publications (about 30). During an expansion of the SI field in the past decade, the number of publications in this domain increased to about 25 annually, but its relative weight reduced to 13% of all publications. In contrast, the weight of Innovation in Business and Service Strategy domain started with 27% in 2003, but increased to 43% during the same time period. The Technolog)' Innovation and Customer Interfaces domain maintained its second place ranking throughout the entire time period investigated. One interesting upward trend in SI research can be identified in the Transformative Services and Public Services domain, whose weight in the SI topic surged from 5% in 2003 to 17% in 2018. Last, the number of publications regarding the Agency Services in Cultural and Entertainment Industries domain remained lower than 2% over this time. (Please see Figure 9.9.)

Figure 9.7 Journal publications by SI domains.

Notes: Dl & FS denotes 'Digital Innovation and Financial Services'; IB St SS denotes 'Innovation in Business and Service Strateg/; Tl Sr Cl denotes 'Technology Innovation and Customer Interfaces'; TS Sr PS denotes 'Transformative Services and Public Services'; SE Sr KS denotes 'Service Ecosystem and Knowledge-Based Services'; AS St El denotes 'Agency Services in Cultural and Entertainment Industries.'

Figure 9.8 SI publications by year from 2003 to 2018.

Figure 9.9 Domain evolution from 2003 to 2018.

This investigation into the structure of the SI mindscape provides a novel perspective for mapping the principle thoughts and their related methods and industrial relevance. Beyond the assessment of the developmental status of the field, this graphic approach also helps with visualizing the structural holes (white spaces) within and, more importantly, between conceptual domains, which sheds light on potential directions for future research.

Finding the next edge of SI research

Looking forward to future research in SI, three approaches are proposed to explore novel conceptual advancement in the SI field. These approaches are (1) strengthening loose interdomain connections; (2) creating new associations between disconnected domains; and (3) thickening links among intra-domain themes. The presence of structural holes and weak links between domains is depicted in Figure 9.10. According to network theory, forming associations and relevance between disconnected ideas is a critical source of combinative ideas (Burt, 2004); however, a new bridge built to mend a structural hole often starts out weak because of the lack of shared viewpoints or methods. Compared to the second approach, strengthening loose inter-domain connections is considered less risky because the initial pathways have been set for subsequent works, thus possibly bringing in low hanging fruit for further investigation. Regarding the third approach, there is still much room for further deepening the conceptual connections within each domain, especially for the rising and nascent SI domains (such as Domains 4, 5, and 6). However, thickening intra-domain links falls outside the scope of this chapter, so it focuses on the first two approaches in the discussion of future research directions.

Weak inter-domain links

Three loose ties among the six SI domains were detected. The first weak link is located between the Technology Innovation and Customer Interfaces Domain and the Transformative Services and Public Services Domain. Currently, the ‘innovation adoption’ concept and other shared analytical methods are barely pinning the two domains together. Future research questions should invite direct linkages between themes across domains, such as: ‘How should transformative value and habitual value be balanced or prioritized in creating public services?’ and ‘To what extent are performance concepts (such as customer satisfaction and loyalty) applicable

Figure 9.10 Potential structural holes and opportunities for future SI research.

in nonpublic services?’ Similarly, ‘How can transformative value be mindfully added to the design and delivery of nonpublic services?’ In addition, considering the ubiquitous influence of technology innovation, it is appropriate to ask: ‘What are the roles that technological innovations can play in advancing transformative value in public sendees?’ These are fascinating linkages to explore between domains.

Another weak tie is found between the Digital Innovations and Financial Services Domain and the Transformative Services and Public Services Domain, mainly bounded through the concept ‘technology and society.’ Many further investigation questions can be conjectured along the lines of: ‘How does digital innovation (such as AI and machine learning) facilitate or inhibit financial inclusion, especially to vulnerable social members or unattractive market segments?’ and ‘What should be the principle or framework for designing, applying, and assessing such digital innovations to promote the transformative value in public services, be it healthcare, education, telecommunication, or wealth distribution?’

Last, the Transformative Services and Public Services Domain connects with the Agency Services in Cultural and Entertainment Services Domain only through ‘metropolitan areas’ and ‘transportation.’ Much more research is needed to investigate the impact of creative innovations on public services and social well-beings and vice versa. In addition, it will be theoretically intriguing to probe: ‘How can the distributed innovation model featured in the cultural creation sectors be extended to scattered transformative resources, such as volunteer forces, to achieve broader impact in society?’

As previously argued, early works between these loosely connected domains, though rare or preliminary, have gleaned early evidence on the subjects and tested the interest among the academic audience and practitioners. Thus, this helps to develop future studies that will deepen the conceptual or methodological connections.

Structural holes between domains

Structural holes exist when a focal actor (in this analysis, a central domain) links to a pair of disconnected actors (or two other domains) (Borgatti & Halgin, 2011; Burt, 2004). Mapping of the subject term network reveals seven structural holes, as shown in Figures 9.5 and 9.10. This research found that structural holes are often located surrounding nascent domains (such as the Service Ecosystem and Knowledge-Based Services Domain and the Agency Services in Cultural and Entertainment Services Domain), where conceptual relevance to other domains in the overall network has yet to be proposed or verified. However, structural holes can also exist between mature domains (such as between the Digital Innovations and Financial Services Domain and the Innovation in Business and Service Strategy Domain), implying potentially departing interests, reasoning logics, and even theoretical foundations. The discussion begins with these structural holes, owing to their theoretical significance.

The network graph reveals two substantial structural holes among three mature SI domains (Domain 1, 3, and 4). The Innovation in Business and Service Strategy' Domain, being the most populous cluster, has not yet conceptually connected with the Digital Innovations and Financial Service Domain. The puzzling phenomenon is that, although both domains are closely tied to the Technology Innovation and Customer Interfaces Domain, somehow the direct edge between the two is hard to find. It can be speculated that the central reasoning of digital innovations (such as machine learning from big data) is based on bottom-up, inductive pattern search and detection from accumulated huge datasets, whereas strategic thinking in regard to organizational innovation in services relies primarily on top- down planning and managerial functioning. Thus, the two reasoning logics still need to find ways to creatively integrate and complement each other in the theorization and practices of SI. Researchers may investigate: ‘What are the mechanism(s) for digital innovations to influence organizational innovations?’ and vice versa. For example, ‘How can cloud-based technologies and services change the employee—technology relationship and facilitate global virtual collaboration?’ This chapter considers the need for integrating the inductive and deductive approaches paramount in SI, indicating a promising direction for future research.

The Innovation in Business and Service Strategy Domain also misses a direct connection to the Transformative Services and Public Services Domain. Because most transformative innovations focus on services to individual citizens or consumers, few researchers have proposed the transformative meaning that a service may provide to another business. For instance, ‘Can a business collaborator be the target of transformative service? If so, how?’ To better understand the nature of new product development for transformative services, researchers may ask: ‘To what extent do sustainable or transformative innovations differ from other innovations in terms of orientation? To what extent should they differ?’ For instance, decisions may include the trade-off between customer-centric versus stakeholder-oriented, the organizational learning mechanism (e.g., learning versus unlearning), and performance metrics to be used (e.g., market growth versus well-being versus triple bottom lines). The substantial theoretical and empirical implications of the probes into such strategic-level challenges in innovations of transformative services warrant a critical direction for future SI research.

Apart from structural holes among mature SI domains, new frontiers in SI can be projected around and between the rising and emerging SI domains (such as Domains 5 and 6). This chapter first explores two structural holes around the nascent Service Ecosystem and Knowledge-Based Services Domain, followed by those around the Agency Services in Cultural and Entertainment Services Domain.

As Iansiti and Levien (2004, p. 1) stated, “Strategy is becoming, to an increasing extent, the art of managing assets that one does not own.” Building and leveraging ecosystem strengths to address strategic challenges, be they market growth, resource orchestration, or service deliver)', is a formidable approach to SI success. Research on ecosystem strengths first needs to embrace digital innovations (Domain 3), where important contemporary developments - including cloud, grid, and web services as well as smart, wearable, and mobile technologies — are involved in such transformations as contractually bound business grids, digital ecosystems, and on-demand and availability-based service level agreements. These service developments build upon and extend established core infrastructure and progressive institutionalization of standards (Chesbrough, Vanhaverbeke, West, 2006). Such digital technology not only increases scale, scope, and reach, but also shapes design and delivery, becoming more central to the structuring of services (Orlikowski & Scott, 2015). It is foreseeable that the combinative ideas between digital innovation and service ecosystem domains will bring about novel service theories and coveted insights into the ever-expanding digital economy.

In relation to technology innovation in customer interfaces (Domain 2), there are significant research opportunities in combining foundational ecosystem or platform theories with customer experience research. For example, one tenet of a service platform delineates its externality or network effect, which occurs when the value of the network or service ecosystem increases whenever other actors (such as additional customers) join the network (Van Alstyne, Parker, & Choudary, 2016). Both positive and negative network effects can influence the consumer service experience. For instance, Amazon Prime delivery' service on Black Friday' offers positive network effects and therefore attracts more merchandisers offering goods, which justifies denser warehouse deployment for delivery. Thus, negative network effects take place when too many accepted offers need to be delivered in a limited time window, causing delivery failures and customer dissatisfaction. It is important to understand consumer adoption attitudes and behaviors toward ecosystem-based services, including their mental accounting of the positive and negative network effects in service experiences.

Next, the discussion moves to structural holes found around the new SI domain - Agency Services in Cultural and Entertainment Services — in relation to organizational innovation (Domain 1) and customer interfaces (Domain 2). These structural holes invite high-level questions such as: ‘How can agency models commonly used in dispersed cultural innovation inform a study in organizational innovation and service process innovation?’ Looking at this from the reverse side, ‘What relevance and utility do conceptual models and research methods established in mature SI domains bring to the future development of the cultural creation field?’ Linking the innovation model in agency industries to customer experience management, comparison can be drawn between the sourcing decisions among agencies in cultural creation fields and crowdsourcing management between a service firm and its customer base. Therefore, ‘How is the governance of noncontractual agencies in cultural creation similar to or different from herding crowd-based customer value co-creation?’ Answers to such research questions will help to fill the structural holes around the agency service domain.

The Agency Services in Cultural and Entertainment Industries domain can be theoretically and practically connected with the service ecosystem domain. Based on real world observations, many entertainment and cultural service industries rely heavily on innovation ecosystems (e.g., video game developer markets) and distribution platforms (e.g., iTunes music store). Contractual and noncontractual innovations can complement each other and jointly foster the development of these service sectors. Olleros (2007) comments that innovation contracts tend to be over-specified and thus limit solutions availability or opportunities detection. By contrast, networks of noncontractual innovation can more effectively mobilize distributed local knowledge, not only for problem-solving, but also for problem-seeking (Sunley et al., 2008). Investigation into the collaboration or integration mechanisms between the two types of innovation networks will help to shape more open-ended and fertile innovation processes.

Conclusion

This chapter draws from network theory and complex network analysis methods to map the conceptual structure of SI, by detecting critical conceptual domains within the field and delineating related themes, service industry contexts, and related research methods within each domain. Based on the observed connections and missing links (or structural holes) among the conceptual domains, potential directions for future research on SI are pointed out. Hopefully, these will ignite creative and combinative ideas to promote the research work across otherwise disparate domains in the SI field. Methodologically, this chapter demonstrates how automated SNA methods can be applied on large-size literature databases for conceptual mapping in service research.

References

Agarwal, R., & Selen, W. (2009). Dynamic capability building in service value networks for achieving service innovation. Decision Sciences, 40(3), 431—475.

Alvarez, F., & Lippi, F. (2009). Financial innovation and the transactions demand for cash. Econometrica, 77(2), 363-402.

Alves, H. (2013). Co-Creation and innovation in public services. Service Industries Journal, 33(7—8), 671-682.

Anderson, L., & Ostrom, A. L. (2015). Transformative service research: Advancing our knowledge about service and well-being. Journal of Service Research, 18(3), 243—249.

Antons, D., & Breidbach, C. F. (2018). Big data, big insights? Advancing service innovation and design with machine learning. Journal of Service Research, 21(1), 17—39.

Baron, S., Warnaby, G., & Hunter-Jones, P. (2014), Service(s) Marketing Research. International Journal of Management Reviews, 16(2), 150—171.

Bastian, M., Heymann, S., & Jacomy, M. (2009) Gephi: An open source software for exploring and manipulating networks. Proc. Inti. AAAI Conf on Weblogs and Social Media, San Jose, CA, May 2009.

Blazevic, V., & Lievens, A. (2004). Learning during the new financial service innovation process: Antecedents and performance effects .Journal of Business Research, 57(4), 374.

Blocker, С. P., & Barrios, A. (2015). The transformative value of a service experience. Journal of Service Research, 18(3), 265-283.

Borgatti, S. P., & Halgin, D. S. (2011). On network theory. Organization Science, 22(5), 1168-1181.

Borgatti, S. P., Mehra, A., Brass, D. J., & Labianca, G. (2009). Network analysis in the social sciences. Science, 323(5916), 892-895.

Brammer, C., & Galloway, R. (2007). IEEE transactions on professional communication: Looking to the past to discover the present. IEEE Transactions on Professional Communication, 50(4), 275—279.

Burt, R. S. (2004). Structural holes and good ideas. American Journal of Sociology, 110(2), 349—399.

Carlborg, P., Kindstrom, D., & Kowalkowski, C. (2014). The evolution of sendee innovation research: A critical review and synthesis. Service Industries Journal, 34(5), 373—398.

Chesbrough, H. W., Vanhaverbeke, W., & West, J. (eds.) (2006). Open innovation: Researching a new paradigm. Oxford: Oxford University Press.

Christensen, С. M. (2006). The ongoing process of building a theory of disruption. Journal of Product Innovation Management, 23(1), 39—55.

Corns, C., & Saatcioglu, B. (2015). An intersectionality framework for transformative services research. Service Industries Journal, 35, 415—429.

Cusumano, M. A., & Gawer, A. (2002). The elements of platform leadership. MIT Sloan Management Review, 43(3), 51—58.

Damanpour, F., Walker, R. M., & Avellaneda, C. N. (2009). Combinative effects of innovation types and organizational performance: A longitudinal study of service organizations. Journal of Management Studies, 46(4), 650—675.

Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982—1003.

Dickson, T. J., Darcy, S., Johns, R., & Pentifallo, C. (2016). Inclusive by design: Transformative services and sport-event accessibility. Service Industries Journal, 36(11—12), 532—555.

Doerfel, M. L., & Barnett, G. (1999). A semantic network analysis of the International Communication Association. Human Communication Research, 25(4), 589-603.

Dong, B., Evans, K. R., & Zou, S. (2008). The effects of customer participation in co-created service recover)-. Journal of the Academy of Marketing Science, 36(1), 123—137.

Drejer, I. (2004). Identifying innovation in surveys of services: A Schumpeterian perspective. Research Policy, 33(3), 551.

Finsterwalder, J., Foote, J., Nicholas, G., Taylor, A., Hepi, M., Baker, V., & Dayal, N. (2017). Conceptual underpinnings for transformative research in a service ecosystems context to resolve social issues: Framework foundations and extensions. Sendee Industries Journal, 37(11—12), 766—782.

Forman, C. (2005). The corporate digital divide: Determinants of Internet Adoption. Management Science, 51(4), 641-654.

Freeman, L. C. (2004). 'The development of social network analysis: A study in the sociology of science. Vancouver: Empirical Press.

Gallouj, F., & Savona, M. (2009). Innovation in services: A review of the debate and a research agenda. Journal of Evolutionary Economics, 19(2), 149—172.

Gomber, P., Kauffman, R. J., Parker, C., & Weber, B. W. (2018). On the fintech revolution: Interpreting the forces of innovation, disruption, and transformation in financial services. Journal of Management Information Systems, 35(1), 220—265.

Gondal, N. (2011). The local and global structure of knowledge production in an emergent research field: An exponential random graph analysis. Social Networks, 33(1), 20-30.

Gruber, T. (1993). A translation approach to portable ontology specifications. Knowledge Acquisition, 5(2), 199-220.

Hipp, C., & Grupp, H. (2005). Innovation in the service sector: The demand for service-specific innovation measurement concepts and typologies. Research Policy, 34(4), 517—535.

Hoffman, D., & Novak, T. P. (1996). Marketing in hypermedia computer-mediated environments: Conceptual foundations. Journal of Marketing, 60 (July), 50—68.

Iansiti, M., & Levien, R. (2004). Strategy as ecology. Harvard Business Review, 82(3), 68—78.

Jantunen, A., Ellonen, H.-K., & Johansson, A. (2012). Beyond appearances: Do dynamic capabilities of innovative firms actually differ? European Management Journal, 30(2), 141—155.

Kovacs, B. (2010). A generalized model of relational similarity. Social Networks, 32(3), 197—211.

Lee, S., Close, A. G., & Love, C. (2010). How information quality and market turbulence impact convention and visitors bureaus’ use of marketing information: Insights for destination and event marketing. Journal of Convention & Event 'l’ourism, 11(4), 266—292.

Leonardi, P. M., & Barley, S. R. (2008). Materiality and change: Challenges to building better theory about technology- and organizing. Information and Organization, 18(3), 159—176.

Lumsden, E. (2018). The future is mobile: Financial inclusion and technological innovation in the emerging world. Stanford Journal of Law, Business & Finance, 23(1), 1—44.

Lusch, R., & Nambisan, S. (2015). Service innovation: A service-dominant logic perspective. MIS Quarterly, 39(1), 155-175.

Manning, C. D., Raghavan, P., & Schiitze, H. (2008). Introduction to information retrieval. New York: Cambridge University Press.

Martin, K. D., & Hill, R. P. (2015). Saving and well-being at the base of the pyramid: Implications for transformative financial services delivery. Journal of Service Research, 18(3), 405—421.

Meuter, M. L., Bitner, M. J., Ostrom, A. L., & Brown, S. W. (2005). Choosing among alternative service delivery modes: An investigation of customer trial of self-service technologies .Journal of Marketing, 69, 61-83.

Mu, T., Goulermas, J., Korkontzelos, L, & Ananiadou, S. (2016). Descriptive document clustering via discriminant learning in a со-embedded space of multilevel similarities. Journal of the Association for Information Science and 'Technology, 67(1), 106—133.

Newman, M. E. J. (2010). Networks: An introduction. Oxford, UK: Oxford University Press.

Obstfeld, D. (2005). Social networks, the Tertius Iungens orientation, and involvement in innovation. Administrative Science Quarterly, 50(1), 100—130.

Olleros, F. X. (2007). The power of non-contractual innovation. International Journal of Innovation Management, 11(1), 93-113.

Oney, C. (2018). Fintech industrial banks and beyond: How banking innovations affect the federal safety net. Eordham Journal of Corporate & Financial Law, 23(2), 541—575.

Orlikowski, W. J., & Scott, S. V. (2015). The algorithm and the crowd: Considering the materiality of service innovation. MIS Quarterly, 39(1), 201—216.

Osborne, S., & Brown, L. (2011). innovation, public policy and public services delivery in the UK: The word that would be king? Public Administration, 89, 1335—1350.

Ostrom, A. L., Parasuraman, A., Bowen, D. E., Patricio, L., & Voss, C. A. (2015). Service research priorities in a rapidly changing context .Journal of Service Research, 18(2), 127—159.

Ottenheimer, H. J. (2006). The Anthropology of Language: An Introduction to Linguistic Anthropology. Belmont, CA: Thomson, Wadsworth.

Parasuraman, A. (2000). Technology' Readiness Index (TRI): A multiple-item scale to measure readiness to embrace new technologies. Journal of Sendee Research, 2, 307—320.

Perez-Corona, N., Hemandez-Colin, D., Bustillo-Hemandez, C., & Figueroa-Nazuno, J. (2012). Model of natural semantic space for ontologies construction. International Journal of Combinational Optimization Problems and Informatrics, 3(2), 93—108.

Qi, Y., & Xiao, J. (2018). Fintech: AI powers financial services to improve people’s lives. Communications of the ACM, 61 (11), 65-69.

Rice, R. E., & Danowski, J. A. (1993). Is it really just like a fancy answering machine? Comparing semantic networks of different types of voice mail users. 'The Journal of Business Communication, 30(4), 369-397.

Rosenbaum, M. S. (2015). Transformative service research: Focus on well-being. Service Industries Journal, 35(7-8), 363-367.

Rosenbaum, M., Seger-Guttmann, T., & Giraldo, M. (2017). Commentary: vulnerable consumers in service settings, Journal of Services Marketing, 31 (4/5), 309—312.

Salton, G., Wong, A., & Yang, C. S. (1975). A vector space model for automatic indexing. Communications of the ACM, 18(11), 613—620.

Scheirer, M. A., & Dearing, J. W. (2011). An agenda for research on the sustainability of public health programs. American Journal of Public Health, 101 (11), 2059—2067.

Scott, A. J. (2006). Creative cities: Conceptual issues and policy questions. Journal of Urban Affairs, 29, 1-17.

Steyvers, M., & Tenenbaum, J. (2005). The large-scale structure of semantic networks: Statistical analyses and a model of semantic growth. Cognitive Science, 29, 41—78.

Sunley, P., Pinch, S., Reimer, S., & Macmillen, J. (2008). Innovation in a creative production system: The case of design. Journal of Economic Geography, 8(5), 675—698.

Tax, S. S., McCutcheon, D., & Wilkinson, 1. F. (2013). The Service Delivery Network (SDN): A customer-centric perspective of the customer journey. Journal of Service Research, 16(4), 454—470.

Van Alstyne, M., Parker, G. G., & Choudary, S. P. (2016). Pipelines, platforms, and the new rules of strategy. Harvard Business Review, 94(4), 54—62.

Van Wijk, R., Jansen, J. J. P., & Lyles, M. A. (2008). Inter- and intra-organizational knowledge transfer: A meta-analytic review and assessment of its antecedents and consequences. Journal of Management Studies, 45(4), 830-853.

Verhoef, P. C., & Leeflang, P. S. (2009). Understanding the marketing department’s influence within the firm. Journal of Marketing, 73(2), 14—37.

Voss, C. A., & Hsuan, J. (2009). Service architecture and modularity. Decision Sciences, 40(3), 541-569.

Weigelt, C., & Sarkar, M. B. (2012). Performance implications of outsourcing for technological innovations: Managing the efficiency and adaptability trade-off. Strategic Management Journal, 33(2), 189-216.

West, D. M. (2004). E-govemment and the transformation of service delivery and citizen attitudes. Public Administration Review, 64(1), 15-27.

Witell, L., Snyder, H., Gustafsson, A., Fombelle, P., & Kristensson, P. (2016). Defining service innovation: A review and synthesis. Journal oj Business Research, 69(8), 2863—2872.

Wunderlich, N. V., Wangenheim, F. V., & Bitner, M. J. (2012). High tech and high touch: A framework tor understanding user attitudes and behaviors related to smart interactive services. Journal of Service Research, 11 (1), 3—20.

Xue, M., Hitt, L. M., & Chen, P. (2011). Determinants and outcomes of internet banking adoption. Management Science, 57(2), 291—307.

Zhang, S., Luo, X., Xuan, J., Chen, X., & Xu, W. (2014). Discovering small-world in association link networks for association learning. World Wide Web, 17(2), 229—254.

Zhang, W., & Zinoviev, D. (2018). How North Korea views China: Quantitative Analysis of Korean Central News Agency Reports. Korean Journal of Defense Analysis, 30(3), 377—396.

Zhang, Y., & Li, H. (2010). Innovation search of new ventures in a technology cluster: The role of ties with service intermediaries. Strategic Management Journal, 31(1), 88-109.

Zhu, Z., Nakata, C., Sivakumar, K., & Grewal, D. (2013). Fix it or leave it? Customer recovery from self-service technology' failures. Journal of Retailing, 89(1), 15-29.

Zhu, Z., Nakata, C., Sivakumar, K., & Grewal, D. (2007). Self-service technology effectiveness: the role of design features and individual traits. Journal of the Academy of Marketing Science, 35(4), 492-506.

Zinoviev, D., Stefanescu, D., Swenson, L., & Fireman, G. (2013). Semantic networks of interests in online NSS1 communities. Proceeding of the Workshop “Words and Networks, ” Evanston, IL, June 2012.