The Role of Entrepreneurial Orientation and Modularity for Business Model Innovation in Service Companies

The phenomenon of business model innovation (BMI) is taking up speed in researchers’ attention. Specifically servi c f rms find it difficult to protect their intellectual property and need dynami c capabilities to invent and launch their new business models. This study on ser vice firms’ BMI applies the Dynamic Capability View (DCV) to understand how t o capabilities, entrepreneurial orientation (EO) and modularity pro duce BMI. These two capabilities consider two sides of BMI. BMI relates to opportunity-seeking behavior, creating new values and opportunity explo itati n, that then allows monetizing and capturing value. In this respect, BM I takes advantage of the two dynamic capabilities: EO and modularity. The re sults of our study of 299 international service firms show that EO and modula rity trigger BMI, and that EO is especially favourable for BMI in the context of high environmental uncertainty.


Introduction
Business model innovation (BMI), first mentioned by Papinniemi (1999), receives flourishing attention by researchers (Srinivasan, 2011;Amit and Zott, 2012;Euchner and Ganguly, 2014).BMI incorporates change and renewal through the commercialization of ground-breaking ideas, technologies, and new business concepts (Chesbrough, 2010;Zott et al., 2011).Previous research, often focused on internet businesses, developed the important characteristics of BMI: value-generation processes, value creation, proposition, and capture (ElMaraghy, 2005;Sánchez and Ricart, 2010).In the literature on BMI, internet firms and the manufacturing industry dominate.The focus of our study lies on service firms that have fewer possibilities to protect their innovation by intellectual property rights and thus are constantly in need of dynamic capabilities to search for new business ideas (Vargo and Lusch, 2008;Voss and Hsuan, 2009).
The novel value generation, value creation, value proposition, and value capture of BMI are at the core of new business ideas and ventures into new territories.Yet, it remains unclear what dynamic capabilities underlie the process of value generation, value proposition, and value capture (Osterwalder et al., 2005;Schneider and Spieth, 2013).This is especially important for service firms which are unable to store and to standardize their offering due to the immateriality of services and the simultaneity of production and consumption (Lovelock and Gummesson, 2004;Grönroos and Ojasalo, 2004).
This paper aims to understand dynamic capabilities that determine BMI in service firms.We selected two capabilities: As venturing into new territory requires risk taking, innovativeness, and proactivity, we focus on EO (Schindehutte et al., 2008;Guo et al., 2013;Habtay and Holmén, 2014); second, service firms can select specific service processes as modularized process patterns.These modular service patterns can create new service offerings or recombine service components to propose and capture value.The modularity of product components has already shown strong improvements in efficiency and innovation, in production management, and in product innovation management (Sanchez and Mahoney, 1996;Worren et al., 2002;Bouncken et al., 2014).Yet, this concept is new to service research.We assume that modularization enables strong gains for service firms and shapes a dynamic capability for their BMI.Thus, we selected modularity as a second capability.We argue that EO and modularity are important in situations of high uncertainty, where service firms have to change and/or adapt their business models.
In essence, this paper considers EO and modularity as dynamic capabilities in service firms and analyses how they influence BMI in cases of high uncertainty.Herein we aim to contribute to the new body of research on BMI, which is still fragmented with weak theoretic underpinnings (Vermeulen, 2013;Chen et al., 2015), and to service research, which has not yet embraced the ideas of BMI even though it drives service venturing and service strategy.Further, we highlight two important antecedents to BMI and link the research on business models with strategic management.
Our theoretical part starts with an explanation of business models and BMI.We continue with dynamic capabilities which are subsequently linked to BMI.Thereafter, we present our sample, the method, and our results.The paper concludes with a discussion of our findings and of the contribution our study makes to theory building.

Business model innovation
Recently, business models received great attention, with more than 1177 articles in peer-reviewed journals (Amit and Zott, 2001;Al-Debei andAvison, 2010, Lecocq et al., 2010;Lambert and Davidson, 2013).Business models are a "set of capabilities that is configured to enable value creation consistent with either economic or social strategic objectives" (Seelos and Mair, 2007).Commonly, a business model defines a large variety of a firm's characteristics, resources, and decision variables, which convert opportunities into value creation and capture (Seddon et al., 2004;O'Connor and Yamin, 2011;Zott et al., 2011).Value creation follows no linear or simple method, but rather more complex and interconnected relationships among multiple actors (Zott et al., 2011).
Innovating a firm's business model involves adding and linking novel activities through integration at different levels and in new ways (Witell and Löfgren, 2013).BMI aims at consciously renewing a firm's core business rather than concentrating on particular product or service innovations (Amit and Zott, 2010).It focuses on the exploitation of firms' core competencies (Hamel and Prahalad, 1993) and the creation of alliances (Mowery et al., 1996).BMI uses trends -e.g.increasing customer magnitude (Teece, 2010), service orientation (Gremyr et al., 2010;Nair et al., 2013), and new forms of innovation such as open innovation (Rajala et al., 2012;Huang et al., 2013).
Based on Mitchell and Coles' (2004) link between continuing BMI and sports, uncertainty or "the unknown" appears to be a central challenge to the successful implementing of innovations and specifically business model innovations (Andries and Debackere, 2013).Originally phenomenon-oriented, some researchers explain BMI on the basis of positive effects on a firm's performance and under environmental dynamism.BMI also corresponds to dynamic capabilities (Heij et al., 2014).Informal exchanges in the organization, mutual involvement, and organizational justice, as well as cultural aspects such as innovation orientation, are dynamic capabilities driving BMI (Bouncken and Koch, 2007;Bouncken, 2007;Santos et al., 2009).BMI contributes to improved environmental adaptation and long-term competitive advantage while facilitating and utilising the firm's entrepreneurial capabilities (Schindehutte et al., 2008;Habtay and Holmén, 2014).Research requires further understanding of underlying capabilities, particularly of how to constantly renew a firm's existent business model.For service companies, BMI especially consists of integrating customers in the process of value creation (Clauß et al., 2014).Due to the high level of customer integration, BMI needs to be more comprehensive in service firms than in manufacturing companies in order to increase service quality and profitability (Kindström, 2010).
Based on Barreto's (2010) definitions and descriptions of dynamic capabilities, we first regroup definitions and major papers referring to them.Second, we highlight the key factors and assumptions of each stream of definition(s).DCV stresses that a firm constantly needs to adapt, improve, and recombine its capabilities and their sources to stay competitive.As such, dynamic capabilities can act as drivers of BMI (Bock et al., 2012;Singh et al., 2013;Roaldsen, 2014).
Previous studies also assume that the performance of a business model depends on components (Aspara et al., 2010;Heij et al., 2014) which are rooted in internally or externally oriented capabilities (Dixon et al., 2014).The literature on (product) innovation processes distinguishes between firms' activities to explore new ideas, concepts, or business models, and their capability to exploit those ideas and to realise additional revenues (Bucherer et al., 2012).Table 1 presents our literature review and links dynamic capabilities to BMI.

The twofold nature of BMI
BMI relates opportunity seeking to the creation of new values and opportunity-exploitation to monetise and capture value (Andries and Debackere, 2013).Drawing on Tushman's and O' Reilly's (1996) notion of ambidexterity, we conceive of EO and modularity as two distinct forms of organizational capabilities to foster change and renewal in uncertain market conditions.
With regard to exploration, firms need to be open and risk-affine in order to screen their environment and perceive opportunities (Ellis et al., 2014).To exploit new ideas, firms have to rely on tools to implement new products and/or market-related features into their range of products without risking cannibalism or obsolescence of the existing business model.This may be achieved, for instance, through modularization as a recombination of existing value elements with new features that influence value generation (Huang et al., 2012).EO and modularization will specifically account for BMI under strong uncertainties as typical for dynamic capabilities (Sánchez and Ricart, 2010;Bock et al., 2012;Cautela et al., 2014;Roaldsen, 2014).

Entrepreneurial Orientation (EO)
According to Kraus (2013), the best-researched dimensions of EO are risktaking (Keh et al., 2002;Morris et al., 2008), proactivity (Rauch et al., 2009), and innovativeness (Lumpkin and Dess, 1996;Morris et al., 2008;Rowley et al., 2011;Van Riel et al., 2011).Scholars have argued that inertia is one of the core barriers to BMI (Chesbrough, 2010).In contrast, experimentation with technologies, idea generation, and collaboration drives BMI, resulting in value generation (McGrath, 2010).A firm's EO and reconfiguration capabilities influence performance (Jantunen et al., 2005;Irava and Moores, 2010).This confirms Teece (2007), who puts forward that organizations need a risk-taking behaviour to successfully reconfigure their business.Thus, EO serves as a dynamic capability that drives the value generation and value proposition of a firm's business model.To put it differently: EO is a major trigger of BMI.EO has often been associated with young, rather small firms which are versatile towards changes in market structure and environmental dynamism (Hannan and Freeman, 1984).Their lack of established structures lowers the risk of organizational inertia.As we focus on young and rather small service firms, the role of EO for BMI becomes even more apparent.
Research also indicates that EO has strong relations to uncertainty (Kraus et al., 2012), which in turn is strongly linked to innovation (York and Venkataraman, 2010).Thus, we argue that uncertainty drives the importance of EO as a dynamic capability for BMI.The innovative, risk-taking attitude of EO drives the creation of new ideas and EO will then also drive BMI.

Modularity
Modularity implies the decomposition of a complex system into modules (Schilling, 2000;Conrad et al., 2014).A modular system is composed of units which are designed independently and are related through interfaces (Campagnolo and Camuffo, 2010).Each module covers a specific, but limited array of features and functions (Cavin and Lohse, 2014).The final product or service is developed by combining different customer or need-specific modules.Modularity allows for efficient management of complexity of multifaceted processes (Ethiraj and Levinthal, 2004;Ro et al., 2007;Tsvetkova and Gustafsson, 2012).Research on modularity spans product modularization (Lau and Yam, 2005;Scannella, 2010;Persson and Åhlström, 2013), business networks (van Liere et al., 2004;Van Liere and Koppius, 2007), knowledge management (Sanchez and Mahoney, 1996;Ravishankar and Pan, 2013), and innovation (Ethiraj and Levinthal, 2004;Ethiraj et al., 2008;Voss and Hsuan, 2009;Ozman, 2011).Modularity is a central design principle of the product architecture and enables strategic and operational flexibility (Tsvetkova and Gustafsson, 2012).Pisano and Teece (2007) state that modularity improves the value capture from innovation.Drivers behind modularity are the creation of variety, the reduction of complexity, and the balance of customization and standardization (Miller and Elgard, 1998;Blecker and Abdelkafi, 2006;Mikkola, 2007).Further, modularity is a vehicle for reconfiguring a firm's asset structure (Sirmon et al., 2007).It can be beneficial in dealing with growing complexity and environmental dynamism (Kodama, 2004).Thus, it is a suitable means for coping with uncertainties.
Modularity is a key principle of the architecture of service firms (Voss and Hsuan, 2009).Services are bundles of performance promises.Their production inescapably integrates the customer, who becomes a co-creator (Zeithaml et al., 1985).Take the following simple example of modularity in service processes: A client in a restaurant chooses some food from the menu.The owner of the restaurant provides the range of accessible food but the choice is made by the client.After the client has selected the components (or modules) of his meal, the staff prepares and serves the menu.The case of the restaurant illustrates the role modularity has for services: modularity serves to satisfy distinct customer needs; it allows diversification and provides flexibility.The customer's possibility to choose lowers the service firm's risk of failing to provide what the customer wants (Weng, 1999).Thus modularity makes it possible to develop and offer new services and is a suited means for shaping existing business models.Modularity further allows the recombination of service patterns and packages to develop new offerings and business models.
This stresses the importance that modularization has for service firms in an environment of higher uncertainty for BMI value generation and capture through reconfiguration of value-chain elements (Sánchez and Ricart, 2010;Amit and Zott, 2012;Simmons et al., 2013;Bouncken et al., 2015).Parente, Baack, and Hahn (2011) further point out that modularization drives the efficiency in the innovation process.We assume that modularity is a dynamic capability for developing new service products, because it allows the capture of value from new patterns and re-combinations of service-product components that emerge in a multifaceted and complex environment.Thus, we perceive modularity as a strategic means for creating a "built-in flexibility" (ElMaraghy, 2005) which is suitable for coping with uncertainty and fostering innovation of the business model.

Complementarity of EO and Modularity
Prior research recognized that firms achieve greater performance when they implement "classical" business innovations such as value creation, proposition, and capture by radical innovation at every level of the business model.Overall, BMI can inspire changes during a process of "business model reconfiguration" of only one or two domains of the business model.Although research still lacks a clear definition of the ambiguous and multifaceted concept of BMI (Amit and Zott, 2001;Al-Debei and Avison, 2010;Lecocq et al., 2010;Lambert and Davidson, 2013), we assume that radically innovative BMIs are subject to uncertainty and benefit from a combination of both dynamic capabilities discussed above.
EO includes proactive, risk-averse, and creativity-oriented thinking and behaviour (Vij and Bedi, 2012).Nonetheless, modularity requires fine-grained analyses, precise planning, and the implementation of formal architectures that consist of highly compatible business, process, or service-process modules that can be recombined without much change to another module.Whereas EO helps to explore new opportunities (Coulthard, 2007), modularity serves to exploit flexible product architectures in stages being near to the market (Sanchez and Mahoney, 1996;Buenstorf, 2005).The combination of EO and modularity by service firms has additional positive effects for BMI.The causal complexity inherent in one of the DCs increases with a further social and causal interaction of the DCs and allows the development of imitation barriers (Ethiraj and Levinthal, 2004;Hipp and Bouncken, 2009), creating a sustained competitive advantage (López, 2005).Thus, we argue that the capability to explore new opportunities with risk-affinity and innovativeness (EO) and the capability to reconfigure resources and existing capabilities (modularization) enables a firm to simultaneously explore and exploit new opportunities (O'Reilly and Tushman, 2009).This, in turn, will lead to a surplus of BMI as a precondition for competitive performance.
H3: Both modularity and entrepreneurial orientation enfold an additional joint effect on business model innovation.
The research framework, the hypotheses, and the suggested directions of the effects are summarized in Figure 1.

Sample characteristics and data collection
Our sample consists of 299 German service firms.The data were collected at two different trade fairs in 2013 and 2014.Potential respondents were chosen randomly.We then made an appointment with sales representatives or with a member of each firm's general management in preparation.Second, informants were contacted about four weeks following each trade fair.While first informants received a paper-and-pencil version of the questionnaire, data from second informants were gathered by computer-assisted telephone interviews (CATIs).
A total of 360 questionnaires were returned to us.After deleting outliers and duplicates, we removed all datasets with missing values on our dependent variable BMI.Finally, we had 299 evaluable questionnaires.We then analysed the dataset for missing values and found that there were no systematic missing data.Nevertheless, we account for 11.7 per cent of randomly missing values on our independent variables (or 35 cases, 264 remaining cases).Table 2 presents the descriptive statistics of our sample.
Table 2 further shows that nearly three of four respondents (74 per cent) belong to top, upper, and middle management.The average tenure of the surveyed managers is 6.8 years (median: 4.2 years).More than 80 per cent (84 per cent or 250 persons) are male.
The high level of internationality is a unique feature of our sample.More than 60 per cent (188 firms) are located outside Germany; nearly every third firm (31 per cent or 95 companies) comes from outside the European Union.The top five countries are Germany (112 firms), Italy (33 firms), China (30 firms), Turkey (15 firms), and the United States (12 firms).Table 3 summarizes the firms' origin.The high level of internationality is associated with a relatively high share of exports: the average service firm in our sample realizes 44 per cent (median: 43 per cent) of its turnover abroad.

Measurement
The variables in our study were measured on five-point Likert-type scales (1=''strongly disagree'', 5 =''strongly agree'').The scales were adopted from previously published articles, translated from English into German, and then translated back to check whether meaning and comprehensibility were well maintained.
Entrepreneurial orientation was measured as a second-order construct, consisting of three dimensions: risk-taking, proactivity, and innovativeness, following the approach of Eggers et al. (2013).The first-order constructs consisted of three items each, such as "We encourage people in our company to take risks with new ideas."(risk-taking), "We consistently look for new business opportunities."(proactivity), or "We consider ourselves as an innovative company."(innovativeness).Thresholds were acceptable for statistical criteria on the first-order level.Our self-developed scale to measure modularity consisted of three items and focused on modular products (Sanchez and Mahoney, 1996).We asked for the amount and functionality of modules and interfaces, and the general function of modularity on the product architecture.The scale was self-developed and adapted from ten to three items following a pre-test.
Technological uncertainty was measured with three items from Lewis' (2002) ten-item scale.We chose the three items with the highest factor loadings from this scale and asked, for example, for the level of "technological feasibility" of changes to existing products.
BMI refers to changes in how a firm creates, proposes, or captures value.The multi-dimensional construct was captured by an existing scale from innovation management (Bouncken et al., 2007) which was enlarged by two items regarding the market and the value added for customers.Overall, we captured all three dimensions of BMI with at least one item.
Competitive performance is a three-item scale measuring the performance of a firm compared with its competitors regarding sales volume, market share, and profitability (margin).The scale was taken from Venkatraman and Ramanujam (1986).
We used firm size as a control variable to account for the influence of a firm's size on our independent variables.The result of a confirmatory factor analysis is shown in Appendix 1.

Analysis
We used the co-variance-based structural equation modelling approach (SEM) in the software package Mplus 6 (Muthén and Muthén, 2012) to test our hypotheses.Compared with multivariate regression analysis, SEM has three major advantages: first, SEM uses latent (or unobserved) variables which account for the influence of residuals or measurement errors (Bagozzi and Yi, 2012); second, SEM is capable of incorporating latent variables as well as manifest variables simultaneously (Byrne, 2012); and third, SEM allows for path modeling which is of great importance for social sciences.
In a first step we validated our measurement model via a confirmatory factor analysis (CFA) with maximum likelihood estimation.The CFA resulted in an acceptable to good overall model fit: Despite a significant chi-square test (t=1.590***), the stricter and more appropriate fit-indices were all below their thresholds (CFI=.934,RMSEA=.044,SRMR=.068).On the construct-level we scrutinized the validity and reliability for each latent variable.Construct validity of the latent variables, expressed by the standardized factor loading of each indicator item, exceeded the minimum threshold of .5 (Hair et al., 2010) for all variables.To assess the convergent validity of our latent variables we applied Cronbach's α (Cronbach, 1951) and the average variance extracted.All constructs exceed the minimum thresholds of .7 and .5, respectively (Nunnally, 1978;Fornell and Larcker, 1981).We assessed the reliability of our latent variables on the item-level and on the construct-level.Indicator reliabilities are greater than .4for all indicator items except one from our BMI-scale, which slightly missed the threshold of .4(Bagozzi et al., 2004).But as the BMI-scale, like all other scales, showed a good reliability on the construct level with composite reliabilities greater than .6 (Bagozzi, 1981), we decided to keep this item in our measurement model.Finally, we tested whether the constructs met the Fornell-Larcker-criterion indicating that each latent variable's variance is better explained by its indicator items than by other constructs (Fornell and Larcker, 1981).None of our variables violated this assumption, which is expressed by a variable's ratio of the variance explained and its maximum correlation with other (latent) variables smaller or equal to one.Appendix 2 presents the results of our confirmatory factor analysis.
To cope with non-response bias we offered each firm a report of the study's results.We hoped to convince some initial non-respondents to support our study by providing this incentive to the participants.As we used single informants for our independent measures, common-method bias might have occurred.To account for this statistical problem we followed the recommendations of Podsakoff et al. (2003) using Harman's single-factor test.The computed factor explained less than 25 per cent (24 per cent) of the variance of all variables used in this study.Additionally, we distributed a short questionnaire containing our most important dependent variables (e.g.firm performance) and some questions about the firms' structure to second informants from the same company.We used first informants' data for our independent variables EO, modularity, and uncertainty.Data from second informants were chosen for our dependent variables BMI and competitive performance.We also analysed the correlations between first and second informants' responses and found medium to high correlations for every item (e.g.firm size .963***).Overall, common-method bias does not seem to be an issue in our data.
Our hypotheses were tested with two different SEMs.Our first model comprised hypotheses one and two and was compiled as a path model with indirect effects to test our mediation hypotheses.The model revealed good overall fit (cf.table 3).The second model contained the moderation effect in hypothesis three.We tested whether the interaction term of the two capabilities EO and modularity yields to a significant increase in the explained variance of our dependent variable BMI (Muthén and Muthén, 2012).All results are presented below.

RESULTS
The correlation matrix of the constructs shows the positive relation between competitive performance and business model innovation.Although not explicitly hypothesized, we find that competitive performance is positively associated to BMI, EO, and modularity (cf.table 4).We also find a medium correlation between EO and modularity (r=.489***), indicating that deployment of the explorative EO is accompanied by exploitative modularity.Interestingly, the control variable 'firm size' does not correlate significantly with any other variable of our model.We can therefore assume that BMI, EO, or modularity is not a matter of size.

DISCUSSION
This paper was motivated by the recently increasing attention paid to business model innovation and the concept's still immature theoretical basis, particularly for service firms that need dynamic capabilities to achieve BMI.
As previous research already hints at a relation to the dynamic capabilityview, we aim to explain capabilities that drive BMI by drawing upon DCV.Research so far has not analysed how dynamic capabilities affect business model innovation.Thus this paper looks how dynamic capabilities such as entrepreneurial orientation and modularity jointly influence BMI in service firms.

Contribution to Theory
The results contribute to service research and to BMI research.First we enlarge the research body of BMI for service firms.Specifically, we show that service firms can take advantage of dynamic capabilities: each of the dynamic capabilities (EO and modularity) enhances BMI in service firms.We suggest that these important triggers serve as antecedents of BMI and should be included in future research on BMI.Moreover, we support previous BMI research that sees a link between dynamic capabilities and BMI.Further, uncertainty is not a thread but a trigger for BMI, especially when combined with EO.In line with Andries and Debackere (2013), we find that the capability to reconfigure resources and existing capabilities through modularization positively influences BMI but does not mediate the effect of technological uncertainty.Even though the risk-prone and creativity-oriented concept of EO is antithetical to the precise planning of modularization embedded in the service firms' process and product architectures, the combination of both does not lead to additional benefits for BMI.However, the central challenge for a successful implementation of BMI to use exploration capabilities such as EO and the ability to exploit new ideas through modularization was not supported by our data.Further we have shown that national culture shapes the impact of EO and modularity on BMI but does not differ significantly between Germany and all other nations.

Limitations and Further Research
Fundamentally, there is strong agreement among scholars that empirical research on dynamic capabilities should be intensified (Eriksson, 2014;McAdam et al., 2014).We believe that further research should analyse the connection between BMI and dynamic capability in particular.While qualitative studies allow a better understanding of processes and managerial and environmental issues, as well as reconfiguration mechanisms, further quantitative research could offer a more concrete identification of the factors involved, their nature, and their interactions (Easterby-Smith et al., 2009).
We did not find a combinative effect of EO and modularization -two somehow antithetical capabilities.Thus, the question arises whether other ambidextrous capabilities may foster BMI.Ambidexterity, the combination of two antithetical activities, is stressed as a driver of product innovation.Does this apply to BMI in service firms as well?
We processed our research with an international sample and controlled for country effect by differentiating a German-only and an international-only model.We did not find great differences between the two models.In contrast to research on the link between culture and firms' innovativeness (Taylor and Wilson, 2012;Černe et al., 2013), we cannot support a significant influence of national culture on BMI.Further research should be conducted to analyse whether our findings are industry-specific or whether the objects of innovation (products, processes, and business models) differ more fundamentally.Finally, not only does research on BMI lack a clear and accepted definition of the construct; established scales do not exist.We would like to encourage others to overcome this obstacle.
As the existence of time was outlined above as a constituent element of any form of dynamic concept, there is a strong requirement for further multiperiodical or longitudinal research to gain a better understanding of dynamic capabilities.Following previous research (Zahra et al., 2006), further investigations are needed regarding how dynamic capabilities differ between entrepreneurial, rather small firms and larger, saturated companies.This research gap gains even more importance for dynamic capabilities underlying BMI.

Managerial Implications
Backed by our results, the promotion of proactive and risky innovative behaviour of employees should be encouraged by managers.EO is a suitable means for contributing new ideas to firms' existing business models.If new ideas or concepts for creating and monetising value are born, they have to be implemented and "translated" into products before being marketed.However, we propose that EO and modularity be pursued consecutively and not simultaneously as hypothesized.BMI is most effective if exploration and exploitation occur together but supposedly one after another.Bagozzi and Baumgartner (1994).c Nunnally (1978).d Bagozzi and Yi (1988).e Fornell and Larcker (1981).

Figure 1 :
Figure 1: Proposed research model (with hypothesized direction of effects)

Table 1 :
Classification of DCs