Influence of relationship networks on the internationalization process: the moderating effect of born global.

For some decades researchers have studied the internationalization strategy of businesses from different perspectives and using various theories. However, the literature on internationalization is fragmented and further studies are needed to analyze holistically and quantitatively how different types of companies develop their internationalization strategies. In this study we intend to analyze holistically how relationship networks affect the internationalization strategies adopted by businesses in the wine sector. We use PLS-SEM and multigroup analysis to analyze the moderating effect that being a born global company can have on the influence of relationship networks in gradually developing an internationalization strategy. The results show that some factors, such as bargaining power, number of distributors, the variety of distribution channels, company prestige, and brand awareness positively influence the gradualness of the internationalization strategies of all companies in the sector, while relationships with clients and national and international competitors have a negative influence on the gradualness of the development of the strategies of businesses characterized as born global.

Introduction

Internationalization strategies have been studied for some decades. However, few studies have approached the various theories from a holistic point of view and analyzed company strategies quantitatively and in depth. At present there is a confrontation between the gradualist Uppsala model and the whole paradigm surrounding born global businesses (Love et al., 2016). Although the Uppsala School's own researchers developed a new model (Johanson and Vahlne, 2009) to incorporate the importance of relationship networks to explain the early internationalization of born global companies, the majority of studies consider the theories as opposed (Casillas et al., 2015).

The networking concept was introduced into the new Uppsala internationalization model to explain how these new, born global companies internationalized despite their lack of experiential knowledge. In the original model this lack of knowledge hindered the explanation of the internationalization process, given that experience is one of the main variables that explain success in internationalization. However, the inclusion of relationship networks in the model has not been validated empirically, nor has there been an analysis of whether these relationship networks are important also for traditional companies.

Born global companies are defined as enterprises that internationalize their activities within three years of their founding and which have 25% of their total sales abroad (Knight and Cavusgil, 2004). This characterization takes into account the age of companies when they enter the international market and their percentage of foreign sales. It does not, however, take into account scope, such as the number of countries in which they operate, or how they evolve after they internationalize (Romanello and Chiarvesio, 2017) and if they develop gradually, as do traditional companies.

Thus, although the internationalization process has been studied since the 1970s, there is still a research gap (Musso and Francioni, 2015). There is a need for more research into born global companies (Ughetto, 2016), especially with respect to the development of explanatory models and theoretical perspectives (Knight and Liesch, 2016). Born global companies have attracted much attention in recent years (Braunerhjelm and Halldin, 2019; Falahat et al., 2018), but most studies focus only on the initial stages of internationalization; very few focus on its later stages (Jones et al., 2011; Gallego and Casillas, 2014). The whole growth process of born global companies has not, therefore, been analyzed in depth or with quantitative methods (Ughetto, 2016; Johanson and Martín, 2015); this represents a gap in the born global empirical research (Zander et al., 2015; Braunerhjelm and Halldin, 2019).

The objective of this work is, on the one hand, to bridge the gap in the literature on the understanding of the main existing theories, in a holistic and complementary way and, on the other, to analyze quantitatively the effect that relationship networks have on both traditional and born global companies. The analysis takes into account both the characterization of the companies as born global or traditional and measures the process that they followed in their entry into the international market to determine if it was gradual or not, regardless of whether they are defined as born global or traditional.

In summary, our main contribution is the two new insights we make into the current research gaps. First, as few studies have examined the Uppsala theory and the born global approach holistically, and the majority of studies consider the theories as being opposed, the research remains fragmented (Cavusgil and Knight, 2015; Love et al., 2016; Casillas et al., 2015; De Clerq et al., 2005; Keupp and Gassmann, 2009; Jones et al., 2011). Thus, the present work proposes a model which uses measures of gradualism and born global to show that: (1) relationship networks, which are normally considered as valuable resources with which to start early internationalization (i.e., solely for born global companies) and (2) gradualism, the approach normally taken by traditional companies, are complementary and together can explain the internationalization process for both born global and traditional companies.

Second, more quantitative research is needed into born global companies because many previous studies are exploratory and descriptive and focus only on the initial stages of internationalization (Ughetto, 2016; Johanson and Martin, 2015; Zander et al., 2015; Braunerhjelm and Halldin, 2019; Jones et al., 2011; Gallego and Casillas, 2014; Knight and Liesch, 2016). The present study, which uses the PLS-SEM/MG quantitative technique, provides a rigorous empirical analysis of: (1) the effect that different kinds of relationship networks have, from a process viewpoint (not only the initial stages), on internationalization, and; (2) if there are differences between the effect on born global and traditional companies. The empirically validated model can be used in other contexts.

In the next section we provide the frameworks of the various recent theories of internationalization. Thereafter, we describe the methodology used for the analysis and present the results. In the final section we present a discussion of the results, the limitations and future study and, the main conclusions.

Theory

Throughout its history the Uppsala, or gradual, model (Johanson and Vahlne, 1977) has been regarded as one of the best means of explaining companies’ internationalization processes (Oviatt and McDougall, 1999); it one of the most cited references in the international business literature (Johanson and Vahlne, 2009) and is the most durable (Knight and Liesch, 2016). During the 1970s researchers from the Scandinavian, or Uppsala, school, undertook various studies to obtain empirical evidence of the processes followed by companies in their internationalization strategies. The results of these studies (Johanson and Wiedherseim-Paul, 1975; Johanson and Vahlne, 1977) showed that companies developed their foreign activities gradually, and that their basic pillars were knowledge of the markets and the level of commitment of the resources and capabilities of the company.

The Uppsala gradualist model proposes that, when companies launch their internationalization strategies, as they lack experience and accurate information, they will start with sporadic, irregular export activities, with only a small commitment of resources. As the company acquires experience and knowledge, it will move on to further stages with different market entry modes, and greater international involvement, that will require increased commitment of company resources (Johanson and Wiedersheim-Paul, 1975).

The theory posits that companies will operate within their own national markets before moving into the international market (Knight and Liesch, 2016). They will, thereafter, begin their internationalization in countries at only small psychic distance from their own before moving into countries psychically more distant (O'grady and Lane, 1996), as entering countries that are psychically closer reduces uncertainty (Johanson and Vahlne, 1990; Leonidou and Katsikeas, 1996).

Later, coinciding with the globalization phenomenon, internationalization behaviors or processes were observed in companies that diverged to a great extent from the gradualist model. A new paradigm arose around a new typology of companies, known as international new ventures, born global, global start-ups, etc., that are born internationalized (Knight and Cavusgil, 1996; Oviatt and McDougall, 1994).

Until recent times, gradualist approaches to the development of the internationalization of companies assumed that the knowledge acquired from their activities in foreign markets was the main explanatory variable of the success of the internationalization (Johanson and Wiedersheim-Paul, 1975). However, born global companies operate in the international market at early stages of their development and, a priori, lack foreign experience and are often characterized by a shortage of resources (Cavusgil and Knight, 2009). Thus, the assumptions of the original Uppsala model are called into question, as there is evidence that the gradual approach has not been followed by all companies (Oviatt and McDougall, 1994; Gallego and Casillas, 2014).

Since the emergence of these type of companies numerous studies have been undertaken to determine the factors that allowed them to internationalize quickly and early (Aspelund et al., 2007; Romanello and Chiarvesio, 2017). However, many of the studies are exploratory and descriptive (Knight and Liesch, 2016) and the results are not consistent. Controversies have arisen over the difficulty of classifying the factors that explain why these companies developed in this way (Jones et al., 2011; Romanello and Chiarvesio, 2017; Casillas et al., 2015). Many of the studies are inconclusive and contradictory (Knight and Liesch, 2016) and the theory and research remains underdeveloped and fragmented (De Clerq et al., 2005; Keupp and Gassmann, 2009; Jones et al., 2011).

Johanson and Mattsson (1988) introduced the networking concept, which argues that the success of companies in entering into international markets cannot be understood only in terms of the companies themselves; consideration must be given to the environments in which they operate and to their networks. The internationalization trajectory of companies is not the result of only their own efforts but also of their relations with other companies in their environments (Cavusgil and Knight, 2015). In short, companies learn from their relationships, through interaction with other agents, about their needs, resources and even strategies and business contexts, which allows them to enter new countries and develop new relationships to enter other countries (Sharma and Johanson, 1987; Johanson and Vahlne, 2006). Thus, the knowledge that born global companies lack is obtained through interaction with other companies in their networks (Johanson and Vahlne, 1990, 2003). It has been observed that relational resources or relationship networks have begun to have importance in internationalization strategies.

Various authors have noted that these relationship networks generate social capital for new international companies (Andersson and Wictor, 2003; Arenius, 2002; Autio, 2005; McDougall and Oviatt, 2003; Sharma and Blomstermo, 2003). The networks facilitate the acquisition and the use of resources for early internationalization. Thus, an increase in the company's social capital provides better access to international business opportunities (Arenius, 2002). This also provides invaluable sources of information about foreign markets (Ellis, 2000; Sharma and Blomstermo, 2003), which helps to overcome barriers to export (Ghauri et al., 2003; Ibeh and Kasem, 2011). In short, networks have been shown to be critical assets in the birth and development of born global companies (Andersson and Wictor, 2003; Arenius, 2002; Oviatt and McDougall, 1994) and, as Yli-Renko et al. (2002) noted, external social capital in the form of relationship networks positively impacts on their knowledge of foreign markets and, therefore, their rapid internationalization.

Some studies have shown the importance of networks to born global companies as mechanisms of entry and development in international markets (Bell, 1995; Coviello and Munro, 1995; Evangelista, 2005), for staff recruitment (Evangelista, 2005), for gaining the financial resources to support the development of new products (Coviello and Munro, 1997), and for accessing resources that they lacked; thus, these networks give companies competitive advantage (Coviello and Cox, 2006).

Uppsala school researchers subsequently developed a revised model to incorporate the networking concept (Johanson and Vahlne, 2009). In the new model they established that a framework of inter-company interactions is essential for the acquisition of the knowledge needed to internationalize, and that the commitment that companies dedicate to the market will be based on the position that they occupy in the network. Similarly, in this new model, the learning exchanged, and the trust created between the companies in the network, are the main determining factors of the knowledge acquisition and the decisions that the new company makes as to the level of commitment it makes to its relationships within the network. Rapid internationalization can be explained in this new model by the presence in the company of significant relationship resources.

According to the network internationalization model, a company might have relationships with various agents, including customers, distributors, suppliers, competitors, nonprofit organizations, public administrations, etc (Ojala, 2009). Compared to the traditional internationalization model, which focuses mainly on the international market and modes of entry, the network approach focuses more on the impact of relationship networks on market entry decisions (Musso and Francioni, 2015).

The born global phenomenon, thus, is conceptualized by the establishment of more explicit relationships than the original Uppsala model and other, more recent, theoretical approaches, such as the network approach applied to company internationalization. This new internationalization model, which takes into account relationship networks, accords with the new born global company paradigm (Madsen and Servais, 1997). However, some authors have established that, despite the fact that born global companies enter the international market at an early stage, and that the process can be characterized by an increase in the commitment made to foreign markets, following evolutionary or gradual steps (Romanello and Chiarvieso, 2017; Gabrielsson et al., 2008), the speed of the process may not accord with the predictions of the Uppsala theory (Hashai and Almor, 2004), as these companies tend to internationalize faster than traditional models (Knight and Liesch, 2016; Hilmersson and Johanson, 2016), and even that the process may not be at all gradual (Hashai and Almor, 2004).

Thus, most of the existing international business literature contrasts the process, or stage model, and the born global approaches (Cavusgil and Knight, 2015; Love et al., 2016), despite the fact that both models have much in common (Prashantham, 2005; Zhou, 2007; Casillas et al., 2015) and can be complementary (Bai et al., 2017; Casillas et al., 2009; Chetty et al., 2014; Johanson and Martín, 2015). The main objective of the present study, therefore, is to consider the main theories on internationalization holistically. This will allow us to analyze the processes followed by companies entering the international market, their characterization as born global and the effect that relationship networks have on the gradualness of the process, in an independent way, for both born global and traditional companies. Thus, our research questions are:

Q1: What networks have a significant effect on the gradualness of the internationalization process?

Q2: How can there statistically significant differences in the effect that relationship networks have on the gradual process of internationalization of born global and traditional companies?

Methodology

Sample and data

The population under study is the Spanish wine sector, due to its economic importance for the country and its global significance. Spain, with 0.98 million hectares of grape production, internationally has the largest area of vineyards, followed by China and France (OIV, 2017). It is also the leading global exporter of wine by volume. To obtain the population under study (Spanish wineries) we used the SABI database (Iberian Balance Analysis System). This is a directory with contact data and financial information on more than 1,080,000 Spanish companies, classified by CNAE code (National Classification of Economic Activities). This gave us 640 Spanish companies assigned to CNAE code 11.02: wine-making process, which formed the population under study.

The data gathering instrument was a questionnaire. We attempted to survey key informants in all the 640 listed wineries. The respondents were senior export managers. The survey was conducted first by email, later supplemented by a telephone survey. The questions were identical in both surveys. The number of valid questionnaires returned was 185, a response rate of 28.9%. The sample has a confidence level of 95%, which we consider adequate for the study.

Consistent with the information available from the Spanish Wine Federation and ICEX, the sample's main international markets are primarily in the European Union (64.79% of the exporting wineries) and America (52.57% of the exporting wineries), leaving Asia (37.94% of the exporting wineries) and Oceania (7.60% of the exporting wineries) far behind. The average age of the wineries in the sample is 26.63 years, with an average international experience of 13.78 years. The descriptive statistical values of the study variables related to relational resources are shown below (Table1):

Table 1

Descriptive statistics.

Variable	Mean	S.D
Prestige of the company	3.72	0.96
Brand awareness	2.97	1.20
Negotiating power with distributors	2.91	1.11
Number of distributors	2.97	1.19
Variety of channels	2.78	1.31
Relations with national clients and competitors	4.06	0.80
Relations with international clients and competitors	3.76	0.88

Finally, to test the reliability of the sample and eliminate potential non-response bias, we applied the test suggested by Armstrong and Overton (1997). We performed a variance analysis between the answers given to the email and telephone questionnaires, obtaining a p-value above 0.05. This determined that there were no significant differences in the items of the two groups of questionnaires, which confirms that the data obtained do not have non-response bias, or bias due to conditioned responses given as a consequence of the data gathering method.

Operationalization of the variables

To measure relationship resources, or networks, we adopted variables previously used by other authors (e.g., Chander and Vishakha, 2011; Clavel et al., 2017). These variables are intended to capture, on the one hand, the relationships with the three principal agents with which companies interact in the market and, on the other, the position that the company establishes in its relationship networks.

To measure the gradualness of the companies' internationalization processes we used Clavel San Emeterio et al. (2018)'s methodology. This uses an indicator (quantitative measure) that implicitly incorporates psychic distance from the host country to other countries, using the variables “export width” and “export depth”, to obtain the international priority index, which describes the gradualness of a company's internationalization process, regardless of whether it is born global or traditional.

Finally, to measure the born global concept we must first define the born global variable. This variable dichotomously classifies between born global companies and traditional companies. Knight and Cavusgil (2004)'s characterization is the most used by researchers (Kuivalainen et al., 2007) and is consistent with the definition in the literature (e.g., Cavusgil and Knight, 2015; Gonzalez-Perez et al., 2016; Gerschewski et al., 2016). Thus, we identify born global companies as those with a ratio of foreign sales exceeding 25% of total company sales and which developed their international activities within three years of their founding.

Reliability and validity of the scales

Having defined the measurement variables, an exploratory factor analysis was carried out, using SPSS 24.0, for the relationship resource measurement variables. This resulted (Table 2) in two factors (relationship resources 1 and relationship resources 2), whose factor loads in all cases exceeded the minimum required value of 0.7 (Hair et al., 2010) and explained, together, 59.71% of the variance. The KMO statistic value was 0.760 (higher than the minimum required) and the Bartlett sphericity test, analyzed by the Chi-square statistic, had a value less than 0.05, so we can affirm that the analyzed data are correct and the variables introduced are correlated. Therefore, the model has appropriate goodness of fit.

Table 2

Definition of the variables.

Factor	Indicator	Variable
F1. Relationship resources 1	Q1	Prestige of the company
	Q2	Brand awareness
	Q3	Negotiating power with distributors
	Q4	Number of distributors
	Q5	Variety of channels
F2. Relationship resources 2	Q6	Relations with national clients and competitors
	Q7	Relations with international clients and competitors
F3. Gradualness	Q8	International priority index
F4. Born Global	Q9	Definition of born global

We next applied PLS-SEM covariance analysis methodology (Partial Least Squares - Structural Equation Modeling), using SmartPLS 3.0, to carry out a confirmatory factorial analysis of the dimensions, to validate and analyze the theoretical model and to perform a multigroup analysis (MGA) (see Figure 1). PLS-SEM is a structural equation modeling technique based on variance (Hair et al., 2012); it is a useful tool for theorizing in management research in general (Hair et al., 2018). The use of PLS-SEM in the present study is supported by the research objectives, the non-normal distribution of some indicators and the sample size (Hair et al., 2012, 2014; Blasco-Lopez et al., 2019; Chien and Chi, 2019). In addition, using this nonparametric method technique is very appropriate, first, because it allows us to simultaneously analyze all the inter-variable relationships, including measurement and structural components, in a conceptual model (Henseler et al., 2016) and, second, it is highly suitable for MGAs (Hair et al., 2014; Henseler et al., 2016; Sarstedt et al., 2011).

Figure 1

Specific analysis model.

We used various evaluation methods to measure the reliability and convergent validity of the measurement scale, among them Cronbach's alpha, the composite reliability index and average variance extracted. As can be seen in Table 3, the constructs show a high degree of internal consistency. For all factors, the Cronbach's alpha value exceeded the recommended level of 0.7 (Nunnally and Bernstein, 1994). All the constructs had composite reliability statistics higher than 0.6 (Bagozzi and Yi, 1988) and an average variance extracted higher than 0.5 (Fornell and Larcker, 1981; Hair et al., 2010).

Table 3

Confirmatory factorial analysis.

Factor	Indicator	Load	t-value	C.A	CR	AVE
F1	Q1	0.7***	11.06	0.8	0.9	0.5
	Q2	0.8***	18.63
	Q3	0.8***	16.28
	Q4	0.8***	16.57
	Q5	0.7***	10.26
F2	Q6	0.9***	4.16	0.7	0.9	0.7
	Q7	0.9***	3.80
F3	Q8	1***	-	1	1	1

***p < 0.01.

The significance of the loads was measured using a bootstrap resampling procedure (5,000 subsamples of the original sample size). As shown in Table 3, all the items are significantly related to each of their factors (p < 0.01), with loads greater than 0.7 (Hair et al., 2010).

The Fornell-Lacker criterion and the ratio between the Heterotrait-Monotrait correlations were used to confirm discriminant validity (Henseler et al., 2015). As can be seen below (Table 4), all AVE square root values exceed the inter-factor correlations and the Heterotrait-Monotrait ratio has values < 1, which confirms the discriminant validity of the measurement model.

Table 4

Discriminant validity.

Factor	F1	F2	F3
F1. Relationship resources 1	0.73	0.30	0.37
F2. Relationship resources 2	0.22	0.86	0.09
F3. Gradualness	0.34	0.08	1

Values on the main diagonal represent the square root of the AVE.

Values below the diagonal represent inter-factor correlations.

Values above the main diagonal: Ratio HTMT.

Results

The estimation of the structural model was made using the bootstrap resampling procedure (5,000 subsamples of the original sample size). As shown in Table 5, the model has satisfactory explanatory capacity, measured through variance explained (R2), which here is higher than the minimum value of 10% suggested by Falk and Miller (1992). To analyze predictive relevance, the Stone-Geisser test (Stone, 1974; Geisser, 1975) was used, with a blindfolding procedure. A value of Q2 > 0 was obtained for both dependent variables, which allows us to affirm that the model has predictive relevance.

Table 5

Analysis of the global model.

Hypothesis	Standardized Beta	t value (bootstrap)
H1: Relationship Resources 1 → Gradualness	0.30***	5.13
H2: Relationship Resources 2 → Gradualness	-0.02	0.41

R2 (Gradualness) = 0.12.

Q2 (Gradualness) = 0.08.

***p < 0.01.

The results show (Table 5) that there is a statistically significant relationship between relationship resource factor 1 (β = 0.30, p < 0.01) and the gradualness index, but that there is not a statistically significant relationship (β = -0.02; p > 0.01) between relationship resource factor 2 and the gradualness index.

We posit that there are differences in the effect that networks have on the gradualness of internationalization based on whether companies are characterized as born global, or not born global. To analyze this moderator effect between the two groups we carried out a multigroup analysis, using SMART PLS (Hair et al., 2014), as multigroup analysis is a special case of moderation in which the moderator variable is categorical (Henseler et al., 2012), as in this case, in which the moderator variable represents the condition of the company as born global company or traditional.

Results of the multigroup analysis

We used Henseler's MGA (Henseler et al., 2009) nonparametric approach in the analysis, as previous research has shown that parametric approaches can lead to type I errors (Hair et al., 2018).

The results (Tables 6 and 7) show that there is a statistically significant relationship between relationship resources factor 1 (β = 0.36, p < 0.01) for both types of companies and the gradualness index, while relationship resources factor 2 shows a statistically significant relationship (β = -0.25, p < 0.05) only for born global companies.

Table 6

Test of the multigroup hypothesis.

	Standardized Beta		t value (bootstrap)
Hypothesis	BG	NBG	BG	NBG
H1: Relationship Resources 1 → Gradualness	0.36***	0.36***	2.52	5.02
H2: Relationship Resources 2 → Gradualness	-0.25**	0.09	1.86	1.41

***p < 0.01.

**p < 0.05.

Table 7

Differences between groups.

Hypothesis	Path Coefficients - diff	t value (bootstrap)
H1: Relationship Resources 1 → Gradualness	0.00	0.01
H2: Relationship Resources 2 → Gradualness	0.34***	2.42

***p < 0.01.

Discussion

The research into born global companies has attracted great interest in recent years (Keupp and Gassmann, 2009), particularly because it has identified a factor that the dominant theories and paradigms in the field of international business have been unable to explain (Gallego and Casillas, 2014).

Although the Uppsala school model was modified to incorporate the networking concept (Johanson and Vahlne, 2009), some studies have argued that the Uppsala and the born global theories are opposed (Cavusgil and Knight, 2015; Love et al., 2016). However, the results of the present study show that our model is empirically valid for both born global and traditional companies. As the results show that these types of relationships (company prestige, brand awareness, negotiating power, number and variety of distribution channels) are significant for both company forms, it can be argued that the two theories are complementary (Casillas et al., 2015; Bai et al., 2017).

As to the analysis of the effect of relations with clients and national and international competitors (relationship resources 2), in contrast to the findings of other authors (Yeoh, 2004; Musteen et al., 2014), we observed that when the companies were analyzed independent of whether they were born global or traditional, that these relationships did not have a significant effect. However, when we carried out a differential analysis of the model for those companies characterized as born global and traditional, we observed that there were statistically significant differences. These results complement those obtained by Musteen et al. (2014), who established that international networks facilitate rapid internationalization.

These results are, again, consistent with the revised Uppsala model (Johanson and Vahlne, 2009), which incorporated the networking concept to explain the gradualness of the internationalization of born global companies. However, we must bear in mind that these relationships with competitors and customers are valid for explaining the gradualness of the internationalization of only born global companies.

Gabrielsson and Kirpalani (2012) established that the born global approach has neither been used in conventional models nor been empirically tested (Braunerhjelm and Halldin, 2019). However, the fact that the effect of these relationships is not significant for traditional companies means that the model defined by Johanson and Vahlne (2009), that incorporated the networking concept, cannot be applied to these companies with respect to relations with national and international competitors and customers.

The results of this work demonstrate empirically that it is necessary to incorporate the networking concept into internationalization models to analyze the process of both born global and traditional companies. In addition, the present research showed that the most commonly-used theories in the international business literature (born global and gradualist models) should not be seen to be in opposition, as they can be complementary, in as much that born global companies can develop their internationalization processes gradually, even if their first foreign moves are made at early stage in their overall development.

However, we should not incorporate all the relationships for the born global companies, as the relationships that mainly ensure that the internationalization process is less gradual for born global companies are with clients and national and international competitors, relations that for traditional companies do not have a significant effect on the speed of the process. On the contrary, factors such as brand awareness, company prestige and relationships with distributors, are necessary for both born global and traditional companies if they want to internationalize gradually.

Limitations and future study

The present study has two main limitations, which in turn provide opportunities for future research lines and that must be taken into account when evaluating our conclusions. First, our sample included only Spanish wineries, so an important research opportunity exists to apply the empirically validated model in other countries and sectors. Second, the study used a cross-sectional research design focused on a given moment in time with enterprises operating in different export stages or with different numbers of years of international experience, thus no longitudinal analysis was performed. Future studies might apply longitudinal designs to illustrate the dynamics of exporting.

Conclusions

We showed that, in our sample of wine sector companies, the relationships between brand awareness, company prestige, the bargaining power with distributors, the number of distributors and the variety of distribution channels (relationship resources 1) and gradualness are positive for both born global and traditional companies. These results suggest that the more of these resources that companies possess, the more gradual will be their entry into international markets.

No significant differences were observed for relationship resources 1 in the segmented analysis of companies characterized as born global or traditional. This result implies that, although the presence of these type of resources will involve greater gradualness in the internationalization process, the impact they have is not different between born global and traditional companies. Thus, if a company's strategy is to internationalize gradually, it must increase these type of resources regardless of when it intends to internationalize.

For traditional companies, that is, those that have not internationalized three years after their founding, and/or with foreign sales below 25% of total sales, the impact that relations with clients and national and international competitors have on their gradual internationalization is not significant. However, for born global companies, the results show that the impact that relations with clients and national and international competitors have on their gradual internationalization is statistically significantly negative. This result supports that, where relationships with customers and competitors, both national and international, are extensive, the internationalization process is less gradual for born global companies. Regardless of when the internationalization takes place, which, by definition, will be within three years, the process that companies without this type of relationship networks will follow will be gradual.

Declarations

Author contribution statement

Mónica Clavel San Emeterio, Emma Juaneda-Ayensa, Rubén Fernandez-Ortiz: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.