The typology of entrepreneurial exporters: has it all been said? An empirical approach using latent class segmentations

Abstract The main aim of this work is to close the research gap that is the lack of a dominant typology of exporting firms in the area of international entrepreneurship (IE). It combines the principles of gradual or stage approaches, scale and scope, with those associated with international entrepreneurial behaviours: time to internationalisation and speed. Latent class segmentation is achieved using a sample of 136 Spanish exporting firms to uncover the nature and characteristics of the different exporters. The results enable us to identify four groups of exporters: true-born global exporters, early international and market concentration exporters, traditional exporters, and true born-again globals. The differences between these four groups are the earliness of their internationalisation, their pace of international expansion, their development of strategies for diversification or market concentration, and their level of international propensity. The results produce both academic and managerial contributions to the field of export activities.


Introduction
The literature on international businesses has long sought to uncover the nature and characteristics of the organisations that compete internationally (Coudounaris, 2018;McDougall et al., 2003;Vissak et al., 2018). But much work remains to be done in this area as there is yet no dominant, widely accepted typology (Vissak & Masso, 2015). One reason for this is the widespread adoption of a partial approach to the conception of the processes of internationalisation, one which regards the gradual approaches and those associated with IE as incompatible (Uppsala and innovation) (e.g., Johanson & Vahlne, 1977, 1990Johanson & Wiedersheim-Paul, 1975;Reid, 1981) (e.g., Cavusgil & Knight, 2015;Coviello, 2015;Oviatt & McDougall, 1994, 2005. Another reason is that the four variables currently regarded as the basic factors that distinguish between types of international firmstime to internationalisation, speed, scope and scalehave been treated as separate phenomena (Kuivalainen et al., 2012). The motivation for the present work is to help to bridge this research gap by linking them together.
The main aim is to offer a typology of exporting firms that combines the principles of the gradual or stage approaches (scale and scope) with those associated with international entrepreneurial behaviours (time to internationalisation and speed). We also incorporate other variables that have been used in previous studies to typify international firms. These include age (e.g., Baum et al., 2011;Johanson & Mart ın, 2015;Knight & Cavusgil, 2004), size (e.g. Aspelund & Moen, 2005;Kuivalainen et al., 2007;Maciejewski & Wach, 2019), international experience (Baum et al., 2011;Johanson & Mart ın, 2015;Kuivalainen et al., 2007), having an export department (e.g., Navarro-Garc ıa, 2016) and being part of an international network (e.g., Baum et al., 2011). In order to do so, we employed a sample of exporting firms located in Andalusia, a region of Spain, using latent class segmentation as a typification tool.
The present paper makes two main contributions. The first is that the degree to which exporting firms can be typified as internationally entrepreneurial depends on their having previously developed rapid internationalisation processes that are wideranging and/or large-scale. We therefore classify exporting firms into four groups.
(a) The first is 'True-born global exporters'; they are entrepreneurial exporters par excellence, firms centred exclusively on exporting as an entry mode abroad. Their internationalisation was early ( 3 years), fast (entry into 1-2 countries per year), high-scope (ratio international sales/total sales >50%) and broad (>15 countries).
(b) The second is 'Early internationals and market concentration exporters', which is a mixed category. They have achieved an international propensity of over 25% and they were early in beginning their internationalisation processes ( 4 years) and their market concentration strategies. But their scope is narrow (<10 countries) and the pace of their international expansion is slow.
(c) The third is 'traditional exporters': conservative firms that are not very internationally entrepreneurial. They started to internationalise over 25 years ago, but their scope is small (<10 countries) and the speed and scale of their internationalisation are low ( 20%).
(d) True born-again globals internationalise very late (>30 years), but when they go international they do so with a global orientation, very quickly (3-4 countries per year) and achieve an international propensity of over 65% and wide international scope (>35 countries). Also, true born-again globals coordinate the other activities of the value chain as they undertake foreign direct investment (FDI) in the form of business or production subsidiaries.
The second contribution is methodological. This work uses latent class segmentation to identify different types of exporting firms, whereas previous studies have used other, less robust methods, which casts doubt on the results obtained.
The following section reviews the most relevant literature in order to determine the typology and denominations used for international firms in previous studies. Then, the key dimensions are presentedtime to internationalisation, speed, scope and scalewhich must be considered in identifying international entrepreneurial behaviours. Next, the methodology is set out, explaining the sample, the data collection, the measurement scales and the tools used: latent class segmentation. The last section includes the results, the discussion, the main conclusions and, finally, the limitations and suggested lines of future research. Table 1 presents a summary of the main approaches considered in the identification of international company typologies. The first works that offered a typology of international firms accepted the gradual, or stage, internationalisation model (Uppsala approach; innovation model). They regarded a firm's ratio of international sales to total sales as the most critical relative measure of its internationalisation (Vissak et al., 2018). They established three groups of exporting firms, broadly defined: experimental exporters, whose international propensity is <10%, active exporters, with an international propensity between 10 and 40% and committed exporters, with an international propensity >40%.

Theoretical framework
The seminal article of Oviatt and McDougall (1994) changed the paradigm in the typology of international firms. It questioned the foundations of gradual approaches Table 1. Approaches used in the identification of international company typologies.

Approach Main contributions
Uppsala model; innovation model Cavusgil (1980, 1984) Czinkota (1982 Czinkota and Johnston (1981) Gankema et al. (1997) Lukason and Vissak (2017) Samiee and Walters (1991) Accepts the gradual or stage internationalisation model, considering the ratio of international sales to total sales as the primary variable in determining the stage of internationalisation a firm has reached.  Oviatt and McDougall (1994) typology of international firms also includes the international scope (number of countries) and the coordination of distinct value-chain activities.
Born Globals: BGs; the traditional approach Andersson and Wictor (2003) Aspelund and Moen (2005) Autio and Sapienza (2000) Cavusgil and Knight (2015) Chetty and Campbell-Hunt (2004) Coviello (2015) Freeman and Cavusgil (2007) Knight and Cavusgil (1996; Rasmussen and Madsen (2002) The identification of BGs includes both the time to internationalisation and the degree of internationalisation. BGs have a high degree of internationalisation from early on. There are different types of BG. For a correct typification of BGs, we must include the following variables: time to internationalisation, speed, scope and scale.
by introducing time as a variable, distinguishing between organisations that went international at, or very soon after, their inception from those that did not. The term 'international new ventures', INVs, was coined: organisations which from very early on have oriented themselves towards foreign markets in search of new opportunities. They need fast growth and are willing to take on the risks and uncertainty of very early internationalisation. Oviatt and McDougall (1994) typology of international firms also involved international scope and the coordination of distinct value-chain activities. These factors have been at the centre of much research, for example, Autio (2005), Coviello (2006), McDougall et al. (2003) and Mudambi and Zahra (2007). A further kind of international firm, the 'born global' (BG), was identified by Knight andCavusgil (1996, 2004). Their definition became another of the dominant concepts in the literature and was discussed by a wide range of authors, such as Autio and Sapienza (2000), Cavusgil and Knight (2015), Chetty and Campbell-Hunt (2004), Coviello (2015) and Kuivalainen et al. (2007). BG firms achieved a high degree of internationalisation early. They are defined as 'entrepreneurial start-ups that, from or near their founding, seek to derive a substantial proportion of their revenue from the sale of products in international markets' (Knight & Cavusgil, 2004). A widely accepted definition proposed by Knight and Cavusgil (2004) suggests that BGs export at least 25% of their sales within three years of inception. Therefore, the identification of BG must consider both the time to internationalisation and the degree of internationalisation.
Although Knight and Cavusgil (2004) recognise that BGs can be similar to INVs, they point out that they are different in some ways, one of which is that they are young firms with limited resources, using exporting as the main form of international entry. BGs exhibit a high degree of international entrepreneurial orientation. They seek better international performance by applying knowledge-based resources to the sale of outputs in multiple countries (Knight & Cavusgil, 2004). Therefore, internationalisation scope must also be taken into account in identifying BGs, which tend more to market diversification (number of countries ! 10) than to market concentration strategies (number of countries < 10) (Aspelund & Moen, 2005). Later works, adopting a post-initial foreign market entry view, have typified organisations as BG on the basis of the rate at which they internationalised (e.g., Chetty & Campbell-Hunt, 2004;Coviello, 2015;Kuivalainen et al., 2012). BGs, then, are firms whose internationalisation was both early and rapid. Some authors, however, have questioned whether BG is a unique typology. Bell et al. (2001) and Bell et al. (2003) introduce the category of born-again global firms, characterised by a sudden, swift internationalisation after being active only in the domestic market for a specific period: typically less than five years. This process of sudden and quick internationalisation by born-again globals can be as structured, proactive and flexible as that of newly created firms (Baldegger & Schueffel, 2010). Kuivalainen et al. (2007), combining degree (ratio international sales/total sales) and international scope (distances of markets), distinguish between 'true-born globals' and 'apparently-born globals/born internationals'. True-born globals are firms which, having internationalised in the first three years after inception, attain a high degree of internationalisation (international propensity > 25%), spanning distant markets and multiple regions/continents. They are called 'apparently-born globals/born internationals' when their exporting activity is limited to nearby or geographically concentrated markets. Late international starters/born-again internationals are similar to this but have restricted their activities to their continent of origin.

Key dimensions of entrepreneurial exporters typology
Four variables must be considered key to identifying the international entrepreneurial behaviours of exporting firms (Navarro-Garc ıa, 2016): time to internationalisation and the speed, scale and scope of internationalisation. These dimensions must be considered together in order to obtain a valid typology of international firms (Kuivalainen et al., 2012). The following subsections will describe each of them and establish the cut-off points used for the typification.

Time to internationalisation and speed
The time to internationalisation is the period between the firm's inception and when it goes international, that is when it begins to obtain sales from one or various foreign markets (Acedo & Jones, 2007). During this interim period the firm generates essential knowledge of external markets, so it is a time of commitment and learning (Chetty et al., 2014). But how many years must be taken as a reference? In the literature we find a range from two to 15 years (Dzikowski, 2018). Rennie (1993) and Moen and Servais (2002) suggest two years, while McDougall et al. (2003) and Zahra et al. (2000) argue for six years for an INV. A proposal widely accepted in the literature is that by Knight and Cavusgil (2004): BGs export at least 25% of their sales within three years of inception. Three years is also the time to internationalisation considered in other studies (Andersson & Wictor, 2003;Glaister et al., 2014;Rasmussen & Madsen, 2002;Zhou et al., 2007). It will be the cut-off point taken as the reference point by this work.
On the other hand, speed is the dimension of IE that has attracted the attention of researchers and academics in the last decade (Schwens et al., 2018). It has been considered essential in the definition of INV, fast internationalisation processes and BGs (Paul & Rosado-Serrano, 2019). Speed is a key aspect of firms' international strategies, and firms need to balance their resources and the international opportunities available (Chetty et al., 2014). Definitions of speed need to take into account post-entry internationalisation; that is to say, the pace at which the firm grows and is developed in foreign markets (Kuivalainen et al., 2012). The most commonly used measurement is the quotient of the number of countries where the firm is present and the years of internationalisation, although some authors have instead used the number of subsidiaries created annually (Li et al., 2012;Vermeulen & Barkema, 2002). A faster international expansion forces the firm to develop organisational structures, routines and processes that allow it to adapt more effectively to the requirements of the international markets. How well the firm adapts determines the levels of competitive and sustainable advantages it achieves, and therefore the business results it gains (Schwens et al., 2018). Although the literature offers no cut-off reference point, the pace of international expansion will depend on the rate at which the firm develops strategies for market diversification or concentration. We consider that a rate of international expansion of one to two countries per year can be considered high, and three or more very high.

Scale and scope of internationalisation
The scale of internationalisation determines the firm's level of international propensity. It is measured by the ratio of international sales to total sales and is an indicator of the firm's orientation towards external markets relative to the domestic market (Kuivalainen et al., 2007). A greater scale of internationalisation enables a firm to have a more global view of international businesses and encourages the development of sophisticated market intelligence capabilities (Navarro-Garc ıa et al., 2016). It also tends to increase management's commitment and proactiveness in the search for business opportunities in foreign markets (Gallego & Casillas, 2014), contributing to the development of structures (e.g., export department) required for international expansion. All this will help sharpen the firm's international competitiveness, reducing costs via economies of scale and driving the achievement of better business results (Zhou & Wu, 2014).
There is no consensus in the literature on the appropriate level of this ratio for a firm to be considered internationally entrepreneurial or for them to be considered BG, although Oviatt and McDougall (1997) argue that it must be set at a 'significant' level. But how much is significant? The criterion most often mentioned is 25% of total turnover (Knight & Cavusgil, 1996Kuivalainen et al., 2007Kuivalainen et al., , 2012Madsen et al., 2000;Moen, 2002). This percentage of 25% is adopted as our reference in the current research, although clearly there are strong grounds for suggesting 50%.
The scope of internationalisation is a measure of the number of foreign markets in which the firm sells; it is referred to in the literature as equivalent to the firm's geographic extension or diversification (Ruzo et al., 2011). It is an essential variable in distinguishing the born globals from the rest (Kuivalainen et al., 2007;. A broader international scope allows a firm to diversify its international risks and generate economies of scale in production and commercialisation, which helps in terms of the structure of its costs (Fernhaber, 2013). It also forces the internationalised firm to organise and plan its international movements more effectively, which is evident in the creation of specific structures (e.g., export departments) which support decisionmaking in the international area (Navarro-Garc ıa et al., 2016). Yet, how many countries must be taken as a reference to identify real international entrepreneurs? There is, again, no consensus in the literature, and in defining BG, Oviatt and McDougall (1994) did not establish a number, although they clearly considered that the geographic scope must be broad. It seems reasonable, nevertheless, to assume that a truly global orientation requires the development of a market diversification strategy covering at least ten countries (Aspelund & Moen, 2005;Ruzo et al., 2011). In the current work, therefore, we will take ten or more countries as the cut-off point for defining wide high international scope.

Sample, data collection and measurement scales
An empirical study was undertaken using a database provided by the Agencia Andaluza de Promoci on Exterior (Trade Promotion Agency of Andalusia) (EXTENDA), made up of 1635 exporting firms located in Andalusia (Spain). A questionnaire was designed in which respondents answered a wide range of questions on entrepreneurship and the internationalisation of firms. The questionnaires were sent online between January and March 2019 to firms' executives in charge of international activities. One hundred and thirty-six valid responses were obtained: a response rate of 8.3%. Table 2 shows the variables analysed and their measurement scales. All the variables included in the latent class segmentation are objective (number of employees, number of countries, etc.). Subjective variables 1 have been considered for the discriminant analysis, associated with managerial perceptions (5-point Likert scale; 1 -Very low … .5 -Very high).

Latent class segmentation
Latent class models are techniques based on statistical segmentation models. They involve minimal arbitrariness in the choice of the segmentation criteria and the possibility of generalising the model to other contexts. They are very useful and effective tools for market segmentation because they allow us to estimate both the usefulness of the segments and the probability of each person belonging to each segment (Wilson-Jeanselme & Reynolds, 2006).
Whereas traditional models (e.g., regression, cluster or logit analysis) contain only parameters that describe the relationships between the observable variables, latent class models include additional parameters that describe the relationships between the observable variables and one or more latent variables (Vermunt & Magidson, 2003). Latent class models are a type of mixed model. They deal with the heterogeneity of the parameters of a population model, imposing a 'mixed probability distribution' in some or all of its parameters. It is assumed that the parameters of the model are heterogeneous between individuals and that they follow a distribution of the population which can be inferred as continuous or discrete (Wedel & Kamakura, 2002). To sum up, latent class models enable the identification of groups that gather the cases that share similar interests and characteristics. They also include a latent variable (the K-category), and each category represents a different cluster. The classification cases within each group are based on the probability of the member belonging to this group according to the direct estimation of the model (Bond & Morris, 2003;Vermunt & Magidson, 2003).

Description of the sample
The majority of the firms included in the sample are small and medium-sized enterprises -SMEs. Small firms predominate: 70.37% have less than 50 workers, and there is a significant number of micro-SMEs, with less than ten employees (45 firms; 33.34%). The firms' age or general experience is notable: 45.19% are more than 20 years old, the average age being around 23 years. The average international experience is 12.76 years; some firms started to internationalise more recently (37.78% have an international experience of fewer than six years), and another group is more experienced in foreign markets (31.11% have over 15 years of international experience). The average time taken to go international is 10.35 years, although 60% of the sample (81 firms) went international within five years of their inception. At the other extreme, another group of firms (34; 25.29%) were slow, having taken more than 15 years to go international. The average international scope is 12.52 countries. 48 firms À35.55%are present in more than ten countries, while 62 firms À 45.93%are centred on country-market concentration, doing business in less than six countries. The average pace of internationalisation is 1.63 countries/year with, again, a wide spread. A group of 40 proactive firms (29.63%) set a pace of 2þ countries per year, and another group are much more conservative: 49.63% internationalise at a pace of less than one country per year. Similarly, there are notable differences in the intensity of degree of internationalisation, whose average is 24.75%. While a small group of firms (26 firms; 19.26%) show a high level of international propensity, over 50%, another group, the majority, have a degree of internationalisation below 25% (87 firms; 63.97%). Most of the firms in the sample do not have business or production subsidiaries, which generally are found in larger firms.
Having a structure dedicated to supporting international decision-making is important for most of the sample's firms: 56.30% have an export department. Likewise, it is important to be able to use an international network that helps the firm seek and exploit opportunities in foreign markets (77.78% have such a network of professional contacts).

Results of the latent class segmentation
The first stage of the latent class segmentation determines the number of groups or segments selected. To do so, we look for the lowest BIC number (Table 3). The segment with four groups is the most appropriate, having a lower BIC and providing a better fit (Vermunt & Magidson, 2003). The model of 5 clusters is similarly low, but that of 4 offers a better fit and more parsimony as it has fewer parameters to estimate, with 86 against the 109 in the 5-cluster model.
In the second stage, the significance level of the variables used in the segmentation is checked. As Table 4 shows, all the variables are significantly distinct in each group, as the p-value of the Wald test is less than 0.05 in all the cases. Table 5 shows the Wald test for the covariables. In this case, none of the three covariables consideredsector, export department and international networkis a discriminant factor in the groups found.
The latent class segmentation enables us to identify four clusters or groups of exporters according to their levels of international entrepreneurship and the variables that influence this (Table 6). We identify the clusters and group the firms according to which latent class segment they belong to.

Discussion
In general, the 4 clusters identified could be divided into two large groups. The first of these would include firms that respond to the principles of IE and to the development of early internationalisation processes: Clusters 1 (true-born global exporters) and 2 (early internationals and market concentration exporters). The second group includes those organisations that internationalised later, following a more traditional, gradual approach: Clusters 3 (traditional exporters) and 4 (true born-again globals). Table 7 summarises the characteristics of each group.
6.1. Clusters associated with early internationalisation processes 6.1.1. Cluster 1: true-born global exporters This cluster includes small and young exporting firms, created less than ten years ago and with less than 20 employees. They went international in the first three years after their inception. Given their youth, they are firms with little international experience: their average international experience is 6 years, but their speed of international expansion is very high, annually entering almost two countries per year. They have therefore attained wide international scope in just a few years (they are present in an average of 16 countries) and have developed market diversification strategies. Their degree of internationalisation is high and their international propensitythe ratio of international sales to total salesis above 50%. They concentrate on exporting as the primary way of selling in foreign markets, not undertaking FDI, so they do not have business or production subsidiaries, mainly because of a relative lack of resources.   They belong to several sectors, although the industrial/technological sector predominates (49%). They have created a specific structure, i.e., an export department, to organise and plan their international activity, and they recognise the importance of a network of international contacts in their development.
In brief, globally born and oriented exporters are part of this cluster. They have reached a high degree of internationalisation, quickly and with broad scope. They are international entrepreneur exporters par excellence.

Cluster 2: early internationals and market concentration exporters
This cluster is made up of medium-sized firms: 85 employees on average. They are not so young: their average age is around 30 years, and they are very experienced internationally, with the highest average À 26 yearsof international experience. The firms in this cluster developed internationalisation processes early on, going international in the first four years after inception. However, we consider their speed of international expansion medium-to-low as they enter a new country once per almost two years. This low or moderate pace of international expansion can result from their having adopted a market concentration strategy with a limited international scope: the average number of countries in which they are active is eight. They are firms in which exporting is the main form of foreign market entry, although in some cases they undertake direct investments with business or production subsidiaries. Despite their moderate international pace and scope, their international propensity is over 25% (33%). They are in several sectors, although agri-food predominates (41%). To develop the external activity, they use both an export department and a network of international contacts. To sum up, this cluster comprises exporters that, despite developing an early internationalisation, maintain quite a low pace of internationalisation. They are exporters that develop market concentration strategies. This means that their international scope is limited to less than ten countries. We call this group 'Early internationals and market concentration exporters'.
6.2. Clusters not associated with early internationalisation processes 6.2.1. Cluster 3: traditional exporters This cluster is mainly made up of small exporting firms, with an average size of 42 employees. They are mature firms, created more than 36 years ago, although they have nine or fewer years of international experience. Their internationalisation process is slowthey tend to enter a new country every 2 or 3 yearsand late: they usually take an average of 28 years to go international. This can be because they focused on their domestic market, giving them a low international propensity (19%) and a limited international scope: they tend to be present in an average of 7 countries. Exporting is their primary entry mode into foreign markets, with few production or business subsidiaries. They do not usually have an export department or a network of international contacts facilitating international transactions. This is a multi-sectoral cluster.
In brief, this cluster comprises exporting firms that are not very internationally entrepreneurial, showing very conservative or not very proactive behaviours in the development of their internationalisation process. They are firms of late and slow internationalisation with limited international scope and propensity. We call this group 'traditional exporters'.

Cluster 4: true born-again globals
This cluster comprises large firms (more than 300 employees) and mature firms, created more than 50 years ago. They have broad international experience, 18 years on average, although their internationalisation process has been the longest of all the clusters: from their inception they took an average of 33 years to go international. However, once they go international, they do so very rapidly, with an average of almost four countries per year. So they are the cluster with the highest speed of internationalisation. They are firms with strong market diversification and have the greatest international scope: 36 countries on average. The average international propensity of this cluster is well above 50% (66%), which reflect a strong international orientation and possibly a global focus. These firms have gone furthest in developing FDI in the form of business or production subsidiaries: they have an average of almost three subsidiaries, which again reflects their high international commitment.
To sum up, the firms in this cluster are late but fast internationalisers, achieving a very high scope and scale of internationalisation. They are global firms, very internationally oriented and willing to take on the risks of committing significant resources to developing their internationalisation process. Their profile is similar to the 'true born-again globals' of Kuivalainen et al. (2012), and that is what we call them.

Conclusions, limitations and future lines of research
Previous studies have shown that international firms vary widely (Aspelund & Moen, 2005;Kuivalainen et al., 2007Kuivalainen et al., , 2012Olejnik & Swoboda, 2012;Vissak & Masso, 2015). However, no previous study has offered a typology of international firms centred on exporting activity that encompasses both the gradual models (scale and scope) and the IE and born-global models (time to internationalisation and speed). We therefore cover a significant research gap.
This study makes several academic contributions. Firstly, a correct typification of exporting firms must include all the variables we cover: time to internationalisation, speed, scope and scale. The majority of previous studies have focused on some of these variables but not on all of them. Most have ignored the speed or pace of international expansion, a good measure of which is the number of countries per year. Secondly, we demonstrate that there are true born-global exporters that not only go international early -( 3 years) but also do so at a high pace (1-2 countries/year), developing genuine market diversification strategies (international presence in more than 15 countries) and achieving a very high international propensity (!50%). They are firms whose managers are proactive towards internationalisation, finding more advantages than barriers in developing early and fast internationalisation processes. They are international entrepreneurs par excellence. Thirdly, this study demonstrates that, even if an exporter was not born global, it can have a truly global orientation. Thus the 'true born-again globals' internationalised very late (>30 years), having initially had a domestic orientation. But, once they start, they do so rapidly, attaining great scope (>35 countries) and a very high scale of internationalisation (>65%). They are the most internationally committed firms, as they tend to undertake FDI, obtaining business or production subsidiaries. This group of firms is the clearest example that the gradual approach and the born-global model are supplementary and not incompatible, and we argue that this deserves serious academic attention. Fourthly, early internationalisation does not mean, per se, that a firm undertakes international entrepreneurial behaviours: it is a necessary condition but not a sufficient one. What happens after the beginning of internationalisation must also be considered: how, where and why the exporting firm expands internationally. Thus, though early internationalisers and market concentration exporters are early in their internationalisation ( 4 years), once they do so they can be rather conservative and not very entrepreneurial in their international development, expanding at a slow pace (<1 country/year) and concentrating their activities in few country-markets (<10 countries). This does not mean that the firm is not internationally oriented, but rather that it has opted for a market concentration strategy. When the firm develops this appropriately, it tends to have a high international propensity (>25%). Finally, in all sets of exporters there is always a group that is more reactive than proactive in going international. These are the traditional exporters whose orientation is usually more domestic than international, perceiving more barriers than advantages in exporting. They expand internationally late (>25 years) and slowly (1 country every two or three years). They also concentrate their activities geographically (<10 countries) and display a low international propensity (<20%). Many of them might well abandon their international activity or develop it spasmodically (Vissak & Masso, 2015). Their managers are usually not very committed to internationalisation, and this is reflected in the lack of international entrepreneurial behaviours.
From the managerial point of view, this work also has two important implications. The first is that entrepreneurial orientation and managerial perspectives will be key for all firms wishing to adopt a global approach in their internationalisation process. This result accords with previous studies (Knight, 2001;Knight & Cavusgil, 2004;Kuivalainen et al., 2007). The development of public policies which foster managerial training in international businesses can make decisive contributions. Secondly, firms wishing to adopt a truly global approach will need to think beyond the resources of size and experience: they will need to seriously develop their international organisational capabilities. A firm's market orientation capabilities and its skills in adapting its marketing mix to the needs of different countries can determine its level of exporting entrepreneurship and therefore its success in foreign markets. The creation of specific structures, particularly an export department, is key in facilitating the planning and development of the exporting activity (Navarro-Garc ıa et al., 2015). It is also imperative that firms whose sole entry mode to foreign markets is exporting try to develop a network of international contacts: agents, distributors, etc., in order to gain the benefits they offer in increasing exporting entrepreneurship.
Although this study offers new, important contributions in the area of IE, it is not without limitations, which must be the starting point for future research. The first limitation is that the sample belongs to a single region, Andalusia, in a single country, Spain. If the results are to be generalised, studies must be undertaken with samples in other countries, in the European Union and beyond. Secondly, our study was developed asynchronically, that is, with a sample that was taken at a moment of time. Only longitudinal studies will enable us to explore the evolution over time of a firm's process of internationalisation, including those firms that abandon and re-enter international markets. A third limitation refers to the environment, the strategic options, competitive advantages and resultsmanagerial satisfactionof the exporting activity having been subjectively considered in our studymanagerial perceptions. Studies are needed of the different types of exporters, including parameters that are quantitatively more objective. Note 1. Only the variables which significantly differentiate according to the clusters obtained have been reported.

Disclosure statement
No potential conflict of interest was reported by the authors.