Entrepreneurial orientation among sales professionals: performance implications and mediating mechanisms

Entrepreneurial orientation (EO), a “top management style” (W. J. Wales et al., 2020, p. 3), reflects how a firm operates and is generally measured via CEO perceptions (Putniņš and Sauka, 2020). Although the literature documents the positive effect of EO on performance, the underlying mechanisms through which EO is translated into superior performance have not been sufficiently studied (Kollmann and Stöckmann, 2014; Wales et al., 2013). EO is not a direct measure of entrepreneurial activities; it is a disposition – a willingness to innovate, to take risks and to engage in proactive behaviours (Wiklund and Shepherd, 2005). Scholars have called for studies that illuminate the chain from this attitudinal disposition to tangible performance outcomes (Jiang et al., 2021; Wales et al., 2013). This is the first objective of the present study.

Secondly, managers in the firm clearly have the potential to act entrepreneurially (Jones, 2005). Among the many units that make up a firm, one unit clearly has the greatest opportunity to engage in entrepreneurial activities but has, counterintuitively, not received scholarly attention in entrepreneurship research. Which unit? Sales – of course. Decades ago, the literature argued that “sales management should represent one of the more entrepreneurial areas within firms” (Morris et al., 1990, p. 3), that “salespeople are entrepreneurs on the front line” (Jones et al., 2000, p. 45). Salespeople are “unsupervised” (Zoltners et al., 2008, p. 115) and have the opportunity to act like entrepreneurs, but “selling has been largely absent from entrepreneurship theorizing” (Matthews et al., 2018, p. 703). Attention to entrepreneurial selling is “sporadic” (Dalecki, 2019, p. 103). “Research on entrepreneurship and B2B sales has remained relatively limited” (Edwards et al., 2023, p. 1), essentially because these two literatures are “disconnected scholarly camps” (Pidduck and Epler, 2026, p. 265).

There is, to this date, not a single empirical study examining the effect and pathway of EO of sales professionals on their performance. The second objective of the present study is thus to bridge the literature of entrepreneurship and selling to examine the EO–performance relationship in this specific context. I also conduct a post hoc test to explore the effect of entrepreneurial intent (EI) on performance.

After polling 252 sales professionals in the UK in early 2020, I find significant effects leading from EO to sales performance, but no direct effect of EI on performance. I further explore the pathways leading from EO to sales performance, a pathway that hitherto is a “black box” (Jiang et al., 2016, p. 103), identifying ambidexterity and role ambiguity as mediating variables. In a nutshell, the conclusions are that thinking like an entrepreneur – having a high EO – directly improves individual performance, while intending to become an entrepreneur – having a high EI – does not. These findings have significant practical implications for organisations that so far have uncritically encouraged entrepreneurial behaviour among sales professionals or that have hired aspiring entrepreneurs in sales functions. This study has important theoretical implications because it illuminates the pathway leading from EO to performance, identifying ambidexterity, i.e. joint farming and hunting, and role ambiguity as important mediators; it also sheds light on the very distinct effect of EI and EO on individual sales performance.

In a pioneering paper, Miller (1983) defines the characterising attributes of an entrepreneurial firm as “one that engages in product-market innovation, undertakes somewhat risky ventures, and is first to come up with ‘proactive’ innovations, beating competitors to the punch”. This view of entrepreneurship is the basis for the subsequent and now prevailing definition of EO by J. G. Covin and Slevin (1989) as a firm-wide strategic posture in which CEOs have management styles that emphasise innovation, proactivity and risk-taking. In this view, EO is a composite construct represented by what innovation, proactivity and risk-taking have in common. Lumpkin and Dess (1996) take a different perspective: EO is a multidimensional construct in which the dimensions function as independent predictors. To the original three dimensions, Lumpkin and Dess (1996) add autonomy and competitive aggressiveness. This redefinition of EO is “radical” (Wales et al., 2013, p. 359), leading to two different perspectives of EO: a composite and a multidimensional view.

Several meta-analytic reviews of EO research indicate that the overwhelming majority of scholars use the three traits of innovation, proactivity and risk-taking to define EO as a unidimensional construct (Covin and Miller, 2014; Wales et al., 2013). These three aspects are the “core, fundamental dimensions” of EO (Wales, 2016, p. 6). This conceptualisation of EO underlies the present study.

The literature predominantly defines EO as a firm-level construct (Bouguerra et al., 2023; Covin and Slevin, 1991) such that the firm, by virtue of behaviours on the part of the CEO and top management, engages in entrepreneurial actions. Only fairly recently have scholars begun to question and criticise this firm-level view of EO. Frese and Gielnik (2014, p. 422) write that “entrepreneurial orientation is a curious concept”, insofar as it assumes that self-perceptions of CEOs regarding their firms’ strategic orientation have any relevance for the firm as a whole. Several researchers therefore urge scholars to “reintroduce the individual in EO research” and to define EO as “interindividual difference” in behaviourally relevant characteristics (Krauss et al., 2005, p. 216). Indeed, a meta-analytic review of more than 100 entrepreneurship studies finds significant relationships between individual personality traits and performance, such as new venture creation and business success (Rauch and Frese, 2007). A subsequent meta-analysis of 60 studies confirms that individual personality traits – in order: conscientiousness, openness to experience, emotional stability and extraversion – are positively related to entrepreneurial intentions and subsequent performance (Zhao et al., 2010). Individual-level analysis is very fruitful.

This individual-level view of EO has characterised entrepreneurship research from its very foundation: Schumpeter (1911) certainly uses an individual-level approach to entrepreneurship: entrepreneurs are those who periodically “revolutionize the economy”, “create new products” and “reorganize” (1911: 492). Studying EO at the individual level is essential because it captures how individual entrepreneurial dispositions translate into entrepreneurial behaviour (Clark et al., 2024). In line with Schumpeter, the founding father, and with other recent studies (Covin et al., 2020; Kollmann et al., 2017; Krauss et al., 2005; Perry et al., 2016), this study takes an individual-level view of EO.

I examine whether EO and, in a post hoc analysis, EI are associated with higher individual sales performance. The positive effect of individual EO on performance is well documented for entrepreneurs (Krauss et al., 2005; Rauch and Frese, 2007) – but not for employees, i.e. salaried individuals working in a subordinate position. Furthermore, we do not yet understand whether EI has any direct effect on individual performance. Table 1 lists representative studies and positions the current study in the context of ongoing research.

Among practitioners, a few question (Kahle, 2014), but most authors advocate the benefits of encouraging entrepreneurship among sales professionals (Hanan, 1973; Zoltners et al., 2009). Among academics, no studies, to the best of my knowledge, examine the relationship between individual EO and salesperson performance, though some examine the role of related traits, e.g. self-efficacy (Edwards et al., 2022) and motivation (Rajabi et al., 2018). The EO–performance relationship has certainly been investigated at the firm level, where a meta-analysis of 51 studies finds a moderately large correlation of 0.242 (Rauch et al., 2009). At the individual level, the strength and nature of the EO–performance relationship have yet to be examined.

EO is a disposition, i.e. an “individual’s durable orientation that influence their context-specific behavioural choices” (Clark et al., 2025, p. 670) for innovation, proactivity and risk-taking. For these three traits, extant research is not uniformly developed, much less so individual-level research: innovativeness, at the department level, increases sales department performance by promoting novel problem-solving approaches and by fostering a climate of autonomy and trust that increases salesperson engagement and effort (Matsuo, 2009). At the individual level, proactivity is, according to several meta-analytic reviews, a valid predictor of job performance (Fuller and Marler, 2009; Spitzmuller et al., 2015; Thomas et al., 2010; Tornau and Frese, 2013): proactive individuals tend to shape their work environments by actively seeking opportunities and taking initiative, which results in them creating more demanding jobs with higher autonomy and responsibility (Li et al., 2014). At the individual level, extant research conceptualises risk propensity as a domain-specific behaviour shaped primarily by perceived benefits and, to a lesser extent, by perceived risks (Hanoch et al., 2006). Risk taking is associated with goal achievement insofar as individuals may accept substantial risks to get what they want (Nicholson et al., 2005), but we do not yet know how “risk-taking propensity affects work, life, and educational outcomes” (Joseph and Zhang, 2021, p. 8). Risk taking is linked to innovativeness (Hernández-Méndez et al., 2025), and, for salespeople, creativity – related to innovativeness – is linked to improved sales performance (Miao and Wang, 2016). At the firm level, risk taking has a positive effect on performance (Saini and Martin, 2009). Thus, while the individual-level research on innovativeness, risk-taking and proactivity is very patchy, proactivity shows the strongest link to job performance. Even so, the evidence still suggests a likely positive association between both innovativeness and risk-taking and performance. Finally, in line with the literature (Miao and Evans, 2007), this study captures performance as comprising both qualitative (behavioural) and quantitative (financial) indicators. The core hypothesis of this study is, therefore:

An important question concerns the mechanisms between EO and performance that “have remained assumed rather than tested” (Wiklund and Shepherd, 2011). Scholars point to the limited research on mediating variables – constructs explaining why events take place – in the EO–performance relationship (Wales, 2016; Wales et al., 2013): “Little consideration [is] being given to such matters as the identification and assessment of mechanisms and mediators through which subsequent performance occurs” (J. Covin and Wales, 2019, p. 12). This view is shared (Wales et al., 2021).

The pathways of this relationship can be examined by using the orientation-behaviour-performance view (Chernetsky et al., 2022; Donbesuur et al., 2020; Jiang et al., 2021; Kollmann and Stöckmann, 2014). According to this view, EO is a predisposition towards entrepreneurial activities (Lumpkin and Dess, 1996), which does not necessarily lead to improved performance, but which will do so if actions are taken. Kollmann and Stöckmann (2014, p. 1004) write that “the EO–performance relationship might be jeopardised by the fact that the EO is not converted into appropriate action”. I use available meta-analytic evidence to identify the intermediary mechanisms bridging the EO–performance relationship: ambidexterity and role ambiguity.

Sales professionals sell – a trite statement. With respect to any activity that impacts performance – strategising, innovating and also selling – scholars have argued that individuals must manage the impossible tension between exploitation and exploration (March, 1991). March (1991) posits that exploration involves searching for and experimenting with new possibilities (e.g. risk taking, discovery and innovation), whereas exploitation involves refining and efficiently executing what is already known (e.g. implementation and execution). Ambidexterity refers to the development of an organisational context that encourages individuals to perform the conflicting strategic actions of efficiently managing what is known and creatively experimenting with the unknown (Gibson and Birkinshaw, 2004). At the level of individual sales professionals, the literature offers three distinct perspectives on ambidexterity. Ambidexterity means selling products and services (Mullins et al., 2020a; Schaarschmidt et al., 2022), selling existing and new products (Nijssen et al., 2017) and, finally, selling to existing and acquiring new customers (Van der Borgh et al., 2017). In line with the prevailing view in the literature, this paper conceptualises ambidexterity as selling to existing and acquiring new customers, as joint farming and hunting (DeCarlo and Lam, 2016; Lam et al., 2019). This distinction between farming and hunting thus captures two key, but distinct, selling activities (DeCarlo and Lam, 2016): farming involves developing and retaining existing customer relationships (e.g. relationship maintenance, cross-selling and up-selling), whereas hunting involves acquiring new customers (e.g. prospecting, lead generation and securing initial orders).

The dominant organisational design in sales is a hybrid hunter-farmer sales role (Ingram et al., 2020; Weinberg, 2013), where salespeople are paid to farm and hunt at the very same time. Ambidextrous behaviour is about managing this impossible – at least, very difficult – tension, since salespeople overwhelmingly farm, leading some researchers to suggest that hunting is more important “than any other issue today” (Weinberg, 2013, p. 35). Several studies provide robust evidence that salespeople who exhibit a high farming and a high hunting orientation realise substantially higher results (DeCarlo and Lam, 2016; Lam et al., 2019).

What leads to ambidexterity? In their study of roughly 200 adolescent companies, Kollmann and Stöckmann (2014, p. 1005) identify EO as an antecedent of both exploitation and exploration, explicitly warning against the “normative bias” of naively assuming that EO would affect only exploration. In their study, EO is linked to both exploitative innovations and explorative innovations, although the explained variance of exploration is substantially higher than that of exploitation. But the conclusion is clear: EO affects firm performance via both exploitation and exploration. In a similar vein, Ireland and Webb (2007, p. 50) argue that EO is linked to ambidexterity since it “balances between opportunity-seeking (i.e. exploration) and advantage-seeking (i.e. exploitation) behaviours”. Furthermore, a recent study of 260 start-ups finds strong relations between entrepreneurial passion and both exploration and exploitation (Ahsan et al., 2023), thus providing indirect evidence that the thought-worlds of entrepreneurs affect both exploration and exploitation. Finally, EO is linked to innovation ambidexterity, i.e. exploratory and exploitative innovation (S. Singh and Singh, 2024). Taken together, these studies thus justify the following set of hypotheses:

As of this writing, there exist three major meta-analytic studies examining drivers of sales performance (Churchill et al., 1985; Claro et al., 2024; Verbeke et al., 2011), in addition to two focused meta-analyses, one examining only personality traits (Vinchur et al., 1998), the other only drivers in business-to-business (B2B) (Ohiomah et al., 2020). These meta-analytic studies provide converging evidence regarding the importance of one key factor driving sales performance: role ambiguity (Churchill et al., 1985; Ohiomah et al., 2020; Verbeke et al., 2011).

Role ambiguity refers to the perceived lack of information to perform the job adequately and the perceived uncertainty about job expectations (Singh, 1998). Role ambiguity in sales is a consequence of the ongoing tectonic shift of mechanisms through which the sales force creates value. DeVincentis and Rackham (1998, pp. 35–36), prominent researchers in sales management, wrote already in 1998: “Unfortunately, generations of salespeople have been brought up with the notion that they create value by bringing in revenue. But bringing in revenue means collecting value, not creating it. And that is not enough to survive in today’s competitive markets”. More recently, some authors question the very relevance of a salesforce: “In a world where anybody can find anything with just a few keystrokes, intermediaries like salespeople are superfluous” (Pink, 2012, p. 13). The role of the salesforce has changed dramatically, from passively collecting orders, to providing information, to understanding and developing customer needs (Rackham, 1988), to co-creating solutions (Macdonald et al., 2016; Storbacka, 2011), to quantifying value (Hinterhuber, 2017), to, finally, increasing customer competitiveness via education (Dixon and Adamson, 2011).

Even under the assumption that role expectations vis-à-vis salespeople are – over a brief period – fixed, role ambiguity remains an important performance differentiator: sales professionals span boundaries that are inherently ill-defined. They are exposed to conflicting demands by customers, supervisors and other departments; objectives, available resources, decision rights, customer needs, product–market fit and competitor strategies are all subject to rapid change (A. A. Zoltners et al., 2008). Structuring and reducing the inherent ambiguity, focusing on the few outcomes that truly matter, is what makes sales professionals great. Aaron Ross, sales director at Salesforce.com, puts it this way: “uncertainty/ambiguity = waste; simplicity/clarity = productivity” (Ross and Tyler, 2011, p. 215).

Over the past decades, scholars have conducted four major meta-analytic reviews on antecedents and consequences of role ambiguity (Abramis, 1994; Gilboa et al., 2008; Jackson and Schuler, 1985; Tubre and Collins, 2000). These studies provide converging evidence regarding the significant negative correlation between role ambiguity and job performance, with a true score correlation of −0.21 (Tubre and Collins, 2000), comparable in size to the effect of “some personality correlations” (2000, p. 164) widely used in personnel selection. These meta-analytic reviews identify several determinants or, at least, antecedents of role ambiguity: in addition to leadership factors, they point out the important role of individual characteristics, such as self-esteem, autonomy, participation in decision-making, external locus of control, job involvement and task identity (Jackson and Schuler, 1985), all of which are negatively correlated with role ambiguity. Autonomy, self-efficacy and internal locus of control are traits that the entrepreneurship literature identifies as antecedents to EO (Frese and Gielnik, 2014; Krauss et al., 2005).

Furthermore, the literature indicates that the facets of EO have a direct role in shaping perceived role ambiguity: being proactive refers to “the extent to which the individual takes self-directed action to anticipate or initiate change in the work system or work roles”. This is “important whenever a work context involves uncertainty and some aspects of work roles that cannot be formalized” (Griffin et al., 2007, p. 329). Innovativeness refers to the ability to challenge the status quo and seek alternative ways to solve problems and to gain advantages from opportunities via experimentation (Baum and Bird, 2010). Risk-taking implies “a tendency to take bold actions” such as venturing into the unknown and making sense of ambiguous conditions (T. Lumpkin and Dess, 2001, p. 431). Taken together, these studies suggest that high levels of EO are associated with low role ambiguity.

Role ambiguity, i.e. the perception of uncertainty and of insufficient information, is, in this view, not an immutable role characteristic but an attribute that innovative, proactive and risk-seeking individuals reduce through actions that shape the external environment. EO is an antecedent of role ambiguity, allowing one to make sense of ill-structured information:

The hypothesised research model is in Figure 1.

Researchers specifically recommend online panel providers when the collected data is sensitive (Porter et al., 2019). In this study, I measure individual sales performance, data that, for sales professionals, has a direct effect on target achievement and thus, typically, on individual variable compensation. The data is extremely sensitive. Furthermore, a meta-analysis indicates the convergence between online panel data and conventionally sourced data, concluding that online panel data are “suitable” for exploratory research (Walter et al., 2019, p. 425).

I commissioned Cint, a leading global provider of online panel data, to recruit sales professionals in the UK. Prior to data collection, I set a target sample size of about 260 complete responses based on considerations of statistical power and available budget. I use strict a priori criteria for participant exclusion to ensure data quality: At the start of the survey, I use a filter question asking about their current role as an additional step to ensure data quality ^[1]. Only individuals employed in one of four selling roles – customer-facing sales role, managerial role in sales, customer-facing account management role or managerial role in account management – were eligible to participate. Participants who indicated “Other” were prevented from taking the survey. Hence, respondents are qualified. Data collection takes place in a two-week period in the second half of March 2020.

To further ensure data quality, I include an attention check question in the online survey. Specifically, I use an item embedded in a survey block that was formatted the same as the preceding and succeeding items and ask respondents to give a particular response to indicate that they are paying attention. Before data analysis, respondents who did not pass the attention check are automatically excluded and replaced. I also added one redundant question on participant gender, eliminating participants with diverging answers to this simple question. Similarly, I ask the panel provider to replace inattentive respondents (e.g. straight liners and speeders) before data analysis.

I am able to retain responses from 252 sales professionals for the final analysis. The sample size is sufficient to test the hypotheses. The sample is diverse in terms of both individual and organisational characteristics: The sample consists mainly of sales professionals from business-to-consumers (B2C) companies (47%), with additional representation from B2B (21%) and mixed B2B/B2C firms (28%). The majority of respondents hold customer-facing sales roles (60%), followed by managerial roles in sales (24%), customer-facing account management roles (11%) and, finally, managerial roles in account management (5%).

The typical respondent in this survey is a female, customer-facing sales professional in the mid-thirties, working for a privately-owned firm with fewer than 250 employees, active in the B2C service industry in the UK. This typical respondent has around five years of work experience and a background education in marketing or sales. Table 2 has the descriptive statistics.

All scales are from the current literature. Adaptations are indicated. The constructs, items and sources are in the  Appendix.

Individual EO is the self-assessed proclivity towards innovativeness, proactivity and risk-taking (J. G. Covin and Slevin, 1989), measured with a ten-item scale (Bolton and Lane, 2012).

EI is the self-acknowledged conviction to start a new venture and to plan to do so at some future moment (Thompson, 2009), a construct with good predictive ability for subsequent new venture creation (Kautonen et al., 2015), measured with a three-item scale (Kautonen et al., 2015).

Ambidexterity refers to joint exploration and exploitation, i.e. joint hunting and farming orientation, each measured with two four-item scales (DeCarlo and Lam, 2016; Lam et al., 2019). Farming orientation measures sales professionals’ preference for developing existing customers, hunting orientation measures sales professionals’ preference for acquiring new customers. Assuming that hunting and farming are non-substitutable and interdependent, I follow prior studies and combine the scales into a single ambidexterity scale by multiplication (Lam et al., 2019; Mura et al., 2021; Zimmermann et al., 2020). Ambidexterity thus refers to the joint preference for developing existing customer relationships and simultaneously pursuing new customer acquisition.

Role ambiguity refers to the perceived lack of information to perform the job adequately and the perceived uncertainty about job expectations (J. Singh, 1998), measured with a six-item scale (Rizzo et al., 1970).

Controls include age, gender, education, job experience, industry, seniority (managerial vs non-managerial role), job function (sales vs account manager), company size, company age, emotional attachment to company, price-value versus competitors (single-item scales) and the two variables resistance to change, a five-item scale (Inyang and Jaramillo, 2020; Oreg, 2003) and competitive intelligence (Rapp et al., 2011), a three-item scale, used as marker variable and instrument variable, respectively, in prior related research.

A review of the sales performance literature critically notes that 97% of studies use a single measurement approach, capturing either results or behaviours (Kerr and Marcos-Cuevas, 2022). Sales performance is multifaceted and spans activities, results, conversions and relationships (Bolander et al., 2021). As in prior studies (Hinterhuber and Khan, 2026; Miao and Evans, 2007), I assess performance using both a qualitative and a quantitative indicator.

Behaviour performance measures the results, relative to colleagues, of qualitative indicators of sales effectiveness, such as relationship quality with customers, ability to solve customer problems and active listening. Seven-item scale, expanded from a four-item scale (Miao and Evans, 2007) based on the original 31-item scale (Behrman and Perreault, 1982).

Selling performance measures the results, relative to colleagues, of quantitative indicators of sales effectiveness, such as sales, orders, win rates, market share and profits, using a nine-item scale (Terho et al., 2015), adapted from the original three-item scale (Homburg et al., 2011).

This research is exploratory: the objective is not to test extant theory but to build new theory. Hence, partial least squares structural equation modelling (PLS-SEM) is an appropriate analytical technique: “In situations where theory is less developed, researchers should consider the use of PLS-SEM” (Hair et al., 2021, p. 15). In contexts where the focus is more on exploration than confirmation, PLS-SEM is preferable to covariance-based SEM (Hair et al., 2021). In contrast to covariance-based models that minimise covariances between the sample and those predicted by the assumed true model, PLS-SEM is a variance-based model that maximises R² values of the target, endogenous constructs. PLS-SEM is particularly indicated for small sample sizes, for non-normally distributed data and for calculating predictive scores, making it the analytical technique of choice for a rapidly increasing number of studies in marketing and other domains (Hair et al., 2012, 2017).

Firstly, it is necessary to assess construct reliability (i.e. Cronbach’s alpha and composite reliability) and validity (i.e. average variance extracted [AVE]). This study follows the recommendations (Hair et al., 2021) on cut-off values for reliability and validity for indicators. All constructs have Cronbach’s alpha values and composite reliabilities above 0.70 and AVE values above 0.50 (see Table 3). Hence, the constructs display satisfactory reliability and validity according to standard assessment criteria.

Next, discriminant validity must be assessed. Firstly, none of the indicators’ cross-loadings exceeds their outer loadings. Secondly, the Fornell–Larcker criterion (Fornell and Larcker, 1981) is met, given that none of the constructs’ correlations with another construct exceeds the square root of their AVE (Table 3).

Finally, since cross-loadings and the Fornell–Larcker criterion do not always accurately assess discriminant validity, I follow the literature and also employ the heterotrait-monotrait ratio (HTMT) of correlations (Henseler et al., 2015). The HTMT evaluates “the average of the heterotrait-heteromethod correlations (i.e. the correlations of indicators across constructs measuring different phenomena), relative to the average of the monotrait-heteromethod correlations (i.e. the correlations of indicators within the same construct)” and is used to examine whether it is below a specific threshold (Henseler et al., 2015, p. 121). The HTMT threshold is 0.90. I examine all indicators and find that none exceeds the threshold value. Taken together, these results indicate discriminant validity.

Common method bias is a possible concern in cross-sectional data collected from a single informant (Podsakoff et al., 2003). It could appear that the use of self-assessed performance data constitutes a fatal flaw of the study. Researchers have, however, long argued that subjective performance evaluations by sales professionals are valid: “In summary, self-report evaluations are most appropriate when responses can be confidential, when much of the effort is not directly observable by the manager, when aspects of performance are not reflected in quantitative data, when multicompany samples are used, and when a reliable scale has been developed to tap different aspects of performance” (Behrman and Perreault, 1982, p. 357). This argument is plausible, but for sure not empirical. We need evidence, ideally meta-analytical.

In their review of 40 studies, Bommer et al. (1995, p. 597) find that objective and subjective employee performance assessments are significantly related, and that objective and subjective measures are “reasonably substitutable” when the same underlying performance construct is being measured. The meta-analysis of sales performance studies reviewed earlier (Churchill et al., 1985) finds no evidence that self-reports are any better or worse than managerial evaluations at measuring performance. Furthermore, one of the meta-analytic reviews on role ambiguity and performance reviewed earlier finds that the results for self-rated performance are “similar” to results based on supervisor ratings or objective performance data: “researchers and practitioners may obtain some useful information from self-report data on stress and performance” (Gilboa et al., 2008, p. 257).

I follow the guidelines and apply the procedural remedies against common method bias: participant anonymity, temporal separation between the dependent and independent variables and unambiguous scale items (Podsakoff et al., 2003). Furthermore, a meta-analysis of the accuracy of key informants involving over 100 studies finds that the average accuracy is “adequate” for identifying medium and large effects (Homburg et al., 2012, p. 605), and, noting the lower reliability in large organisations, concludes it is “particularly useful for research on small and medium-sized enterprises” (2012, p. 606). I note that in this study, about 50% of respondents work in small- and medium-sized companies. Crucially: “Key informants are significantly more reliable for constructs that refer to the present, […] and that address salient events. In addition, there is some evidence that they are less reliable for constructs referring to the firm environment” (2012, p. 606). It should be noted that this study’s focal constructs all capture present, role-proximal perceptions, rather than judgements about the broader environment.

In addition, I also examine the effect of two variables, used separately in two prior studies as instrument variable and marker variable – competitive intelligence, a three-item scale (Rapp et al., 2011), used as instrument variable in a study on value-based selling (Mullins et al., 2020b); resistance to change, a five-item shortened scale (Oreg, 2003), used as marker variable in a study on salesperson implementation of sales strategy (Inyang and Jaramillo, 2020) – on the dependent variables. I do not find significant effects of these two variables (see Table 4).

Taken together, the data suggest that common method bias does not substantially affect the conclusions of the current study or those of other studies relying on self-reported data by sales professionals (Miao and Evans, 2013; Román and Iacobucci, 2010).

Endogeneity occurs when the independent variable correlates with the error term as a result of measurement errors, omitted variables, simultaneity or other factors (Ullah et al., 2018). In the context of the current study, I address endogeneity via control variables. “A straightforward way of handling, or at least reducing, endogeneity is to specify a set of control variables” (Hult et al., 2018, p. 3). This is because “[u]sing several control variables can also help in controlling for omitted variable bias” (Ullah et al., 2021, p. 3). More specifically: “Best practice dictates the inclusion of all theoretically relevant covariates in a statistical model” (Anderson et al., 2022, p. 9). I include a very broad set of control variables, including industry, firm size, firm age, respondent job experience, age, gender, educational background, emotional attachment to company and the relationship between value and price of own products/services versus competitive offers. Control variables are all non-significant, except for industry (see below).

In sum, the available evidence indicates that common method bias and endogeneity do not appear to be present in the current study.

This study uses 5,000 bootstrapping re-samples and one-tailed tests to estimate the significance of the structural model’s standardised path coefficients (ß) (Figure 2). Because PLS-SEM does not make distributional assumptions, bootstrapping is used to allow for significance testing of these path coefficients (Hair et al., 2021). Path coefficients are listed in Table 4.

Individual EO is positively and significantly related to behaviour performance (H1a, ß = 0.19, p  < 0.01) and selling performance (H1b, ß = 0.14, p  < 0.1). EO thus increases behaviour performance and selling performance. EO is also positively and significantly related to ambidexterity (H2, ß = 0.64, p  < 0.01), and ambidexterity, in turn, is positively and significantly related to behaviour performance (H3a, ß = 0.168, p  < 0.05) and selling performance (H3b, ß = 0.30, p  < 0.01).

As expected, EO is significantly and negatively related to role ambiguity (H5, ß = −0.41, p  < 0.01) and, finally, role ambiguity is significantly and negatively related to both behaviour performance (H6a, ß = −0.41, p  < 0.01) and selling performance (H6b, ß = −0.24, p  < 0.01).

Following prior studies that also examine interaction effects (Ingenbleek et al., 2003), the moderation analysis reveals that price-value versus competitors negatively moderates the ambidexterity-selling performance relationship (ß = −0.10, p  < 0.05), indicating that the positive effect of ambidexterity on selling performance is weaker when sales professionals perceive their offering’s price-value relationship as more favourable relative to competitors. Stated differently, a strong competitive price-value relationship partly substitutes for ambidexterity. When the offer is already attractive relative to competitors, the marginal benefit of jointly hunting and farming is smaller; when the firm’s price-value position is less favourable vis-à-vis competitors, ambidexterity matters more for selling performance (Table 4). Because price-value versus competitors is measured with a single-item scale, the observed moderation effect must be interpreted very cautiously, in line with Hair et al. (2021, p. 246), who “advise against” single items for abstract, unobservable moderators.

All controls are non-significant except industry, coded as discretionary versus non-discretionary, which is significant for one dependent variable, consistent with meta-analytic findings (Hizarci et al., 2023; Rauch et al., 2009).

I follow the standard analytical guidelines (Nitzl et al., 2016) and report data for mediation analysis in Table 5. As per guidelines (Hair et al., 2021), this study tests the direct path alongside the indirect paths. The total coefficient for the EO-behaviour performance is 0.475 with a 90% bias corrected confidence interval [0.376;0.560], the total coefficient for the EO-selling performance is 0.437, interval [0.330;0.649]. The mediation from EO is therefore complementary or partial for both paths (Chin, 2010). H4a, H4b, H7a and H7b are thus supported.

The final structural model is in Figure 2.

The model explains a good amount of the variance: 47% in behaviour performance and 39% in selling performance (Table 4). The effect size f² indicates the change in the R² value of the endogenous constructs when a specific exogenous construct is omitted (Hair et al., 2021): the effect size is largest for ambidexterity and role ambiguity and above 0.02 for all other constructs, indicating good model quality (Table 6).

Finally, PLS allows an assessment of the predictive relevance of the exogenous constructs on the endogenous variables by estimating Stone–Geisser’s Q² (Stone, 1974; Geisser, 1974). The structural model displays good predictive relevance (Table 4). Specifically, the Q² value for individual EO on ambidexterity is large (0.395), for EO on role ambiguity is medium (0.1) and is medium again for EO on behaviour performance (0.27) and selling performance (0.25).

Finally, addressing unobserved heterogeneity is important, as results could be biased by heterogeneity resulting from unobservable group differences (Sarstedt and Ringle, 2010). The most prominent latent class approach for ascertaining that heterogeneity does not influence results, and thus supporting analysis of a single model, is finite mixture PLS (FIMIX-PLS) (Hair et al., 2021). This study performs FIMIX-PLS to detect whether an unobserved group biases path coefficients in the model. FIMIX-PLS shows that for a two-segment solution, the first segment (∼80%) is much larger than the second segment (∼20%), indicating that an unidentified group does not materially bias model estimates (Hair et al., 2018; Hirsch et al., 2018).

I also test an alternative, apparently related, yet substantially distinct construct: EI. Based on the theory of planned behaviour that conceptualises intentions as antecedents of behaviour (Ajzen, 1991, 2011), several meta-analytic reviews find that intentions explain about 28% of the variance in subsequent behaviour across domains (Armitage and Conner, 2001; Sheeran, 2002).

Does EI – the “self-acknowledged conviction by a person that they will set up a new business venture and consciously plan to do so at some point in the future” (Thompson, 2009, p. 687)– lead to subsequent new venture creation? Several studies find that entrepreneurial intentions explain about 30% of the variance in subsequent start-up activities (Delanoë-Gueguen and Liñán, 2019; Kautonen et al., 2015; Kibler et al., 2014; Liñán and Fayolle, 2015). EI is thus a “good predictor of subsequent entrepreneurial actions” (Kautonen et al., 2015, p. 668). Stated differently: EI identifies future entrepreneurs quite reliably.

I thus test, analogously to the hypothesised relationships for EO, the effect of EI on ambidexterity, role ambiguity and sales performance. Table 7 lists the path coefficients.

EI is not significantly related to behaviour performance (ß = 0.02, p  = n.s.) or selling performance (ß = 0.09, p  = n.s.). EI is positively and significantly related to ambidexterity (ß = 0.128, p  < 0.01) and ambidexterity, in turn, is again positively and significantly related to behaviour performance (ß = 0.246, p  < 0.01) and selling performance (ß = 0.35, p  < 0.01). EI is not significantly related to role ambiguity (ß = 0.044, p  = n.s.) and, finally, role ambiguity is significantly and negatively related to both behaviour performance (ß = −0.438, p  < 0.01) and selling performance (ß = −0.277, p  < 0.01). Also, here the controls are not significant except for industry. Thus, the post hoc analysis indicates that EI has no direct effect on either behavioural performance or selling performance.

A very thoughtful paper argues that there are three ways to generate courageous research ideas: observe, bridge and challenge (Kock et al., 2020). This study is the result, as the opening quote suggests, of the “observe the world” approach (p. 1142), where a vivid real-world phenomenon sparked the curiosity of this writer. A top manager’s aspiration for salespeople to embody entrepreneurial traits seems reasonable – after all, entrepreneurs are often linked to innovation, risk-taking and self-determination. However, one critical question remains unasked and unanswered by this manager: if salespeople exhibit entrepreneurial traits, does this really lead to high performance? This is the question that, first of all, a master’s thesis student of this writer asked and that sparked the interest in this study.

A foundational principle of this study is the idea that EO should be analysed at the individual level – and not at the level of the firm by polling CEOs. Practising managers seem to concur: Louis Gerstner (Gerstner, 2015), credited with turning IBM around in the face of imminent bankruptcy, commented: “Why did four CEOs not turn around AT&T and it had to be sold to one of its former childs? [Sic: Author comments: SBC, a former subsidiary, purchased AT&T in 2005]. Why did three CEOs not turn around Eastman Kodak? Did they not see digital photography? They invented it! And at AT&T, did they not see mobile? They invented it! […] What happens is: CEOs say: ‘We have to change, there is where we are going’. And they point somewhere else and nobody turns and follows them”. Gerstner thus seems to suggest that any EO of the CEO is irrelevant unless individual employees think and act differently. Prior research has called for “reintroduc[ing] the individual in entrepreneurial orientation research” (Krauss et al., 2005, p. 315). This study answers this call.

By bridging the entrepreneurship and sales management literatures, this study addresses the important research question of whether and how EO matters to individual performance by identifying the underlying mechanisms while examining, post hoc, also the role of EI for individual performance. This is an important research question, since entrepreneurial behaviour by salespeople is today still a “black box” (Y. Zhao et al., 2025, p. 985). The empirical findings have important theoretical and practical implications.

Firstly, I show, to the best of my knowledge, for the first time, that EO benefits sales professionals, thus extending the literature that has documented the positive effects for CEOs (Rauch et al., 2009), academics (Clarysse et al., 2011) and entrepreneurs (Krauss et al., 2005). Addressing a call for such studies (Kollmann and Stöckmann, 2014), this study examines how EO is converted into specific activities that drive tangible performance outcomes. Specifically, I document that both ambidexterity and low role ambiguity play a fundamental role in converting an individual entrepreneurial disposition into behavioural and financial performance outcomes. In doing so, I answer calls for further research illuminating intermediary mechanisms in the EO–performance relationship (Jiang et al., 2021; W. J. Wales et al., 2013).

The two mediating variables that convert EO into superior performance are ambidexterity and role ambiguity: EO is inherently about bringing about change, and ambidexterity operationalises the specific actions that individuals can implement to bring about change. Yet, while EO-induced change is beneficial, scholars call for the examination of elements related to the fundamental need of all organisations of “stability, focus and control” (J. Covin and Wales, 2019, p. 11). Low role ambiguity is one such factor. The two mediating variables in this study are thus like yin and yang, complementing and reinforcing each other: yin, the stable element, is represented by low role ambiguity and yang, the ever-changing element, is represented by high ambidexterity, i.e. the ability of sales professionals to acquire new customers while at the same time harvesting existing customer relationships. This study, therefore, also answers a call for empirical studies examining the so far neglected static elements of EO (Covin and Wales, 2019).

Furthermore, there is some potential evidence suggesting that a more favourable price-value relationship vis-à-vis competitors weakens the positive effect of ambidexterity on selling performance: a strong competitive price-value proposition may partly substitute for ambidexterity, although the finding should be interpreted very cautiously given the single-item measurement of the moderator (Hair et al., 2021).

Secondly, this study highlights the differential impacts of EO and EI on individual performance. EI, a construct measuring an action orientation, surprisingly has no direct effect on performance; EO, a construct measuring a disposition, instead has a direct effect. From a theoretical perspective, these findings document that inside established companies, thinking like an entrepreneur is more beneficial than acting like one. So far, studies have nearly exclusively explored the effect of EI on new venture creation, finding strong direct effects (Meoli et al., 2020; Zhao et al., 2010). Inside a company, the relationships are more complex: a high EI increases ambidexterity, which, in turn, increases performance – there is no direct relationship to performance. This is in stark contrast to EO, which has strong direct and indirect performance effects. EI, like EO, is associated with ambidexterity, indicating the ability to efficiently harvest opportunities available today and to creatively build opportunities available only in the future. This disparity in the chain between EO/EI and performance points towards the very different nature of individuals that think like entrepreneurs (high EO) versus those who report specific entrepreneurial intentions (high EI): Entrepreneurship is an “extreme setting” and an “emotional rollercoaster” where success depends, according to a meta-analytic review, more on emotional intelligence than on analytical intelligence (Allen et al., 2021, p. 352). Individuals high in EI are the likely entrepreneurs of tomorrow – willing to assume personal risks (Schumpeter, 1911), emotionally intelligent and probably nurturing ambitions beyond merely selling (Åstebro and Thompson, 2011), traits that are all beyond the scope of this study – but they are not high-performing sales professionals. Individuals with high EO, by contrast, are. In this light, EO is a dynamic capability (Wales et al., 2021): it has strong direct effects on individual performance, it is associated with ambidexterity as well as with role ambiguity, and it thus helps individuals in sense-making, in shaping their external environment, and in encouraging them to exploit and explore at the very same time.

Thirdly, this study is one of the first to empirically link the entrepreneurship and selling literatures, two streams that have long evolved but, counterintuitively, with little theoretical overlap. “We contend that selling is a fundamental entrepreneurial activity, but that it is rarely an area of focus within core entrepreneurship theories” (Matthews et al., 2018, p. 691). Sales is arguably the function of a company that has the largest opportunity for entrepreneurial action: “Senior sales leaders, field sales managers, and salespeople are essentially ‘running a business’ within the overall company business operation” (Ingram et al., 2005, p. 150). It is thus surprising that we know so little about the effect of EO in the context of selling. Selling consists of exploitative actions to capture opportunities at hand and explorative actions to create future opportunities. This study documents that research at the intersection of entrepreneurship and selling is fruitful and helps uncover a set of mechanisms that contribute to our understanding of entrepreneurial processes in established firms, thereby advancing theory. Results show that viewing processes – here: the sales process – in established firms through the lens of the entrepreneurship literature is extremely fruitful and could thus serve as a benchmark for future studies.

The profession of selling formally developed in the 1800s when John Patterson, founder of National Cash Register Company, created one of the first formal sales training systems, consisting of four phases – approach, proposition, demonstration and close – based on rigid scripts with the objective of creating a market for the nascent cash register technology and of overcoming customer objections (Inks et al., 2019). Today, sales scripts are everywhere: Sandler, a sales training company calling itself the world’s largest, bases its training essentially on a modern version of Patterson’s scripts. Scripts are, needless to say, the death of innovativeness, proactivity and risk-taking. Results of this study suggest letting each sales professional find his or her voice in the sales process, stimulating as much as possible the individual EO and using scripts only inasmuch as they help reduce role ambiguity or increase ambidexterity. For many organisations, this could well mean abandoning scripts altogether.

There is likely no senior business leader who does not wish to make her or his organisation more entrepreneurial. It is tempting to think that the best salespeople are aspiring entrepreneurs. There are examples: Howard Schultz, founder of Starbucks, was a “star employee” as Xerox sales rep (Krass, 1999, p. 300), Ross Perot, founder of EDS, a “top salesman” at IBM (p. 382); Thomas Watson, IBM’s founder, was a “selling genius” working for John Patterson (p. 424); Alfred Fuller, founder of the Fuller Brush company, “excelled” as a salesman (p. 365); Herman Lay, founder of Frito-Lay, was a salesman (p. 434); and Samuel Goldwyn, founder of the eponymous movie studio, was a “top salesman” at a glove factory (p. 486). It is indeed tempting to conclude that the best-performing salespeople are all aspiring entrepreneurs. Tempting, but wrong – since the evidence, littered with some of the world’s most successful entrepreneurs, is anecdotal only. This quantitative study highlights the crucially important distinction between EO and EI. Thinking like an entrepreneur increases sales performance; behaving like an entrepreneur preparing to set up their own venture does not. The key drivers for increased performance in the sales organisation are thus innovativeness, proactivity and risk-taking, ambidexterity and role clarity, all of which substantially increase sales performance. The opening quote of this paper (“Being a salesperson is like being an entrepreneur in your own territory”), delivered by a sales manager of a multibillion-dollar company to newly hired sales managers, one of whom was a former graduate student of this writer, is best qualified as follows: Salespeople should think like entrepreneurs, but not aspire to become entrepreneurs.

The sales profession is highly relevant to national economies. In the USA, there are currently more than 13 million persons working in sales, amounting to roughly 9.4% of all salaried positions (Bureau of Labor Statistics, 2021). Sales positions have the largest overall variability in output: across all jobs, the top 1% are more than twice as productive as the average worker, but for sales positions, the ratio is as high as 4:1 (Hunter et al., 1990). This high output variability implies a high economic and social cost of underperformance. Behind every great company, there is a high-performing salesforce. “When salespeople do well, the organization is likely doing well […] salespeople still represent the most visible signs of corporate success” (Rich et al., 1999, p. 41). When organisations do well, the overall economy is likely doing well. There is thus a public policy interest in companies and in improving the performance of the sales function.

What drives sales performance? Sales performance is driven by the development of an entrepreneurial mindset – innovation, proactivity and risk-taking – by developing capabilities to efficiently manage today and to creatively prepare for the future and, finally, by clarifying the changing role of sales professionals: all these measures improve performance, thereby creating societal value.

Several choices underlie the present study. Some of these constitute potential limitations. Future studies should assess their importance and impact on the validity of the findings.

This is a single-country study, like other pioneering EO studies (Clarysse et al., 2011; J. G. Covin et al., 2006; de Jong et al., 2015; Engelen et al., 2014; Ferreras-Méndez et al., 2022; Jiang et al., 2021; Kollmann and Stöckmann, 2014; T. Lumpkin and Dess, 2001; Palmer et al., 2019; Perry et al., 2016). This focus, typically on countries like the USA, the UK, China and Germany, seems to characterise overall research patterns: a recent bibliometric review finds that research in international EO has “lagged behind” and that the overall research output is “quite limited” (Gupta et al., 2021, p. 75). In addition, research on the relationship between EO and national culture was found to be equally “very limited” (J. G. Covin and Miller, 2014, p. 22). Research examining the interaction between EO and national culture could be highly fruitful, thus exploring, not unlike Hofstede’s country scores (Hofstede, 2001), how innovativeness, proactivity and risk-taking differ across countries and how these differences translate into overall performance. Equally interesting for future studies is the question of whether R² values and the path coefficients in this model – values that are quite high by any standard (Chin, 1998; Hair et al., 2021) – hold true in other contexts.

This study collects a single data point from a single respondent across hundreds of companies, an approach taken by literally all other studies of the chain between EO and performance. The next logical step is multi-level studies where, in a single company, members of the top management team, middle managers and managers with operational responsibilities are polled on EO, which would then allow examining the relative contribution of EO transmitted through supervisors vis-à-vis the EO resulting from individual personality traits. At the moment, research has examined individual psychological traits as key antecedents of individual EO (Rauch and Frese, 2007). True multi-level studies – data collected from three different hierarchical levels – are literally non-existent in EO and, more broadly, in management research, but they would make meaningful contributions. In this sense, the study by Liu and Xi (2022), collecting data from CEOs and middle managers and finding that CEOs influence the confidence of the latter, is an important preview of the design of studies to come.

As outlined in the introduction, a near dogmatic obsession with defining EO at the firm level based on CEO self-assessments has characterised research for decades. This is one important self-afflicted limitation: several other disciplines, first of all entrepreneurship, but also strategic management and others, have witnessed an increased interest in micro-foundations, in the exploration of relationships between individual personal traits or individual biases and outcomes at the organisational level (Felin and Foss, 2005; Helfat et al., 2007; Hinterhuber and Liozu, 2017; Rauch and Frese, 2007). In this study, I define EO as an individual-level propensity, and one could view this as a limitation given that the literature predominantly defines the construct at a higher level of abstraction. I contend that the logical connection between individual performance outcomes and individual activities requires a definition of EO at exactly this level.

Several studies assert that salespeople have more freedom to act entrepreneurially than those fulfilling any other function in the company (Jones et al., 2000; Morris et al., 1990). Examining the impact of EO on the performance of salespeople specifically is thus warranted. Future studies could explore the effect of EO on the performance of other important organisational functions, such as new product development, R&D, HR or finance. It is not at all clear, but it is theoretically very interesting, whether for other organisational functions this effect is positive.

Individual EO is a fascinating research domain with many promising avenues for future research.

A further fruitful avenue for future studies is the examination of the moderating role of context – who, when and where – on key variables identified in this and other studies. The authors of a prior meta-analytic review on EO call for further studies examining “additional moderators” (Zhao et al., 2010, p. 398). Newbert et al. (2022), p. 2 write: “We have tended to employ theories meant to predict context-free phenomena to propose and/or test conceptual models that are not context-free”. Key, according to this line of thought, is the examination of “whether and how our theories hold up across distinct locations, at various times, and for different types of entrepreneurs”. We need to examine ‘boundary conditions’ to better understand when key relationships “no longer hold or reverse direction”. Combs et al. (2021), p. 347. In sum, the separate or, ideally, joint examination of micro-foundations and contextual variables will help to identify boundary conditions of our theories and contribute to outcomes at the organisational level.

A further fruitful avenue is the development of multi-level frameworks on antecedents of EO. On one level, external factors are antecedents: extant research points out that external pressure is beneficial for EO (Williams et al., 2024). The next level is organisational factors: external focus (Brettel et al., 2015) and empowering leadership behaviours (Kim and Beehr, 2023) are conducive to an EO. Finally, micro-foundations or individual-level characteristics. The literature in marketing, strategy and entrepreneurship has all begun to explore causal links between individual mental schemas and outcomes at the organisational level (Barney and Felin, 2013) – but research has only just begun, and future opportunities are substantial. Fayolle and Liñán (2014, p. 664) write: “Significant progress” could be made if future research investigated “the role and the importance of mental prototypes, cognitive scripts, mental schemas”. This interest in the psychology of entrepreneurial actions and dispositions is widely shared: “Future research could further integrate constructs from I/O psychology to provide a more comprehensive approach towards entrepreneurship” (Frese and Gielnik, 2014, p. 431). Future studies could thus integrate these factors into a single multilevel framework to assess the relative importance of external, organisational and individual drivers in fostering EO.

This study is the result of a very long process that benefited from the contribution of many individuals and organisations. Ester Zanatta, one of the first master’s thesis students after joining Cà Foscari in 2018, had the original idea of examining the effect of entrepreneurship among sales professionals in her thesis. The catalyst, by her account, was an international sales meeting she attended, as an intern, at PTC, where a sales manager delivered the opening quote. This opening quote by this unnamed PTC sales manager led then to the main study, conducted in 2020. Hinterhuber and Partners GmbH provided generous research funding for data collection and analysis. Evandro Pollono, Hinterhuber and Partners, provided assistance in the programming of the online survey. Christian Nitzl, Universität der Bundeswehr, München, provided advice and consulting for the PLS-SEM data analysis. Earlier drafts of this paper have been accepted as conference papers at the 83rd Academy of Management conference in Boston (USA) in 2023, at the 8th Biennial Enhancing Sales Force Productivity Conference in Tampere (Finland) in 2024 and at the 22nd Academy of Innovation, Entrepreneurship and Knowledge Conference in Alcala de Henares (Spain) in 2025. Conference participants and reviewers provided ideas that helped to improve the paper over all these years. Finally, Wesley Friske, Editor in Chief and two anonymous reviewers provided superb suggestions for improving the positioning, clarity and theoretical contribution of this paper. The author thank you all.