Intrapreneurship, as an evolving field of research, is conceptually derived from entrepreneurial literature. However, the concept of intrapreneurship is intrinsically different from that of entrepreneurship. An entrepreneur is a person ‘who organizes and operates a business through taking personal and financial risk’, while an intrapreneur is ‘an individual within an organization or company who creates and/or takes responsibility for transforming an idea into a profitable venture through taking an innovative approach’ (Haller 2015:10). Intrapreneurship can also be seen as

Intrapreneurship is therefore the practice of entrepreneurship within an existing organisation, characterised by behavioural intentions departing from the customary (Coulthard & Loos 2007; Kuratko & Hodgetts 1995).

Despite large volumes of research on the topic, prominent authors in the field of intrapreneurship have noted a lack of consistency between definitions of intrapreneurship, suggesting that it is still in the process of determining its conceptual identity (Sandberg 2000; Sharma & Chrisman 1999; Zahra 1991). Intra-organisational activities that can be classified as ‘intrapreneurial’ have been described through such terms as ‘corporate entrepreneurship’, ‘intra-corporate entrepreneurship’, ‘corporate venturing’ and ‘internal corporate entrepreneurship’ (Antončič & Hisrich 2004; De Villiers-Scheepers 2012). Despite this array of differing terminologies attempting to label intrapreneurial activity, the scope can broadly be categorised into (1) the pursuit of entrepreneurial opportunity, (2) new market entry or (3) new venture creation (Gartner 1989; Lumpkin & Dess 1996; Shane 2003). Common characteristics that can be found across these diverse terms include innovation, risk-taking and proactiveness. These traits are also traditionally associated with entrepreneurial actions (Antončič & Hisrich 2001; Azami 2013; Stewart 2009). Although innovation is an integral component of both entrepreneurship and intrapreneurship, with Covin and Miles (1999:49) describing innovation as being ‘at the center of the nomological network that encompasses the construct of corporate entrepreneurship’, ongoing innovations within organisations should not necessarily be used synonymously with intrapreneurship. While maintaining a constant stream of product, service and process innovation is necessary to thrive in an increasingly competitive business environment, intrapreneurship is not simply equivalent with innovation. Intrapreneurship particularly concerns itself with the execution component of turning an innovation into a profitable business venture (Herve 2015). In addition, not all innovations originating from an organisation can be considered intrapreneurial, as this is dependent on where the idea originates from. Traditionally, in a corporate setting, ideas and product decisions are filtered from the top down, while in intrapreneurial organisations the idea for a new venture is likely to originate from the bottom up, that is, from any employee within the organisation (Haller 2015).

Despite these nuances, intrapreneurship can be the source of valuable contributions (Antončič & Hisrich 2004). These contributions, as described in numerous studies, include improvements in organisational productivity and performance, increased profitability, increased revenue, higher growth, enhanced organisational innovativeness, development of new products, processes or technologies, as well as improved competitive positioning (Antončič & Hisrich 2001; Bosma et al. 2012; Brizek 2014; Covin 1991; De Jong & Wennekers 2008; Felício, Rodrigues & Caldeirhina 2012; Gawke, Gorgievski & Bakker 2017; Hastuti et al. 2015; Kacperczyk 2012; Zahra & Covin 1995).

The dynamic nature of marketplaces, consumer behaviour and the rapidly changing competitive landscape place increasing importance on promoting intra-organisational processes that allow employees to transform opportunities into innovation (Hisrich & Kearney 2012). Intrapreneurial orientation is vital for any business, but particularly so for small businesses, as it positively affects survival and competitiveness in the marketplace (Kakati 2003). An organisation’s ability to harness intrapreneurial talent over a period of time builds internal capabilities, allowing the organisation to reinvent itself permanently, ultimately leading to improved performance and above-normal returns (Covin & Slevin 1989; Felício et al. 2012; Hecker 2017; Lumpkin & Dess 1996; Olokundun et al. 2017; Zahra 1991).

The increased focus on the employee as a source of innovation results in the recognition of the employee acting as an intrapreneur, with the skills, abilities and temperament to create value for the organisation by exploiting opportunities (Ma, Liu & Karri 2016). The tendency of employees to act in an intrapreneurial manner is referred to as IO. Although IO can be characterised by a multitude of attitudes and behaviours, it is most commonly expressed in risk-taking propensity, innovativeness and proactiveness (Bolton & Lane 2012; Matsuno, Mentzer & Özsomer 2002; Sinha & Srivastava 2015). As little research has been conducted on individual traits of IO, other studies propose alternative traits such as achievement orientation, autonomy, competitive aggressiveness and managerial skills (Goosen, De Coning & Smit 2002; Luchsinger & Bagby 1987; Lumpkin & Dess 1996). While no universal agreement exists as to which individual-level factors foster IO, early evidence nonetheless indicates that IO, in a varying combination of factors, has the potential to influence organisational growth. Accordingly, the following proposition is made:

In 2017, Schachtebeck conducted a systematic review of prominent Entrepreneurial Orientation (EO) and IO constructs, supported by a Delphi study with subject matter experts and semi-structured interviews with SME employees. Based on the study findings, he categorises IO into managerial factors and employee-related or ‘personal elements’ factors.

Schachtebeck (2017) identifies six managerial characteristics, which are similar to those proposed by Goosen et al. (2002). These characteristics include (1) intracapital (the provision of resources in spite of risk), (2) goals (future-oriented goal-setting), (3) intrapreneurship championing (support for intrapreneurship through structure, systems, processes and financial support), (4) rewards and innovation systems (rewarding desired goals and behaviours), (5) intrapreneurial freedom (empowerment of employees) and (6) communication (open channels of communication and information sharing).

The importance of support for IO at management level through championing, empowerment, autonomy, rewards and communication has been stressed in a number of studies (Fasnacht 2009; Goosen et al. 2002); however, at an individual level, the growth impact has not been explored. Based on these factors, the following propositions are made:

In terms of personal elements factors, Schachtebeck (2017) identifies the following six qualities: (1) risk-taking propensity, (2) innovativeness, (3) proactiveness, (4) personal control (PC), (5) self-esteem (SE) and (6) achievement orientation. These IO-related elements are seen as bolstering SME growth. Risk-taking refers to the acceptance of uncertainty and tolerance of potential financial and reputational loss (Aarakit 2010; Goosen et al. 2002; Rauch et al. 2009). Innovativeness means willingness to engage in creative thinking in pursuit of novel discovery (Antonites & Nonyane-Mathebula 2012; Miller 1983). Proactiveness is a future-oriented, opportunity-seeking mindset built on adaptability, in pursuit of competitive advantage (Alarape 2014; Urban 2012). Personal control refers to an employee’s self-perception over the control of his or her work (Robinson et al. 1991). Self-esteem is an individual employee’s self-belief in his or her skills (Krishnan & Kamalanabhan 2015; Robinson et al. 1991). Achievement orientation (AO) is an individual’s interpretation and reaction to the success and growth of a venture (Choe & Loo 2013; Jain, Ali & Kamble 2015; Robinson et al. 1991). The proposed IO elements of risk-taking, innovativeness and proactiveness overlap with those used in a number of studies, most notably by Urban and Oosthuizen (2009), Bolton (2012), Bolton and Lane (2012) and Farrukh et al. (2017). On account of this literature, the following propositions are made:

The remaining personal elements factors of IO, namely PC, SE and AO, feature prominently in previous studies, notably those of Van Wyk and Boshoff (2004), Shetty (2004), Jain and Ali (2012) as well as Sinha and Srivastava (2013, 2015, 2016). On the basis of these studies, the following propositions are suggested:

The adapted measuring instrument is based on an integrated, conceptual framework crafted as part of an empirical study by Schachtebeck (2017). The conceptual framework, reflecting the propositions, personal elements and managerial factors described above, is depicted in Figure 1.

Intrapreneurial employees typically possess particular traits, such as creativity, risk-taking, patience, perseverance and proactiveness (Azami 2013). While the existence of these attitudes and traits in employees is a good point of departure for intrapreneurship, an enabling work environment is nonetheless necessary to foster IO. To harness these traits, management needs to resource, encourage and empower intrapreneurial employees (Azami 2013; Goosen et al. 2002; Sinha & Srivastava 2015). Coetzer (2016) argues that an organisation that designs managerial strategies to foster IO in employees improves its chances of organisational success. The integrated conceptual framework captures the interplay between managerial and personal factors. The framework therefore proposes that for an SME to grow, both managers and employees contribute to developing an individual-level IO.

The measuring instrument is based on the integrated conceptual framework presented in Figure 1. The instrument contains two sections. Section A investigates key demographic variables, while Section B is concerned with IO variables. The instrument measures key factors of IO as per the conceptual framework, categorised according to managerial or personal factors. Managerial factors include intracapital, goal-setting, intrapreneurship championing, reward and innovation systems, intrapreneurial freedom and communication. Personal factors include risk-taking, proactiveness, innovativeness, PC, SE and AO. The instrument contains eight growth-related items in order to relate IO factors to SME growth, which are operationalised in five growth variables. The instrument, in total, contains 50 items on a five-point Likert scale to accurately measure strength of perception. The instrument adapts items from other well-known instruments used in previous EO and IO studies. Management items (M1–M14) were adapted from Goosen et al. (2002); risk-taking (R1–R4), innovativeness (I1–I4) and proactiveness (P1–P5) were adapted from Bolton and Lane (2012); intrapreneurial orientation (IO1–IO3) was adapted from Stewart (2009); and personal control (PC1–PC4), self-esteem (SE1–SE4) and AO (AO1–AO4) were adapted from Robinson et al. (1991).

The population of the study included SMEs in nine provinces of South Africa. While no reliable data currently exist as to how many SMEs exist in South Africa, it is estimated that between 700 000 and 960 000 registered businesses exist, with around 434 000 of these reporting a loss or no taxable income (Absa SME Index 2014; Davis Tax Committee 2014; Finmark Trust 2010;). However, as the estimated number of businesses includes the informal sector, it is difficult to accurately gauge the exact size and extent of formal sector SMEs.

Between November 2017 and January 2018, questionnaires were distributed both online, in Google Forms, and in hard copy. Hard copy questionnaires were administered in five South African provinces, while online questionnaires were sent to SMEs listed in a number of SME databases. The sampling frame included SMEs listed in YelloSA, Ezee-Dex and the Small Business Directory.

For both survey approaches, a probability sampling approach was followed in the form of simple random sampling. A total of 1421 questionnaires was distributed, with 332 responses being received, resulting in an overall response rate of 23.3%. Of the 1421 questionnaires, 1091 were distributed online, yielding 72 responses, thereby resulting in a response rate of 6.6% for online responses. The remaining 330 questionnaires were distributed physically, yielding 260 responses, thereby resulting in a response rate of 78.8% for physical distribution. Twenty physical questionnaires were excluded from the analysis because of missing pages, while a further 22 questionnaires were regarded as invalid and thus excluded from final analysis as they were completed by owners of SMEs, therefore not falling within the target population. A total yield of 290 valid responses was therefore obtained.

Data were captured on Microsoft Excel and imported into the Statistical Package for Social Science (SPSS) version 25. SPSS software was used to analyse the data and generate statistics. Data analysis techniques included tests for reliability, sample adequacy, descriptive statistics, exploratory factor analysis (EFA), correlation analysis and regression analysis.

Descriptive statistics include means, standard deviation (SD) and percentages of key demographic variables. In the factor analysis, principal component analysis (PCA) was used to transform the original variables into a smaller set of linear combinations, ‘with all the variance in the variables being used’ (Pallant 2016:183). This yielded a new group of related factors. To determine correlation, Pearson’s product–moment correlation coefficient (r) was utilised, as it is the most frequently utilised technique as a measure of association, thereby indicating the strength of relationships between two variables (Pallant 2016; Tabachnick & Fidell 2007). A standard multiple regression analysis was performed, indicating how a set of independent variables predicts a dependent variable, namely SME growth (Pallant 2016). To test sample adequacy and the factorability of data, Bartlett’s test of sphericity and the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy were utilised. Bartlett’s test of sphericity should be significant (p < 0.05) for factor analysis to be deemed appropriate, while KMO should be > 0.6 (Pallant 2016).

Reliability of the measuring instrument was determined by means of Cronbach’s alpha (α), which was used to analyse the internal consistency of the sub-scales. A value of 0.7 is most commonly regarded as an acceptable level of internal reliability (DeVellis 2012; Malhotra 2010). Content validity of the measuring instrument was established by pretesting the questionnaire with four academics and seven employees of SMEs at various levels to ensure representativeness of the study population. Question wording and formatting was adapted before commencement of the study.

The front page of the questionnaire informed respondents about their rights, emphasising the voluntary nature of the research, the right of withdrawal at any time as well as the assurance of confidentiality and anonymity. No identifying information, other than basic demographic data such as age and gender, was collected in the course of the research. Ethical clearance for the study was obtained from the College of Business and Economics at the University of Johannesburg (ethical clearance number: FOM2017-BM065).

The results of the demographic profile of the respondents are presented in Table 1.

The demographic variables indicate that, in terms of gender, the sample of 290 respondents was almost equally distributed, as 149 were male (51.4%) and 141 were female respondents (48.6%). The majority of respondents were between the ages of 18 and 30 years (37.2%) and 31 and 40 years (36.6%), with the remainder being above the age of 41 years. No respondents were below the age of 18 years. While responses were received from all nine provinces, the majority originated from the economic powerhouse of South Africa, Gauteng (60.7%), followed by Mpumalanga (17.2%), North-West province (6.9%), Western Cape (5.5%), Free State (3.1%) and Limpopo (2.8%). In terms of race, the majority of the sample were black people (58.3%), followed by white people (25.5%), mixed race people (9.0%) and Indian/Asian population (5.9%). In terms of education, most of the respondents had a matric (school leaving) certificate (49.3%), 22.1% had an undergraduate degree and 15.5% had postgraduate degrees. The remainder of the respondents (13.1%) indicated basic schooling as their highest level of education. For length of service, most of the respondents were employed between 1 and 5 years (41.7%), followed by 6 and 10 years (24.1%) and more than 10 years (15.2%), thereby demonstrating good institutional knowledge. Of the respondents, 19% were employed for less than 1 year. While all the respondents were employees, a majority of them occupied non-managerial positions (48.3%), followed by middle, lower and top management positions, at 22.4%, 18.6% and 10.7%, respectively. From an organisational point of view, most of the SMEs in the sample were small businesses, employing between 5 and 20 employees (36.2%), 21–50 employees (29%) or less than 5 employees (9.3%). A quarter of employees (25.5%) reported working for a medium-sized business with more than 51 employees. Lastly, in terms of the sector the employees worked in, most of the respondents reported working in the catering and accommodation industry (15.5%), followed by manufacturing and finance and business services (both 14.5%). The remainder of the respondents reported working in a number of different industries, such as agriculture or construction.

Cronbach’s alpha was used to measure the internal consistency of the measuring instrument, with the original constructs prior to the factor analysis. Cronbach’s alpha for each underlying construct was computed to determine whether each construct displayed internal consistency. The findings are presented in Table 2.

While a Cronbach’s alpha coefficient of above 0.7 is commonly accepted as indicating good internal consistency, it is sensitive to the number of items in a scale, particularly for scales with fewer than 10 items where low coefficients (0.5) are common (Pallant 2016). In cases where a scale contains fewer than 10 items, reporting the mean inter-item correlation is recommended, with a minimum value of 0.2 considered as acceptable (Briggs & Cheek 1986; Pallant 2016).

As indicated in Table 2, proactiveness (0.760), IO (0.704), growth (0.829), PC (0.708) and management (0.873) all display a Cronbach’s alpha coefficient of greater than 0.7, thereby indicating good internal consistency. Risk-taking, innovativeness, AO and SE record a Cronbach’s alpha coefficient of 0.626, 0.685, 0.662 and 0.528, respectively. Because of the sub-scales each containing only four items, mean inter-item correlation values are reported. The mean inter-item correlation values for risk-taking (0.294), innovativeness (0.354), AO (0.333) and SE (0.278) are all above the 0.2 threshold, thereby indicating adequate internal consistency. A second reliability analysis is presented post-EFA, reflecting reliability scores of newly emerging and combined factors.

Bartlett’s test of sphericity was used to determine the factorability of data, while KMO was used to test for sampling adequacy. The recorded KMO value was 0.825, which was above the threshold minimum value of 0.6, thereby indicating that the data are suitable for factor analysis. Bartlett’s test of sphericity, at 0.000, is significant at p < 0.05, thereby indicating that factor analysis is appropriate for the data set.

Next, factor extraction was performed by means of PCA. In the decision-making process regarding which items or variables to retain, Kaiser’s criterion, known as the eigenvalue rule, was used to only retain factors with an eigenvalue of 1 or more.

In the factor extraction process, an unrotated factor solution was selected. Direct oblimin was chosen as the rotation method. The initial PCA showed the presence of 10 factors, with eigenvalues of 1.14 and above, explaining 19.8%, 9.2%, 5.8%, 4.4%, 3.5%, 3.3%, 2.7%, 2.6%, 2.5% and 2.4% of the variance, respectively. As the pattern matrix indicated significant cross-loading and an unclear pattern structure, a number of items were removed to obtain a clearer pattern. After removing items SE1, SE2, M5, M12, M13, M14, P1, I2, I4, PC4 and A4, a seven-component solution was obtained, explaining 53.140% of the variance.

The rotated solution indicated the presence of a simpler structure, with seven factors displaying strong loadings (above 0.3) and associated items loading against the seven factors. The pattern matrix with the seven-component solution is shown in Table 3.

Table 3 indicates that six of the original factors loaded under three new distinct variables, contrary to the original framework. The three new distinct variables consisted of (1) proactiveness and AO, (2) risk-taking and innovativeness and (3) PC and SE. Furthermore, the ‘management’ construct was split into two variables, consisting of (1) M6-11 and (2) M1-M4. The table further shows commonalities for each item, outlining how much of the variance in each item is explained. Values below 0.3 indicate a poor fit with the other items in a component (Pallant 2016). All values for commonalities are above the 0.3 threshold, thereby indicating adequate fit of items for each component.

Literature suggests that proactiveness and AO are closely related, as the purpose of proactiveness is the pursuit of a ‘first-mover’ advantage, in itself a form of achievement, while achievement implies impact and perseverance through proactive actions (Bateman & Crant 1993, 1999; Frese et al. 2007). This factor is referred to as proactive achievement (PA).

Risk-taking and innovation (RI) are usually deemed distinct concepts in entrepreneurial literature (Lumpkin & Dess 1996; Miller 1983; Stewart 2009). Some studies however suggest that, specifically at an individual level, RI form part of a combination of personality traits and therefore should not be considered separately, as they tend to provide a weak indication of an individual’s propensity to act in an entrepreneurial manner (Bolton & Lane 2012; Zhao & Seibert 2006; Zhao, Seibert & Lumpkin 2010). At an individual level, the pursuit of innovation involves an element of risk and can fluctuate in the individual employee based on past successes and failures (Harris & Gibson 2008; Robinson et al. 1991). This factor is referred to as RI.

Items from an instrument developed by Robinson et al. (1991) were used in this study and it is these authors who, most prominently, postulate that PC and SE are distinct individual factors positively affecting business outcomes. This view is supported by other authors, such as Van Wyk and Boshoff (2004), Sharif and Saud (2009) and Krishnan and Kamalanabhan (2015). In the South African context, this scale presented similar results for PC and SE. Boshoff and Hoole (1998) attribute this finding to intercultural differences between the North American and South African contexts. Similar findings, with constructs deviating from the original instrument, can be seen in other South African studies (Boshoff & Scholtz 1995; Van Wyk, Boshoff & Kruger 2004). This factor is referred to as personal control and self-esteem (PCSE).

Lastly, the split in the managerial items can be attributed to phrasing in the underlying scale, supported by findings seen during initial scale development (Goosen et al. 2002), with items M1–M4 referring to management-controlled activities (such as goal-setting, rewards and innovation systems, and intracapital) and items M6–M11 containing management-controlled activities (such as intrapreneurial freedom, communication and championing). Items M1–M4 contain management-related items that are more prescriptive and management-led in nature, while M6–M11 contain items involving interpersonal activities. The two managerial factors are referred to as management-led activities (MGT) and interpersonal activities (INT), respectively.

A second reliability analysis measured internal consistency of the new factors determined through the EFA. The Cronbach’s alpha for new constructs is shown in Table 4.

For the newly determined constructs, all Cronbach’s alpha coefficients were above the threshold value of 0.7, an improvement from the pre-EFA reliability analysis. All mean inter-item correlations were above the threshold value of 0.2.

The relationship between growth and proposed underlying IO constructs was investigated using Pearson product–moment correlation coefficient. Prior to the correlation analysis, a preliminary analysis was conducted to check for violation of assumptions of homoscedasticity, linearity and normality. The results of the correlation analysis, together with the means and SDs for the underlying factors, are presented in Table 5.

Table 5 indicates that at the 0.01 level, a statistically significant, positive linear correlation exists between growth and the factors PA, MGT, INT, RI, PCSE and IO. It should, however, be noted that while a statistically significant linear correlation between two factors may exist, effect size needs to be considered to interpret practical significance (Elliott & Woodward 2016). Effect sizes are considered trivial when r < 0.10, small for r between 0.10 and 0.29, medium for r between 0.30 and 0.49 and large for r greater than 0.50 (Cohen 1988; Elliott & Woodward 2016). The factors PA, INT, RI, PCSE and IO indicate r of 0.238, 0.260, 0.247, 0.198 and 0.184, respectively. This indicates that while statistically significant, the strength of the relationship between these variables and growth is considered small. The strength of the relationship between MGT and growth can be considered medium because of an r value of 0.313.

An initial standard multiple regression analysis was performed to assess the ability of all independent variables (PA, MGT, INT, RI, PCSE and IO) to predict growth. An analysis to ensure no violation of the assumptions of normality, linearity, multicollinearity and homoscedasticity was conducted. No major deviations from normality could be observed. No multicollinearity was found as all tolerance values were above 0.1 and all variance inflation factor (VIF) values were below 10 (Pallant 2016). Mahalanobis distances, at a maximum of 22.4, were below the critical value of 22.46 (Tabachnick & Fidell 2013), thereby indicating no critical outliers. Cook’s distance at 0.044, below the critical value of 1 (Pallant 2016), confirmed the indications provided by the Mahalanobis distances, indicating no undue influence of outliers on the model. A summary of the initial regression analysis including all constructs is presented in Table 6.

Table 6 indicates that 17% of variance (R² = 0.17) in the dependent variable (growth) is explained by the model (all independent variables). Of the independent variables in the model, only MGT was statistically significant (p < 0.05). The beta coefficients indicate that MGT (B = 0.200) makes the strongest unique contribution in explaining growth. All other variables record p > 0.05, thereby indicating that the other variables do not make a unique contribution to growth.

A further regression analysis was conducted to achieve a parsimonious model. A parsimonious model aims to accomplish a desired level of prediction based on as few predictor variables as possible in order to discriminate the signal from the noise, thereby allowing better predictability and generalisation of a data set (Vandekerckhove, Matzke & Wagenmakers 2015). Constructs were removed from the secondary regression analysis in the order of least significance based on observed p-values in each iteration. In order of removal, the constructs included IO, INT and PA, which were found not to be significant in any of the iterations. P¹⁰ and P¹³ are therefore not supported, as proactiveness and AO (p = 0.303) do not positively relate to internal growth. P⁴, P⁶ and P⁷ are therefore also not supported, as the INT (p = 0.141), containing the elements intrapreneurial freedom, communication and intrapreneurial championing, does not positively relate to internal growth. After three iterations, a regression model highlighting significant IO constructs was obtained, thereby achieving a parsimonious model. The results of the final iteration are presented in Table 7.

The same analysis as for the initial regression analysis was conducted to ensure no violation of the assumptions of normality, linearity, multicollinearity and homoscedasticity. No major deviations from normality were observed. Tolerance values were above the critical level 0.1 and all VIF values did not exceed the critical level of 10. Mahalanobis distances, at a maximum of 15.367, were below the critical value of 16.27 (Tabachnick & Fidell 2013), thereby indicating no critical outliers. Cook’s distance at 0.044, below the critical value of 1, confirmed the indications provided by the Mahalanobis distances, finding no undue influence of outliers on the model.

Table 7 indicates that 17.4% of variance (R² = 0.174) in the dependent variable (growth) is explained by the model (all independent variables), marginally unchanged from the original solution, thereby indicating the quality of fit of the final model. The regression analysis found MGT (p = 0.000), RI (p = 0.033) and PCSE (p = 0.003) to be statistically significant (p < 0.05). The beta coefficients indicate that MGT (B = 0.270) makes the strongest unique contribution in explaining growth, followed by PCSE (B = 0.174) and RI (B = 0.127). With regard to propositions, P⁸ and P⁹ are supported, as RI positively relate to the internal growth of an SME. Furthermore, P¹¹ and P¹² are supported, as PC and SE positively relate to the internal growth of an SME. In addition, P², P³ and P⁵ are supported as goal-setting, rewards and innovation systems, and intracapital positively relate to the internal growth of an SME. Lastly, with regard to P¹, only certain elements of an IO positively relate to the internal growth, namely MGT, PCSE and RI.