SIO is one of the approaches of organizational strategic orientation, which focuses on introducing and enhancing goods and services, as well as optimizing production and organizational processes to gain competitive advantages (Chou & Yang, 2011; Denicolai et al., 2018; Dogan, 2017; Chatzopoulou et al., 2022; Hakala, 2011; Gatignon & Xuereb, 2006; Kumar et al., 2012; Naranjo-Valencia et al., 2011; Yousaf et al., 2020; Mutlu & Sürer, 2016; Adams et al., 2019). Organizations that adopt this approach implement technology surveillance processes, allowing them to leverage practical knowledge to produce new technical solutions. This helps them meet customers’ needs and requirements through new and improved products and services, fostering loyalty and attracting new customers (Do Hyung & Dedahanov, 2014; Yousaf et al., 2020; Mutlu & Sürer, 2016; Adams et al., 2019; Serafim & Veríssimo, 2021; Saqib et al., 2017; Lee et al., 2015; Hortinha et al., 2011; Kocak et al., 2017).

IC is an intangible asset based on knowledge that creates value within organizations (Roos et al., 2001; Bontis, 1998; Umanto et al., 2018; Dang & Wang, 2022; Lo et al., 2020; Quintero-Quintero et al., 2021; Gómez-Bayona et al., 2020). It comprises a set of productive resources that boost a company’s capabilities and competencies, enabling it to stand out within the industry and increase customer loyalty and performance (Morales Clark et al., 2020; Ibarra-Cisneros et al., 2020; Pedraza Melo & de la Gala Velásquez, 2022). Thus, the comparative advantages derived from IC are difficult to replicate, leading to a rise in intrinsic value for the company (Morales Clark et al., 2020; García Garnica & Taboada Ibarra, 2012).

According to the literature, IC includes three main components: (i) human capital, which involves employees’ knowledge, skills, training, and application of this knowledge to benefit the organization; (ii) structural capital, which encompasses the organizational structure, processes, routines, and techniques that influence efficiency and effectiveness; and (iii) relational capital, which considers the quality and quantity of relationships with customers, partners, suppliers, the financial sector, allies, and other stakeholders (Zhang et al., 2019; Gómez-Valenzuela, 2022; Roos et al., 2001; Bontis, 1998; Nazir et al., 2021). Moreover, recent studies have proposed that IC can be operationalized through two dimensions: (i) internal IC, which includes aspects of human and structural capital developed internally and aligned with the company’s strategic orientation, and (ii) external IC, which pertains to the organization’s relationships with stakeholders, including suppliers, customers, and a country’s science and technology system, all of which are related to strategy implementation (Restrepo-Ramírez, 2023).

OP refers to achieving strategic objectives to create value for stakeholders (Masa’deh et al., 2018; Chmielewska et al., 2022; Princy & Rebeka, 2019). Consequently, organizations focus their strategic actions on understanding, analyzing, and managing factors that positively impact their performance (Masa’deh et al., 2018; Ekaningrum, 2021). OP can be measured using different approaches, both financial and non-financial (Ekaningrum, 2021; Gunasekaran et al., 2015). Financial metrics include Return On Investment (ROI), profitability, return on sales, and return on invested capital (Mcgivern & Tvorik, 1998; Ekaningrum, 2021). Other indicators include Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) (Maxim, 2023) and Economic Value Added (EVA) (Tudose et al., 2022; Nagarajan, 2015). Non-financial measures encompass effective, efficient, and innovative performance (Rojas et al., 2018; Ekaningrum, 2021). By measuring performance, organizations can validate their competitive advantages and compare themselves with their environment, enabling them to reassess their strategy (Gehrisch & Süß, 2023; Odiri & Ideh, 2021).

Organizations that adopt an SIO encourage employees to explore promising concepts for implementing new ideas, products, or processes, which makes it possible to achieve innovative impacts and contribute to long-term success (Deutscher et al., 2016). As a result, companies can increase productivity, thus fostering the creation of innovative goods and services, which in turn translates into improved OP (Pertuz & Pérez, 2020; Demuner Flores, 2021). SIO guides organizations to actively adopt new technologies for developing products, services, and processes, potentially enhancing their ability to thrive in competitive environments (Grinstein, 2008; Song & Jing, 2017; Walker et al., 2015).

To achieve superior performance, organizations often embrace an SIO to gain a competitive advantage and reduce the performance gap caused by environmental uncertainty (Azar & Ciabuschi, 2017). Therefore, it is possible to assert that SIO has a significant and positive impact on performance (Azar & Ciabuschi, 2017; Magnier-Watanabe & Benton, 2017; Walker et al., 2015). In fact, previous studies have shown a positive relationship between SIO and OP (Schweiger et al., 2019; Magnier-Watanabe & Benton, 2017; Walker et al., 2015; Suryantini et al., 2023). Nonetheless, although most of these studies demonstrate a strong conceptual and empirical association between the two (Schweiger et al., 2019; Demuner Flores, 2021; Ayinaddis, 2023), others have found no clear positive relationship, suggesting that SIO does not naturally result in improved OP (Mu et al., 2017; Chatzopoulou et al., 2022; Kornelius et al., 2021).

Furthermore, organizational leaders and scholars posit that innovative outcomes are beneficial for organizations, which is supported by empirical research showing that innovation strategies and activities positively influence performance (Magnier-Watanabe & Benton, 2017; Walker et al., 2015; Dinu et al., 2023). Given these considerations, the following hypothesis is proposed:

H₁: Strategic innovation orientation positively influences organizational performance.

The relationship between SIO and IC has not been widely studied in the literature. Most existing research has focused on specific components of SIO, such as technological innovation, process optimization, and product innovation (Bernal González et al., 2020). Hence, the need for a more comprehensive examination of the SIO–IC relationship by considering SIO as a whole. Despite limited research on this relationship, some studies have shown that the use of both tacit and explicit knowledge in SIO has a positive impact on IC (Magnier-Watanabe & Benton, 2017; Giménez, 2021). Additionally, the learning process involved in developing and designing products and services is essential for the activities stemming from SIO and IC (Lin & Chen, 2016; Coad et al., 2016). Furthermore, SIO allows companies to respond quickly and effectively to market demands by optimizing internal resources such as relationships with stakeholders, organizational structure, and employees’ knowledge (Magnier-Watanabe & Benton, 2017; Dang & Wang, 2022).

In short, organizations should aim to develop a balanced portfolio of technological and organizational innovations to navigate changing environments and uncertainties and to strengthen knowledge, organizational structure, and relationships with stakeholders (Azar & Ciabuschi, 2017; Dang & Wang, 2022; Córdoba-Vega & Naranjo-Valencia, 2017; Thien & Hung, 2023). Based on this, the following hypothesis is put forward:

H₂: Strategic innovation orientation positively influences intellectual capital.

Previous research has provided empirical evidence of a positive relationship between IC and OP (Galeitzke et al., 2017; Masood et al., 2023). However, most studies have focused on a specific dimension of IC, i.e., human capital (Díez et al., 2010; Nguyen, 2020; Albertini & Berger-Remy, 2019), structural capital (Lin & Chen, 2016; Verbano & Crema, 2016; Ghlichlee & Goodarzi, 2023), or relational capital (Agostini & Nosella, 2017; Bontis, 1998; Edvinsson & Malone, 1998). This calls for a comprehensive assessment of the impact of IC on OP (Agostini et al., 2017). Also, IC has been studied as a mediating (Agostini et al., 2017; Magnier-Watanabe & Benton, 2017; Javed et al., 2023) or moderating factor (Adams et al., 2019; Schweiger et al., 2019; Agostini et al., 2017) in the relationship between OP and constructs such as dynamism, hostility, tacit and explicit knowledge, and technological innovation (Magnier-Watanabe & Benton, 2017; Schweiger et al., 2019; Agostini et al., 2017; Nguyen, 2020).

By directing resources towards enhancing employees’ skills and knowledge—both individually and collectively—organizations can boost productivity (Ekaningrum, 2021; Agostini et al., 2017). Companies should also strengthen relationships with internal and external stakeholders, including employees, suppliers, and customers, to bridge communication gaps and gain a competitive edge in the market (Inkinen, 2015; Nguyen, 2020; Hussein et al., 2023; Ekaningrum, 2021). Additionally, it is important for companies to identify and manage their key production processes to differentiate themselves in the market and enhance productivity (Khan et al., 2019; Ibarra-Cisneros & Hernández-Perlines, 2019; Handoyo et al., 2023). Given this, the following hypothesis is proposed:

H₃: Intellectual capital positively influences organizational performance.

Data for this study were sourced from the most recent version of Colombia’s Survey on Development and Technological Innovation in the Manufacturing Industry (EDIT for its acronym in Spanish) for the 2019–2020 period. This survey was conducted by the National Administrative Department of Statistics (DANE, 2021), covering a population of 7,769 organizations. The analysis focused on companies classified by DANE into the “strict” or “broad” categories, representing companies that achieved innovative results during the study period in both domestic and international markets. Hence, the sample comprised 1,765 organizations falling under these two categories.

Indicators from the EDIT associated with each of the constructs under study were identified following Restrepo-Ramírez’s (2023) approach. This author developed a conceptual model using observable indicators from the EDIT linked to the constructs of SIO, IC, and OP and their respective dimensions. His approach comprises three main steps: (i) conducting a literature review to understand the theoretical content of the constructs; (ii) identifying possible indicators from the EDIT that reasonably represent the constructs under study; and (iii) validating content with experts.

Considering this, the measurement scales have already been validated in prior studies, including Restrepo-Ramírez’s work (2023), using the Kuder–Richardson test (Durán-Pérez & Lara-Abad, 2021) to assess reliability. This enabled their adoption in the present study.

Table 1 shows the number of indicators for each construct and dimension. Particularly, it outlines the distribution of the 88 indicators from the EDIT associated with each construct under study. Additionally, two example indicators are provided for each construct or dimension.

Table 1
Indicators associated with the constructs

Tabla 1. Indicadores asociados a los constructos

Source. Own work.

Descriptive statistics: Since dichotomous variables (1: presence of the attribute; 0: absence of the attribute) were employed, the frequency of each value (1 and 0) was counted for every variable, and the relative frequency was calculated based on the total to determine the importance of each variable in the behavior of the construct.

Mediation model: The proposed mediation model was tested using Baron and Kenny’s product of coefficients method (1986), which involves the following three sequential steps to verify the relationships between the dependent variable (X), the independent variable (Y), and the mediator (M) (Hayes, 2009; Zhao et al., 2010; Pardo & Román, 2013):

1) Estimating the regression coefficient for the direct relationship between the dependent and independent variables: This first step involves confirming that 𝑋 and 𝑌 are related—i.e., the regression coefficient should be different from zero (Hayes, 2009; Zhao et al., 2010; Pardo & Román, 2013)—following the expected direction, as indicated in Equation 1.

(1)

Here, 𝑖₁ is the intercept, 𝑐 is the regression coefficient relating 𝑋 to 𝑌, and 𝑒₁ are the random errors (i.e., the portion of 𝑌 not explained by 𝑋). These errors are assumed to be normally distributed with constant variance and independent from one another.

2) Determining whether the independent variable has a significant effect on the mediating variable: This second step involves checking if the coefficient estimated by the linear regression between 𝑋 and 𝑀 (represented by 𝑎) is different from zero (Hayes, 2009; Zhao et al., 2010; Pardo & Román, 2013), as shown in Equation 2.

(2)

3) Testing the relationship between the variable of interest and the mediator to estimate the total direct and indirect effects of the model: This third step involves examining whether 𝑀 and 𝑌 are related after controlling for the effect of 𝑋. This means that coefficient 𝑏 should be different from zero (Hayes, 2009; Zhao et al., 2010; Pardo & Román, 2013). This condition is verified by performing a linear regression analysis of 𝑌 explained by 𝑋 and 𝑀, as indicated in Equation 3.

(3)

As a result, the relationship between 𝑋 and 𝑌 should significantly decrease when controlling for the effect of 𝑀. In other words, the estimated coefficient 𝑐 from Equation 1 should decrease when testing the entire mediation model (Equation 3) (Hayes, 2009; Zhao et al., 2010; Pardo & Román, 2013).

Moreover, Baron and Kenny’s (1986) method recommends using the Sobel test (Preacher & Leonardelli, 2001) to estimate the indirect effect of the mediation model between 𝑎 and 𝑏, as shown in Equation 4. Here, 𝑎 is the estimated effect between 𝑋 and 𝑀 from Equation 2, 𝑏 is the regression coefficient between 𝑀 and 𝑌 from Equation 3, and 𝑠_𝑎 and 𝑠_𝑏 are their respective standard deviations (Zhao et al., 2010).

(4)

Where:

𝑧: statistical significance value

𝑎: estimated effect of SIO on IC

𝑏: estimated effect of IC on OP

𝑠_𝑎 and 𝑠_𝑏: standard errors of the respective estimated effects

The entire procedure was conducted using the mediation package in RStudio, performing bootstrapping with 500 estimates. This involves simulating the model multiple times with randomly sampled subsets from the dataset to ensure results within confidence intervals. Bootstrapping allows for the estimation of the impact of one variable on another, as well as its level of significance. The level of significance should be less than 0.05 to reject the null hypothesis that the estimated parameters are equal to zero, concluding that there is a significant relationship between the variables under analysis (Rivas-Ruiz et al., 2013).

Table 2 presents the descriptive statistics of the strategic innovation orientation construct.

Table 2
Descriptive statistics of strategic innovation orientation

Tabla 2. Estadística descriptiva OEI

Source: Own work.

As observed, a vast majority (84 ­%) of manufacturing companies in Colombia invested in scientific and technological activities, demonstrating an interest in focusing their strategy on innovation. Furthermore, a significant percentage of companies directed their strategy towards internal innovation: 47­ % focused on improving production, distribution, and logistics methods; 25 ­% developed or enhanced organizational methods; and 33­ % concentrated on improving and introducing new marketing techniques. In addition, 34­ % and 31­% of the companies developed new or significantly improved services or goods for the company, respectively. Moreover, there was limited intention to pursue outward innovation, as evidenced by the higher focus on creating new goods on a national scale (31 ­%) rather than in the international market (0.4 ­%).

Table 3 reports the descriptive statistics of the internal intellectual capital construct. Particularly, it provides information about the internal IC of manufacturing companies, including the internal personnel assigned to the different areas (based on their education and training) who participate in scientific, technological, and innovation activities, as well as the internal sources of innovative ideas.

Table 3
Descriptive statistics of internal intellectual capital

Tabla 3. Estadística descriptiva CI interno

Source: Own work.

As can be seen, 74 ­% of the companies allocated personnel for scientific, technological, and innovation activities. Additionally, in 75 % of the companies, personnel with higher education participated in these activities. Regarding sources of innovative ideas, all companies primarily resorted to their internal departments.

Table 4 shows the descriptive statistics of the external intellectual capital construct. Specifically, it outlines the components of external IC, including external personnel by area dedicated to scientific, technological, and innovation activities, as well as external sources of innovative ideas and the relationships with actors in the science, technology, and innovation system.

Table 4
Descriptive statistics of external intellectual capital

Tabla 4. Estadística descriptiva CI externo

Source: Own work.

As observed, 28­% of the companies had researchers involved in scientific, technological, and innovation activities. Additionally, approximately 1­% of the companies under analysis had personnel with doctoral qualifications dedicated to these activities. Regarding the most significant external sources of innovative ideas, 55­% of the companies relied on the internet, 25­% used fairs and exhibitions, and 32­% resorted to standards and technical regulations. The most common actors with whom the companies cooperated were customers (15­%), universities (10­%), and consultants (10­%).

Table 5 presents the descriptive statistics of the organizational performance construct. It highlights aspects of organizational performance such as efficiency, effectiveness, and innovative performance.

Table 5
Descriptive statistics of organizational performance

Tabla 5. Estadística descriptiva DO

Source: Own work.

Regarding efficiency, 74­ % of the companies reported an increase in productivity, and 52­ % reduced labor costs thanks to scientific, technological, and innovation activities. As for effectiveness, 76 ­% of companies reported improvements in the quality of goods and services, while 82 ­% maintained their participation in the market. Innovative performance, for its part, showed poor results, as, for example, only 6­ % of the companies registered industrial designs.

Table 6 shows the results of the mediation model.

Table 6
Mediation model

Tabla 6. Modelo con mediación

Source: Own work using RStudio. Note: p-value: *< 0.05; **< 0.01; ***< 0.001.

Based on the significance value of the direct relationship (p < 0.05), there is a positive and significant relationship between SIO and OP, which confirms hypothesis 1. After introducing the mediating variable, the relationship between the independent variable (SIO) and the mediator (IC) was tested. This analysis rejected the null hypothesis that the estimated parameters are zero (p<0.05), confirming that SIO positively influences IC and thus confirming hypothesis 2. Then, the mediation model was analyzed, showing that IC has a positive impact on OP, confirming hypothesis 3. Importantly, the mediating effect was examined by considering IC as a whole, rather than focusing on the mediation of each dimension of the construct.

Thus, according to the results of the mediation model, there is a total, positive, and significant indirect effect as a result of the multiplication of the coefficients of the SIO–IC–OP relationship. In other words, IC absorbs a portion of the effect that SIO has on OP, indicating partial mediation.

Table 7 reports the estimated direct and indirect effects based on the coefficients calculated in the linear regression analysis.

Table 7
Direct and indirect effects

Tabla 7. Efectos directos e indirectos

Source: Own work using RStudio. Note: p-value: *< 0.05; **< 0.01; ***< 0.001.

**Bootstrapping with 500 simulations

As can be seen, the Average Causal Mediation Effect (ACME) of 0.293 indicates the indirect effect of SIO on OP mediated through IC. This effect ranged from 0.204 to 0.40 within a 95% confidence interval. Based on the p-value (p < 0.05), IC significantly explains the relationship between SIO and OP. The Average Direct Effect (ADE) of 0.708, for its part, represents the direct effect of SIO on OP without mediation through IC. This effect ranged from 0.554 to 0.86 within a 95% confidence interval.

The total effect (ACME + ADE) reflects the direct effect of SIO on OP and its indirect effect through IC mediation. Note that the estimated value of the direct relationship before including mediation (1.0009, see Table 6) decreased significantly when IC was introduced as the mediating variable in the model (0.708). This highlights the importance of IC as a mediating mechanism in explaining the relationship between the two constructs. Furthermore, ADE (SIO–OP) was greater than ACME (SIO–IC–OP), which was calculated using the Sobel test (Preacher & Leonardelli, 2001), as shown in Table 8.

Table 8
Indirect effect

Tabla 8. Efecto indirecto

Source: Own work. Note: p-value: *< 0.05; **< 0.01; ***< 0.001.

Figure 1 illustrates the research model with the estimated results. As shown in this figure, introducing mediation into the model reduced the direct effect of SIO on OP to an estimated coefficient of 0.70, which is statistically significant. This suggests that IC partially mediates the relationship between SIO and OP.

Figure 1
Research model

Figura 1. Modelo de investigación

Source: Own work.

Figure 2 depicts the scatter plot of the mediated relationship. As observed, the model exhibits partial mediation, as the indirect effect explains part of the relationship between SIO and OP but not entirely. Nonetheless, including IC as a mediating variable is crucial because it helps explain what the direct relationship does not fully capture. Note the difference between the model without mediation (black line) and the model after introducing the mediating variable (blue line), indicating partial mediation.

Figure 2
Scatter plot of the mediated relationship

Figura 2. Diagrama de dispersión de la relación mediada

Source: Own work using RStudio.