1.1. Value Co-creation

According to ^{Prahalad and Ramaswamy (2000}, ²⁰⁰², ^2004a, ^2004b), value co-creation is about solving needs that influence competitive environments. Therefore, the authors identify a framework made up of four dimensions to co-create value, referred to the acronym DART standing for: Dialogue (leads the exchange between stakeholders); Access (leads the organization to exchange information on value and thus create positive experiences); Risk reduction (requires more information and accountability for value-makers in the management of the risks of co-created assets); and Transparency (refers to the confidence required for the interaction between organizations and clients). Authors present the DART model as the key to value co-creation, where the organizations can interact with their clients or partners and identify opportunities related to their needs. Under this scenario, organizations become structures that generate innovation due to creating value through interactions and continuous relationships (^{Taghizadeh et al., 2016}).

1.2. Value co-creation in sports organizations

Regarding value co-creation in sports organizations, ^{Woratschek et al. (2017)}; ^{Woratschek et al. (2014a}, ^2014b) suggest that the traditional concepts of creating value only through internal resources of nonprofit and for-profit organizations in sports should migrate towards the creating value together concept. A recent study on the benefits of value co-creation in sports organizational settings (^{Woratschek et al., 2017}) analyzes mega sports events (e.g., FIFA World Cup or Olympic Games). The authors identify two new dimensions of motivation for passive sports consumption: freedom to move and intercultural contact, representing new ways to create value in sports events. ^{Hedlund (2014)} states that value co-creation in fanatics happens when they interact among themselves through participation in rituals and traditions associated with the network of fanatics, increasing the intention of consumption and attendance to the games.

^{Horbel et al. (2016)} demonstrate the influence of co-creation in sports services in the 2014 FIFA World Cup context. The authors conclude that suppliers of sports services should know the most crucial value dimensions for consumers in the considered context. The above discussion represents the importance of co-creating value in sports, which could change the way to obtain more significant value creation and benefit sports organizations. However, the studies do not analyze how the benefits of applying value co-creation impact the performance of these organizations.

1.3. Organizational performance in sports organizations

Organizational performance is one of the most relevant constructs when researching the field of management (^{Baruch and Ramalho, 2006}; ^{Richard et al., 2008}). Several studies have conceptualized measuring performance in sports organizations (^{Winand et al., 2014}). Due to the different perspectives on organizational performance in the literature, there is a lack of consistency in measuring performance (^{Nowy et al., 2015}). Organizational performance combines efficiency and effectiveness (^{Perck et al., 2016}; ^{Winand et al., 2010}; ^{Winand et al., 2014}). Effectiveness is the ability of an organization to achieve its objectives, whereas efficiency refers to the relationship between inputs and produced goods.

^{Nowy et al. (2015)} highlight the research that ^{Winand et al. (2010)} carried out as a reference for researching performance measurement. The authors base their argument on the fact that the multidimensional model integrates and develops previous works on performance measurement, includes efficiency and effectiveness measures, enables comparison between organizations of different levels, types, and sizes, and is empirically tested.

1.4. The relationship between value co-creation and organizational performance

The value co-creation approach supports collaboration between organizations and their interest groups. This fact promotes significant improvements in organizations performance, incentivizing their growth (^{Ge et al., 2019}) through market opportunities (^{Ratten et al., 2021}).

One of the critical factors for improving sports organizations’ success is their ability to collaborate with other players through bilateral relationships and multilateral networks for co-creation (^{Gerke et al., 2020a}, ^2020b). Co-creation occurs through a value network that includes employees, citizens, consumers, and government agencies, positively affecting organizational performance (^{Kim et al., 2020}). Value co-creation occurs due to participating in open innovation networks at intergroup, interunit, or inter-organizational levels, generating benefits for sports organizations’ performance (^{Gerke et al., 2020a}). Value co-creation may enhance organizational performance in two ways; first, by emphasizing the role of key interest groups and, second, by promoting good governance practices and suitable environmental conditions (^{Chao, 2019}).

Co-creation capabilities are based on inter-connected operational resources, which, when combined, represent the ability of an organization to co-create value along with other organizations (^{Wilden et al., 2019}). Therefore, combining these capabilities becomes essential for achieving sustainable organizational performance (^{Clauss, 2017}). This means organizations must constantly innovate to enhance performance and sustain competitive advantage (^{Huang et al., 2019}). Consequently, value co-creation capabilities are necessary for enhancing innovation performance, growing in the market, and improving consumer satisfaction, which influences organizational performance (^{Chao et al., 2019}).

^{Nowy et al. (2015)} propose assessing the results of a sports organization against the dimensions presented by ^{Winand et al. (2010)}, considering among them: the sports, the customers, the image and the communication, the financial and the organizational, due to the theoretical and empirical considerations taken into account for the study result.

The theory of value creation argues that organizations, including those of sports, require creating differentiation in products, services, or activities to be sustainable in the long term (^{Gómez et al., 2010}). This situation has ended up in the quest for innovation and the harnessing of market opportunities to enhance performance (^{Winand and Hoeber, 2017}; ^{Winand et al., 2016}). An example of this is sports clubs, which through innovation, are more likely to enhance their performance -efficiency and effectiveness- (^{Weerawardena and Mort, 2012}; ^{Wemmer and Koenigstorfer, 2016}). Thus, organizations can achieve greater efficiency by successfully applying value co-creation (^{Hoyer et al., 2010}). This thesis leads us to the first hypothesis:

H1. Value co-creation directly and positively influences the sports performance dimension of sports organizations.

^{Gómez et al. (2010)} suggest that value creation in sports organizations related to soccer should be assessed, taking into account, among others, the dimensions of sports and communication. Additionally, according to ^{Hoyer et al. (2010)}, the interaction between customers and organizations results in co-creation, intervening in the maximization of organizations’ performance through high-impact collaboration, thus obtaining growth, learning, and innovation at lower costs (^{Evans and Wolf, 2005}). Thus, the second and third hypotheses arise:

H2. Value co-creation directly and positively influences sports organizations’ customer/members’ performance dimension.

H3. Value co-creation directly and positively influences sports organizations’ image and communication performance dimensions.

Sports organizations’ performance, according to ^{Byers et al. (2012)}; ^{Nowy et al. (2015)}; ^{Robinson et al. (2012)}; ^{Treem and Leonardi (2012)}, considers the following measuring variables: I) the efficiency, in which there is a relationship between the utilized resources and the obtained products; and II) the effectiveness, which is related to the capacity of achieving organizational goals (^{Bayle and Madella, 2002}). Likewise, in the performance assessment of sports organizations, financial measurements for evaluating efficiency are combined with non-financial measurements for evaluating effectiveness (^{Herman and Renz, 1999}; ^{Winand et al., 2010}). Therefore, the fourth and fifth hypotheses arise as follows:

H4. Value co-creation directly and positively influences the financial performance dimension of sports organizations.

H5. Value co-creation directly and positively influences the performance dimension of sports organizations.

In order to close this knowledge gap, this study proposes empirically validating that nonprofit sports organizations need to co-create value to improve their organizational performance (^{Taghizadeh et al., 2016}) (Figure 1).

Source: Elaborated from ^{Winand et al. (2010a)}; ^{Taghizadeh et al. (2016)}.

Figure 1 Conceptual model and hypotheses  

2.1. Research context

Amateur football clubs in the Antioquia region of Colombia, recognized at the national level and affiliated with the Antioquia Football League (LAF), were selected for this study. These clubs compete mainly in 2 categories -male and female football- distributed among 750 teams and 13,500 players between 9 and 21 years of age. Currently, the clubs obtain financial resources from players’ sports rights, sponsorships, government subsidies, and donations. Some of these clubs carry out innovation projects, providing a base for researchers interested in investigating value co-creation.

2.2. Measures

Within the DART model (^{Taghizadeh et al., 2016}), twenty-one items were adapted for each construct under four dimensions to measure value co-creation in amateur soccer clubs (i.e., Dialogue 6; Access 5; Risk 5; Transparency 5). Twenty items were adapted to measure the performance dimensions of sports organizations (^{Winand et al., 2010}) (i.e., Sports 4; Customers/Consumers 4; Communication and Image 4; Finance 4; and Organization 4). Since the study focuses on amateur football clubs, the grassroots and elite sport dimensions were merged into one dimension labeled sport. All survey items were measured using a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

2.3. Instrument

A two-part self-administered printed questionnaire was used; in the first part, demographic data were included; and, in the second part, the items measuring the constructs were incorporated. Under the postulates of ^{Czaja and Blair (2005)}, the rigor of the measurement instrument is following the pre-and pilot tests. The rigor tests applied to the questionnaire are presented below. Pre-test 1. Review of the questionnaire by academic peers on the choice, translation into Spanish, wording, and consistency of the 41 items that emerged from the literature review of constructs such as value co-creation and the performance dimension in sports organizations. Pre-test 2. Application of the questionnaire to a panel of 7 experts, composed of managers of the sports sector in Antioquia, to detect possible errors in the wording or interpretation of the items. In this way, the questionnaire was adjusted before applying the pilot test.

Pilot test. Application of the questionnaire to 24 managers of teams participating in LAF tournaments and competitions. This validated the design of the questionnaire, the internal consistency -Indicator: Cronbach’s Alpha (AC) => 0.7 (^{Nunnally and Bernstein, 1994})-, and the reliability of the instrument -Indicator: Composite Reliability (^{Crossan and Apaydin, 2010}) => 0.70 (^{Fornell and Larcker, 1981})-. Finally, after adjusting the design and verifying the internal consistency of the instrument and the initial composite reliability, the final version of the questionnaire was proposed.

Seven hundred and fifty (750) sports and administrative managers of amateur soccer clubs were invited to complete the questionnaire, but 405 (54%) completed and returned it. Once the completed questionnaires were received, 60 questionnaires (15%) were discarded due to lack of completion, following ^{Nunnally and Bernstein’s (1994)} guideline of a minimum of 80% completion. Thus, 345 questionnaires were completed, with a response rate of 48%.

2.4. Data analysis

The literature proposes different approaches to the multivariate method, but Structural Equation Modeling SEM was chosen for this research. This method overcomes the weaknesses of others, such as cluster analysis, exploratory factor analysis, multidimensional scaling, analysis of variance, multiple regressions, and confirmatory factor analysis (^{Hair et al., 2017}). One of the advantages of this approach is that the population distribution is not restricted, so exploratory and confirmatory research can be conducted with small and medium-sized samples (^{Tsao et al., 2016}).

Subjective least squares path modeling (PLS-SEM) hypotheses were tested using SmartPLS 3.0 software (^{Ringle and Sarstedt, 2012}). Partial Least Squares Structural Equation Models [PLS-SEM] are part of the most recent strategies for conducting quantitative research in sports management ^{Zhang et al. (2017}. The minimum sample size was ensured through the result of the power test analysis. (^{Hair et al., 2017}) suggest this test to ensure adequate statistical power, considered the core of the PLS-SEM algorithm.

Thus, by using the G*Power 3 software (^{Faul et al., 2007}), for a standard significance level µ = 0.05, a mean effect size (f2 = 0.15), with 345 validated cases, and the regression of 4 explanatory variables-4 that make up the independent construct of co-creation, the power achieved was 99%. That is a good percentage achieved for studies in social sciences.

3.1. Measurement model

The model presents acceptable Compound Reliability (^{Crossan and Apaydin, 2010}) values for all reflective constructs, demonstrating the internal consistency and reliability of the instrument. Moreover, the Average Variance Extracted (AVE) values of the reflective constructs are more significant than 0.5 (Table 1). Also, all loadings sizes were significant and greater than 0.6. These values show compliance with convergent validity; the items measured the respective constructs robustly.

According to ^{Fornell (1982)}; ^{Fornell and Larcker (1981)}; ^{Hair et al. (2017)}, the discriminant validity is determined as follows: diagonal values were calculated by the SmartPLS software, such as the square roots of the AVE (^{Chin, 1998}) (Table 2). These variances are more significant than the highest quadratic correlation of the constructs listed below the diagonal. Secondly, comparing factorial cross-loadings of indicators in a latent variable against the loadings of all other latent variables is necessary. Factorial loadings had a more excellent value with their variable than other variables assessed in the model, demonstrating discriminant validity.

The relative relevance of the weight load of the dimensions of the value co-creation construct was validated through the weights. Table 3 shows that only the dialogue dimension was significant (P>0.001). The other dimensions (access, transparency, and risk) were not. Therefore, it is required to evaluate the weight of the non-significant dimensions. According to ^{Hair et al. (2017)}, the dimension is maintained if the loading is greater than 0.5 or if the dimension is significant, even if the loading is not greater. Therefore, all dimensions meet the relative importance of the weight load. An additional evaluation was performed to obtain the Variance Inflation Factor (VIF) for the formative dimensions, thus verifying the model’s goodness of fit to be estimated. The values associated with the indicators were lower than five (^{Korkman, 2006}), with a tolerance (^{Di Minin et al., 2014}) higher than 0.20 (^{Hair et al., 2017}). The formative dimensions of the measurement model are validated when their indicators are not highly correlated. Table 3 shows that the dimensions meet the highest loading (> 0.5).

3.2. Structural model

The results show that R² values were statistically significant with relevant and moderate indicators for the dependent variables: finances (R² = 0.484), communication, and image (R² = 0.477); organization (R² = 0.457), costumer/members (R² = 0.422), and sport (R² = 0.413), enabling making inferences about them (^{Hair et al., 2017}). The resampling procedure called Blindfolding was executed to evaluate Q2. The findings demonstrate that all dependent variables had a predictive relevance greater than zero (Q² > 0) (Hair et al., 2017).

Therefore, based on the path coefficient structural model results, the DART model of value co-creation directly and positively impacts the organizational performance in amateurs’ football clubs (Figure 2 and Table 4).

Source: From the results of the SmartPLS software.

Figure 2 Results of the structural model