The study is grounded in relationship marketing and transaction cost analysis theories with reference to the variables under investigation and the different relationships proposed between them. The relationship marketing theory originated in the early 1980s, the concept being positioned in the domain of services and business-to-business marketing. This theory was founded on the principle of building long-term profitable relationships with customers, moving away from the transactional approach towards relationship establishment (Gummesson, 2017, p. 17). The commitment–trust theory of Morgan and Hunt (1994, p. 22) became the foundation of the relationship marketing theory, stating that both trust and commitment are required to secure sustainable long-term relationships. Even today, trust and commitment are being studied as key constructs of the relationship marketing theory, driving customer satisfaction in buyer–seller relationships in both business-to-business and business-to-consumer markets (Verma, Sharma, & Sheth, 2016, pp. 207–208). Gummesson (2002, pp. 50–53) has identified the following five principles that will form the foundation of relationship building in the future: firstly, to value each customer individually; secondly, to secure trustworthy collaboration enabling value creation; thirdly, to understand that it is financially rewarding to invest in long-term relationships; fourthly, to secure win-win relationships for all parties involved; and, finally, to develop relationships based on service values such as ethics, honesty, integrity, reliability and respect for time. Relationships built on such trustworthy principles allow suppliers to become more effective and efficient, resulting in a financial benefit based on customer retention (Samiee, Chabowski, & Hult, 2015, p. 1). Considering that buyer–supplier relationships are continuously evolving, a deep understanding of the relationship marketing theory is required to secure the establishment and management of customised, long-term relationships. These relationships need to be customised and founded on principles that will secure customer commitment and satisfaction in the long term (Sarmento, Farhangmehr, & Simões, 2015, p. 586).

The transaction cost theory provides a theoretical foundation for understanding how business-to-business companies function and survive, especially considering that the taking of risks and opportunities is central to conducting business in a business-to-business environment (Sinnewe, Charles, & Keast, 2016, p. 198). The transaction cost theory originated in fields such as economics, sociology and organisational behaviour. It is significant to the marketing field in describing the advantages of establishing business relationships, with specific reference to corporate sellers and buyers that are engaged in transactions of an ‘arm’s length’ nature (Mpinganjira, Bogaards, Svensson, & Mysen, 2014, p. 3). The transaction cost theory can be used as a theoretical foundation to understand how trust can be strengthened and opportunism can be reduced in the relationship building process. A positive relationship between trust and satisfaction is also secured through the application of the transaction cost theory in the management of business-to-business transactions (San-Martín & Jiménez, 2017, p. 2212). According to Yasuda (2005, p. 763), the transaction cost theory avers that organisational buyers and sellers will commit to a business-to-business relationship if overall fixed and continual transaction costs are reduced (inclusive of opportunism and conflict). By including transactional cost in the assessment of business opportunities, a buyer and supplier can establish the risk of fixed and continual transaction costs (such as opportunism and potential conflict) and better understand the requirements for securing a stable relationship (Meghwani & Thakur, 2017, p. 2). Because of uncertainty in the business-to-business environment, an organisational buyer should increasingly consider the influence of transaction costs when dealing with a supplier, especially when the possibility of opportunism exists, which could reduce overall levels of satisfaction for both parties in the long term (McIvor, 2009, p. 54). Therefore, the transaction cost theory proposes the necessity to safeguard investments that increase transaction costs and thus have a negative influence on the building of long-term relationships between an organisational supplier and buyer (Parida, Wincent, & Oghazi, 2016, p. 1823). However, safeguarding investment could help to lessen the possibility of conflict between the parties that could arise because of opportunism and/or disagreement, influencing the overall level of satisfaction experienced by both parties (Brown, Lusch, & Smith, 1991, pp. 15–16; Standifer et al., 2015, pp. 694–695).

Against the background provided above, the study uses the relationship marketing theory and the transaction cost theory to develop hypotheses for relationships between the constructs of the study in a South African business-to-business environment. No previous study in a South African business-to-business environment has used relationship marketing theory and transaction cost theory to propose that satisfaction is an outcome of trust and commitment, on the one hand, and an antecedent of continuity and conflict, on the other hand.

Trust refers to the preparedness of parties to take on risk and be exposed to another party in the relationship building process (Alhabash et al., 2015, p. 149). It entails the belief by one party that the benefits exceed the risk incurred in establishing and managing a relationship with another party (Sahin, Zehir, & Kitapçı, 2011, p. 1291). Trust can therefore be seen as a wide-ranging assessment at a more advanced level, compared to satisfaction, with satisfaction being perceived as an outcome of trust. Trust secures the supplier’s loyalty to its promised role obligations to secure satisfaction (Wu, Cheng, & Ai, 2018, p. 202). According to Mpinganjira et al. (2014, p. 5), the principle of trust can be studied from the perspective of interpersonal and inter-firm levels. However, for the purpose of this study, the focus is on investigating trust from an inter-firm perspective. According to various research studies (Amin, Rezaei, & Abolghasemi, 2014, p. 262; Chen & Chang, 2013, p. 67; Veloutsou, 2015, p. 407), customer satisfaction is perceived as a critical element in securing a long-term relationship orientation. Satisfaction also entails an assessment of economic or psychological factors that can have a positive influence on the building of relationships with customers. Therefore, trust is perceived as an antecedent of satisfaction in business relationships (Altinay, Brookes, Madanoglu, & Aktas, 2014, 724; Cheng et al., 2017, p. 396; Han & Hyun, 2015, p. 22; Jiang et al., 2011, p. 6). Considering this, the following hypothesis is proposed:

Commitment can be described as the confidence of parties to a relationship, illustrating a willingness to preserve the relationship and make short-term sacrifices to secure long-term advantages (Richard & Zang, 2012, p. 573). Customer commitment encompasses an emotional connection between a buyer and a seller and demonstrates the continuous desire of the buyer to be in a relationship with the seller, or vice versa (Chiu, Kwag, & Bae, 2015, p. 628; Su et al., 2016, p. 3262). Akman and Yörür (2012, p. 220) concur and state that committed buyers are more willing to support a supplier if the latter secures the provision of high-quality products or services and creates opportunities for continuous engagement, resulting in higher levels of satisfaction. Therefore, a feeling of commitment increases the belief that a supplier will secure results leading to the buyer’s satisfaction with the relationship (Fullerton, 2011, p. 95; Payan et al., 2016, p. 66). Considering this, commitment is noted in research literature (Farrelly & Quester, 2005, p. 212; Kasuma, Ung, Kanyan, Kamri, & Yacob, 2016, p. 302; Tsao & Hsieh, 2012, p. 823) as an important antecedent of customer satisfaction and a critical factor in enhancing relationship outcomes such as satisfaction. Therefore, the following hypothesis is formed:

The founding theory of Morgan and Hunt (1994, p. 24) on the relationship between trust and commitment established that trust influences relational commitment. The founders of the trust–commitment theory argue that relationships founded on the principle of trust are so greatly appreciated that partners aspire to be committed to such relationships. As commitment requires vulnerability, partners to a relationship will be interested only in a partner that is trustworthy (Chai & Dibb, 2014, pp. 63–64; Li, Li, & Feng, 2015, p. 1045; Morgan & Hunt, 1994, pp. 24–25). Many research studies (Hong & Cho, 2011; Morgan & Hunt, 1994; Ou, Shih, & Chen, 2015, p. 676; Wang, Ngamsiriudom, & Hsieh, 2015, p. 558) have proven the fundamental relationship between trust and commitment in relational exchange where trust is the focal point of the relationship. Trust is perceived as a critical element in the relationship building process to secure a long-term orientation through the establishment of commitment (Taylor, Donovan, & Ishida, 2014, p. 129). Ng, Fang and Lien (2016, p. 38) argue that the level of trust existing between a buyer and a seller indicates the possibility of a continuing relationship and its potential for success. They also argue that trust has a positive influence on commitment, because relationships characterised by high levels of trust are strongly valued by relationship partners, resulting in an increased willingness to be committed to the relationship. It can therefore be argued that trust as an antecedent of commitment has been well researched and found to be significant as a predictor of commitment in the relationship building process (Theron, Terblanche, & Boshoff, 2008, p. 998). Considering this, the following hypothesis is formulated:

The first specified outcome of satisfaction, as per Figure 1, is opportunism. Dahlstrom, Nygaard, Kimasheva and Ulvnes (2014, p. 271) refer to opportunism as ‘the incomplete or distorted disclosure of information, especially to calculated efforts to mislead, distort, or otherwise confuse a trading partner’. It entails behaviour founded on dishonesty and pretence. Therefore, by being aware of opportunism, an organisation can evaluate its standing in relation to that of a business partner and recognise areas where they might become the victim of opportunistic behaviour (Paswan, Hirunyawipada, & Iyer, 2017, p. 131). Opportunism therefore presents a risk to the building of long-term relationships between an organisational buyer and supplier because it generates insecurity, inadequacies and substantial negative economic effects for exchange relationships (Zhou, Zhang, Zhuang, & Zhou, 2015, p. 147). Chiou and Shen (2006, p. 11) argue that in a case where the supplier is perceived to pose a high risk of opportunism in a relationship with a buyer, the latter will be required to devote a large number of resources to managing the sources of opportunism. Opportunism can be prevalent in various forms, such as through different types of intentional misrepresentation during the initial stage of the relationship building process between the buyer and supplier, as well as numerous types of infringements over the duration of the relationship (Elsharnouby & Parson, 2013, p. 146). Opportunism therefore creates a significant challenge to a buyer or a seller wanting to be in charge of or to manage an exchange partner’s actions where the propensity for opportunism is high. Ping (1993, p. 326) and San-Martín and Jiménez (2017, p. 2212) note that there is a greater chance of opportunism in relationships where satisfaction between two parties declines, and a lower possibility of opportunism in relationships characterised by high levels of satisfaction and rewarding engagements. The following hypothesis is therefore proposed:

The second outcome of satisfaction, as per Figure 1, is conflict. Conflict refers to incompatible activities by two parties. It arises when there is no or limited delivery on promises made between parties (Mazaheri, Basil, Yanamandram, & Daroczi, 2011, p. 236). Considering the limited possibility that the objectives of parties to a relationship will be similar, misunderstandings are bound to arise from the different goals and unexpected emergencies in their daily relationships, which could result in conflict (Zaheer, McEvily, & Perrone, 1998, p. 146). Conflict needs to be managed to ensure that it is beneficial to all parties concerned. In addition, conflict should be countered through open dialogue and professional engagement to enhance positive experience and overall satisfaction (Karnani, 2008, pp. 31–32). Ng (2012, p. 166) concurs, stating that increased levels of mutual satisfaction are critical to secure long-term relationships between parties. Varela-Neira, Vázquez-Casielles and Iglesias (2010, p. 104) also state that the professional management of conflict will lead to increased levels of satisfaction. The inability to address conflict may result in higher levels of dissatisfaction or even aggravate emotions of rage and the need to take revenge. Therefore, is the following hypothesis is formulated:

The target population of the study included the top South African companies by revenue. The study was quantitative in nature and the sample frame included the largest 500 South African companies. The sample frame was obtained from the Topco list, which comprises the top 500 companies functioning in the South African private sector. The Topco list was selected because this list makes available details of the top 500 companies in South Africa. Companies on the list represent a variety of industries in South Africa, ranging from financial services, agriculture, the mining sector and telecommunications to business support services. The study followed a convenience sampling approach by which the purchasing or procurement managers of the selected companies were contacted by telephone to determine their suitability for answering the questionnaire. Purchasing or procurement managers were selected because the study focused on business relationships with suppliers. Purchasing or procurement managers liaise directly with suppliers in a business, making them more informed about relationship issues with their business suppliers. To gather data from respondents, they were interviewed by means of a structured questionnaire and computer-assisted telephonic interviews. When answering the statements in the questionnaire, the purchasing or procurement managers were requested to focus on a single supplier that they were acquainted with. In total, 250 usable questionnaires were obtained, representing a response rate of 50%.

The questionnaire contained two items to establish the competence of the respondents to answer it. These items established their level of knowledge about their company’s view of the study topics and on particular experiences with the supplier. This approach is supported by Campbell (1955), who states that individuals who are selected to complete a questionnaire should be knowledgeable on the topic being investigated. The results illustrate that a total of 94.8% of those who responded were familiar with the viewpoint of their company in terms of the topic under study. In addition, a total of 98.4% respondents had comprehensive knowledge about experiences with a specific supplier. The mean values for these two items were 4.26 and 4.33, respectively. These mean values indicate that the purchasing or procurement managers who responded had a suitable understanding of their company’s perspective on the topics being investigated and the experience that their companies had with a supplier. The statement relating to specific experiences with the supplier sought to determine the level of knowledge that the purchasing or procurement managers had gained through specific, and not simply general, experiences with the supplier whom they considered when responding to the different items in the questionnaire.

Table 1 illustrates the different items used to measure the constructs illustrated in Figure 1. The items used to measure the interrelationship between antecedents, mediator and postcedents were adapted from various sources as indicated below:

Existing scales were used in the study as Peter (1979) and Churchill (1979) argue that existing scales improve validity. The study made use of a five-point Likert-type scale to determine the degree to which respondents agreed or disagreed with the items provided in the questionnaire. The scale points ranged from (1) ‘strongly disagree’ to (5) ‘strongly agree’ (see Table 1).

Confirmatory factor analysis (CFA) and structural equation modelling (SEM) were used to test the measurement model and evaluate the proposed structural relationships shown in Figure 1 (Jöreskog & Sörbom, 1976). The researchers first conducted CFA of the measurement model (19 indicator variables as in Figure 2), founded on five constructs, applying the SPSS/AMOS 22.0 software. The testing of the model produced reliable and acceptable findings. The goodness-of-fit measures were all acceptable (Hair, Black, Babin, Anderson, & Tatham, 2006, pp. 745–749), as shown in Table 2.

Table 3 illustrates the results for construct reliability and validity testing. The results indicate that the variance extracted from items for all constructs exceeds 50% in the study, clearly reflecting convergent validity. The average explained variance is 66.9%, and the composite trait reliability levels of all the constructs included in the study are also above 0.7 (Hair et al., 2006).

In addition, the variance extracted was compared with the squared inter-construct correlations to determine whether the model measured different constructs (Hair et al., 2006). In the study, the variance extracted for the majority of the constructs was equal to or greater than the corresponding squared inter-construct correlations (see Table 3). It can therefore be argued that the tested structural model reflects satisfactory discriminant validity in the study (see Figure 2). The hypothesised relationships of the model were all significant in the study, clearly reflecting nomological validity. The testing of the model in the study with specific reference to South African business relationships attained convergent, construct and nomological reliability. It can therefore be argued that the measurement and structural metrics of the model display fully consistent validity and reliability.

Based on the satisfactory results from the CFA of the measurement model, founded on South African business relationships, the researchers tested the structural model, illustrated in Figure 2.

Once the assessment of the psychometric properties of the measurement model was completed and the checking of the underlying assumptions was done, the structural properties of the model had to be secured. Different goodness-of-fit measures were measured in the structural model, as was the case in the measurement model. These measures include the chi-square, the normed chi-square (χ²/df), the NFI, RFI, IFI, TLI, CFI and the RMSEA. The results reflected by Table 4 clearly indicates that the different goodness-of-fit measures are within the parameters as recommended by Hair Black and Babin (2014). The normed chi square (χ²/df) are below the value of 3, the NFI, RFI, IFI, TLI and CFI are all in excess of the cut-off of 0.9, and the RMSEA is below a value of 0.08. The combination of the goodness-of-fit measures as reported in Table 4 is therefore an indication of a good model fit (Hair et al., 2014).

Furthermore, the structural model’s hypothesised relationships (see Figure 2) were all significant, as illustrated in Table 5. Accordingly, the results support the six hypotheses of the confirmed model on South African business relationships in the study.