Relationship intention and satisfaction as predictors of wholesale and retail customers’ loyalty towards their training providers

3It is not surprising that service providers are increasingly attempting to establish customer loyalty as competition intensifi es in service industries. Building long-term relationships and satisfying customer expectations could be an effective strategy to follow according to research that suggests strong relationships between customer relationships, customer satisfaction and customer loyalty. However, some researchers suggest that service providers should direct their marketing efforts only towards customers who have relationship intentions. It is thus essential for service providers to consider customers’ relationship intentions and satisfaction when drafting strategies aimed at building customer loyalty. The purpose of this study was to determine the extent to which relationship intention and satisfaction predict customer loyalty within the wholesale and retail training sectors. Data were gathered from 185 wholesale and retail skills development decision-makers located across South Africa, who were involved in the selection of their organisations’ training providers. Using hierarchical multiple regression analysis, the fi ndings indicate that relationship intention and satisfaction signifi cantly predict customer loyalty towards wholesale and retail training providers. Wholesale and retail training providers thus need to understand that establishing customer loyalty depends on their ability to develop strong relationships with customers who are receptive to relationship marketing efforts, and to ensure that these customers’ needs are met. 4

Service providers are increasingly trying to build and maintain strong relationships with their customers due to the belief that long-term relationships lead to customer loyalty (Cronze, Bieger, Laesser & Riklin 2010: 52;Mazhari, Madahi & Sukati 2012: 83;Richard & Zhang 2012: 569).Unfortunately, trying to build long-term relationships with customers is not always the most effective strategy, as Kumar, Bohling and Ladda (2003: 667) argue that not all customers have relational intentions towards service providers.It is thus essential to identify and focus on those customers with relationship intentions when attempting to establish a loyal customer base.
When trying to build customer loyalty, research suggests that service providers should ensure that their customers' needs are satisfied (Hansen 2012: 352;Raza & Rehman 2012: 5085;Terblanche & Boshoff 2010: 6), as satisfied customers are more receptive to building long-term relationships (Halimi, Chavosh & Choshali 2011: 51), and increased satisfaction leads to a greater possibility of customers returning in the future (Curtis, Abratt, Rhoades & Dion 2011: 15;Pan, Sheng & Xie 2012: 156).Customer satisfaction should thus be considered a priority to service providers, because empirical evidence suggests a strong link between satisfaction, customer loyalty (Goncalves & Sampaio 2012: 1509) and financial performance (Williams & Naumann 2011: 20).Customer satisfaction can therefore be regarded as the necessary precondition for the development of customer relationships as well as loyalty (Raciti, Ward & Dagger 2013: 615).
This study considered customers' relationship intentions, satisfaction and loyalty specifically among South African wholesale and retail training providers.This decision was based on the notion that the implementation of skills development legislation in the form of the Skills Development Act (No. 97 of 1998) by government, to ensure that training takes place in the wholesale and retail sectors (W&RSETA 2013: 37, 128), has resulted in a significant increase in competition among training providers competing in these sectors.By gaining insights into their customers' loyalty, wholesale and retail training providers stand a greater chance of not only surviving, but also gaining a competitive advantage, reducing costs, increasing customer retention and ultimately profitability (Boora & Singh 2011: 158;Hoissain & Ullah 2011: 7;Ishaq 2012: 26;Islam 2010: 141;Terblanche & Boshoff 2010: 1;Van Vuuren, Roberts-Lombard & Van Tonder 2012: 93).
The purpose of this article is accordingly to determine the extent to which relationship intention and customer satisfaction predict customer loyalty within the wholesale and retail training sectors.

Literature review
Relationship marketing 1Given the increasingly competitive market in which they operate, it is not surprising that more service providers have changed their focus from transactional marketing to relationship marketing (Chaman, Masoumi, Moghadam & Shaabani 2013: 164;Raza & Rehman 2012: 5085).As opposed to transactional marketing, relationship marketing is viewed as a long-term and continuous marketing approach (Dwyer, Schurr & Oh 1987: 13;Kumar et al. 2003: 675).Morgan and Hunt (1994: 22) define relationship marketing as developing, establishing and maintaining of successful relational exchanges between service providers and their customers.Relationship marketing thus centres on customer-seller relationships, benefits both parties and is longitudinal in nature (Hunt, Arnett & Madhavaram 2006: 83;Morgan & Hunt 1994: 34).
However, despite many service providers adopting relationship marketing strategies, not all customers want to build relationships with them (Gilaninia et al. 2011: 797).Since the success of any relationship marketing strategy largely relies on the customer's intention to be involved in the relationship (Raciti et al. 2013: 615-616), Cronze et al. (2010: 60) advocate that relationship marketing strategies should only be practised on customers with relationship intentions.Identifying whether customers have relationship intentions should thus be considered as the starting point of any decision to implement relationship marketing strategies (Kumar et al. 2003: 667;Raciti et al. 2013: 615-616).

Relationship intention
1Relationship intention refers to customers' intentions to build relationships with service providers while purchasing products or services from them (Kumar et al. 2003: 669).It has been argued that customers can be categorised according to their relationship intentions on a continuum ranging from low (thus transactional customers) to high relationship intention, and that five sub-constructs should be considered when establishing customers' relationship intentions, namely involvement, expectations, forgiveness, feedback and fear of relationship loss (Kumar et al. 2003: 668-670).
Involvement refers to customers' willingness to be involved in relationship activities without feeling obliged to do so (Kumar et al. 2003: 670).This willingness is a result of the importance of the product or service to the customer (Bienstock & Stafford 2014: 216;Zaichkowsky 1985: 345).Increased customer involvement enables service providers to better understand customer needs, thereby increasing customer satisfaction (Ashley et al. 2011: 752;Kumar et al. 2003: 670).It has thus been suggested that customers with relationship intentions are more involved with the service provider and its products or services (Ashley et al. 2011: 755;Kumar et al. 2003: 670).
Expectations are beliefs about the level of service against which the actual service delivery performance is measured (Wilson, Zeithaml, Bitner & Gremler 2012: 51).These beliefs are formed by various factors such as implicit and explicit service promises, past experiences and word of mouth (Cronze et al. 2010: 55;Wilson et al. 2012: 61).Customer expectations can be classified into two groups, namely desired service and adequate service.When the level of service experience falls between these groups (referred to as the zone of tolerance), customers are satisfied; however, when the level is below the adequate level of service, customers are dissatisfied (Wilson et al. 2012: 53-54).Since customers develop higher expectations of service providers by investing time and effort when forming a relationship (De Wulf, Odekerken-Schröder & Iacobucci 2001: 34;Liang & Wang 2006: 120), it is believed that customers with high relationship intentions will have higher expectations of service providers than customers with low relationship intentions (Kumar et al. 2003: 670).
A service failure occurs when customers' service expectations are not met (Siddiqui & Tripathi 2010: 122).Forgiveness allows customers to release negative feelings towards service providers, thereby enabling them to overcome occasional service failures (La & Choi 2012:111).Cronze et al. (2010: 55) explain that forgiveness in relationships entails the willingness to accept short-term disadvantages to maintain long-term relationships.This stems from the belief that customers engaging in relationships with service providers are more likely to forgive service failures since they expect to maintain the relationship (Kim, Ok & Canter 2012: 59;Wilson et al. 2012: 340).Kumar et al. (2003: 670) accordingly propose that, unlike customers with low relationship intentions, customers with high relationship intentions will forgive service providers when service failures occur without expecting some form of compensation.Kumar et al. (2003: 670) argue that customers with relationship intentions will provide service providers with both positive and negative feedback regarding their expectations and experiences, thereby enabling service providers to continuously improve their services.When customers do not provide feedback, especially in the case of negative feedback, it robs service providers of the opportunity to perform service recovery efforts (Siddiqui & Tripathi 2010: 122).Customer feedback is thus very important as it enables service providers not only to perform service recovery, but to detect areas of possible service failures before they occur, thereby allowing them to perform prevention techniques (Siddiqui & Tripathi 2010: 135).Kumar et al. (2003: 669) accordingly hypothesise that customers with high relationship intentions are more willing to provide service providers with feedback than customers with low relationship intentions.Kumar et al. (2003: 669) suggest that customers with high relationship intentions are less likely to switch to competitors and will make an effort to use the service provider.This is a result of customers' fear of losing their relationship with their service provider due to the perceived switching costs involved.Switching costs consist of procedural switching costs (i.e. the effort required to find a new service provider), social switching costs (i.e. the loss of the relationship with the service provider or its employees) and lost benefit costs (i.e. the loss of the relationship benefits which customers enjoyed with their service providers) (Jones, Reynolds, Mothersbaugh & Beatty 2007: 337).
Customer satisfaction will be discussed next, as it is believed that customers engage in relationships where they obtain a feeling of satisfaction (Brotherton & Evans 2010: 28;Raciti et al. 2013: 627).

Customer satisfaction
1Customer satisfaction can be defined as the positive feelings customers experience after consuming a product or service (Oliver 1980: 460).Bolton and Christopher (2014: 17) support this perspective by viewing customer satisfaction as customers' positive post-consumption assessment.
When studying customer satisfaction, the expectancy disconfirmation paradigm probably serves as the best explanation thereof (Oliver 1980: 460).According to this paradigm, customers hold certain expectations every time they purchase products and services.When the product or service performance meets customers' expectations, confirmation occurs.However, when the actual product or service performance differs from customers' expectations, disconfirmation occurs.The result of the disconfirmation is either customer satisfaction when the performance exceeds expectations, or dissatisfaction when the performance falls below expectations (Oliver 1980: 460-461;Wilson et al. 2012: 53-54, 75).
When studying customer satisfaction, it is important to note that it is influenced by various factors, such as product or service features, quality, price, the customer's mood, as well as situational factors (Wilson et al. 2012: 75).Moreover, customer satisfaction is not static but dynamic and can vary at different stages of the service experience, mainly due to the fact that customer satisfaction is related to customers' feelings such as fulfilment, contentment, pleasure, happiness, delight, relief and ambivalence (Wilson et al. 2012: 75).
The importance of pursuing customer satisfaction becomes evident when considering the benefits associated with achieving customer satisfaction.Firstly, it is believed that an improvement in customer satisfaction leads to customers spreading positive word of mouth (Tohidinia 2011: 250) as well as an increase in profitability (Halimi et al. 2011: 51;Williams & Naumann 2011: 25-26).Secondly, research determined that satisfied customers are not only more motivated to build relationships with service providers (Halimi et al. 2011: 51), but also have higher repurchase intentions (Bolton & Christopher 2014: 17;Curtis et al. 2011: 15).Finally, it has been established that customer satisfaction leads to the development of customer loyalty, an essential predecessor of customer retention (Goncalves & Sampaio 2012: 1509;Vesel & Zabkar 2009: 402).

Customer loyalty
1Customer loyalty can be viewed from either an attitudinal or behavioural perspective (Pan et al. 2012: 156).Whereas behavioural loyalty refers to customers' actual purchasing behaviour, attitudinal loyalty refers to the emotional bond, favourable attitude and strong preference customers have for a service provider (Oliver 1999: 35).Behavioural loyalty thus considers purchasing history, while attitudinal loyalty refers to a customer's future purchasing intentions (Gee, Coates & Nicholson 2008: 360).When considering the influence of behavioural and attitudinal loyalty on a number of different relationship marketing-related constructs (including customer satisfaction, trust, perceived value and switching costs), Pan et al. (2012: 156) found no significant differences between the different types of loyalty, resulting in the conclusion that attitudinal loyalty could be viewed as "a plausible surrogate of behavioural loyalty".As proposed by Pan et al. (2012: 251), this study accordingly defines customer loyalty as customers' attachment to a service provider and their intent to continually support the service provider in the future.
Considering the above, it becomes clear why it has been argued that having a loyal customer base helps service providers not only to survive, but to thrive in highly competitive markets (Pan et al. 2012: 157).However, despite the view that loyalty can be considered as a very important asset, especially in service industries (Cronze et al. 2010: 53;Richard & Zhang 2012: 568), the attainment thereof has become increasingly difficult (Raza & Rehman 2012: 5091) due to the competitiveness within markets, making it difficult to achieve and maintain customer loyalty (Alrubaiee & Al-Nazer 2010: 155).

Problem statement, objectives and hypotheses
1T raining providers within the South African wholesale and retail sectors have experienced increased competition due to the implementation of the Skills Development Act (No. 97 of 1998) by government (W&RSETA 2013: 37).It has thus become essential for wholesale and retail training providers to understand which customers to focus marketing efforts on in an attempt to increase customer loyalty, as insights gained into customer loyalty could increase their chances of success.
Considering the fact that the majority of existing relationship intention research has focused on customers within a business-to-consumer (B2C) setting (Cronze et al. 2010: 57;Fernandes & Proença 2013: 48;Kruger & Mostert 2012: 44), the need arises to determine customers' relationship intentions also within a business-to-business (B2B) environment, as originally proposed by Kumar et al. (2003: 667).With limited research on the influence of relationship intention on other relational constructs, particularly its influence on customer loyalty (Cronze et al. 2010: 51-62;Vázquez-Carrasco & Foxall 2006: 205-219), research considering the influence of relationship intention on customer loyalty is also warranted.
Similarly, service providers have been striving towards ensuring customer satisfaction, as it has been argued that achieving loyalty is not possible without satisfying customer needs (Hansen 2012: 352;Ishaq 2012: 26;Raza & Rehman 2012: 5085;Van Vuuren et al. 2012: 93;Vesel & Zabkar 2009: 402).Bolton and Christopher (2014: 17) support this view by arguing that an interaction effect exists between customer satisfaction and re-purchases: when customers repeat their purchases with a service provider, they are likely to experience greater satisfaction, which in turn increases the possibility of future spending with the provider.Bolton and Christopher (2014: 17) accordingly encourage service providers to invest in customer relationships earlier in their dealings with them, rather than later.However, since Kumar et al. (2003: 667) warned that not all customers want to build long-term relationships with service providers, it is imperative that service providers identify, focus on, and satisfy the needs of those customers with relationship intentions.The purpose of this study was accordingly to determine the extent to which relationship intention and customer satisfaction predict customer loyalty within the wholesale and retail training sectors.
The following secondary objectives were set for the study, namely to: • Determine the validity and reliability of the relationship intention measurement scale within the wholesale and retail training sectors (thus within a B2B context) • Categorise wholesale and retail skills development decision-makers according to their relationship intentions • Determine whether relationship intention and customer satisfaction predict customer loyalty towards wholesale and retail training providers.

Research methodology
Research design, study population and sampling

1A
descriptive and quantitative research design was used in the study.The study population included wholesale and retail skills development decisionmakers, located across South Africa, who were involved in the selection of their organisations' training providers.A database (used as sampling frame) containing contact details and e-mail addresses of 3800 individuals working in the training industry, including skills development decision-makers, was obtained from a South African training provider.Since the sampling frame did not specifically identify wholesale and retail skills development decision-makers, a decision was made to distribute the invitation to participate in the research study to the entire sampling frame.Screening questions were therefore used to ensure that only wholesale and retail skills development decision-makers participated in the study.In total, 185 of the 192 completed questionnaires could be used for analysis.

Questionnaire and data collection
1This study used self-administered online questionnaires to collect data, as done in similar studies (Akman & Yurur 2012: 217-229;Ishaq 2012: 25-36).The questionnaire commenced with screening questions to ensure that only wholesale and retail skills development decision-makers took part in the study.The screening questions included: "Has your company participated in wholesale and retail training which was provided by an external wholesale and retail training provider?"and "Are you involved in the process of deciding which training provider to use?"If they met the criteria set by the screening questions, eligible respondents were requested to complete the four sections included in the questionnaire.The first three sections included measuring scales using five-point Likert-type scales, where 1 = strongly disagree and 5 = strongly agree, adapted from a number of authors to measure customer satisfaction (Gremler & Gwinner 2000: 95;Zboja & Voorhees 2006: 389), customer loyalty (Dagger, David & Ng 2011: 280;Zeithaml, Berry & Parasuraman 1996: 38) and relationship intention (Kruger & Mostert 2012: 15-23).The final section in the questionnaire captured respondents' demographic details.The questionnaire was pretested (Cooper & Schindler 2011: 89) among 12 wholesale and retail skills development decision-makers prior to data collection, leading to minor language-related changes being made to the questionnaire.
A hyperlink to the questionnaire, hosted on Qualtrics, was e-mailed to the names contained in the database used as the sampling frame.Since the study population comprised business decision-makers, it was decided to limit follow-up emails inviting prospective respondents to participate in the study to two requests.In total, 185 of the 192 completed questionnaires could be used for analysis.

Data analysis
1The data were automatically captured by Qualtrics on to an Excel spread sheet and exported to the Statistical Package for Social Science (SPSS) (version 22) for analysis.To validate that the scales measured what they were intended to measure, exploratory factor analyses were performed, while Cronbach's alpha coefficient values were calculated to test the reliability of each scale.As suggested by Pallant (2013: 104), Cronbach's alpha values of at least 0.7 were regarded as being indicative of acceptable reliability.
To test the hypotheses formulated in the study, hierarchical multiple regression analysis (also called sequential multiple regression) was used to determine the extent to which relationship intention and customer satisfaction (as independent variables) predict the dependent variable, namely customer loyalty.Before performing the regression, the researchers ensured that the assumptions of multiple regression analysis, being normality, linearity, multicollinearity and homoscedasticity, were not violated (see the discussion on the results from the hierarchical multiple regression).R-square was used to determine how much of the change in customer loyalty is explained by relationship intention and satisfaction (Pallant 2013: 167), whereas p-values of ≤ 0.05 were regarded as statistically significant (Pallant 2013: 167).Beta values were analysed to indicate which independent variable made the largest contribution in explaining the dependent variable (Pallant 2013: 167).

Sample profi le
1The sample profile of the respondents participating in the study, including the respondent's position in the organisation, the number of employees working at their organisation and the organisation's area of business, are summarised in Table 1.  1, it is clear that the majority of respondents participating in the study indicated that they had a position other than the options included in the questionnaire.
Since respondents could specify their position within their organisation, further analysis found a number of diverse positions, including general non-descriptive, positions such as 'manager', 'branch manager' and 'regional manager'.Against the listed options that respondents could choose from, most indicated that they were skills development facilitators (21.6%), business owners (20.5%), training managers (16.2%) or human resource managers (15.7%).It is furthermore evident that most of the respondents worked at organisations with more than 150 employees (43.8%), followed by those with 1-50 employees (35.1%).Finally, concerning the organisations' area of business, most respondents indicated that they operated in an area other than the options included in the questionnaire (29.2%).Since respondents could specify their business area, further analysis found a number of diverse areas including 'various', 'procurement', 'medical', 'education' and 'finance'.Against the listed options that respondents could choose from, most indicated that they were general dealers (26.04%),followed by food wholesalers/retailers (17.19%) and fuel wholesalers/retailers (8.33%).

Reliability and validity
1T o determine the construct validity of the measuring scales used in this study, exploratory factor analyses were performed.The measures of sampling adequacy (MSA) were all higher than 0.6 (Pallant 2013: 199), and more than 60% of the mcmiiT able 1 continued variance was explained by the underlying dimensions of each measurement scale used in this study.With eigenvalues for each factor extracted being larger than one (Pallant 2013: 191), the underlying dimensions were uncovered and accordingly labelled.Since the items contained in each scale measured the same underlying construct, composite scores for each scale could be calculated (Pallant 2013: 105).
The reliability of the scales used was determined by calculating Cronbach's alpha coefficient values.All scales included in the questionnaire were considered to be reliable, as the Cronbach's alpha values were above 0.70 (Pallant 2013: 104).Table 2 presents the underlying dimensions uncovered during the factor analyses, the mean scores for each dimension, as well as the realised Cronbach's alpha values.  2 that respondents tended to agree with the scale items included in the customer satisfaction (mean = 4.13; SD = 0.70) and loyalty (mean = 4.00; SD = 0.75) scales.It can thus be concluded that respondents were relatively satisfied with the service they received from, and loyal towards, their training providers.However, respondents displayed mediocre relationship intentions (mean = 3.62; SD = 0.56) towards their training providers.
Categorising respondents according to relationship intention levels 1Based on their overall relationship intention mean scores, respondents were categorised into three groups using the 33.3 and 66.6 percentiles by using the Visual Binning functionality offered in SPSS.The mean scores separating groups were therefore 3.27 and 3.87, differentiating between respondents with low, moderate or high relationship intentions.Table 3 presents the frequencies, mean and standard deviations (SD) for the three relationship intention groups.1F rom Table 3 it can be seen that 62 respondents were categorised as having low relationship intentions (mean = 3.01; SD = 0.24), 63 respondents as having moderate relationship intentions (mean = 3.58; SD = 0.15) and 60 respondents as having high relationship intentions (mean 4.26; SD = 0.34).The number of respondents varied in the groups due to the fact that ties occurred in the continuous data.From Table 3 it can thus be concluded that wholesale and retail skills development decision-makers can be categorised according to their relationship intentions towards their training providers.

Hierarchical multiple regression
1P rior to conducting a hierarchical multiple regression, a number of assumptions first had to be tested.From the preliminary analysis, it became evident from the scatterplot, and was confirmed when checking the Mahalanobis distances, that one outlier was detected (Tabachnick & Fidell 2014: 10).The decision was thus made to remove the case from the analysis and to rerun the hierarchical multiple regression analysis (Pallant 2013: 166).From the analysis it could be concluded that all the required assumptions for conducting a hierarchical multiple regression were met, as illustrated by the following: • Based on the equation proposed by Tabachnick and Fidell (2014: 159-160), the study required a minimum sample size of 66.For the purposes of the regression analysis, the study realised a sample of 184 respondents, which is well above the suggested minimum sample size.• Since no correlations above 0.9 were realised (Pallant 2013: 157), the variable inflation factor index values (VIF) were below 10 (Cooper & Schindler 2011: 533), and the independent variables were not a combination of other independent variables (Pallant 2013: 157), it was concluded that multicollinearity and singularity did not exist in the data.• When calculating the Mahalanobis and Cook's distances (Pallant 2013: 166), it became evident that no outliers were present in the dataset.• The normal probability plot indicated a fairly straight diagonal line from bottom left to top right, and the scatterplot residuals were distributed in a rectangular-like shape with the majority of the scores in the centre (around 0).The assumptions of normality, linearity and homoscedasticity were thus met (Allen & Burnett 2010: 194;Pallant 2013: 165). 1The Pearson's product moment correlations between the constructs, namely relationship intention, customer satisfaction and customer loyalty, were analysed.All the constructs were significantly and positively correlated to one another.The construct with the strongest correlation to loyalty was satisfaction with 0.813.The correlation between relationship intention and loyalty was 0.637, and between relationship intention and satisfaction 0.616.
A hierarchical multiple regression was accordingly performed to, firstly, determine the statistical significance of relationship intention (independent variable) as a predictor of customer loyalty (dependant variable), and secondly, to determine whether adding a second predictor, namely customer satisfaction (independent variable), would significantly improve the model to predict customer loyalty (dependent variable).Table 4 includes a summary of the results obtained from the two models tested in the study, including the coefficient of determination (R-square values) of both models.4, the first model, including only relationship intention as a predictor of customer loyalty, produced a coefficient of determination (R-square value) of 0.406, indicating that relationship intention explained 40.6% of the variance in customer loyalty.However, when satisfaction was added as a second predictor in the regression model, the coefficient of determination improved to 0.690, thus implying that relationship intention and customer satisfaction in combination explained 69% of the variance in customer loyalty.By adding satisfaction to the regression equation, an additional 28.4% of the variance in loyalty was thus accounted for.The results of the ANOVA test are presented in Table 5, indicating that both models were statistically significant with p<0.05.

Dependent variable: Customer loyalty
1F rom Table 6 it can be seen that for the second model, both relationship intention and satisfaction are significant (p < 0.05) predictors of loyalty.It can furthermore be seen that satisfaction recorded a higher beta value (beta value = 0.677, p < 0.05) than relationship intention (beta value = 0.220, p < 0.05).Table 6 confirms that both relationship intention and satisfaction are statistically significant (p < 0.05) predictors of customer loyalty.
The following conclusions regarding the hypotheses formulated for the study can therefore be drawn based on the results from the hierarchical multiple regression: • Hypothesis 1 stating that customers' relationship intentions significantly predict their loyalty towards their wholesale and retail training providers (beta value = 0.637; p < 0.05), is therefore supported.• Hypothesis 2 stating that customers' satisfaction significantly predicts their loyalty towards wholesale and retail training providers (beta value = 0.677; p < 0.05), is therefore supported.• Hypothesis 3 stating that customers' relationship intentions and satisfaction, in combination, significantly predict their loyalty towards wholesale and retail training providers (relationship intention: beta value = 0.220, p < 0.05; satisfaction: beta value = 0.677; p < 0.05), is therefore supported.
Discussion and managerial implications 1Creating and maintaining customer loyalty is of utmost importance to service providers not only due to increased competition, but also because of the benefits associated with customer loyalty.Since customer loyalty can be achieved by focusing on customers with relationship intentions (Cronze et al. 2010: 60), identifying and focusing on customers with relationship intentions could lead to a competitive advantage by enabling service providers to better understand their customers and thereby increase customer satisfaction (Raciti et al. 2013: 626).It is furthermore believed that satisfied customers are more motivated to build relationships with service providers (Halimi et al. 2011: 51), and that customer satisfaction will ultimately lead to customer loyalty (Goncalves & Sampaio 2012: 1509;Vesel & Zabkar 2009: 402).
The purpose of this study was to determine the extent to which relationship intention and customer satisfaction predict customer loyalty within the wholesale and retail training industry.A first finding from this study indicates that the relationship intention measurement scale was reliable and valid within the wholesale and retail training context.The relationship intention measurement scale used in this study can therefore be used to measure wholesale and retail skills development decision-makers' relationship intentions towards their wholesale and retail training providers.This finding implies that the relationship intention measure is not only valid and reliable within a B2C context (Kruger & Mostert 2012: 44), but also within a B2B context.This finding thus supports Kumar et al.'s (2003: 667) view that relationship intention is applicable to both B2C and B2B customers.Wholesale and retail training providers, as well as other service providers, within a B2B environment can thus use the relationship intention measure to determine customers' relationship intentions.
As proposed by Kumar et al. (2003: 667), the results from this study found that wholesale and retail skills development decision-makers can be categorised according to their relationship intention levels, namely low, moderate and high.The results also suggest that respondents participating in this study had mediocre relationship intentions towards their training providers, and therefore wholesale and retail training providers cannot bargain on building long-term relationships with all of their customers, thus supporting Kumar et al.'s (2003: 669) point of view.If wholesale and retail training providers want to retain their customers and be successful in the market, they must focus their relationship marketing efforts on building relationships with those customers who have higher relationship intentions.
The results also showed that relationship intention, satisfaction and customer loyalty were all significantly and positively related to one another.In particular, it was found that respondents' relationship intentions towards their wholesale and retail training providers significantly predict their loyalty towards their wholesale and retail training providers.This finding is in line with previous research findings, indicating that relationship intention is significantly related to customer loyalty (Bloemer, Odekerken-Schröder & Kestens 2003: 239;Cronze et al. 2010: 51-62;Vazquez-Carrasco & Foxall 2006: 215).It is thus recommended that wholesale and retail training providers identify customers with relationship intentions and build long-term relationships with these customers in order to maximise the outcomes of their relationship marketing strategies and in particular to achieve customer loyalty.
Previous research studies (Goncalves & Sampaio 2012: 1521;Raza & Rehman 2012: 5085;Terblanche & Boshoff 2010: 6;Van Vuuren et al. 2012: 93) found that customers' satisfaction significantly influences their loyalty.The results from this study support previous findings, since customer satisfaction was found to be a predictor of customer loyalty.It is therefore essential that wholesale and retail training providers ensure customer satisfaction by managing expectations and ensuring that the service provided to customers fulfils and exceeds customers' expectations.It is recommended that wholesale and retail training providers conduct customer expectation and satisfaction research among their customers to determine customers' current expectations and satisfaction levels.Furthermore, research could be conducted to forecast customers' future expectations, thereby allowing training providers to timeously adapt their service offerings to ensure greater customer satisfaction.
Finally, the results showed that relationship intention and satisfaction, in combination, predict customer loyalty.Service providers should thus focus not only on customers with relationship intentions, but also ensure these customers' satisfaction, as doing so could result in greater customer loyalty and possibly a competitive advantage (Islam 2010: 141).Focusing on these customers is furthermore important, since it has been established that greater levels of customer satisfaction lead to higher profitability (Hoissain & Ullah 2011: 7) and customer retention (Terblanche & Boshoff 2010: 1).Thus, despite the importance of considering customers' relationship intentions as well as customer satisfaction, training providers should not look at these two variables in isolation, but should specifically ensure the satisfaction of customers with relationship intentions, as this combination offers a greater prediction of customers' loyalty.
Limitations and future research 1The limitations associated with this study should be highlighted.Firstly, this research focused on only one wholesale and retail training provider and its customers.This limits the generalisation of the findings to the entire training and service industry.Secondly, as the study was conducted among business customers, the response rate obtained was relatively low.The low response rate was due to the chosen online survey method, as the researchers were not able to conduct the fieldwork in person.The cost associated with the fact that wholesale and retail skills development decision-makers are geographically dispersed prohibited interviewer-administered data collection.
Future research might find it valuable to investigate the extent to which relationship intention and customer satisfaction predict customer loyalty across the training sector and in other B2B service industries.It is suggested that future research should also investigate other potential predictors of customer loyalty in the wholesale and retail training sector, such as perceived value, relationship quality, switching costs and relational benefits.To improve on the sample size, more efficient methods of gathering data and encouraging participation among business customers should be explored.
1The following hypotheses were formulated to support the secondary objectives: H 1 : Customers' relationship intentions significantly predict their loyalty towards their wholesale and retail training providers.H 2 : Customers' satisfaction significantly predicts their loyalty towards their wholesale and retail training providers.H 3 : Customers' relationship intentions and satisfaction, in combination, significantly predict their loyalty towards wholesale and retail training providers.

Table 1 :
Sample profi le

Table 2 :
Overall mean scores and Cronbach's alpha values lii Construct 1It is evident from Table

Table 3 :
Relationship intention groups

Table 4 :
Model summary a

Table 5 :
ANOVA a