According to Sozen and O’Neill (2020), the graphical IPA model ranks service constructs from most to least important. The IPA, which Martilla and James formulated in 1977, is a tool that is used to evaluate product qualities in order to ascertain which attributes are most critical in terms of performance. The information obtained can then be used to prioritise product and service enhancements (Stergiou, 2018). As an expectation-disconfirmation model, the IPA model suggests that customer satisfaction is a function of how well various product or service features perform and how significant they are to the customer (Cai & Chi, 2021; Sever, 2015; Stergiou, 2018). This is a simple but useful way to explore quality attributes on two levels: performance and importance. Studies based on the IPA model typically use either a stated or derived measure of importance. Stated measures are collected directly from the target population via a Likert scale or direct rating scales, ranking or constant sum-scales, while derived measures are collected inferentially by correlating the attribute-level data of performance with an outcome variable such as customer satisfaction (Mikulić et al., 2012).

Policymakers can gain a snapshot that aids in effort concentration by plotting survey importance and performance ratings onto a two-dimensional plot (Figure 1), where performance is displayed on the horizontal axis and importance is on the vertical axis (Stergiou et al., 2018). This plot is divided into four quadrants with varying management conclusions by crosshairs, which are horizontal and vertical lines. For example, if an attribute is located in the ‘concentrate here’ quadrant, it may require management’s attention in order to improve performance when the relevance of the attribute is high and the performance is poor. These quadrants assist managers in creating quality-based marketing plans by demonstrating how to assess currently implemented marketing tactics.

The four quadrants depicted in Figure 1 are briefly described as follows.

Festivals form a significant part of the world’s leisure industry and are becoming increasingly important to society and the economy. As a result, festival organisers and festival experts rely heavily on information pertaining to attendees and their spending behaviour (Wong et al., 2015). Apart from the financial side, there is also a need for a better understanding of attendee satisfaction. What the attendees want, their previous experiences and how they respond to marketing messages all play a role (Haverila et al., 2021). For example, Rajaguru and Hassanli (2018) state that the expectancy-disconfirmation theory of Oliver (1980) is one of the best ways to explain the connection between the calibre of service and the level of customer satisfaction. The comparison can be evaluated using this theory to assess the comparison between what a customer would expect and the perceived service actually received. This theory serves as the foundation of service quality research and the hypothesis that a higher level of service perception when compared to preconceived expectations would lead to satisfaction (Hermann & Nemaorani, 2023). Most of the time, a positive or negative opinion of the event is formed by comparing expectations and perceptions (Rajaguru & Hassanli, 2018). The level of satisfaction could change because of factors that cannot be controlled, such as how good the service is seen to be perceived by others, how visitors feel, social interactions and other events (Koo et al., 2014; Liu et al., 2024). Some factors relating to an event that can be changed are its location, staff, parking, effective queue management, the variety and quality of food and beer, the price of the ticket and the quality of the event programme (Yuan & Jang, 2008), which includes entertainment such as live shows and demonstrations. These factors affect how satisfied people are and how likely they are to recommend and return to the event (Hattingh & Swart, 2016).

Torres (2014) and Hall et al. (2016) argue that festivals can only succeed if they provide high-quality service to their attendees, and that researchers and festival organisers should be cognisant of service quality and customer satisfaction to improve the experience they offer attendees. As the advent of the service quality theory in the early 1980s (Collins, 2017), the correlation between high-quality service, satisfied customers and repeat business has been the subject of much research in the field of tourism. Perhaps the best-known tool for gauging perceptions of service quality is the SERVQUAL instrument, developed by Parasuraman et al. (1988). This instrument measures the perceptions of service quality based on the variables reliability, assurance, tangibility, empathy and responsiveness. Customers’ impressions, according to Bowdin et al. (2011, p. 387), ‘are based on the technical (performance outcomes) and functional (process-related) qualities of the experience’, which implies that SERVQUAL’s functional areas of concern are where satisfaction with service quality should be prioritised. Destination marketing, which influences consumer purchases, relies heavily on customer satisfaction (Pizam et al., 2016; Torres, 2014; Welthagen and Lötter, 2020). Torres (2014) and Pizam et al. (2016) assert that Oliver’s (1980) expectation-disconfirmation model forms the theoretical foundation of the phenomenology of customer satisfaction. The expectations (E) and evaluation (P) of the festival service provider are both measured by the SERVQUAL model, which is based on the disconfirmation paradigm, and are useful for identifying service gaps and areas for service improvement.

To enhance the quality and keep customers content, numerous researchers (Cole & Chancellor, 2009; Tkaczynski & Stokes, 2010; Yoon et al., 2010) have evaluated the festival quality dimensions, which are made up of several service attributes. Welthagen and Lötter (2020) propose, however, that a deeper investigation into festival satisfaction is required. According to Tanford and Jung (2017), customer satisfaction is an aggregate rating of the quality of the product or service received. It is generally accepted that enhancing the quality of a product or service would boost customer satisfaction, which, in turn, would increase the likelihood of various positive outcomes (Yoon et al., 2010), such as repeat business, revenue, market share, favourable recommendations and product combinations (Hermann & Nemaorani, 2023). With a deeper understanding of the elements and factors affecting perceived service quality, there is a greater ability for service providers, specialists and marketers to improve the festival’s quality (Wong et al., 2015). Once this has been achieved, it is possible to earn repeat business from satisfied attendees.

Ethical approval to conduct this study was obtained from the Tshwane University of Technology Faculty of Management Sciences Research Ethics Committee (No. FCRE2022/FR/05/014-MS (2)).

The findings revealed an equal (50/50%) split between male and female respondents and that a significant portion of the respondents fell in the age range of 33–42 years. Most respondents indicated that their home language was Afrikaans (44.09%) and that they resided in Pretoria. Furthermore, the results indicated that most of the respondents (66.13%) were employed full-time. Most respondents (38.66%) indicated that their annual income was between R0 and R100 000, ranging from R10 000 to R30 000 per month. The majority of respondents were visiting the event for the first time, while most attended the festival with friends or family (54.95%). The respondents had mostly heard about the festival through social media (46.33%), while most indicated that their reasons for attending the festival were to spend time with family and friends (30.35%), followed by those who went there to listen to and enjoy live music (25.56%).

In all, 27 service attributes were measured, summarised and interpreted using the IPA framework to determine their contribution to customers’ dissatisfaction or satisfaction. Table 1 displays the mean importance and performance ratings of these attributes as well as a descriptive comparison between the means. The P-I score (performance minus importance) displays a descriptive gap. Cronbach’s alpha scores range between 0.81 and 0.97, indicating a high internal consistency and reliability of the data.

The importance-performance grid for the 27 festival attributes is displayed in Figure 2. The grid is divided into five dependent variables, which are as follows: pricing (18–21), staff (22–27), live entertainment (1–5) and site elements (6–12), amenities and services (13–17). Although the characteristics are presented in this order in Table 1, they were arranged in a random order in the questionnaire. The outcomes of the data analysis are presented descriptively in Table 1. The gap analysis measured the average performance deficit by comparing the average importance. When the average performance is higher than the average importance, a positive gap results. When the gap between mean performance and mean importance is negative, however, event organisers and administrators should act.

As seen from Table 1, the following festival attributes received positive gaps: Sticking to the programme schedule (P > I gap = –0.278), promptness of artists (P > I = –0.252), variety of music (P > I = –0.093), availability of staff (P > I = –0.019), and control of alcohol (P > I = –0.016). These results, highlighted in Table 1, indicate that the mean performance of these attribute items exceeded the mean importance from the results obtained from respondents of the event. Management should strive to keep the festival’s qualities performing at their current level because the findings show that they are functioning well in these areas. Table 1 further indicates that the following festival attributes received negative gaps: Adequate seating and spacing (P < I = 0.211), toilet cleanliness (P < I = 0.201), food and beverage availability (P < I = 0.166), overall safety and security (P < I = 0.153), and food and beverage variety (variety of beer/brewery) (P < I = 0.150). These results show that the mean importance is lower than the mean performance, which means that the recommendation about these attributes to CCBF management is that they signify a waste of scarce resources that are used inefficiently and might be reassigned elsewhere.

Crosshairs (vertical and horizontal lines) then separate this gap analysis map into four sections, each of which reveals a unique result (Figure 2). The ability to portray the results of an IPA graphically on a two-dimensional grid that is simple to understand is one of the attractive attributes of this type of study. The 27 attribute ratings for Table 1 are represented as 27 points on the importance-performance grid that is shown in Figure 2. The statistics are relative to the characteristics that are presented in Table 1. Marketing efforts are referred to by the labels that are placed on the quadrants A, B, C and D. An example of this would be the phrase ‘Concentrate here’, which refers to a region (A) in which qualities are significant and also where performance can show improvement. Concentrating constructive action in this one area would result in the best possible outcomes.

The 27 items were plotted on the IPA matrix (Figure 2) after collecting the implicitly derived importance and satisfaction performance of all tourism satisfaction variables. The grid was based on the average importance and performance scores. The ‘Keep up the good work’ quadrant contained 15 attributes, as illustrated in Table 2 and Figure 2. These attributes indicate that both the importance and performance of these attributes were high; these are the strengths of the festival. These attributes include responsiveness of staff, availability of staff, friendliness of staff, promptness of staff, reliability of staff, efficiency of staff, sound quality, quality of artists or entertainers, adequate signage and communication, overall safety and security, food and beverage availability, food and beverage variety, hygiene and/or cleanliness, control of alcohol and food and beverage quality. The ‘Concentrate here’ quadrant captured only three attributes (toilet cleanliness, availability and food and beverage prices). These indicate that the importance exceeds the perception of performance. The eight attributes in the ‘Low priority’ quadrant, which represents the attributes that respondents rated as matching the importance, but not exceeding it by a significant margin, include the (variety of music, promptness of artists, quality of viewing, adequate seating and spacing, parking availability, entrance ticket prices, parking price and vendor, exhibit and crafts prices). Sticking to the programme schedule is the only attribute included in the ‘Possible overkill’ quadrant.

Although most of the elements related to service quality fall in either Quadrant B or C, both of which have some correlation between importance and performance, the actual importance and performance ratings should be investigated. Qualities that fall in Quadrant B show how well the festival met the guests’ expectations for performance in the areas they thought important. Those qualities stand for significant advantages and a competitive edge that must be preserved.

The qualities in Quadrant C are seen as minor deficiencies that do not necessarily call for additional work if better performance is not attained. These traits also represent an urgent danger from competitors. The items in the other quadrants also require close inspection before being categorised as performing satisfactorily or unsatisfactorily. On closer inspection, several problem areas were revealed.

In order to further analyse the results and to determine whether any significant gaps existed in service delivery, a regression analysis, Wilcoxon signed rank test, and Spearman’s rank order correlation were conducted, as depicted in Table 3.

Quantile regression (QR) was used to examine the relationship between the independent variables, importance and performance, and the dependent variable comprising 27 festival attribute items. Quantile regression offers a robust alternative to traditional least squares regression by enabling analysis across the entire conditional distribution of the dependent variable, not merely its mean (Shasha et al., 2022). This approach is particularly effective in uncovering relationships at various quantile levels, capturing a broader range of potential impacts (Davino et al., 2014). Table 3 presents the QR model outputs, reporting the 50th percentile (median), interquartile range (P25 – P75), and the range of observed values. The subsequent section details QR findings specific to attendees’ evaluations of the CCBF’s performance attributes. These results precede and are complemented by an IPA, offering a comprehensive view of the determinants of attendee satisfaction.

To assess the discrepancies between perceived importance and performance of festival attributes, the Wilcoxon signed rank test was used. This non-parametric test is appropriate for large paired samples (n > 10) and does not assume a normal distribution (Chatterjee & Suklabaidya, 2021). It evaluates whether median differences exist between paired observations without relying on the assumption of normality (Harris & Hardin, 2013). Table 3 presents the test results, including Z-scores and two-tailed p-values.

The analysis revealed that most paired comparisons yielded statistically significant results (p < 0.05), with the exception of performance versus importance for price-related variables and for staff attributes. Significant Z-scores ranged between –3.748 and 0.479, exceeding the critical threshold of ± 1.645 for a 95% confidence interval, thus leading to the rejection of the null hypothesis for most attributes. The notable exceptions – price and staff – did not show significant discrepancies, suggesting minimal perception gaps in these areas.

Attributes related to live entertainment, site elements and amenities and services exhibited statistically significant differences between importance and performance (Z > 1.645), indicating meaningful perception gaps. Despite these differences, the overall magnitude of the effect was small, suggesting that the CCBF generally met attendee expectations. This implies that the event delivery closely aligns with what attendees consider important.

In addition to the Wilcoxon analysis, Spearman’s rank order correlation was used to explore the relationship between attendee satisfaction and various festival attributes. This non-parametric technique is ideal for assessing ordinal relationships without normality assumptions (Astivia & Zumbo, 2017). The correlation analysis was computed based on the range of Pearson correlation coefficient (r) strengths presented next. The Pearson correlation coefficient measures the linear correlation between two variables, X and Y. The coefficient can range from –1 to +1, where –1 indicates a perfect negative linear correlation, 0 means no linear correlation, and +1 indicates a perfect positive linear correlation. The interpretation of Pearson correlation coefficient values is as follows: no correlation: a value of zero implies no relationship; low degree: 0.00 – 0.199 = very weak correlation; 0.20 – 0.399 = weak; 0.40 – 0.599 = moderate; high degree: 0.60 – 0.799 = strong; 0.80 – 1.000 = very strong (Astivia & Zumbo, 2017). Table 3 presents Spearman’s rho values, revealing significant positive correlations across most attribute dimensions:

The findings of this study generated several suggestions for enhancing the sustainability of the CCBF. To increase visitor retention, the organisers should strengthen digital customer relationship management through their website. This includes integrating an online feedback mechanism, promotional campaigns, loyalty programmes and regular e-newsletters. The facilitation of ticket purchases via the official website or linking to a secure external platform, alongside features such as Google Maps navigation, virtual tour videos and information on safety protocols, would improve the user experience. In addition, reservation systems for both attendees and vendors, and information on community engagement initiatives, would enhance transparency and accessibility.

Given the socio-economic diversity of the attendees, it is essential to maintain affordable pricing. Organisers are advised to avoid excessive price increases while considering differentiated ticketing strategies – such as bundling entry with premium amenities (e.g., food, beverages, VIP access) – to appeal to higher-income segments without alienating core patrons.

Site infrastructure significantly influences attendee satisfaction. Parking, in particular, warrants attention, as most visitors arrive in private vehicles. Expanding parking availability and promoting alternatives like the Gautrain could alleviate pressure. Social media remains the most effective promotional channel, reflecting the influence of peer recommendations. On the other hand, traditional media (radio, television, print) has a minimal impact. To retain the existing market and grow others, especially in Gauteng, targeted advertising through Afrikaans-language platforms, such as Jacaranda FM, is recommended.

The festival experience could be further improved by enhancing the live entertainment components. Sound quality, programming coherence, and artist scheduling directly impact attendee satisfaction. While staff-related aspects were not the most critical to satisfaction levels, they remain important to the overall experience. Thus, continuous investment in staff training, responsiveness, and professionalism is essential.

Collectively, these strategies could assist festival organisers in aligning service delivery with visitor expectations, thereby fostering long-term event viability and growth.