Digital Engagement and Visitor Loyalty at Cultural Festivals: Evidence from Vilnius and Implications for Financial Sustainability

Abstract

Cultural festivals increasingly depend on digital channels to reach audiences, which makes the allocation of limited promotional resources a recurring management problem. This study examines how digital engagement shapes visitor satisfaction and revisit intention at a cultural tourism event, using the Vilnius Pink Soup Festival as a case. Five factors (social media, public website quality, electronic word of mouth (eWOM), festival expectations, and ICT usability) are tested as predictors of visitor satisfaction and revisit intention. The decision to model visitor-level rather than adoption-level outcomes follows the human-centric premise of Industry 5.0, which motivates the choice of dependent variables but is not itself operationalized or tested. Five factors (social media, public website quality, electronic word of mouth (eWOM), festival expectations, and ICT usability) are tested as predictors of visitor satisfaction and revisit intention. A quantitative survey of 384 attendees was analyzed in IBM SPSS Statistics 29 by ordinary least squares regression on composite construct scores, with mediation tested through the PROCESS macro (Model 4) using 5000 bootstrap resamples. Four of the five factors significantly influence satisfaction, with festival expectations exerting the strongest effect, followed by social media, public website quality and electronic word of mouth; ICT usability was not significant once website quality was accounted for. Satisfaction did not mediate the relationships between these antecedents and revisit intention; with the antecedents controlled, they predicted revisit intention directly, explaining 57.5% of its variance. No ticket revenue, visitor expenditure, marketing expenditure or sponsorship income was collected, and no financial outcome is tested here. The findings are therefore reported as evidence on visitors’ response, and their financial relevance—which channels organizers might prioritize when allocating promotional budgets—is developed as a managerial implication rather than as an empirical result.

1. Introduction

Information and communication technology (ICT) has become an important field in the study of cultural-event promotion, expanding rapidly since the arrival of the World Wide Web (Darwish & Lakhtaria, 2011). Over time, the integration of websites and social media has given cultural-event organizers powerful tools for increasing attendance and audience contact and for building stronger ties among the participants who sustain an event. By reaching broad online audiences at low cost, these platforms are best understood as marketing investments whose returns accrue indirectly, through attendance, visibility and loyalty (Rust et al., 2004; Srinivasan & Hanssens, 2009). Digital marketing gives organizers access to platforms such as Facebook, Instagram, X, YouTube, TikTok, and review sites, where visitors actively engage, share content, and exchange experiences, and this active participation by organizers and visitors is central to promoting cultural-tourism products, sustainability, and services (Darwish & Lakhtaria, 2011; Li et al., 2021). Recent studies emphasize the role of social media in marketing and sustainable development (Li et al., 2021), yet the specific contribution of social media to the sustainability of heritage festivals, particularly in tourism, remains inadequately addressed. Closing this gap requires examining how public websites and social media serve cultural events and tourist destinations, and how consumer activity, interaction, and social relationships are shaped through digital technologies (He, 2022), which matters for promoting sustainable tourism and, in turn, the wider development of the sector (Moric et al., 2021; Aripin et al., 2023). Financial sustainability is increasingly recognized as a central dimension of cultural tourism, where well-promoted local events attract visitors, stimulate spending, and strengthen local economies (Chong & Balasingam, 2019). To differentiate their offerings and compete globally, cultural-tourism destinations must market themselves effectively through social media (Khan et al., 2021), and many countries now treat websites as a key tool for marketing festivals, since they generate greater exposure and interaction for such events (Arasli et al., 2021).

The motivation for this framing is financial, although no financial quantity is measured in this study. For a cultural festival, as for a firm, digital promotion is a budget-allocation problem under uncertainty: promotional spending creates value only insofar as it converts audiences into attendance, loyalty and repeat visitation (Fischer et al., 2011; Katsikeas et al., 2016), and festivals, which typically operate with narrow margins and volatile revenue, have little room for misallocation (Getz, 2002; Andersson & Getz, 2009). The human-centric pillar of Industry 5.0 places people—visitors and host communities rather than technology alone—at the center of value creation (Breque et al., 2021). We use this pillar in a deliberately bounded way, and it is worth stating precisely what it does and does not do here. It does not generate hypotheses that would not also follow from technology-acceptance or service-quality theory; the five antecedents tested below are drawn from the literature, no Industry 5.0 construct is operationalized, and the paper makes no claim to extend Industry 5.0 theory. What the framing does is govern the choice of dependent variables: an adoption-centered account would evaluate digital investment through what the organizer deploys—platform uptake, posting frequency, site traffic—whereas the human-centric premise relocates the criterion to what the visitor experiences, which is why satisfaction and revisit intention serve as the outcomes of the model. Industry 5.0 therefore functions here as a normative frame for cultural-tourism research rather than as a theory under test, and the contribution of the study is empirical: a tested model of how five forms of digital engagement relate to visitor satisfaction and loyalty at a cultural festival.

Cultural-tourism destinations increasingly promote local community cultures and traditions as a way of enriching visitor experiences and fostering deeper cultural encounters (Martins et al., 2025). The UK government, for example, has actively marketed the Notting Hill Carnival, an annual London celebration of the Caribbean diaspora promoted widely through social media (Gugolati & Klien-Thomas, 2022), while the Lithuanian government has similarly promoted the Vilnius International Film Festival through various websites and social platforms. Social media and public websites have not only eased communication but also reshaped how communities engage with one another, enabling real-time connection and the sharing of experience, which sustains cultural events by fostering long-term engagement and awareness (Achmad & Yulianah, 2022).

Local festivals are significant cultural-tourism events, uniting visitors and residents in shared pride and celebration of common heritage (Hassanli et al., 2021). Websites and social media have become primary tools for promoting such festivals (Utami et al., 2024), and within sustainable-festival management they have transformed both how and what destinations communicate to their target markets (Arasli et al., 2021), which makes them especially valuable in tourism. Through Facebook, Instagram, TikTok, and X, among other channels, social media has emerged as a dynamic technology contributing substantially to the marketing of cultural-tourism products (Litovka-Demenina et al., 2025). Public websites and social media also shape consumers’ intentions to visit particular destinations; as Torabi et al. (2022) note, overall satisfaction at a destination rests on emotional and physical experiences formed across dimensions such as positivity, surprise, and usability, and these experiences deepen attachment to the destination and foster long-term engagement.

The novelty of this study lies in the absence of empirical research on the specific effect of public websites and social media on the sustainability of cultural events such as the Vilnius Pink Soup Festival in Lithuania. This study aims to examine how digital engagement shapes visitor satisfaction and revisit intention at a cultural festival. Satisfaction and loyalty are treated as visitor-level outcomes that prior research identifies as antecedents of festival revenue. The financial sustainability of the event is not measured here; it enters the paper as a managerial interpretation of the behavioral findings, not as a tested outcome. The gap points to a broader need to understand how digital platforms can be used effectively to strengthen the long-term sustainability and success of cultural festivals.

To meet these objectives, the study reviews the existing literature on the use of websites and social media in destination marketing. Its contributions are twofold.

First, the study contributes to the literature on the sustainability of heritage festivals in the previously under-researched setting of the Vilnius Pink Soup Festival in Lithuania. By focusing on this event, it highlights the value of presenting cultural traditions and historical sites through public websites, social media, and other information sources.

Second, the proposed factors offer insight into consumer revisit intentions, providing useful guidance for festival organizers and participants, particularly at the Vilnius Pink Soup Festival. Using the festival as a case, the research examines how digital communication advances social, economic, and cultural sustainability in line with SDGs 11 and 12, so that, by analyzing consumer responses, event managers can design more engaging and relevant experiences that meet visitors’ needs while securing the event’s long-term sustainability and growth.

1.1. Digital Engagement as a Marketing Investment

Treating digital promotion as an investment rather than a communication expense follows a long line of work on marketing productivity and marketing return on investment. This literature argues that promotional spending should be evaluated by the value it creates for the organization—through customer acquisition, retention and customer equity—rather than by exposure metrics alone (Rust et al., 2004; Gupta et al., 2004; Katsikeas et al., 2016), and that the link between marketing action and financial performance runs through intermediate, perceptual outcomes such as satisfaction, attitude and loyalty (Srinivasan & Hanssens, 2009; Hanssens & Pauwels, 2016). The present study operates at precisely that intermediate level: it measures the perceptual returns of digital engagement, which prior work identifies as the pathway through which promotional spending eventually reaches revenue, rather than measuring revenue itself.

A second strand concerns how promotional budgets are allocated across channels when returns are uncertain and channel costs differ. Budget-allocation studies show that reallocating a fixed budget toward higher-elasticity channels can raise returns substantially without increasing total spending (Fischer et al., 2011), while research on earned media finds that word-of-mouth-driven acquisition produces longer-lived and more cost-efficient effects than paid promotion (Villanueva et al., 2008; Trusov et al., 2009; Stephen & Galak, 2012). This provides the basis for comparing earned channels (eWOM, social media) with channels that require sustained expenditure (public website development and maintenance) within a single model.

Third, festivals are typically non-profit or mixed-funding organizations with concentrated, volatile revenue and thin reserves, which makes them financially vulnerable and makes promotional efficiency a question of survival rather than of optimization (Tuckman & Chang, 1991; Getz, 2002; Andersson & Getz, 2009; Bowman, 2011). Low-cost digital channels are attractive to such organizers precisely because they lower the fixed cost of reaching an audience; whether they do so effectively depends on which channels visitors actually respond to (Gorokhova & Simanavičienė, 2026). Consistent with this literature, financial terminology is used here in a specific and bounded sense: the study examines the perceptual antecedents of financial outcomes, not realized revenues or costs.

1.1.1. Customer Satisfaction and Sustainable Cultural Event Promotion

Tourism sustainability requires balancing social, cultural, and environmental considerations (UNWTO, 2025). Cultural events matter for social and cultural sustainability because they help communities reinforce their identities, maintain traditions, and strengthen local economies (Hassanli et al., 2021). Digital promotion through official websites and social media advances this aim by bringing audiences together and stimulating interest in local culture and its continuity (Thelen & Kim, 2024), and the Vilnius Pink Soup Festival offers a telling example of how online resources enhance the sustainability of cultural tourism through integrated marketing.

International tourist arrivals returned to roughly pre-pandemic levels in 2024 (UNWTO, 2025), and this recovery has intensified competition among destinations for visitor attention, placing a premium on cost-effective promotion for the small, locally organized events that make up much of the cultural-tourism offer. The Vilnius Pink Soup Festival is one such event: it draws more than 93,000 attendees to a single Lithuanian city over two days (Go Vilnius), a scale that is negligible in global terms but substantial for a municipal cultural budget, and that depends almost entirely on digital channels rather than paid international advertising to reach its audience. By foregrounding traditional Lithuanian cuisine and sustainable food practices, it offers a distinctive cultural experience, supports local agriculture, and promotes environmental responsibility. As Chong and Balasingam (2019) observe, such initiatives align with the wider integration of sustainability into tourism, preserving cultural heritage for future generations while benefiting the local economy.

Tourist satisfaction, as a post-consumption evaluation of a product or service, captures the visitor’s overall attitude toward the holiday experience (Amissah et al., 2021), and this evaluation is central to how tourists choose among services. Festival satisfaction is commonly measured through social media, festival expectations, website quality, online word of mouth, and usability. Most existing work focuses on how technology raises customer satisfaction by improving the quality of tourism services at cultural destinations (Amissah et al., 2021; M. K. Rahman et al., 2022), while value creation, particularly through sustainable practices, has also emerged as a determinant of a festival’s long-term sustainability (Li et al., 2021). Tourism information centers and government organizations play an important part in planning and promoting festivals that serve the cultural needs of diaspora communities while allowing residents and visitors alike to benefit from a destination’s heritage (Hassanli et al., 2021).

Like any cultural activity, the festival uses social media and other websites to build community and connect visitors with local culture. Many festival visitors now incorporate social media into their festival experience, which both enriches their engagement and creates potential economic benefits for organizers.

1.1.2. The Role of Social Media as a Contributing Factor

The internet and social media have proven to be among the most effective means of shaping the visibility and brand awareness of cultural events (M. K. Rahman et al., 2022), and in the twenty-first century, social media has become a powerful, everyday instrument for marketing cultural events and tourism organizations (He, 2022). Because these platforms reach wide audiences without high promotional costs, they offer low-cost access to large audiences; evidence from other settings suggests that such earned reach can support attendance at lower promotional cost (Trusov et al., 2009), although no revenue data are examined here. Destination management organizations (DMOs) and tourism information centers now rely heavily on these platforms to market cities, regions, and national events and to communicate festival messages, since social media offers a space for engagement that conventional communication cannot match, making it easier to share ideas, opinions, and creative content and to reach audiences directly (Utami et al., 2024).

Studies by Chen et al. (2022) confirm that tourists increasingly turn to social media to meet their information needs, easing the burden on traditional tourist information centers. Like all cultural festivals, the Vilnius Pink Soup Festival faces the challenge of placing its promotional material before potential visitors, and social media has become an increasingly important channel for advertising tourism destinations and festivals tied to regional culture. Most marketers now use social media and spend more each year on advertising through these platforms (Aripin et al., 2023); the global number of users is projected to reach 5.3 billion by 2025, about 64.7% of the world’s population. Social media has thus become the most effective means for tourist destinations to engage their audiences and promote cultural activities (Arasli et al., 2021), allowing organizers to connect directly with potential visitors. To build a strong online presence, festival organizers must understand why and how tourists take part in social media discussions and how this participation can be used to raise the visibility of events such as the Vilnius Pink Soup Festival.

The literature identifies two drivers of social media adoption among tourism consumers. First, the nature of tourism leads potential travelers to rely heavily on others’ experiences on social media when deciding (Arasli et al., 2021), and many treat the information shared in these online communities as being as credible as advice from family and friends (Khan et al., 2021; Arasli et al., 2021). Second, A. Rahman et al. (2023) note that social media lets tourists share their experiences continuously, around the clock, and this constant sharing produces a sense of belonging and enjoyment that draws tourists further into online communities and heightens the appeal of destinations such as the Vilnius Pink Soup Festival through authentic user-generated content.

Beyond connecting friends and relatives, social media also enables interaction between organizers and consumers at tourist destinations (Gulati, 2024). By drawing on user-generated content, organizers can promote sustainable festivals, attract tourists, and encourage both repeat visitation and wider festival promotion. This leads to the following hypothesis:

H1. 

Social media has a positive and significant impact on tourists’ satisfaction with sustainable cultural events.

1.1.3. Festival Expectations

For cultural events such as the Vilnius Pink Soup Festival, festival expectations set the standard for tourism quality and shape the provision of products and services. High-quality activities determine the value of an experience because they influence visitors’ attitudes and behavior toward the destination (Utami et al., 2024). Festival expectations reflect the quality or standard that visitors anticipate (Gündüz et al., 2024) and indicate how well a product or service performs against those imagined standards (Borges et al., 2021). When festival offerings meet or exceed expectations, visitors tend to share positive feedback online, stimulating demand and participation, so that alignment with expectations may support revenue indirectly by encouraging repeat visitation (Getz, 2002; Andersson & Getz, 2009).

Rodríguez-Campo et al. (2022) argue that festival expectations form early, so that a high-quality festival raises visitor satisfaction, and that expectations and perceptions can broaden the diversity of festival offerings, leading tourists to develop positive attitudes toward the event and the destination. In the case of the Vilnius Pink Soup Festival, organizers need to establish the quality of their products and services on social media and public websites, which raises festival quality and visitor expectations, which prior work links to attendance (Yoon et al., 2010). Perceived quality rests on standards and expectations, and prioritizing these can strengthen the festival’s reputation and visitors’ intention to return (Garg & Modi, 2025).

Meeting consumer expectations is the key to any festival’s success, though much depends on how visitors evaluate their experience. High standards are needed for a cultural event to develop, grow in popularity, and leave a lasting positive impression on tourists and residents, and continuous fulfillment of expectations is the central requirement for success (Borges et al., 2021). Enhancing festival quality and shaping visitor expectations and perceptions can widen visitation and revenue, offering a competitive advantage to the destination (Akhundova, 2024). Yoon et al. (2010) identify several components of festival quality, including interaction quality, physical quality, outcome quality, access quality, and promotion quality, all of which form visitor expectations, and how visitors experience these components is central to effective festival management. The satisfaction of festival tourists directly affects an event’s future success, and for cultural events, festival expectations must meet visitors’ standards and secure the event’s long-term survival.

Overall, meeting festival expectations produces a positive emotional experience that can spur further engagement and revisitation (Borges et al., 2021; Akhundova, 2024; Garg & Modi, 2025). At high-quality festivals, expectations should be satisfied enough for visitors to spread positive word of mouth, whereas poor quality invites negative word of mouth, a distinction that carries particular weight for festivals designed to attract both residents and tourists. These considerations lead to the following hypothesis:

H2. 

Festival expectations have a positive and significant impact on tourists’ satisfaction with sustainable cultural events.

1.1.4. Information Communication Technology Usability

Rapid technological change has reshaped many sectors in recent years, including tourism, redefining the visitor experience through more interactive and personalized engagement. Tourism now depends heavily on technology, with specialized systems such as self-check-in kiosks, travel-management platforms, and in-room entertainment, alongside general tools such as mobile applications and agency websites, all contributing to the delivery and enhancement of tourism services (Slivar et al., 2019).

Greater global connectivity and the spread of digital technologies now allow consumers to access information and complete transactions across cultural, linguistic, and national boundaries (Saridakis et al., 2024). Mobile technology, and smartphones in particular, is central to this shift: no longer mere communication devices, smartphones have become essential travel companions that offer navigation, travel advice, payment, and language assistance (Fatema et al., 2024), delivering multiple services on a single platform through built-in software environments. As smartphone ownership grows, tourists increasingly use their devices as personal guides, which underlines the importance of mobile technology in tourism and the need for destinations and service providers to make full use of it (Srinivasan et al., 2024). By integrating mobile solutions, event organizers can offer a smoother, richer experience that adds convenience and interactivity for visitors.

Mobile technology plays an important part in trip planning, making it easier for tourists to organize their travel and enjoy the experience (Vujko et al., 2025). With accommodation bookings, transport reservations, maps, and other services available directly on their smartphones, tourists have what they need at hand, removing the need to queue at ticket offices or ask for directions and allowing for a more convenient, less stressful journey.

Easy-to-use digital platforms improve efficiency in festival planning and information delivery while reducing administrative and distribution costs. By streamlining communication and purchasing, ICT systems may reduce administrative and distribution costs, although such cost effects are not measured in the present study. Supporting both the planning and the navigation of travel, mobile technology widens access to tourism and encourages tourists to undertake journeys, and features that enable destination discovery and review-based decisions help travelers make informed choices and avoid poorly rated events. Mobile technology thus allows tourists to optimize their trips and immerse themselves in the experience without logistical worry, attracting more visitors and enabling fuller engagement with the destination.

Regarding this, we propose the following hypothesis:

H3. 

Information communication technology usability has a positive and significant impact on tourists’ satisfaction with sustainable cultural events.

1.1.5. Public Website Quality

Darwish and Lakhtaria (2011) describe perceived website quality as a customer’s overall judgment of a website’s excellence and efficiency. The quality of a public website strongly affects consumer satisfaction, fosters loyalty, and shapes the intention to return, and it also supports electronic word of mouth, which improves the success of online destination marketing (A. Rahman et al., 2023). In leisure and tourism management, the need to evaluate website quality, particularly for public websites, is widely accepted (Vîlcea et al., 2024). Web quality is usually measured through core notions such as service, system, and content quality, or through the usability and influence of the site (Liu et al., 2024), and the concept derives from the adaptation of established service-quality models such as SERVQUAL, WebQual, and E-S-Qual. In tourism, the emphasis falls on how consumers use digital channels and technologies, which increasingly complement direct engagement.

For cultural festivals such as the Vilnius Pink Soup Festival, a clear, well-designed public website is central to conveying essential information, attracting visitors, and improving their experience. By ensuring high website quality, organizers can raise festival awareness, drive online engagement, and increase both attendance and satisfaction (Liu et al., 2024). The Lithuanian government has responded positively to internet-based solutions for destination marketing, including for cultural events such as the Vilnius Pink Soup Festival, and by combining social media with data-processing techniques, these systems give tourists access to valuable information and strengthen their engagement with the destination (Slivar et al., 2019; A. Rahman et al., 2023). Artificial-intelligence technologies further support informed decisions by offering personalized recommendations (Yuan & Vui, 2023).

Maintaining strong, consistent communication with online consumers is critical for organizers. The experiential nature of tourism, together with the rapid growth of online networks, has created an environment in which travelers increasingly base destination choices on shared experiences (Huang et al., 2024). As more tourists engage with travel websites and social media, they generate distinctive content that influences others’ decisions (Thelen & Kim, 2024). Much attention has focused on the impact of user-generated content, particularly electronic word of mouth (eWOM), yet its effect on tourism decision-making in culture-based events still merits study, since the content tourists share shapes perceptions, attracts new visitors, and ultimately contributes to a festival’s sustainability and success.

A high-quality website is therefore essential to a good consumer experience: easy to navigate, convenient, and accessible across devices (Garg & Modi, 2025). Such websites provide essential festival information and incorporate user-generated content through tagging, sharing, and social media engagement on platforms such as Facebook, Instagram, and TikTok (Liu et al., 2024), which raises the festival’s online visibility and enables attendees to take part in its promotion by sharing their experiences. Organizers likewise recognize the importance of adding value to visitors’ experiences so that cultural festivals are not only enjoyable but memorable (Wood et al., 2024). Huang et al. (2024) stress that festival destinations must provide high-quality facilities and services to create a lasting experience essential to visitor satisfaction, and Semenda et al. (2024) add that a key challenge is aligning social media promotion with festival marketing strategy, noting that website quality plays a significant, positive role in shaping consumer decisions, building trust, and creating perceived value. A high-quality public website supports the conversion of online interest into attendance; the present study does not measure conversion or expenditure, and this link is therefore not tested here. These considerations lead to the following hypothesis:

H4. 

Public website quality has a positive and significant impact on tourists’ satisfaction with sustainable cultural events.

1.1.6. Electronic Word of Mouth

Since the early 2000s, a growing number of event organizers, developers, and marketers have recognized electronic word of mouth (eWOM) as a powerful tool for promoting cultural events (Dowell et al., 2019). Through eWOM, organizations have built strong connections with prospective tourists using cost-efficient, time-efficient marketing for festival destinations, and websites have become a dynamic medium for eWOM, spreading information not only to potential consumers but across their broader social networks (Bui et al., 2025). eWOM has thus become a vital element in shaping perceptions, marketing destinations, and encouraging participation in festivals such as the Vilnius Pink Soup Festival.

eWOM, or online word of mouth, spreads to large numbers of consumers through social platforms (Bui et al., 2025). It is spontaneous communication via the internet that provides feedback on the quality of goods and services to friends, family, and wider audiences (Tirendi & Gargiulo, 2024). As consumers and businesses make greater use of social media, understanding its effect on consumption behavior becomes more important than ever, and it is especially valuable to examine how eWOM on these platforms influences decisions such as the intention to revisit (Dowell et al., 2019).

A key advantage of electronic word of mouth (eWOM) is its capacity to reach a large audience with minimal resources while remaining highly effective in promoting tourism destinations (M. Rahman & Mia, 2025). M. Rahman and Mia (2025) emphasize that eWOM is particularly effective within social networks, whose large followings give it wide reach, making it a powerful tool for promoting cultural events and shaping public perception and visitation. As promoters of the Vilnius Pink Soup Festival increasingly build social media into their marketing, it becomes more important to examine what drives consumer engagement in eWOM through social platforms. Several researchers (Dowell et al., 2019; Bui et al., 2025; Tirendi & Gargiulo, 2024) identify eWOM as a significant antecedent of tourist behavior, particularly for destination choice and satisfaction, and in the festival context, tourist behavior is heavily shaped by user-generated content, with many visitors planning their attendance around the experiences of others.

Unlike traditional advertising, eWOM is a form of unpaid promotion that gives cultural events access to vast networks of prospective visitors at no cost, reducing dependence on paid media. Evidence from other settings indicates that customers acquired through word of mouth are more valuable and longer-lived than those acquired through paid channels (Villanueva et al., 2008; Trusov et al., 2009; Stephen & Galak, 2012). Tourism information centers and event managers engaged in creative tourism marketing share responsibility for shaping the festival experience and building a strong brand identity (Khan et al., 2021). Word of mouth, particularly through social media and official websites, has proven a highly influential source of information for prospective tourists, often prompting them to book travel on the basis of past attendees’ experiences and recommendations (Tirendi & Gargiulo, 2024). Pratama and Astarini (2023) further argue that eWOM can generate referrals on platforms such as TikTok and Instagram, allowing events like the Vilnius Pink Soup Festival to reach a wide audience. From the visitor’s perspective, negative eWOM can seriously damage a destination’s appeal, since untrustworthy or critical content makes consumers less willing to take the perceived risk of attending. This leads to the following hypothesis:

H5. 

Electronic word of mouth (eWOM) has a positive and significant impact on tourists’ satisfaction with sustainable cultural events.

1.1.7. Consumer Satisfaction and Revisit Intention

Satisfaction is one of the most important factors in the marketing of tourism and cultural festivals, where public and private organizations at every geographical level use festivals and events to reshape and promote cultural identities (Lopes & Hiray, 2024). As Akhundova (2024) notes, satisfaction is central to cultural-festival and tourism marketing, in which institutions across geographical scales draw on festivals to recreate and promote cultural values. In tourism and festival marketing, satisfaction is commonly defined as the consumer’s overall response, a post-purchase evaluation of whether a service met, exceeded, or fell short of expectations (Torabi et al., 2022). At the Vilnius Pink Soup Festival, visitor satisfaction is typically assessed through overall evaluations of the event, with key quality factors such as the experience, food, organization, and atmosphere taken into account. Singh et al. (2023) argue that satisfaction is a fundamental component of promotional effort and that the ability to measure and manage it is essential for survival and growth.

Positive experiences at cultural events can substantially raise tourist satisfaction, with lasting effects on individuals’ well-being and future attitudes (Su et al., 2023). Satisfaction is also shaped by factors such as event quality, pricing, and the interaction between them (Yoon et al., 2010), when offerings such as traditional cuisine, cultural performances, sustainability features, and related services are seen as desirable, tourists are more likely to judge them high in quality and to intend to return (Rasoolimanesh et al., 2025).

In preparing this research, we identified specific factors that significantly influence tourists’ intentions to return to festivals. On this basis, we propose the following:

H6. 

Consumer satisfaction has a positive and significant impact on revisit intention with sustainable cultural events.

H7. 

Festival satisfaction mediates the relationships between social media (H7a), festival expectations (H7b), ICT usability (H7c), public website quality (H7d), electronic word of mouth (H7e) and revisit intention.

Because the design is cross-sectional, H6 and H7 are tested as associations rather than as causal effects, and the mediation hypothesis is examined against the alternative that the antecedents relate to revisit intention directly.

1.1.8. Scope of the Model and Level of Analysis

The hypotheses above are specified at the demand side of the festival’s financial system. Marketing-productivity research establishes that promotional spending reaches financial performance indirectly, through intermediate customer responses—satisfaction, attitude and loyalty—which then translate into attendance, repeat visitation and revenue (Srinivasan & Hanssens, 2009; Katsikeas et al., 2016; Hanssens & Pauwels, 2016). The present study models the first two links in that chain and does not test the third. No cost, revenue, expenditure or budget data were collected, and consequently, no hypothesis here concerns marketing return on investment, promotional cost efficiency or realized festival income. What the model can establish is which forms of digital engagement are most strongly associated with the visitor responses that prior research identifies as the antecedents of festival revenue; the financial implications drawn in the Discussion are therefore interpretive extensions of these behavioural findings rather than tested propositions. Testing the final link would require organizer-side accounting data, which we identify as a priority for future research.

As shown in Figure 1, the theoretical model emphasizes the influence of several key factors, including social media, festival expectations, information and communication technology usability, public website quality, and electronic word of mouth, on consumer satisfaction.

Solid lines indicate direct effects (H1–H6); dashed lines indicate indirect/mediated effects on revisiting intention via consumer satisfaction (H7a–H7e). Each of these factors, represented by H1 through H5, is hypothesized to affect consumer satisfaction directly, which in turn influences the intention to revisit (H6). Consumer satisfaction serves as the central mediating variable, linking the initial predictors to revisit intention. By mapping these relationships, the model captures how digital engagement and consumer perceptions shape the likelihood that tourists will revisit a festival.

2. Materials and Methods

This study focuses on the Vilnius Pink Soup Festival, an annual event held in Vilnius, Lithuania. It adopts a quantitative approach centered on the satisfaction and revisit intention of local and international tourists in response to digital engagement and festival evaluation. Although the study concentrates on a single case, it offers lessons relevant to other small-scale cultural events, and for emerging cultural-tourism destinations, such case studies show how digital sustainability can be pursued at the community level.

Data were collected after the festival through a structured questionnaire designed to capture overall visitor satisfaction. The questionnaire had two sections: demographic questions

Table 1. Demographic characteristics.

Demographic Variables		Frequencies	Percentage
Gender	Male	180	46.9
	Female	203	52.9
	Other	1	0.3
Age	18–29	279	72.7
	30–42	77	20.1
	42 and more	28	7.3

and construct-specific questions drawn from the literature. A 5-point Likert scale ranging from “strongly disagree” (1) to “strongly agree” (5) measured perceptions of social media, expectations, technology use, website quality, eWOM, and consumer satisfaction.

Participants were reached through publicly accessible digital channels, including festival-related platforms and social media. To ensure the relevance and quality of responses, participants had to meet two inclusion criteria: they had to have attended the Vilnius Pink Soup Festival and they had to be familiar with promotional activity for the event on social media or other online platforms. The questionnaire was freely accessible online and, to screen out random or ineligible responses, included a control question about attendance. Festival volunteers and participants supported distribution by sharing the questionnaire online. The study used convenience sampling, which suits exploratory research during short-term festivals where random sampling is not feasible (Amissah et al., 2021). Because no visitor list was available for the more than 93,000 attendees, Cochran’s formula for large populations was used as a rule of thumb to establish a target sample size, and the achieved sample (n = 384) meets that target. Since recruitment was by convenience rather than probability sampling, this calculation indicates only that the sample is large enough for the analyses performed; it does not establish that the sample is representative of the festival’s visitors, and the results should not be read as generalizable to the full attendee population on that basis. The response rate could not be calculated, since the questionnaire was openly distributed and the number of people who saw it is unknown. The composition of the achieved sample is described below and compared with what is known about the festival’s audience.

Convenience sampling, a form of non-probability sampling, was chosen given the annual, short-duration nature of the festival. The method involves selecting accessible and willing participants, which was appropriate for gathering insights from attendees within the festival’s limited timeframe.

Established and validated measurement scales were used for each construct, all adapted to the festival context and measured on a 5-point Likert scale. Each construct was measured with four items, giving 28 items in total. Social media engagement was adapted from Primananda et al. (2022), festival expectations and ICT usability from Pai et al. (2025), public website quality from Almakayeel (2023), electronic word of mouth and revisit intention from Azhar et al. (2022) and festival satisfaction from Molina-Gómez et al. (2021). The full wording of all items, together with the original scale from which each was adapted and the wording of the original item, is provided in Appendix A. No items were removed during analysis; all 28 items administered were retained, and the item numbering in

Table 2. Reliability and validity of measurement scales.

Constructs and Items	Factor Loading	AVE	CR	Cronbach’s Alpha	FVIF
Social media (SM)		0.498	0.798	0.799	2.170
SM1	0.643
SM2	0.749
SM3	0.662
SM4	0.759
Festival expectations (FE)		0.588	0.850	0.849	2.129
FE1	0.661
FE2	0.753
FE3	0.807
FE4	0.835
Information and communication technology usability (ICTU)		0.563	0.837	0.836	1.995
ICTU1	0.715
ICTU2	0.707
ICTU3	0.755
ICTU4	0.819
Public website quality (PWebQ)		0.584	0.849	0.849	1.855
PWebQ1	0.748
PWebQ2	0.743
PWebQ3	0.779
PWebQ4	0.786
Electronic word of mouth (eWOM)		0.439	0.757	0.758	2.587
eWOM1	0.649
eWOM2	0.656
eWOM3	0.684
eWOM4	0.660
Festival satisfaction (FS)		0.432	0.752	0.749	2.084
FS1	0.640
FS2	0.585
FS3	0.700
FS4	0.697
Revisit intention (RI)		0.367	0.698	0.696	2.361
RI1	0.575
RI2	0.587
RI3	0.598
RI4	0.661

Note: Factor loadings are reported at the item level; AVE, CR, Cronbach’s alpha and FVIF are construct-level statistics and are therefore shown only in the construct summary rows. Cells left blank are not missing values. The reliability and validity results show that the loadings of all items for social media (SM), festival expectations (FE), information and communication technology usability (ICTU), public website quality (PWebQ), electronic word of mouth (eWOM), festival satisfaction (FS), and revisit intention (RI) exceeded the 0.5 threshold, with statistically significant p-values, indicating good convergent validity.

runs consecutively within each construct.

All analyses were conducted in IBM SPSS Statistics 29. Descriptive statistics and Pearson correlations were computed in SPSS.

The measurement properties of the seven scales were examined in three steps. First, sampling adequacy for the 28-item correlation matrix was confirmed by the Kaiser–Meyer–Olkin measure (KMO = 0.947; item-level values ranging from 0.920 to 0.968) and by Bartlett’s test of sphericity (χ²(378) = 4882.07, p < 0.001). Second, because all seven constructs were measured with previously validated scales, and because the study does not aim to discover a factor structure, unidimensionality was assessed by extracting a single common factor within each construct separately (principal axis factoring, one factor, no rotation) rather than through a joint exploratory solution across all items. Every item loaded on its intended construct, with loadings ranging from 0.568 to 0.835, and the proportion of item variance accounted for by the single factor ranged from 36.8% for revisit intention to 58.9% for festival expectations; these loadings are reported in Table 2. Two supplementary two-factor exploratory factor analyses were additionally run to examine the separation of social media from eWOM and of satisfaction from revisit intention, using the same extraction settings. Third, internal consistency was assessed with Cronbach’s alpha, composite reliability (CR) and average variance extracted (AVE) were computed from these loadings using the formulae of Fornell and Larcker (1981), and discriminant validity was assessed with the Fornell–Larcker criterion and with the heterotrait–monotrait ratio of correlations (HTMT), the latter computed from the item-level Pearson correlation matrix following Henseler et al. (2015), with 95% percentile bootstrap confidence intervals based on 5000 resamples.

CR, AVE, the Fornell–Larcker criterion and HTMT originate in the CFA and PLS-SEM traditions. They are reported here as descriptive evidence on the adequacy of the composite scores that enter the regression models, not as the output of an estimated measurement model, and no measurement model was estimated.

The structural relationships were estimated by ordinary least squares (OLS) regression on standardized composite scores, each construct being represented by the unweighted mean of its items. Model 1 regressed festival satisfaction on the five antecedents with age and gender entered as controls (H1–H5); Model 2 regressed revisit intention on festival satisfaction (H6). Multicollinearity was assessed through variance inflation factors, and common method bias through Harman’s single-factor test (Podsakoff et al., 2003) and the full collinearity test proposed by Kock (2015), the latter adapted from its original PLS-SEM context to the present composite-score regression setting.

The mediating role of festival satisfaction (H7a–H7e) was tested with the PROCESS macro for SPSS (Hayes, 2022), Model 4, using 5000 bootstrap resamples and percentile confidence intervals. Because Model 4 accommodates a single focal predictor, five separate analyses were estimated: in each, one antecedent served as the focal predictor (X), festival satisfaction as the mediator (M), revisit intention as the outcome (Y), and the remaining four antecedents together with age and gender were entered as covariates. This yields a partial mediation specification in which direct paths from all five antecedents to revisit intention are estimated simultaneously; the direct effects (c′) so obtained are numerically identical to those from a single OLS regression of revisit intention on the five antecedents, festival satisfaction and the controls, and the b path from satisfaction to revisit intention is common to all five analyses.

No latent-variable structural equation model was estimated. Neither covariance-based SEM nor partial least squares SEM was used; consequently, no confirmatory factor analysis is reported, and no SEM fit indices (χ², CFI, TLI, RMSEA, SRMR) or PLS-SEM criteria such as f², Q² or the saturated-model SRMR are reported. The analysis is a path-analytic regression model, and this term is used consistently throughout the manuscript.

A total of 384 participants were recorded. Of these, 180 identified as male, 203 as female, and one as non-binary or another gender category, indicating a slight predominance of female respondents (52.9%) over male respondents (46.9%). Most participants (n = 279; 72.7%) fell within the 18–29 age group, suggesting strong engagement of younger visitors with the festival and with the digital platforms used to promote it. Participants aged 30–42 numbered 77 (20.1%), while those aged 42 and above formed the smallest group, at 28 (7.3%). This distribution is markedly younger than the general population and probably younger than the festival’s audience, although no official visitor demographics are published for the event, so the two cannot be compared directly. Two features of the design are likely to have contributed: recruitment took place mainly through digital channels, and eligibility required familiarity with the festival’s online promotion. Both favour younger, digitally engaged attendees. The findings should therefore be read as describing this segment in particular rather than festival-goers generally, which matters for a study whose predictors are themselves forms of digital engagement.

As shown in Table 2, Cronbach’s alpha exceeded 0.70 for six of the seven constructs (SM 0.799, FE 0.849, ICTU 0.836, PWebQ 0.849, eWOM 0.758, FS 0.749), with revisit intention marginally below at 0.696, and composite reliability followed the same pattern (0.698 to 0.850). The average variance extracted exceeded the 0.50 threshold for festival expectations (0.588), public website quality (0.584) and ICT usability (0.563), and fell just short for social media (0.498). For eWOM (0.439), festival satisfaction (0.432) and revisit intention (0.367) the AVE was below 0.50. Fornell and Larcker (1981) note that convergent validity may still be regarded as adequate where AVE falls below 0.50 provided composite reliability exceeds 0.60, a condition met by all seven constructs here. The lower AVE values for these three constructs reflect the restricted response variance evident in

Table 3. Descriptive Statistics of Study Constructs.

Construct	Average (M)	Standard Deviation (SD)
Social media	4.15	0.46
Festival expectations	3.80	0.85
Information and communication technology usability (ICTU)	3.65	0.72
Public website quality	3.33	1.02
Electronic word of mouth (eWOM)	4.56	0.38
Festival satisfaction	4.86	0.26
Revisit intention	4.43	0.36

, where satisfaction (M = 4.86, SD = 0.26), eWOM (M = 4.56, SD = 0.38) and revisit intention (M = 4.43, SD = 0.36) all cluster near the upper end of the scale; such ceiling effects attenuate item covariances and therefore depress extracted variance. This is acknowledged as a limitation. All full collinearity variance inflation factors were below the conservative 3.3 threshold (Kock & Lynn, 2012).

3. Results

Table 3 presents the descriptive statistics. Among the seven constructs, festival satisfaction recorded the highest mean (M = 4.86, SD = 0.26), indicating that visitors were highly satisfied with their experience. Electronic word of mouth (M = 4.56) and revisit intention (M = 4.43) were also rated highly, suggesting that satisfied attendees are inclined both to recommend the festival online and to return, and social media (M = 4.15) was likewise viewed positively, confirming its role as an important channel for engaging audiences. Festival expectations (M = 3.80) and ICT usability (M = 3.65) received more moderate scores with larger standard deviations, pointing to mixed experiences regarding expectations and the ease of using digital tools.

Public website quality received the lowest mean score (M = 3.33, SD = 1.02), indicating that visitors saw the website as less effective than other aspects of the festival experience. This points to a gap in the festival’s promotional strategy, since public websites are expected to provide reliable information, enhance usability, and support positive experiences. The comparatively low and variable ratings suggest scope to improve design, accessibility, and content, and strengthening the website could improve ICT usability, help meet or exceed expectations, and ultimately raise satisfaction, positive eWOM, and revisit intentions, supporting the long-term sustainability and success of cultural events such as the Vilnius Pink Soup Festival.

Three constructs show a pronounced ceiling effect. Festival satisfaction is the most affected: 71.1% of respondents recorded the scale maximum, no respondent scored below 4.0, and the distribution is strongly left-skewed (skewness = −1.91). Electronic word of mouth (30.5% at the maximum, minimum 4.0) and revisit intention (16.4% at the maximum, minimum 3.5) are similarly compressed, whereas social media, festival expectations, ICT usability and public website quality retain their full response range.

This restriction has consequences that should be stated plainly. The compression is present at the item level as well as at the construct level: between 84.6% and 88.0% of respondents selected the maximum on each satisfaction item, between 53.6% and 58.6% on each eWOM item, and between 39.8% and 45.8% on each revisit-intention item. The three constructs have standard deviations of 0.26, 0.38 and 0.36 respectively, against 1.02 for public website quality.

Three consequences follow. First, attenuated variance lowers the covariance available for estimation, so the coefficients involving these three constructs are conservative rather than inflated, and the explained variance in revisit intention in the satisfaction-only model (R² = 0.256) is depressed for the same reason. Second, the compression depresses the average variance extracted for these constructs, as reported in Table 2, which is the mechanical reason the Fornell–Larcker criterion is not met for eWOM and revisit intention; it also inflates the heterotrait–monotrait ratio, since that ratio corrects observed correlations for attenuation. The discriminant-validity difficulties reported below are therefore partly a property of these response distributions rather than of the constructs themselves. Third, a measure that discriminates poorly at the upper end cannot order respondents reliably within a near-universally satisfied sample, which bears directly on how the mediation result should be read.

Three explanations are plausible and cannot be separated from the present data. The first is substantive: the festival is a free, single-day community event with a strong local following, and near-universal satisfaction among self-selected attendees is not implausible. The second is social desirability, since respondents evaluated an event they had chosen to attend, using an instrument distributed in part through the festival’s own channels. The third is sample selection, since the inclusion criteria favoured visitors already engaged with the festival’s online promotion, as discussed in the limitations. The most likely reading is that all three contribute. Whichever dominates, the practical implication is the same: the satisfaction measure discriminates poorly at the upper end, and the estimates involving it should be treated as a lower bound.

Table 4. Correlation analysis.

	SM	FE	ICTU	PWebQ	eWOM	FS	RI
Social media (SM)	1	0.52 *	0.42 *	0.33 *	0.67 *	0.55 *	0.61 *
Festival expectations (FE)		1	0.46 *	0.42 *	0.59 *	0.62 *	0.62 *
Information and communication technology usability (ICTU)			1	0.62 *	0.56 *	0.50 *	0.51 *
Public website quality (PWebQ)				1	0.46 *	0.48 *	0.50 *
Electronic word of mouth (eWOM)					1	0.59 *	0.65 *
Festival satisfaction (FS)						1	0.51 *
Revisit intention (RI)							1

* Correlation is significant at the 0.001 level (2-tailed).

reports the correlation analysis among the study’s main constructs. Social media, festival expectations, ICT usability, public website quality, eWOM, festival satisfaction, and revisit intention are all positively and significantly correlated, indicating that these factors are interrelated and jointly contribute to visitor satisfaction and the intention to revisit.

The strongest associations are those between social media and electronic word of mouth (r = 0.67) and between electronic word of mouth and revisit intention (r = 0.65), indicating that the two earned channels operate together and are closely tied to the intention to return. Public website quality correlates positively with satisfaction (r = 0.48) and revisit intention (r = 0.50) but less strongly than social media and eWOM, and its high correlation with ICT usability (r = 0.62) suggests that the two capture overlapping aspects of the festival’s digital infrastructure. Discriminant validity was assessed using the Fornell–Larcker criterion in

Table 5. Fornell–Larcker criterion.

	SM	FE	ICTU	PWebQ	eWOM	FS	RI
Social media (SM)	0.706
Festival expectations (FE)	0.522	0.767
Information and communication technology usability (ICTU)	0.416	0.464	0.750
Public website quality (PWebQ)	0.330	0.418	0.624	0.764
Electronic word of mouth (eWOM)	0.674	0.591	0.562	0.457	0.663
Festival satisfaction (FS)	0.553	0.622	0.501	0.483	0.587	0.657
Revisit intention (RI)	0.613	0.624	0.508	0.503	0.646	0.506	0.606

Diagonal elements (bold) are the square roots of the average variance extracted; off-diagonal elements are construct correlations.

.

Discriminant validity was assessed with two criteria in

Table 6. Heterotrait–monotrait ratio (HTMT).

	SM	FE	ICTU	PWebQ	eWOM	FS
Festival expectations (FE)	0.632 [0.538, 0.714]
Information and communication technology usability (ICTU)	0.508 [0.403, 0.607]	0.550 [0.450, 0.640]
Public website quality (PWebQ)	0.398 [0.290, 0.503]	0.492 [0.389, 0.582]	0.740 [0.669, 0.803]
Electronic word of mouth (eWOM)	0.866 [0.809, 0.923]	0.736 [0.658, 0.809]	0.707 [0.629, 0.781]	0.570 [0.469, 0.663]
Festival satisfaction (FS)	0.712 [0.612, 0.797]	0.779 [0.695, 0.856]	0.633 [0.537, 0.721]	0.604 [0.511, 0.687]	0.779 [0.695, 0.860]
Revisit intention (RI)	0.821 [0.748, 0.895]	0.811 [0.747, 0.876]	0.666 [0.575, 0.756]	0.655 [0.564, 0.740]	0.889 [0.816, 0.964]	0.698 [0.612, 0.786]

Note: Values in brackets are 95% percentile bootstrap confidence intervals based on 5000 resamples.

. Under the Fornell–Larcker criterion (Fornell & Larcker, 1981), the square root of the AVE exceeded all inter-construct correlations for social media, festival expectations, ICT usability, public website quality and festival satisfaction. For eWOM and revisit intention the criterion was not met: the square root of their AVE (0.663 and 0.606) fell below their correlations with social media (0.674) and with eWOM (0.646), respectively. Because the Fornell–Larcker criterion is directly sensitive to the level of AVE, and because the AVE for these two constructs is attenuated by the ceiling effects described above, the heterotrait–monotrait ratio was also examined, together with percentile bootstrap confidence intervals based on 5000 resamples, since a point estimate alone does not convey the precision with which the ratio is estimated (Henseler et al., 2015).

All 21 HTMT point estimates fell below the 0.90 threshold recommended for conceptually related constructs. The confidence intervals, however, distinguish two situations. For 19 of the 21 pairs the upper bound also lay below 0.90, and discriminant validity is supported on this criterion. This includes satisfaction and revisit intention (HTMT = 0.698, 95% CI [0.612, 0.786]), for which the separation is unambiguous. For two pairs involving eWOM—with social media (HTMT = 0.866, 95% CI [0.809, 0.923]) and with revisit intention (HTMT = 0.889, 95% CI [0.816, 0.964])—the upper bound exceeded 0.90. Discriminant validity for these two pairs is therefore not established at the conventional criterion, and we report it as a measurement limitation rather than as a resolved question.

Because the two elevated ratios involve eWOM, the separation of the constructs was examined directly in two supplementary exploratory factor analyses, one on the eight social media and eWOM items and one on the eight satisfaction and revisit-intention items. For satisfaction and revisit intention, two eigenvalues exceeded unity (3.303 and 1.115), the two-factor solution assigned every item to its intended construct with no cross-loading above 0.30, and the factors correlated at 0.652; the distinction between these constructs is therefore well supported.

For social media and eWOM, the evidence is weaker. Only one eigenvalue exceeded unity (4.033, the second being 0.832), although a forced two-factor solution did assign every item to its intended construct, with a single cross-loading of 0.303 (SM2) and a factor correlation of 0.773. Two considerations support retaining the constructs as separate. Conceptually, the social media items concern the informational usefulness of the festival’s own channels, whereas the eWOM items concern the respondent’s reliance on other visitors’ reviews; these are distinct behaviours even when they occur on the same platform. Empirically, the two constructs relate differently to the rest of the model: eWOM is more strongly associated with ICT usability (HTMT = 0.707 versus 0.508) and with public website quality (HTMT = 0.570 versus 0.398) than social media is, and the two exert different effects on satisfaction (β = 0.132 and β = 0.202). A single underlying dimension would not produce this divergence. We nonetheless regard the empirical separation of official-channel engagement from user-generated recommendation as insufficiently sharp in the present instrument and identify it as a priority for measurement development.

The hypothesized relationships were estimated by multiple regression on standardized construct scores, controlling for age and gender. Model adequacy was assessed through explained variance, the significance of the individual paths, and collinearity diagnostics; all variance inflation factors were well below conventional thresholds, indicating that the estimates are not distorted by multicollinearity.

Because all constructs were measured with a single self-report instrument administered to one respondent per observation, common method bias was assessed in two ways. First, Harman’s single-factor test was applied (Podsakoff et al., 2003): all 28 items were entered into an unrotated principal components analysis, and the first unrotated component explained 37.18% of the total variance, below the 50% threshold conventionally taken to indicate a dominant method factor. Second, the full collinearity test proposed by Kock (2015) was applied: all full collinearity variance inflation factors reported in Table 2 fell between 1.855 and 2.587, below the 3.3 threshold at which a model may be considered contaminated by common method variance. The two tests together suggest that common method bias is unlikely to account for the relationships reported here, although, as with any single-source cross-sectional design, it cannot be ruled out entirely.

In

Table 7. Hypothesis testing.

Hypothesis	Path	Estimate (β)	S.E.	t	p-Value (p)	Result
H1	SM → FS	0.202	0.050	4.06	<0.001	Supported
H2	FE → FS	0.332	0.047	7.11	<0.001	Supported
H3	ICTU → FS	0.091	0.051	1.80	=0.073	Not supported
H4	PWebQ → FS	0.162	0.047	3.44	<0.001	Supported
H5	eWOM → FS	0.132	0.056	2.35	=0.019	Supported
H6	FS → RI	0.506	0.044	11.48	<0.001	Supported
Age		0.023	0.059	0.39	0.697
Gender		0.043	0.072	0.60	0.548
Model 1 (DV: FS)		R2 = 0.518		adj. R2 = 0.509	F(7, 376) = 57.67, p < 0.001
Model 2 (DV: RI, satisfaction only)		R2 = 0.256		adj. R2 = 0.254	F(1, 382) = 131.68, p < 0.001

Note: Coefficients are standardized OLS regression estimates. Model 1 (DV: festival satisfaction) includes the five antecedents with age and gender entered as controls; Model 2 (DV: revisit intention) includes festival satisfaction as the sole predictor. SM = social media, FE = festival expectations, ICTU = information and communication technology usability, PWebQ = public website quality, eWOM = electronic word of mouth, FS = festival satisfaction, RI = revisit intention.

the hypothesized relationship between social media and festival satisfaction (H1) was positive and statistically significant (β = 0.202, p < 0.001), supporting H1 and indicating that greater engagement with the festival’s social media raises visitors’ satisfaction. Festival expectations (H2) emerged as the strongest predictor in the model (β = 0.332, p < 0.001), so H2 was supported: satisfaction depends above all on whether the event delivers the quality and content that visitors anticipated.

Public website quality (H4) was positive and significant (β = 0.162, p < 0.001) and electronic word of mouth (H5) was positive and significant at the conventional level (β = 0.132, p = 0.019), supporting both hypotheses. The path from ICT usability to satisfaction (H3) was positive but did not reach significance (β = 0.091, p = 0.073), so H3 was not supported. This result should be read in light of the strong correlation between ICT usability and public website quality (r = 0.62), which suggests that the two constructs capture overlapping aspects of the festival’s digital infrastructure; when both are entered simultaneously, the website-quality measure absorbs most of the shared variance. ICT usability is significantly correlated with satisfaction at the bivariate level (r = 0.50), so the non-significant coefficient indicates redundancy with website quality rather than irrelevance to the visitor experience.

The direct relationship between festival satisfaction and revisit intention (H6) was positive, strong and highly significant (β = 0.506, p < 0.001), supporting H6 and confirming that satisfied visitors are substantially more likely to return. As reported below, however, this association does not survive the inclusion of the five antecedents, and H6 should therefore be read as a bivariate rather than an independent relationship.

Taken together, the five antecedents explain 51.8% of the variance in festival satisfaction (adj. R² = 0.509). Satisfaction alone explains 25.6% of the variance in revisit intention (adj. R² = 0.254), while the five antecedents entered directly explain 57.5% (adj. R² = 0.567), as reported in

Table 8. Direct, indirect and total effects on revisit intention.

Path	a (X→FS)	Total Effect (c).	Direct Effect (c′)	Indirect Effect	95% CI	Mediation
SM → FS → RI	0.202 ***	0.251 ***	0.266 ***	−0.015	[−0.036, 0.002]	Not supported
FE → FS → RI	0.332 ***	0.274 ***	0.298 ***	−0.024	[−0.055, 0.004]	Not supported
ICTU → FS → RI	0.091	0.048	0.055	−0.007	[−0.022, 0.002]	Not supported
PWebQ → FS → RI	0.161 ***	0.183 ***	0.194 ***	−0.012	[−0.030, 0.002]	Not supported
eWOM → FS → RI	0.131 *	0.204 ***	0.214 ***	−0.010	[−0.025, 0.002]	Not supported

Note: Coefficients are standardized; age and gender were included as controls in all models. The b path from festival satisfaction to revisit intention, estimated with all five antecedents included, was β = −0.073 (SE = 0.048, t = −1.50, p = 0.133). Indirect effects were estimated with 5000 bootstrap resamples and percentile confidence intervals; all intervals include zero. Total effects model (DV: RI, five antecedents and controls): R² = 0.575, adj. R² = 0.567, F(7, 376) = 72.54, p < 0.001. Direct effects model (DV: RI, five antecedents, satisfaction and controls): R² = 0.577, adj. R² = 0.568, F(8, 375) = 63.97, p < 0.001. * p < 0.05, *** p < 0.001.

. The ordering of the significant coefficients is notable: the strongest determinant of satisfaction is not a promotional channel but the extent to which the festival meets prior expectations, with the digital channels contributing significantly but more modestly. This pattern suggests that digital engagement operates on satisfaction primarily by shaping and then fulfilling expectations, rather than as an independent source of value.

Festival satisfaction was then examined as a potential mediator of the relationships between the five antecedents and revisit intention, with direct paths from each antecedent to revisit intention estimated simultaneously.

Table 8 shows that festival satisfaction did not mediate any of the five relationships. With the antecedents included in the model, the path from satisfaction to revisit intention was small, negative and non-significant (β = −0.073, p = 0.133), and every bootstrapped indirect effect had a confidence interval spanning zero. H7a to H7e are therefore not supported.

The direct effects tell a different story. Four of the five antecedents predicted revisit intention directly and significantly: festival expectations (β = 0.298, p < 0.001), social media (β = 0.266, p < 0.001), electronic word of mouth (β = 0.214, p < 0.001) and public website quality (β = 0.194, p < 0.001), with ICT usability again not reaching significance (β = 0.055, p = 0.251). For each antecedent, the direct effect is close to the total effect, confirming that satisfaction accounts for none of the association. The five antecedents together explain 57.5% of the variance in revisit intention (total-effects model: R² = 0.575, adj. R² = 0.567), substantially more than the 25.6% explained by satisfaction alone. Adding satisfaction to the model raises this only marginally (R² = 0.577), which is consistent with the absence of an independent contribution from satisfaction.

This outcome should be read alongside the ceiling effect documented above. Satisfaction is correlated with revisit intention at the bivariate level (r = 0.506), and H6 was supported when the relationship was estimated on its own. Once the antecedents are controlled, however, satisfaction adds nothing, which indicates that the bivariate association reflects their shared antecedents rather than a distinct contribution of satisfaction. With 71.1% of respondents at the scale maximum and no observation below 4.0, the satisfaction measure retains too little variance to carry the effects of its antecedents, whatever its role might be in a sample with a wider response range.

4. Discussion

This study examines the use of public websites and other online platforms to promote sustainable cultural events, taking the Vilnius Pink Soup Festival as its case and building on research that shows the growing importance of social media and ICT in cultural-tourism promotion (e.g., Srinivasan et al., 2024; Chen et al., 2022). It contributes to the literature by empirically testing a path-analytic regression model of five determinants of festival satisfaction—social media, festival expectations, ICT usability, public website quality, and electronic word of mouth—and their onward effect on revisit intention. The contribution is empirical rather than theoretical: the study tests a model of digital engagement, visitor satisfaction and revisit intention, and does not advance or test Industry 5.0 as a theory. Five of the six hypotheses concerning direct relationships were supported: social media (H1), festival expectations (H2), public website quality (H4) and electronic word of mouth (H5) each raised satisfaction, and satisfaction in turn strongly predicted revisit intention (H6). The mediation hypotheses (H7a–H7e) were not supported. The path from ICT usability to satisfaction (H3) was positive but did not reach significance. Together the five antecedents accounted for 51.8% of the variance in satisfaction. These results speak to the visitor-side conditions that prior work associates with the social and cultural sustainability of festivals; community participation and cultural preservation were not measured here, and the link to them is interpretive rather than tested.

Festival expectations emerged as the strongest determinant of satisfaction (β = 0.332, p < 0.001), consistent with work emphasizing that the fulfilment of anticipated experience quality is the primary driver of visitor evaluation (Torabi et al., 2022; Borges et al., 2021; Rodríguez-Campo et al., 2022). Social media followed (β = 0.202, p < 0.001), confirming its role in shaping the visitor experience (Arasli et al., 2021; Gulati, 2024), while public website quality (β = 0.162, p < 0.001) and electronic word of mouth (β = 0.132, p = 0.019) contributed significantly but more modestly. These results echo findings on web marketing at global events such as the Notting Hill Carnival in the UK (Gugolati & Klien-Thomas, 2022) and food festivals in Turkey and Italy (Gündüz et al., 2024), where digital engagement helped sustain culture and encourage return visits. Two findings depart from expectations set by prior work. First, electronic word of mouth, which studies of tourism eWOM often identify as a dominant influence (M. Rahman & Mia, 2025; Azhar et al., 2022), was the weakest of the significant predictors of satisfaction in this sample, even though it correlates strongly with revisit intention at the bivariate level (r = 0.65). A plausible reading is that eWOM operates principally on attention, expectation formation and the decision to attend, rather than on the post-consumption evaluation captured by the satisfaction construct; its influence would then reach loyalty largely through the expectations it sets, which are themselves the strongest predictor here. Second, ICT usability did not reach significance (β = 0.091, p = 0.073) despite a substantial bivariate correlation with satisfaction (r = 0.50). Its strong association with public website quality (r = 0.62) indicates that the two constructs tap overlapping features of the same digital infrastructure, so the result points to conceptual redundancy in the measurement model rather than to the irrelevance of usable technology. Public website quality remained significant while recording the lowest mean of all constructs (M = 3.33, SD = 1.02), a combination that identifies it as the clearest area of unrealized potential. This aligns with evidence that websites, though essential, tend to be less developed than social media in tourism promotion (Tham et al., 2020; Pan et al., 2021), and underlines the need to improve functionality, accessibility and design in support of festivals’ sustainability goals.

The mediation analysis did not support festival satisfaction as a mediator. Once the five antecedents were entered alongside it, the association between satisfaction and revisit intention disappeared (β = −0.073, p = 0.133), and none of the indirect effects reached significance. Instead, four of the antecedents related to revisit intention directly, and the direct-effects model accounted for 57.5% of its variance. This departs from studies that report satisfaction as an intervening variable between digital engagement and loyalty (Khan et al., 2021; Arasli et al., 2021), and the divergence is best explained by the properties of the satisfaction measure in this sample rather than by a substantive contradiction. Where satisfaction is near-universal, as it is at this free, single-day community festival, the construct cannot discriminate between visitors and therefore cannot transmit the influence of anything else. What the data support is a direct association between digital engagement and the intention to return, with satisfaction sharing the same antecedents rather than standing between them and loyalty.

The study shows that online strategies are significantly associated, alongside the quality of the experience itself, with the visitor responses that prior research links to the long-term success of cultural events, with the fulfilment of visitor expectations rather than any single promotional channel emerging as the dominant lever. One managerial implication follows, though it is not something the data can test: if greater loyalty and lower dependence on paid promotion translate into cost savings, as work in other settings suggests they can (Trusov et al., 2009; Villanueva et al., 2008), digital engagement would also improve financial efficiency. Costs, conversion and revenue were not measured here, so this remains a proposition for organizers to evaluate against their own accounts rather than a finding of this study. In practical terms, the findings underline the value of investing in public-website development, aligning social media content with visitor expectations, and fostering authentic electronic word of mouth (eWOM) to attract tourists and build loyalty. For scholarship, the study adds empirical evidence to models of ICT use, visitor satisfaction, and sustainable cultural tourism, offering a foundation for further research in comparable settings.

Read through a marketing-investment lens, these results suggest that promotional spending alone does not determine visitor response. The dominant predictor of satisfaction is the fulfilment of expectations—an outcome produced by the event itself rather than by any communication channel—which implies that digital promotion raising expectations without a matching improvement in delivered experience risks depressing rather than raising satisfaction. The channels themselves display a trade-off between effect size and cost. Electronic word of mouth exerted the smallest significant effect on satisfaction, yet it is an earned channel typically carried at low marginal cost, and it shows the strongest bivariate association with revisit intention of any predictor (r = 0.65). Because no expenditure data were collected, its cost efficiency cannot be assessed here; if the cost patterns reported in other settings hold, however, earned channels of this kind deliver more durable and cost-efficient effects than paid promotion (Trusov et al., 2009; Stephen & Galak, 2012), which would make eWOM an attractive channel for organizers to examine against their own budgets. The public website, by contrast, exerts a larger effect on satisfaction but typically requires continuing development and maintenance expenditure; that it is simultaneously significant and the lowest-rated construct suggests an asset that is currently under-developed relative to its demonstrated contribution. For organizers, this frames digital spending as a budget-allocation decision under uncertainty rather than a communications afterthought (Fischer et al., 2011), in which the first claim on resources is the experience that expectations are set against, and the residual is best directed toward the asset with the widest gap between contribution and current quality. These inferences concern relative effectiveness in producing the visitor responses that prior research identifies as antecedents of festival revenue; because no cost, revenue or expenditure data were collected, they should not be read as estimates of return on investment. The study thus offers research on marketing productivity and on the financial sustainability of non-profit and mixed-funding events a demand-side account of how digital engagement relates to the visitor responses that prior work links to festival survival. Whether those responses translate into revenue or into lower promotional costs at this festival is a question for organizer-side data that this study did not collect.

5. Conclusions

This study developed and empirically tested a path-analytic regression model integrating five determinants of festival satisfaction—social media, festival expectations, ICT usability, public website quality, and electronic word of mouth (eWOM)—based on established models of ICT use, consumer satisfaction, and revisit intention. Four of the five antecedents significantly influenced satisfaction, which in turn correlated with revisit intention at the bivariate level. Satisfaction did not, however, mediate the relationships between the antecedents and revisit intention: with the antecedents controlled, the association vanished, and the antecedents predicted revisit intention directly, explaining 57.5% of its variance. Digital engagement therefore appears to relate to visitor loyalty directly rather than by way of satisfaction, at least where satisfaction is measured near its ceiling. For managers, the results suggest a way of framing digital spending rather than a measured return on it: if loyal visitors do lower future promotional requirements and support steadier attendance, as prior work on word-of-mouth acquisition indicates, then digital engagement can be treated as a marketing investment whose returns accrue through visitor response. No ticket revenue, visitor expenditure, marketing spend or sponsorship data were collected in this study, so this framing is offered as a managerial implication and not as an empirical result. The results qualify this in an important way: promotional funds appear to create value mainly where they set expectations that the event can actually meet, a pattern that sits comfortably with the human-centric premise that motivated the choice of visitor-level outcomes, although the study provides no test of that premise and the result follows equally from expectancy-disconfirmation accounts of satisfaction. The study also connects to the sustainable development agenda: digital promotion of cultural events is widely argued to strengthen cultural identity and community ties, in line with SDG 11 (Sustainable Cities and Communities) and SDG 12 (Responsible Consumption and Production), although the present design measures visitor satisfaction and revisit intention rather than these broader outcomes.

This study has limitations. The model was estimated as a partial mediation specification, with direct paths from each antecedent to revisit intention. Because the design is cross-sectional, the absence of mediation cannot be taken as evidence that satisfaction plays no role in loyalty formation; it shows only that, in this sample, satisfaction carries no variance beyond that already shared with its antecedents. A further limitation concerns construct discriminants: ICT usability and public website quality were strongly correlated, and their overlap most likely accounts for the non-significant coefficient of the former. Instruments that separate infrastructure usability from website content and design more sharply would allow the two contributions to be estimated independently. A similar caveat applies to electronic word of mouth. Its HTMT ratios with social media (0.866) and with revisit intention (0.889) fall below the 0.90 threshold as point estimates, but their bootstrap confidence intervals extend above it, so discriminant validity is not established for these two pairs at the conventional criterion. The constructs are retained on conceptual grounds and because they relate differently to the rest of the model, but future instruments would benefit from a sharper operational separation between engagement with official channels and user-generated recommendation behavior. The low variance in the satisfaction, eWOM and revisit-intention items, all of which cluster near the top of the response scale, additionally depressed the extracted variance for these constructs and weakened the Fornell–Larcker assessment; instruments with wider effective response ranges, or scales with more discriminating item wording, would improve measurement precision in future replications. The evidence is cross-sectional and drawn from a single festival, so relationships should be read as associations rather than causal effects, and generalization to other events and countries calls for caution. The ceiling effects reinforce this caution. Because satisfaction, eWOM and revisit intention are all measured near the top of their scales, the finding that satisfaction carries no variance beyond that shared with its antecedents should be read as specific to a sample in which satisfaction is close to universal, and not as evidence that satisfaction is unimportant to loyalty formation in general. Replication in settings with greater variation in visitor evaluation—a ticketed event, a multi-day festival, or a sample not recruited through the festival’s own digital channels—is needed before the pattern can be generalized. No financial outcome was examined. The study measures visitor perceptions only—satisfaction and revisit intention—and no ticket revenue, visitor expenditure, marketing expenditure or sponsorship income was collected. Any financial implication drawn from these results is inferential and rests on prior research linking satisfaction and loyalty to revenue and to marketing costs, not on evidence from this festival’s accounts. Connecting these perceptions to objective financial data, such as ticket and concession revenue, marketing expenditure, and the return on digital-promotion budgets, and examining how cultural events finance their digital and sustainable transition through public support, sponsorship, and green-event financing instruments, would strengthen the evidence and is a priority for future research. Accordingly, terms such as return, allocation and investment are used throughout in a bounded, descriptive sense, referring to the perceptual antecedents of financial performance rather than to measured financial quantities. A further limitation concerns sample selection. Respondents were required to have attended the festival and to be familiar with its online promotion, and most were recruited through digital channels. This is likely to over-represent visitors already engaged with and favourably disposed toward social media and online promotion, which may inflate the estimated associations between digital engagement and satisfaction. The recruitment channel was not recorded for individual respondents, so the online- and on-site-recruited subsamples cannot be compared statistically; future studies should record this information and stratify accordingly, or recruit through channels independent of the festival’s own digital presence. Notably, the strongest predictor of satisfaction in this sample was festival expectations rather than either digital channel, and eWOM was the weakest of the significant predictors—a pattern that runs against the direction the selection mechanism would be expected to produce, though it does not eliminate the concern. Relatedly, the use of Cochran’s formula to set the target sample size should not be taken to license population-level inference, since the formula presumes probability sampling; it addresses statistical power, not representativeness. Because satisfaction was measured at the top of its range for the majority of respondents, the instrument cannot distinguish degrees of satisfaction among highly satisfied visitors; scales with wider effective ranges, or measures anchored to specific aspects of the experience rather than to overall evaluation, would improve discrimination in future studies. Relatedly, 72.7% of respondents were aged 18–29. Because official demographic data for the festival’s audience are not available, it cannot be established whether this reflects the visitor population or the recruitment method; the estimates may accordingly be specific to younger, digitally engaged visitors, and replication with age-stratified recruitment would establish whether the relationships hold across age groups.

Future research could extend this work in several directions. Longitudinal studies would show how digital engagement and satisfaction evolve across successive festival editions, while cross-country designs covering multiple event types would test whether the relationships observed here hold in different cultural and economic contexts. Incorporating additional factors, such as mobile-application use, the personalization of digital services, or emerging technologies like virtual and augmented reality, could enrich the model and capture new dynamics in visitor behavior. Finally, examining sustainability-related outcomes, including the environmental and social impacts of ICT adoption at festivals, could yield further insight into the long-term success of cultural-tourism initiatives. A further extension concerns the circular-economy dimension of festival promotion. Substituting digital channels for printed and physical marketing, and foregrounding local sourcing in event catering, may in principle decouple a festival’s visibility and value from material throughput (Geissdoerfer et al., 2017; Kirchherr et al., 2017). This proposition was not operationalized in the present instrument and is therefore untested here; measuring it would require material-flow and procurement data alongside visitor perceptions and would be a valuable direction for future work.