Neurotourism Aspects in Heritage Destinations: Modeling the Impact of Sensory Appeal on Affective Experience, Memory, and Recommendation Intention

Abstract

This study models how designable cues in digital heritage promotion shape advocacy through affect and memory. Relying on the stimulus–organism–response paradigm, we argue that three stimuli, Visual Sensory Appeal (VSA), Narrative Immersion (NI), and Perceived Authenticity (PA), trigger Emotional Engagement (EE) and become Destination Memory (DM), leading to Intention to Recommend (IR). A cross-sectional quantitative design with an online self-report survey was employed. Using Structural Equation Modeling (SEM) we modeled 653 usable responses to test hypothesized stimulus–organism–response processes and Multi-Group Analysis (MGA) tested heterogeneity across gender, age, education, recent contact, cultural-travel frequency, preservation interest, prior heritage experience, and technology use. Direct associations revealed VSA was a strong predictor of IR, and EE and DM predicted IR positively. NI and PA were not incrementally directly affecting IR. Mediation tests revealed partial mediation for VSA (through EE and DM) and complete mediation for NI and PA; across all stimuli, DM far surpassed EE, suggesting memory consolidation as the overall mechanism. MGA revealed systematic segmentation: women preferred visual and authenticity approaches; men used affective conversion, narrative, and authenticity-to-memory more; young adults preferred story/memory levers; higher education made authenticity pathways legitimate; exposure, experience, sustainability interest, and technology use further conditioned strength of paths. Results sharpen S–O–R accounts by ranking visual design as a proximal driver and placing EE on DM as the central channel through which narrative and authenticity have their influence. In practice, the research supports visually consistent, memory-backed, segment-specific strategies for sustainable, inclusive heritage communication.

1. Introduction

Increased urban growth and concomitant pressures upon built heritage have also created issues of preservation, interpretation, and effective use of architectural heritage [1,2]. World heritage tourism has also been a tool used to encourage conservation and development but whose effectiveness increasingly rests on places’ ability to convey the value of heritage through digital media and site interpretation. A developing body of evidence suggests that certain sensory and narrative design factors influence affective response and intentions: semantic–differential analysis integrated with eye tracking in Macau’s Historic Centre linked physical visible items—door ornament, proportional balance between buildings, proportional balance of windows, and wall type—to behavioral intention, with distinct positive, strong, and enthusiastic feelings strongly correlated with those intentions [1,2,3]. Parallel research threads in immersive technologies emphasize the same processes. Critical examinations of AR/VR in heritage focus on virtual museums, heritage travel, immersive narrative, and mobile AR “revival,” all systematically reporting gains in presence, interaction, and learning benefits but warning that results are more a function of interaction design and expressive affordances than of hardware [1,4,5,6]. Accordingly, design quality and communicative structure emerge as first-order levers for experience and intention.

Recent mobile AR syntheses can be found in UX patterns (uncover hidden objects, “overlay” the past, direction-based advice) that link perception/tracking to meaningful interaction design, as well as in reference to neglected sonic modalities (soundscapes, spatialized audio) and enduring sound-rendering deficiencies [7,8,9,10]. New digital-storytelling techniques emphasize the use of platform-native features (social media sharing, multi-media interaction, user comments, familiar culture) in promoting their content, in a bid to attract younger readers. Inclusive-design checks value accessibility to older and mobility-impaired visitors and move away from device-centered to people-centered experience design. Throughout this work, there is a persistent highlighting of the role that authenticity plays as a driver of behavior, both directly and indirectly, through the experience of memorable experiences and place attachment while, despite ambitious requirements for increased conceptualization and measurement across its traversing contexts [9,10,11], our own recent-horizon review articulates a call for sharper conceptualization and measurement. In brief, sensorial and narrative craftsmanship, inclusivity, and authenticity-connected communication make what the visitors feel, remember, and do [6,9,10].

Unresolved issues remain. First, the effects of “immersion” versus presence are not clear, and when clarity, engagement, and authenticity are held constant, immersion affordances alone are weak predictors. It may well be that perceptual fluency and affective resonance matter more than technological intensity [12,13]. Second, “memory” itself is defined variably (MTE, memorability, autobiographical memory, total experience), making synthesis and construct equivalence problematic. Third, authenticity serves as both a stimulus (curatorial/design cue) and an appraisal (visitor judgment) mechanism: while in general connected with satisfaction, attachment, and revisit via imagery and memory, some negative emotions are associated with a reduction in perceived authenticity, and domestic audiences might apply different authenticity standards from their international equivalents [4,5,6]. Lastly, despite advancements in XR, there are remaining gaps in expressive/interaction affordances (e.g., sound simulation and spatialization), inclusive design, and youth-focused social media narrative [1,13]; the field has yet to firmly tie content features to psychological mechanisms rather than to hardware limitations [11,14,15]. Taken together, these threads speak to the central role of well-designed visual/auditory and narrative cues—infused within approachable, authenticity-sensitive design—in evoking emotional response, coalescing meaning, and fostering advocacy and sustained heritage engagement and motivate mechanism-oriented models that connect designable cues to organismic processes and action in audience segments [4,5,6].

These trends demand an approach that (i) operationalizes stimuli at the communicative content level (visual appeal, narrative immersion, authenticity), (ii) frames emotion → memory as the fundamental organismic process, and (iii) aims for advocacy (recommendation intention) as the key response of interest for destination competitiveness and sustainable heritage consumption. By operationalizing emotional engagement and destination memory using validated perceptual measures and estimating their mediating effects between designable cues and recommendation intention using survey-based SEM, the study calls for theoretical clarity in neurotourism and for generalizable results beyond laboratory-based trials. It also positions authenticity as input and as assessment, enabling tests of its immediate and mediated impacts in a composite model and subgroup analyses where audience heterogeneity (e.g., local vs. foreign; age groups) is plausible. The following section develops research objectives and hypotheses on this conceptualization, detailing anticipated relations among stimuli, organismic mediators, and behavioral responses.

With the above as context, this paper specifically addresses the psychological mechanisms through which heritage communication informs advocacy-relevant behavior. Building on the Stimulus–Organism–Response (S-O-R) paradigm, it outlines three designable stimuli shared in digital heritage promotion: visual sensory appeal (perceived richness and aesthetic quality of image and layout), narrative immersion (extent of mental engagement within a destination narrative), and perceived authenticity (appraisals of authenticity and cultural fidelity) [2,3,6]. Such stimuli are believed to evince emotional engagement (e.g., inspiration, awe, nostalgia), which is supposed to intensify into destination memory (vividness, clarity, retrievability) and eventually into recommendation intention (offline WOM/eWOM advocacy). This emphasis is attested to in empirical trends: image and memorability have been found to mediate the rejuvenation program effects on repeat visit intention in heritage experience quality studies [10], Memorable Tourism Experience (MTE) has been found to be a direct predictor of loyalty, as well as indirectly through satisfaction and attachment in mass tourism and coastal tourism studies [16,17], and impact factors and rehearsal have been found to drive revisit and recommendation through place attachment in autobiographical memory studies [15,18]. Social-digital settings are important too: social return—predicted positive appraisals of one’s posts—increases MTE and advocacy [19,20], and vicarious nostalgia drives behavioral intention through MTE and is enhanced by social return [19,20]. Together, this evidence suggests that designable cues work less through explicit persuasion and more through affect-to-memory pathways that provoke advocacy. Consequently, testing emotion–memory channels offers the most promising route to explain recommendation behavior.

The study was conducted in Greece through a bilingual (Greek/English) online survey distributed nationally to adults who have been targeted by digital heritage promotions during the last six months. Recruitment occurred through Greek social media groupings, research emailing lists, and tourism/heritage groupings; the measure included validated scales adapted and pilot-tested for linguistic equivalence. While the questionnaire mentioned heritage promotions more widely (rather than one site), the cultural frame of the sample is almost exclusively Hellenic/European. The empirical estimates will thus be interpreted as locally referenced and the underlying S–O–R mechanisms (stimulus and emotion on memory and advocacy) hypothesized to be transportable. We thus position the present evidence as context-specific but methodology-generalizable, and we issue calls for cross-cultural replications (e.g., multi-nation samples, language versions) and formal tests of measurement invariance to shed light on if path strengths and the contribution of authenticity and narrative differ between cultural settings.

Foreshadowing our hypotheses, variance-based SEM on 653 surveys revealed that visual sensory beauty directly and strongly impacted intention to recommend, whereas organismic states—primarily destination memory—also influenced advocacy. Story immersion and feeling of authenticity did not predict recommendation in itself after controlling for variables but secondarily through emotion and more strongly through memory (full mediation for NI and PA; partial mediation for VSA through EE and DM). Multi-group analysis provided systematic heterogeneity by gender, age, education, exposure/experience before now, interest in sustainability, cultural-travel frequency, and technology use, which suggests segment-specific processes which are unique. Cumulatively, evidence supports an S–O–R explanation whereby images operate proximally, with authenticity and narrative influencing advocacy primarily through the creation of emotional identification which solidifies into enduring representations.

The remainder is organized as follows: Section 2 summarizes the sensory/narrative cues, authenticity, emotion–memory, and advocacy literature and formulates hypotheses. Section 3 details the S–O–R model, data, measures, and methods (PLS-SEM, MGA). Section 4 reports measurement and structural findings, mediation, and subgroup differences. Section 5 outlines the findings. Section 6 outlines practical implications. Section 7 concludes with contributions, limitations, and future research.

2. Literature Review and Research Hypotheses

2.1. Sensory and Narrative Cues in Digital Heritage Tourism

Digital heritage communication more and more uses sensorially rich and narratively organized content to attract attention, emotion, and downstream behavioral response. Within museum, heritage place, and virtual heritage contexts, the evidence as a whole directs towards aesthetic quality, presence-eliciting design, and narrative storytelling as essential engagement levers; but findings are coy about how some cues track to affect, cognition, and intention. In virtual museum environments, Li et al. [21] showed that affective engagement, cognitive engagement, and perceived presence all boost learning motivation, but immersion is not an effective motivator on its own—while visual aesthetics do boost immersion, affect, and presence. This distinction between presence (psychological) and immersion (usually regarded as a technical or configurational attribute) persists: Wan et al. [22] established that, in non-immersive VR tours, design aesthetics, interactivity aesthetics, and presence aesthetics are predictors of hedonic and utilitarian values, attitudes, and visit intentions, but interactivity aesthetics will not enhance utilitarian value. At the same time, Han et al. [23] discovered that AR visual beauty, pleasure, and amusement augment perceived experiential authenticity and happiness and thus supportive behavior but not escapism, previously considered affective motivation for immersion. Taken together, these findings warn against interpreting “immersion” as a one-size-fits-all mechanism; presence, aesthetic structure, and emotionally engaging structure seem safer routes than pure technological intensity.

In physical design for museum and exhibition, Hu et al. [24] integrated eye tracking with self-report rating and established that 3D-object-plus-physical displays are most enjoyable, picture-plus-physical displays are faster in cognition, and solo displays lag behind in preference and memory. Their fixation measures lead to opposite directions on perceived fun/richness/comfort/attractiveness and authenticity scales and suggest different perceptual pathways to hedonic vs. evaluative judgments. Spatial structure is also involved: Nubani et al. [25] used visibility graph analysis to demonstrate that spatial location and visibility predict contact and navigation and, when combined with exhibit saliency, account for group behaviors regarding engagement. For exterior galleries, Yang et al. [26] enumerated dimensions of environmental stimuli (atmosphere; space/layout; communication/service; landscape aesthetics) and find that most dimensions elicit positive affect leading to commitment; surprisingly, landscape aesthetics also elicit negative affect, and tourist type moderates aesthetics on emotion. These findings highlight that aesthetics are strong but not necessarily positive; valence of aesthetic reactions depends on visitor type and context.

Narrative messages are also effective. Zins et al.’s [27] review on website content analysis highlights a persistent planning and managerial shortfall: research continuously investigates effects of narrative but has comparatively sparse advice regarding how DMOs ought to organize narrative aspects on behalf of heritage assets. Results of experiments fill part of this gap: Chang et al. [28] demonstrated that narrative transportation is enhanced by digital storytelling, which in turn enhances persuasiveness and travel intention, with persuasiveness mediating to some extent the transportation on intention relationship. With millennial museum art visitors, Hyun demonstrates that hedonic value—evoked by ambiance and aesthetics—exerts more effect than utilitarian value in satisfaction and loyalty driving in accordance with a narrative–sensory path to affective consequences. At the semiotics of landscape scale, Chang et al. [29] showed that visual attention and emotional experience are connected but distinct; aesthetic experience and perceptual fluency are the boundary conditions which make possible the passage from visual stimulation to emotional involvement. This is a vital boundary condition: attention is sufficient but not necessary for emotion. Sensory fluency and narrative structure rather than looking regulate affect.

XR and virtual heritage studies validate the precedence of information and emotion in multi-sensory design. Zhang et al. [13] discovered that immersion is strongly predicted by informational clarity while emotional resonance underpins personalization; visual–auditory modalities are more effective than richer sensory stacks, indicating diminishing returns when sensory load outstrips cognitive bandwidth. A sizable survey by Lin et al. [30] remarks on a bias for interface-based visual stimulation for tangible heritage, comparative neglect of intangible heritage, and minimal concern with the maintenance of long-term emotional engagement; they recommend emotional design and gamification to enhance and extend engagement. In the context of architectural heritage photos, the authors of [22] integrate eye tracking with semantic differentials and demonstrate that certain visual attributes (proportional balance, window balance, wall style, door decoration) and discrete positive emotions together predict behavioral intention. Demographic variables influence these processes: Yuan et al. [31] described age differences in scan paths and text–image integration in promotional brochures, and Ren [32] demonstrated that natural attributes draw early visual attention and high aesthetic appreciation whereas cultural ones, while less immediately attention-grabbing, are admired on historical grounds, with age, sex, nationality, and expertise moderating these appreciations.

Throughout this corpus, there are a couple of rough generalizations and exclusions [21,24]. Aesthetics and narrative transportation always pre-figure affect and intention, but the reports are more in terms of presence, intelligibility, pleasure, and emotional effect than anything on the order of generalized immersion or escapism; interactivity effects are context-dependent and occasionally asymmetric (gaining hedonic but not utilitarian value) [13,25,29]. Second, attention does not necessarily indicate emotion: visual prominence has to be processed smoothly and narratively segmented in order to convert it into felt involvement. Third, authenticity always contributes affective connections and contentment (AR and heritage environments), but the literature is bewildered as to whether authenticity is a stimulus (designable cue) or an emergent appraisal—a conceptual ambiguity that renders measurement challenging [14,33,34]. Fourth, managerial advice for narrative building for heritage brandness is scarce, and intangible heritage and long-term emotional connection are understudied. Lastly, heterogeneity is an issue: demographic and tourist-type moderators intervene in aesthetic–emotion relationships, but multi-group testing systematically is not common [27,35].

This work fills these gaps through direct pairings of sensory (visual sensory appeal) and narrative (narrative interest) stimuli with felt authenticity in one S–O–R process in which emotional interest and destination memory are mediators for effects on recommendation intent. By pre-empting presence-like affect and memory consolidation of undifferentiated “immersion,” the model advances reconciliation of conflicting evidence for immersion and interactivity and consonance with evidence of operation of clarity, aesthetics, and storytelling through affective resonance and recalled meaning. Moreover, survey-based SEM design with validated perceptual and affective measures and specified multi-group analyses (e.g., by age or previous heritage experience) meets calls for managerial utility and accommodation to audience diversity. In the process, the paper develops theoretical accuracy in neurotourism by determining what sensory–narrative stimuli are relevant, how they operate (emotion on memory), and to whom and ensuing design implications for the communication of digital cultural heritage. To this end, the following hypotheses were formed:

H1. 

Visual Sensory Appeal (VSA) influences Intention to Recommend (IR).

H2. 

Narrative Immersion (NI) influences Intention to Recommend (IR).

H3. 

Perceived Authenticity (PA) influences Intention to Recommend (IR).

2.2. Emotional Engagement and Destination Memory Formation

Throughout heritage tourism settings, there is strong thread between affective responses provoked by designable stimuli and enhancing strong experiences and, ultimately, behavioral outcomes [36,37,38]. From destination, museum, religious, and virtual heritage settings, evidence shows that emotional engagement is a mediating factor through which authenticity indicators, beauty qualities, and narrative qualities are retained as destination memory and conveyed into loyalty, revisit, recommend, or pro-social intentions [39,40]. Concurrently, scholarship points to conceptual and methodological tensions—i.e., what is authenticity being measured against, whether emotions necessarily enhance perceptions of authenticity, and how “memorability” is definably conceptualized—that drive towards a more nuanced, theory-driven model of emotion-to-memory pathways to heritage communication [16,41].

A vast amount of evidence positions Memorable Tourism Experiences (MTEs) as the leading mnemonic construct that makes experience quality correspond to downstream behavior. For sun-and-beach and rural destinations, Moliner-tena et al. [38] established that environmental sustainability can enhance MTEs in terms of moderation by age effects, while in Santorini Stavrianea et al. [16] established that MTEs are a predictor of destination loyalty either directly or indirectly through satisfaction. Within cultural heritage itself, Zhou et al. [42] suggest that heritage “rejuvenation” experience quality enhances MTEs and destination image and that these in turn enhance revisit intentions; Rasoolimanesh et al. [43] similarly discovered that authenticity, nostalgia, and sacredness influence MTEs, which mediate the impacts on subjective well-being and destination image (with sacredness having indirect but not significant direct effects). Complementing these symmetric SEM results, Rasoolimanesh et al. [39] integrated PLS-SEM and fsQCA to expose equifinal processes whereby hedonism and novelty—diverse affective and cognitive experiential elements—also generate satisfaction and behavioral intentions. Coupled together, these studies indicate a two-stage process: affectively rich appraisals and values are synthesized first as stable experience, which subsequently underlies loyalty-relevant intentions.

Authenticity—as perhaps the paradigmatic heritage assessment—has strong, albeit indirect, relationships to emotion and memory. In religious tourism, Huynh et al. [40] found object-based, constructive, and existential authenticity creates emotional solidarity and higher MTEs and commitment; multi-group results show these are different for domestic and overseas visitors. In festival contexts, Thi Kim et al. [44] included emotional involvement in an intermediary between authenticity (specifically existential authenticity) and enjoyment, placing felt meaning over object-oriented “realness” as the more powerful affective pathway. Nonetheless emotion does not always justify authenticity judgments, for example, in analyzing user-generated content with machine learning, Calderón-Fajardo et al. [45] demonstrated that some emotions actually decrease perceived authenticity, with domestic tourists using higher standards of authenticity than foreign tourists. This paradox fruitfully redefines authenticity as an affective judgement, not just as a stimulus; positive affect can amplify authenticity if culturally framed, but culturally misframed emotion can destroy it—an asymmetry with obvious implications for content and audience design.

Aside from authenticity, varied affective antecedents always trigger memory formation and intention. In heritage structures, Manley et al. [46] associate certain visual features with certain enjoyable emotions that precede behavioral intention, whereas in poster promotion Pan et al. [47] established that mystery-based visual attractiveness elicits curiosity and desire to visit, which is increased when textual introductions are tangible—proof that affective arousal and processing fluency form stronger approach tendencies. In non-immersive VR, Atzeni et al. [48] concluded that object-based authenticity generates affective response to satisfaction, attachment to VR, and intention to visit; existential authenticity modulates both cognitive and affective routes. Dağ et al. [41], in AR-experience museums, observed that immersion augments satisfaction and perceived authenticity by virtue of user engagement, once more locating engagement as the fulcrum of emotion–outcome chains. Taking the discussion through to conservation tourism, Yang et al. [49] outline cultural traits that enable pleasure and pro-conservation consequences in an S-O-R framework, while in “welcome-back” tourism, Pan et al. [47] demonstrated that recollection of earlier experience influences revisit by way of a mediated destination attachment that takes place through an implicit memory-to-attachment procedure. Specifically, Kim [37] fashions a scale for negatively memorable experience and proves that affect can be memorable but not necessarily positive—a considerate reorientation to a literature trending towards equating MTEs as always enjoyable—and Ref. [36] found that authenticity dimensions support positive feelings and revisit intention and are mediated by nostalgia propensity. These strategies complement each other to caution against valence bias: negative or ambivalent emotions are durable and behaviorally salient as well, particularly when integrated with heritage connotations.

Methodologically, the majority of studies use cross-sectional SEM and self-report measures, which are effective for construct validation but weak for the capture of temporal dynamics of consolidation of memory or attention differentiation from affect [43,44,46,50]. Numerous contributions go beyond that: Calderón-Fajardo et al. [45] utilize machine learning and SHAP values to estimate the relative contribution of emotion vs. demographics to judgments of authenticity; Lv et al. [50] combined behavioral experiments, ERP (P2/LPP) findings, and EEG-based decision trees in order to link cultural landscape types to recommendation through place attachment and restorativeness and explain early affective processing; and Manley et al. [46] used an integrated qualitative methodology in order to uncover an ambition for emotive and sensual engagement with intangible and tangible heritage among Chinese consumers of a Western exhibition. But Hosany et al. [51] warn that MTE studies remain geographically imbalanced, over-measured, and scale-dependent and suggest negative experiences, cross-cultural examination, and mixed methods—each with direct applicability to heritage communication.

Situated within this context, this paper adopts a stimulus–organism–response structure wherein visual sensory appeal, narrative interest, and subjective authenticity represent stimuli that lead to emotional involvement, which solidifies to destination memory and arouses recommendation intention. By placing focus on emotion-to-memory mediation, authenticity differentiation as stimulus and appraisal, and salience to valence diversity (remarking the likelihood of negatively memorable affect), the model bridges discrepant findings regarding the role of emotion in authenticity assessments and explains the psychological mechanisms involved in supportive persuasion in heritage communication. The strategy eliminates methodological problems by separating confirmed constructs appropriate to SEM and by justifying multi-group analyses (e.g., domestic vs. international, susceptibility to nostalgia) to explain repeatedly found audience heterogeneity in the literature. Thus, the corresponding hypotheses were formulated:

H4a. 

Emotional Engagement (EE) influences Intention to Recommend (IR).

H4b. 

Destination Memory (DM) influences Intention to Recommend (IR).

H5a. 

Emotional Engagement (EE) mediates the relationship between Visual Sensory Appeal (VSA) and Intention to Recommend (IR).

H6a. 

Emotional Engagement (EE) mediates the relationship between Narrative Immersion (NI) and Intention to Recommend (IR).

H7a. 

Emotional Engagement (EE) mediates the relationship between Perceived Authenticity (PA) and Intention to Recommend (IR).

2.3. Memory and Behavioral Intention in Heritage Tourism

Heritage tourism studies increasingly shift attention to the hypothesis that what the visitor remembers—rather than what they simply experience in the here-and-now—maps downstream behavioral intent such as recommendation, eWOM, revisit, and even on-site consumption [52,53,54]. But the literature varies as to how “memory” is theorized (as Memorable Tourism Experiences (MTEs), as autobiographical dimensions of memory, or as domain-specific recall) and what role intervening appraisals like satisfaction, destination image, attachment, or preference have been assigned [17,55,56]. Across different contexts, the research repeatedly demonstrates indirect models wherein memory-like constructs are mediators of perceptual antecedent–response relations, whereas direct influences of perceptual antecedents on intention are context-dependent and fluctuate [57,58,59].

One of the early streams treats MTE as the mnemonic engine of advocacy. In cultural and heritage settings, MTE comes out as a mediator and a predictor of behavioral intention. In the UNESCO city of Kashan, Rasoolimanesh et al. [55] illustrated that authenticity and tourist experience generate revisit and eWOM via MTE, with destination image also enabling eWOM; their mixed-methods design (PLS-SEM and fsQCA) also identifies equifinal configurations wherein different combinations of experiential ingredients have equivalent effects. To this, Stavrianea et al. [16] adds that MTEs have direct and satisfaction-mediated impacts on loyalty in a mass-tourism destination, whereas Arina et al. [17] conceptualized MTE as a higher-order construct that increases perceived economic value and place attachment and, eventually, recommendation, with destination competitiveness serving to moderate the strength and even direction of these impacts. These findings converge on the status of MTE as a proximal cognitive–affective summary that integrates experience into behaviorally relevant memory.

A second stream more directly measures memory. Zhao et al. [18] defined autobiographical memory as impact and rehearsal in rural tourism and demonstrate that both predictors predict revisit and recommendation with place attachment as the mediator. In a two-wave panel design, Cao et al. [56] establish that aesthetic experiential attributes (scenery, cleanliness, harmony, art/architecture, genuineness) increase memorability, which completely mediates most effects on loyalty; some aesthetic attributes have no direct impact on loyalty once memorability is controlled. Similarly, Wang et al. [59] connected destination reputation both directly and indirectly to consumption behavior on-site through enjoyment and memorability and show a chain from reputation signals to affect to memory to economic action. In addition, these works strengthen the contention that memory is a causal link between designable causes and effects on behavior and that not modeling memorability masks actual effect sizes.

A third stream of work adds social and web context to the memory–intention link. Mittal et al. [20] demonstrated that social return—expected good evaluation of the user’s posts—raises both MTE and behavior intention, with the former effect being partially mediated by MTE; but social return has weak direct influence on revisit, and social-evaluative motivation drives advocacy rather than frequent visitation. Building on this, Bhogal et al. [19] showed vicarious nostalgia shapes behavior intention solely through MTE and that social return enhances the MTE–intention relationship—ample proof that social delight frames the memorability–advocacy route. In food experiential consumption, Lai et al. [57] proved that, while among experiential value dimensions, emotional value has the greatest impact on satisfaction and memory, and memory is a significant driver of eWOM generation. In empirical studies on UGC and publicity, Xu et al. [52] also revealed that affectively framed UGC builds image and satisfaction, which in turn create loyalty behavior, while publicity and eWOM combine to create awareness and preference that propel intention. Together, these findings indicate that social and media ecologies play a stronger role in affecting the impact of memory on advocacy-led intentions (recommendation, eWOM) than revisit, a distinction replicated in city contexts where Khairani et al. [54] observed that MTE significantly predicts satisfaction and eWOM but not revisit.

Affective pathways to memory also differ by emotion family. According to Keskin et al. [58], nostalgia emotions enhance MTE and satisfaction, thereby revisit and recommendation; Lai et al. [60] rendered cultural memory a mediator of cultural contact and revisit under a stimulus–response process with attitude to culture moderating contact on memory and contact on revisit. These findings indicate that identity-congruent emotions (nostalgia, cultural identification) are especially efficient at leaving enduring traces in memory and coordinating behavioral commitment. Meanwhile, the literature cautions against valence biases: even as an overwhelming number of studies indicate positive affect, the duration of memory may also be the result of bittersweet or even negative events, and it is a consideration highlighted elsewhere within the MTE corpus and one to which more weight should be given in heritage settings.

There are a few boundary conditions and contradictions to be resolved. First, the terminus of behavior is important. Memory always precedes recommendation and eWOM irrespective of context, but its effect on revisit is variable, at times mediated by satisfaction, attachment, or image (e.g., Rasoolimanesh et al. [55]; Zhao et al. [18]; Khairani et al. [54]). Second, moderators change path strengths: competitive context (Arina et al. [17]), social return (Bhogal et al. [19]), and cultural attitude (Lai et al. [57]) change memory’s conversion to intention. Third, methodological design shapes inference. Most are cross-sectional and self-report based (raising common method problems), with Cao et al.’s [33] two-wave design and Rasoolimanesh et al.’s [43] mixed/symmetric–asymmetric modeling being stronger causal and configurational evidence. Lastly, operationalization is contradictory: research employs MTE, memorability, autobiographical memory, or general experience inconsistently, sometimes synonymously. Though the variety of measures parallels the richness of the construct, it makes meta-comparison problematic and indicates the importance of proper construct definition and validly validated measurement in mediating models of memory.

Positioned in this corpus, a conceptual framework wherein emotional involvement–destination memory–recommendation intention is the operative path provides theoretical beauty and empirical consistency. It subsumes evidence that aesthetics, reputation, and social pleasures tend to operate through memorability, explains inconsistency in revisit effects (via satisfaction, attachment, and image), and encompasses social-context moderators (e.g., social return) that amplify advocacy. Employing well-established, discrete measures of destination memory and emotional engagement, defining recommendation as the central behavioral outcome, and building multi-group analyses (e.g., domestic–international, experience level, nostalgia-proneness), this study bridges persistent gaps: asymmetrical operationalization of memory, restricted testing of indirect vs. direct effects between behavioral endpoints, and under-specification of social-digital ecologies structuring the memory–intention relationship in heritage tourism. Thus, the following hypotheses were formulated:

H5b. 

Destination Memory (DM) mediates the relationship between Visual Sensory Appeal (VSA) and Intention to Recommend (IR).

H6b. 

Destination Memory (DM) mediates the relationship between Narrative Immersion (NI) and Intention to Recommend (IR).

H7b. 

Destination Memory (DM) mediates the relationship between Perceived Authenticity (PA) and Intention to Recommend (IR).

Building on these hypotheses, the next section translates the theoretical relationships into an empirically testable model, detailing construct operationalization and path specifications to be estimated via PLS-SEM.

3. Research Methodology

3.1. Conceptual Model and Rationale

This subsection formalizes the operationalization of the hypothesized relations into a structural equation model that can be tested empirically. While Section 2 integrated prior research and derived the hypotheses, here we detail (i) the latent constructs and corresponding reflective indicators, (ii) the one-way directional paths for H1–H7, and (iii) the estimation plan and identification decisions that align with PLS-SEM. Placing the conceptual model into the methodology explains how theoretical conjectures get translated into the measurement and structural specifications and offers an transparent connection to the procedures employed on reliability, validity, and hypothesis assessment.

With the shifting paradigm of digital tourism marketing, heritage sites are increasingly struggling to create emotionally appealing and cognitively memorable content that will appeal to prospective visitors. Traditional informative approaches are no longer sufficient; instead, visually appealing, narratively appealing, and culturally authentic promotion techniques are needed to generate significant connections and behavioral responses [61,62,63]. In spite of increased scholarly focus on sensory and narrative characteristics in tourist communications, there is a missing link to understanding how these characteristics individually, or as a set, affect internal psychological processes of emotional involvement and memorability of destinations and how these affect behavioral tendencies like word-of-mouth. This research fills this void by creating and empirically validating a Structural Equation Model (SEM) based on the Stimulus–Organism–Response (S-O-R) theory that describes how external media stimuli elicit internal affective and cognitive responses that direct behavioral outcomes [64,65,66].

The theoretical model identifies three external drivers of interest in heritage destination promotion: visual sensory appeal, narrative immersion, and perceived authenticity. The constructs are complementary but differing design elements employed consistently in destination branding and storytelling. Visual sensory appeal refers to the perceived richness, coherence, and aesthetic beauty of visual components of imagery, color scheme, and composition. Past research has already shown that visually engaging content heightens attention and affective arousal [67,68], which are also the precursory requirements for engagement and encoding. Under heritage tourism, sensory attractiveness can lead to the experience of being amazed and sentimental when visual attributes illustrate historical or cultural richness.

Narrative immersion refers to the degree to which people are psychologically and emotionally engaged in the story or information presented by the advertising message. Aligned with narrative transportation theory [69,70], the scale has been proven to impact attitudes and intention through identification, empathy, and vicarious experience. In the case of tourism situations, narrative immersion enables potential visitors to project themselves within the destination environment and thus enhance affective engagement and increase the chance of long-term memory [1,2,3].

Perceived authenticity, the third stimulus variable, is the extent to which the destination is presented as being authentic, culturally realistic, and uncommercialized. Authenticity has been shown to be a predictor of trust and affective attachment in heritage tourism [6,7,11]. Travelers are more likely to develop strong affective bonds with material they perceive to be authentic, respectful of cultural values, and uncommercialized. Realistic portrayals not only strengthen affective reactions but also credibility, which can further contribute to inducing positive cognitive and behavioral impacts.

These external stimuli are posited to impact two internal organism-level processes: emotional engagement and destination memory. Emotional engagement captures the affective reactions—be it joy, awe, inspiration, or nostalgia—created by the promotional material. Emotions are key to the formation of tourist experience and choice. Hosany et al. [51] showed that emotional intensity during the first exposure to tourism communication strongly predicts memory strength and behavioral intention. Since emotion strengthens attention, encoding, and retrieval, emotional engagement is hypothesized to be a key mediator in the pathway between stimulus cues and downstream consequences.

Destination memory refers to the vividness, clarity, and retrievability of the mental image created of the destination. Tourism memory is a crucial precursor of future behavior; destinations that are well-remembered are more likely to be revisited or recommended [67,68,69]. Earlier findings in cognitive psychology also corroborate that emotionally stimulating and personally relevant stimuli have a higher chance of being committed to long-term memory, particularly when supplemented with sensory and narrative prompts.

The final outcome in the model is intention to recommend, which is the intention behavior to promote or spread the destination via interpersonal or online word-of-mouth. In accordance with the theory of planned behavior, intention behaviors are a result of internal attitudes and perceptions, which here are developed on the basis of emotional and memory-based pathways. It is thus anticipated that both emotional engagement and destination memory will directly predict intention to recommend and also mediate the influence of the three external stimuli [17,20,71].

By combining these constructs into a single model, the research is advanced both theoretically and practically. Theoretically, it progresses the utilization of the S-O-R paradigm in heritage tourism promotion, illustrating how design and content factors invoke internal psychological responses that result in advocacy behavior. In practice, the research provides destination management and marketing guidance on how to create content that engages individuals not only to command their attention but to create meaningful emotional experiences and long-term memory—two of the major predictors of recommendation behavior. By doing this, the research connects visual storytelling to quantifiable visitor response, with a psychological basis for more efficient digital heritage marketing based on campaigns. Figure 1 illustrates the conceptual model developed in the present research and presents the hypothesized relationships between the external stimuli, internal mediators, and behavioral outcome.

3.2. Data Collection and Sampling

This research utilized a quantitative, cross-sectional design with an online self-completion survey to investigate the effects of sensory and narrative elements of digital heritage tourism marketing materials on tourists’ emotional arousal, memory construct, and word-of-mouth intention [72,73,74]. Sampling strategy and data collection were meticulously designed to be commensurate with the theory underpinning the research—Stimulus–Organism–Response (S-O-R)—and the psychometric demands of Structural Equation Modeling (SEM).

Purposive sampling was utilized, augmented with snowball sampling in an effort to access individuals who were most recently exposed to electronic promotion materials for cultural or heritage tourism sites [75,76,77,78]. Purposive sampling was determined to be the most suitable sampling in view of the particular experiential and perceptual needs of the study [75,77]. Participants were not drawn from the general population but from a specifically selected subset, that is, those with the most recent exposure to applicable tourist media. This method meant participants had the experiential and cognitive grounding necessary to meaningfully analyze such constructs as visual sensorial appeal, narrative interest, and perceived authenticity. Snowball sampling also increased the scope of the survey by allowing initial participants to pass the questionnaire around among their networks, in this case, people known to have travel, culture, or online media interests [77,78].

Inclusion criteria were as follows: participants must be above the age of 18, have adequate English or Greek language skills, and must state that they viewed digital content encouraging heritage or cultural tourism destinations in the last six months. Criteria in this fashion guaranteed that participants were capable of giving valid consent, understanding questionnaire material, and giving useful feedback to take into account their perceptual and emotional responses towards milieu of tourism media. Exclusion criteria were used in order to maximize data quality and conceptual relevance. Those who said they had no contact with heritage promotion content during the previous six months, those who did not pass attention check questions, those who completed the survey within less than two minutes, or those who demonstrated patterns of careless response sets such as straight-lining were excluded. These were the controls utilized in order to maximize response validity and internal reliability, especially since study constructs involved perceptions and affective nature.

Data were gathered through an online template questionnaire created using Google Forms for standardization, wide reach, and effective data capture. It was made available in English and Greek to facilitate access among a multicultural sample as well as enhance representativeness. Recruitment was conducted through social media platforms, academic mailing lists, and online tourist discussion forums. Each participant was given an opening page that included an informed consent form stating the purpose of the study, ethical protections, use of data, and that volunteering was not obligatory. Continued participation in the survey proper was possible only after electronic consent.

The survey questionnaire consisted of four sections: (1) eligibility and screening questions, (2) measurement items for the six SEM model constructs, (3) background and demographic questions, and (4) quality control and attention check items. All major constructs—visual sensory appeal, narrative immersion, perceived authenticity, emotional engagement, destination memory, and intention to recommend—were assessed using Likert-type items adapted from already validated scales from tourism, marketing, and consumer psychology literature.

There were 653 usable responses collected. The required sample size was computed based on SEM practice and principally the practice of Hair et al. [79], who suggest a minimum ratio of 10 observations for every parameter estimated within the model. With 30–24 items across six latent factors, at least 300 was considered adequate to provide statistical power, model stability, and practicability for multi-group analysis (e.g., by gender, age, or past travel experience) [80,81,82,83]. The sample size achieved was thus in excess of the minimum and adequate for sound SEM estimation and hypothesis testing [84]. Gender was assessed descriptively and at the multi-group level. For this wave, the survey employed a binary measure (male/female); thus, no participant was categorized outside the gender binary. We recognize that this operational definition of gender limits inference on gender diversity and may obscure non-binary realities. In the future, data collections will incorporate inclusive measures of gender (e.g., non-binary, self-identify, prefer not to answer; multi-select) and will account for continuous measures of gender identification where necessary.

For the purpose of guaranteeing instrument validity and reliability, the questionnaire was pilot-tested among 20 participants before using it at full capacity. Pilot feedback guided minimal changes in wording of items and interface design. Measurement items were developed from scales with known psychometric properties, and internal consistency was assessed using Cronbach’s alpha and Composite Reliability (CR). Convergent and discriminant validity were also assessed through Confirmatory Factor Analysis (CFA) with measures such as Average Variance Extracted (AVE) and the Heterotrait–Monotrait (HTMT) ratio.

All the academic research guidelines involving human subjects were strictly adhered to. The survey was completely anonymous, and there was no Personally Identifiable Information (PII) gathered. Electronic informed consent was gotten, and the participants were made aware that they had a right to withdraw at their convenience. The research followed the ethical guidelines of Democritus University and was in line with the General Data Protection Regulation (GDPR) regarding research data gathering and retaining. Vulnerable groups, including children or people with compromised consent capacity, were not targeted or engaged in the sampling.

In conclusion, the sampling and data gathering approach was clearly defined to facilitate the study validity, reliability, as well as ethical integrity, yet simultaneously remain consistent with the study’s theoretical underpinnings and the analytical goals.

3.3. Measurement Scales

All latent constructs were operationalized reflectively via items taken from validated existing scales and reworded in the context of digital heritage promotion (see Appendix A,

Table A1. Measurements used for data analysis.

Visual Sensory Appeal (VSA)
VSA1	The images used were visually appealing and immersive.	Yue Gong et al. [67]
VSA2	The colors and composition were attractive and well-balanced.
VSA3	The visuals captured my attention immediately.
VSA4	The overall design was aesthetically pleasing.
Narrative Immersion (NI)
NI1	The story behind the destination drew me in.	Lele Xue et al. [62]
NI2	I felt emotionally involved in the message of the promotion.
NI3	I imagined myself being at the destination.
NI4	The promotional content felt like a vivid narrative.
NI5	I was fully immersed in the message presented.
Perceived Authenticity (PA)
PA1	The content portrayed the destination’s heritage authentically.	Kolar et al. [85] and Park et al. [86]
PA2	The cultural elements seemed genuine and not staged.
PA3	The destination felt true to its identity.
PA4	The cultural aspects appeared realistic and believable.
PA5	The promotion felt sincere and trustworthy. (deleted)
Emotional Engagement (EE)
EE1	I felt emotionally connected to the destination.	Ahmed et al. [68]
EE2	I experienced joy while viewing the content.
EE3	I felt a sense of awe while watching or reading the material.
EE4	The promotional content moved me emotionally. (deleted)
Destination Memory (DM)
DM1	I can vividly recall the destination shown in the promotion.	Oh et al. [65] and Jorgenson et al. [87]
DM2	The destination details stayed in my mind after viewing.
DM3	I can clearly remember the visuals and message.
Intention to Recommend (IR)
IR1	I would recommend this destination to others.	Huang et al. [63]
IR2	I would encourage others to visit this location.
IR3	I feel confident suggesting this destination to others.

). For Visual Sensory Appeal (VSA), four items assessed visual richness perceived, coherence of composition, salience of attention, and overall aesthetic appeal (e.g., “The images were visually appealing and immersive”). Items were taken from measures of aesthetic/visual appeal (e.g., Yue Gong et al. [67]). For Narrative Immersion (NI), five items measured transportation and story immersion (e.g., “The story behind the destination drew me in”; “I imagined myself being at the destination”), adapted from narrative immersion/transportation scales (Lele Xue et al. [62]). For Perceived Authenticity (PA), four measured estimates of cultural authenticity and genuineness (e.g., “The cultural elements seemed genuine and not staged”) operationalized from heritage authenticity scales that distinguish object/constructive/existential types (Kolar et al. [85]; Park et al. [86]) were used. For Emotional Engagement (EE) items, three of them addressed affective connection and specific emotions (joy, awe) elicited by the content (e.g., “I felt emotionally connected to the destination”), according to affective engagement measures (Ahmed et al. [68]). One was excluded because of low loading (EE4). For Destination Memory (DM), three items assessed vividness, retention, and recall clarity of the destination being advertised (e.g., “I can vividly recall the destination shown in the promotion”). Three items were taken from tourism memorability and destination memory scales (Oh et al. [65]; Jorgenson et al. [87]). Finally, for Intention to Recommend (IR), three items assessed advocacy intentions (e.g., “I would recommend this destination to others”), drawn from recommendation/eWOM intention scales (Huang et al. [61]). All the components of each construct were combined to generate composite scores. Reliability and validity were determined before structural testing.

3.4. Sample Profile

No respondents were non-binary in the present data set. Results should not thus be generalized to non-binary populations; subsequent waves will have inclusive response options. Participants (n = 653) were 53.9% female (n = 352) and 46.1% male (n = 301), as presented in

Table 1. Sample profile.

		Frequency	Percentage
Gender	Female	352	53.9%
	Male	301	46.1%
Age	18–24	199	30.5%
	25–34	197	30.2%
	35–44	125	19.1%
	45–54	78	11.9%
	55+	54	8.3%
Highest Level of Education Completed	High school diploma	122	18.7%
	Undergraduate student	195	29.9%
	Bachelor’s degree	161	24.7%
	Master’s degree or higher	175	26.8%
In the past 6 months, have you seen any digital content (ads, videos, social media posts) promoting heritage or cultural tourism destinations?	No	375	57.4%
	Yes	221	33.8%
	Not sure	57	8.7%
Have you ever visited a heritage or cultural tourism site (e.g., archaeological site, historical town, museum)?	Yes	301	46.1%
	No	248	38.0%
	Not sure	104	15.9%
How often do you typically travel for cultural or heritage-related reasons?	Never	109	16.7%
	Rarely (once every few years)	176	27.0%
	Occasionally (once a year)	260	39.8%
	Frequently (2–3 times a year or more)	108	16.5%
How often do you use digital platforms (e.g., websites, YouTube, social media) to explore or plan cultural travel?	Never	94	14.4%
	Rarely	42	6.4%
	Sometimes	179	27.4%
	Often	147	22.5%
	Always	191	29.2%
To what extent are you personally interested in cultural heritage preservation or sustainable tourism?	Not at all interested	107	16.4%
	Slightly interested	163	25.0%
	Moderately interested	153	23.4%
	Very interested	106	16.2%
	Extremely interested	124	19.0%

. By age group, 30.5% were 18–24 years old (n = 199), 30.2% were 25–34 years old (n = 197), 19.1% were 35–44 years old (n = 125), 11.9% were 45–54 years old (n = 78), and 8.3% were 55+ (n = 54). Educationally, 18.7% (n = 122) possessed a high school diploma, 29.9% (n = 195) were undergraduate students, 24.7% (n = 161) possessed a bachelor’s degree, and 26.8% (n = 175) possessed a master’s degree or higher. Recent exposure to digital heritage/cultural promotions (past 6 months) was as follows: 33.8% (n = 221) said yes, 57.4% (n = 375) said no, and 8.7% (n = 57) did not know. Additionally, 46.1% had made a visit to heritage/cultural sites (n = 301), 38.0% (n = 248) had no previous visit, and visit status for 15.9% (n = 104) was unknown. Cultural/heritage tourism frequencies were never (16.7%, n = 109), rarely—once every few years (27.0%, n = 176), occasionally—approximately once a year (39.8%, n = 260), and often—2–3 times a year or more (16.5%, n = 108). Use of internet sites to visit/browse cultural tourism responses were never (14.4%, n = 94), rarely (6.4%, n = 42), sometimes (27.4%, n = 179), often (22.5%, n = 147), and always (29.2%, n = 191). Conservation of cultural heritage or ecotourism was of no interest (16.4%, n = 107), little interest (25.0%, n = 163), some interest (23.4%, n = 153), much interest (16.2%, n = 106), and a great deal of interest (19.0%, n = 124).

4. Data Analysis and Results

Data were treated with structural equation modeling using SmartPLS 4 (version 4.1.1.4). In accordance with Nitzl et al. [88], variance-based SEM is properly applied in business and social-science studies. PLS-SEM was utilized because it evaluates causal models on the basis of maximum variance explained in endogenous latent variables, thus highlighting predictive relevance [89]. To explore possible heterogeneity, Multi-Group Analysis (MGA) was used to equate structural paths across subpopulations and identify differences in context that conventional regression models will miss [89]. Estimation and testing were according to Wong’s [90] recommendations for computing path (beta) coefficients, standard errors, and reliability diagnostics. In reflective measurement models, indicator reliability should have sufficient convergence with a similar construct; minimum outer loadings of 0.70 were acceptable. This analytical approach was used to test the structural relationships rigorously while facilitating high-quality measurement, allowing for close inspection of the postulated mechanisms both within and across appropriate respondent groups in the sample.

4.1. Common Method Bias (CMB)

Potential Common Method Bias (CMB) was evaluated to determine the validity and reliability of the research according to steps developed by Podsakoff et al. [91]. Harman’s single-factor test was used to determine if one underlying factor explained most of the covariance structure. In principal factor analysis without rotation, it was revealed that the first factor explained only 30.726% of the overall variance—well below the traditional 50% rule—and suggests that CMB is not likely to contaminate the results. Through the suggestion of low CMB, direct identification and reporting enhance construct validity and the interconstruct relation demonstrated reliability by limiting worries about systematic measurement bias [91,92].

4.2. Measurement Model

The PLS-SEM process starts with reflective measurement model importance evaluation. In line with Hair et al. [79], assessment aimed at Composite Reliability (CR), indicator reliability, convergent validity, and discriminant validity to show sufficient psychometric quality prior to interpreting structural relationships. Following Vinzi et al.’s [93] recommendation, indicator reliability was conceptualized as the variability of an item that would be explained by its latent construct and was measured in terms of outer loadings. According to Wong [90] and Chin [94], loadings of 0.70 or higher were regarded as being sufficiently indicative of items’ quality. However, as Vinzi et al. [93] warn, social-science applications tend to return a set of indicators less than that threshold. Deletion decisions were thus not mechanistic but incrementally based on improvement in model quality: items were retained as long as retaining them enhanced CR and AVE and could be dropped only if dropping them had a substantial positive effect upon these indexes in order to prevent loss of potentially useful measures prematurely.

Based on the suggestions of Hair et al. [95], loadings of indicators between 0.40 and 0.70 were dropped only if this resulted in a considerably larger CR or AVE for the respective construct. After applying these criteria, and based on the decision-making rules of Gefen et al. [96], purification of the measurement model was obtained via dropping of two indicators—PA5 and EE4—with factor loadings less than 0.50. This parsimonious refinement, reported in

Table 2. Factor loading reliability and convergent validity.

Constructs	Items	Factor Loadings	Cronbach’s Alpha	rho_A	CR	AVE
Destination Memory	DM1	0.903	0.860	0.861	0.915	0.781
	DM2	0.873
	DM3	0.875
Emotional Engagement	EE1	0.696	0.614	0.696	0.785	0.553
	EE2	0.860
	EE3	0.659
Intention to Recommend	IR1	0.701	0.783	0.848	0.873	0.699
	IR2	0.905
	IR3	0.887
Narrative Immersion	NI1	0.860	0.817	0.829	0.857	0.549
	NI2	0.687
	NI3	0.755
	NI4	0.628
	NI5	0.752
Perceived Authenticity	PA1	0.887	0.902	0.904	0.932	0.774
	PA2	0.826
	PA3	0.909
	PA4	0.895
Visual Sensory Appeal	VSA1	0.770	0.793	0.794	0.865	0.616
	VSA2	0.809
	VSA3	0.788
	VSA4	0.772

, enhanced overall quality of measurement at the cost of no reduced construct coverage and thus established a justifiable platform for further estimation of the structural paths and hypothesis testing in the hypothesized S–O–R framework.

Reliability was measured by Cronbach’s alpha, rho_A, and Composite Reliability (CR). Following Wasko et al. [97], the 0.70 threshold was achieved for DM, EE, IR, NI, PA, and VSA; other constructs also showed moderate to high reliability, as in previous evidence [98,99]. Since the coefficient conceptually falls somewhere in between alpha and CR, rho_A was over 0.70 in the majority of the instances, supporting the reliability findings presented by Sarstedt et al. [100] and upholding the consistency requirements postulated by Henseler et al. [101].

Adequate convergent validity was considered since the Average Variance Extracted (AVE) was more than 0.50 for the majority of the constructs [102]. Fornell et al. [102] also observe that AVE values marginally less than 0.50 may still be acceptable when CR > 0.60, which is met where relevant to this study. Discriminant validity was also tested by means of the Fornell–Larcker criterion: each construct’s AVE square root was greater than its correlations with other constructs, which was found satisfactory. This result was further supported using the Heterotrait–Monotrait (HTMT) ratio, where every value was below the conservative 0.85 cutoff suggested by Henseler et al. [101]. Together, these diagnostics suggest sound construct validity and strong internal consistency throughout the measurement model.

Table 3. HTMT ratio.

	DM	EE	IR	NI	PA	VSA
DM
EE	0.532
IR	0.594	0.559
NI	0.200	0.155	0.101
PA	0.554	0.734	0.523	0.126
VSA	0.494	0.674	0.654	0.175	0.731

Note: This table presents all pairs of latent constructs’ HTMT ratios. HTMT values below the threshold value of 0.85 are indicative of acceptable discriminant validity. All values in this analysis satisfy this requirement, in support of the notion that each construct is empirically distinct.

and

Table 4. Fornell and Larcker criterion.

	DM	EE	IR	NI	PA	VSA
DM	0.884
EE	0.383	0.744
IR	0.493	0.439	0.836
NI	−0.207	−0.067	−0.066	0.741
PA	0.488	0.576	0.455	0.028	0.880
VSA	0.408	0.520	0.537	0.096	0.615	0.785

Note: Diagonal values (in bold) are square roots of AVE of each construct and must be larger than interconstruct correlations in corresponding rows and columns. This condition is true for all the constructs, and it implies discriminant validity in the measurement model.

present detailed indices of reliability and validity such as alpha, rho_A, CR, AVE, interconstruct correlations, and HTMT statistics and offer open documentation of the psychometric sufficiency that underlies subsequent structural analyses.

4.3. Structural Model

Reliability was determined by Cronbach’s alpha, rho_A, and Composite Reliability (CR). In line with Wasko et al. [97], the cut-off of 0.70 was achieved for DM, EE, IR, NI, PA, and VSA; moderate to high reliability was also achieved for the other constructs, as upheld by prior evidence [99,103]. Being a coefficient theoretically located between alpha and CR, rho_A exceeded 0.70 for most cases, supporting the reliability findings presented by Sarstedt et al. [100] and confirming the consistency criteria established by Kock et al. [83].

The structural model was assessed with coefficient of determination (R²), predictive relevance (Q²), and the significance of structural paths [79]. The model accounted for 31.2% of the variance in destination memory (R² = 0.312), 38.7% in emotional engagement (R² = 0.387), and 39.5% in intention to recommend (R² = 0.395), reflecting moderate explanatory power. Predictive relevance was also achieved, with cross-validated redundancy values of Q² = 0.302 for destination memory, Q² = 0.378 for emotional engagement, and Q² = 0.318 for intention to recommend, in line with moderate-to-strong out-of-sample prediction.

Hypotheses were examined to determine the statistical significance of construct relationships. Path estimates and standard errors were derived through non-parametric bootstrapping, as per Hair et al. [95]. Indirect effects were examined through a bias-corrected, one-tailed bootstrap procedure with 10,000 resamples, as described by Preacher et al. [104] and Streukens et al. [105]. In combination, these diagnostics provide evidence for the structural adequacy and predictive ability of the model. Complete estimates are presented in

Table 5. Hypothesis testing.

Hypothesis	Path	Coefficient (β)	SD	t-Value	p-Value	Results
H1	VSA → IR	0.334	0.040	8.402	0.000	Supported
H2	NI → IR	−0.032	0.034	0.945	0.172	Not Supported
H3	PA → IR	0.036	0.051	0.706	0.240	Not Supported
H4a	EE → IR	0.134	0.039	3.418	0.000	Supported
H4b	DM → IR	0.281	0.043	6.467	0.000	Supported

.

Direct effects for Intention to Recommend (IR) were estimated through PLS-SEM non-parametric bootstrapping (two-tailed, α = 0.05). Visual Sensory Appeal (VSA) was positively and significantly related to IR, β = 0.334, SE = 0.040, t = 8.402, p < 0.001, confirming H1. Of organismic variables, Emotional Engagement (EE) and Destination Memory (DM) were strong positive predictors of IR: EE → IR, β = 0.134, SE = 0.039, t = 3.418, p < 0.001 (H4a confirmed); DM → IR, β = 0.281, SE = 0.043, t = 6.467, p < 0.001 (H4b confirmed). Conversely, however, Narrative Immersion (NI) was not statistically significant in having a direct effect on IR, β = −0.032, SE = 0.034, t = 0.945, p = 0.172, and neither was Perceived Authenticity (PA), β = 0.036, SE = 0.051, t = 0.706, p = 0.240, so H2 and H3 cannot be confirmed. Overall, the pattern indicates that visually prominent content has the greatest direct impact on advocacy intentions, followed by destination memory and emotional involvement. Storytelling engagement and authenticity did not provide incremental explanation to IR when estimated together with these predictors. These results validate the framework to some extent by making visible design and organismic reactions leading drivers in guiding recommendation behavior.

4.4. Mediation Analysis Results

Mediation was confirmed in PLS-SEM through non-parametric bootstrapping, testing whether Emotional Engagement (EE) and Destination Memory (DM) mediate the impacts of Visual Sensory Appeal (VSA), Narrative Immersion (NI), and Perceived Authenticity (PA) on Intention to Recommend (IR). For the direct models, VSA was statistically significantly, positively correlated with IR, β = 0.334, SE = 0.040, t = 8.402, p < 0.001, confirming the argument that visually attractive content indirectly affects advocacy by promoting it directly. Conversely, NI and PA did not have significant direct effects on IR (NI: β = −0.032, SE = 0.034, t = 0.945, p = 0.172; PA: β = 0.036, SE = 0.051, t = 0.706, p = 0.240), meaning that, if they have an effect, it would be indirect. As to be expected, the direct specific indirect effects through EE and DM were all significant. For VSA, the indirect effects were VSA → EE → IR: β = 0.038, SE = 0.014, t = 2.775, p = 0.003, and VSA → DM → IR: β = 0.057, SE = 0.014, t = 3.953, p < 0.001; the total indirect effect (through both mediators) was β = 0.095, SE = 0.019, t = 4.920, p < 0.001. For NI, the indirect pathways were also large (NI → EE → IR: β = 0.014, SE = 0.005, t = 2.648, p = 0.004; NI → DM → IR: β = 0.066, SE = 0.015, t = 4.393, p < 0.001), and the overall indirect effect was β = 0.081, SE = 0.016, t = 4.975, p < 0.001. PA also caught up (PA → EE → IR: β = 0.055, SE = 0.017, t = 3.279, p = 0.001; PA → DM → IR: β = 0.104, SE = 0.022, t = 4.733, p < 0.001), with the overall indirect effect being β = 0.159, SE = 0.030, t = 5.205, p < 0.001 (

Table 6. Mediation analysis.

Hypothesis	Direct Effects	Coeff. (β)	SD	t-Value	p-Value	Results	Mediation Type
	VSA → IR	0.334	0.040	8.402	0.000
	NI → IR	−0.032	0.034	0.945	0.172
	PA → IR	0.036	0.051	0.706	0.240
	Total Effects	Coeff. (β)	SD	t-Value	p-Value
	NI → IR	0.081	0.016	4.975	0.000
	PA → IR	0.159	0.030	5.205	0.000
	VSA → IR	0.095	0.019	4.920	0.000
	Specific Indirect Effects	Coeff. (β)	SD	t-Value	p-Value
H5a	VSA → EE → IR	0.038	0.014	2.775	0.003	Supp.	Partial Mediation
H5b	VSA → DM → IR	0.057	0.014	3.953	0.000	Supp.	Partial Mediation
H6a	NI → EE → IR	0.014	0.005	2.648	0.004	Supp.	Full Mediation
H6b	NI → DM → IR	0.066	0.015	4.393	0.000	Supp.	Full Mediation
H7a	PA → EE → IR	0.055	0.017	3.279	0.001	Supp.	Full Mediation
H7b	PA → DM → IR	0.104	0.022	4.733	0.000	Supp.	Full Mediation

).

Across predictors, the DM route was always greater than the EE route, suggesting that memory consolidation is the dominant psychological route from stimuli to recommendation. Since VSA had large direct and also large indirect effects, partial mediation of the influence of VSA on IR by EE and DM is suggested by the findings. However, NI and PA had non-significant direct but significant indirect effects through the two mediators, as in complete mediation. Together, these findings imply that, while strong visual appeal has the ability to influence recommendation both directly and indirectly, narrative interest and authenticity both exert effects on advocacy mainly through the formation of emotional connection and, in particular, long-lasting destination representation.

4.5. Multi-Group Analysis (MGA)

Two-tailed multiple-group tests (α = 0.05) detected segment-specific stability in cue interpretation into advocacy; only significant differences are noted. By gender, women had stronger visual routes to recommendation—both direct (VSA → IR) and indirect through memory (VSA → DM)—and stronger authenticity influences on emotion and recommendation (PA → EE; PA → IR). Males showed greater VSA → EE, EE → IR, NI → IR, and PA → DM, reflecting greater use of affective conversion, narrative, and authenticity-to-memory processes.

Age trends also evidenced various levers: 18–29-year-olds applied mostly narrative immersion and memory (NI → IR; DM → IR), 25–34-year-olds evidenced the greatest PA → IR and greater NI → EE than 35–44-year-olds, and VSA → IR was greatest for 45–54-year-olds; 25–34-year-olds drove emotion to advocacy less strongly than neighboring cohorts (EE → IR weaker). Education reinforced distinct paths; compared to high-school-educated participants, bachelor’s participants reported higher authenticity-based internalization (PA → DM; PA → EE; NI → DM). High-school-educated participants, however, reported higher visual paths (VSA → DM/EE/IR) than bachelor’s and undergraduate participants. Unmediated authenticity impacts on recommendation rose with education (master’s+ > bachelor’s/high school), but the effect of high school was higher than that of master’s+ on VSA → DM/EE. Exposure to digital heritage promotion strengthened upstream and downstream relationships. When exposure was unsure (“Not sure”), the visual direct route (VSA → IR) and upstream NI → DM and PA → EE were stronger than those of non-exposed respondents. However, when exposed, conversion relationships to advocacy were stronger (DM → IR vs. Not sure; EE → IR vs. No). The NI → IR direct narrative route was weaker for exposed respondents compared to no/uncertain exposure groups. Cultural tourist frequency was also important. Very high-frequency cultural tourists returned lower NI → EE and VSA → DM but higher PA → DM scores, reflecting that experienced travelers are more concerned about authenticity than narrative or visual richness. For interest in preservation/sustainability of culture, very interested participants used memory less for advocacy purposes (DM → IR weaker than low interest) but were more motivated by authenticity (PA → IR stronger than low and moderate interest). Low-interest participants depended more on visual and memory-based processes (VSA → EE/DM and PA → DM more for low than moderate interest). Previous heritage tourism experience also polarized responses even more: uncertain experience strengthened VSA → DM, PA → EE, and PA → IR over the experienced group, while experienced respondents were more robust on VSA → EE than the uncertain group. Lastly, use of technology in trip planning distinguished paths: high tech users showed more PA → EE and PA → IR but less VSA → EE and EE → IR than low users; they also showed more NI → IR than low and moderate users, indicating that tech-knowledgeable planners map story and authenticity more directly to recommendation, while low tech users apply visual affect and emotion-to-advocacy mapping more. All statistically significant between-group differences (two-tailed, α = 0.05) are summarized in

Table 7. Significant MGA results with group comparisons.

Path	Group Comparison	Difference (Δβ)	p-Value
EE → IR	Female vs. Male	−0.285	0.000
VSA → DM	Female vs. Male	0.367	0.000
NI → IR	Female vs. Male	−0.387	0.001
PA → IR	Female vs. Male	0.322	0.001
VSA → EE	Female vs. Male	−0.261	0.002
PA → EE	Female vs. Male	0.244	0.003
VSA → IR	Female vs. Male	0.212	0.007
PA → DM	Female vs. Male	−0.186	0.017
PA → IR	18–29 vs. 25–34	−0.485	0.000
PA → IR	25–34 vs. 35–44	0.384	0.003
PA → IR	25–34 vs. 45–54	0.394	0.016
EE → IR	18–29 vs. 25–34	0.263	0.002
EE → IR	25–34 vs. 35–44	−0.396	0.002
DM → IR	18–29 vs. 25–34	0.216	0.011
DM → IR	18–29 vs. 35–44	0.263	0.008
NI → IR	18–29 vs. 25–34	0.244	0.035
NI → IR	18–29 vs. 45–54	0.416	0.021
VSA → IR	18–29 vs. 45–54	−0.290	0.035
VSA → IR	25–34 vs. 35–44	0.237	0.019
VSA → IR	35–44 vs. 45–54	−0.369	0.012
NI → EE	25–34 vs. 35–44	0.199	0.025
PA → DM	Bachelor’s vs. High school	0.396	0.000
PA → DM	Bachelor’s vs. Master’s+	0.283	0.019
PA → DM	High school vs. Undergraduate	−0.287	0.009
PA → EE	Bachelor’s vs. High school	0.406	0.001
PA → EE	High school vs. Master’s+	−0.349	0.004
PA → EE	High school vs. Undergraduate	−0.299	0.013
VSA → DM	Bachelor’s vs. High school	−0.304	0.003
VSA → DM	High school vs. Master’s+	0.222	0.034
VSA → DM	High school vs. Undergraduate	0.211	0.033
VSA → EE	Bachelor’s vs. High school	−0.256	0.011
VSA → EE	High school vs. Master’s+	0.213	0.038
VSA → EE	High school vs. Undergraduate	0.238	0.030
NI → DM	Bachelor’s vs. High school	0.196	0.025
VSA → IR	Bachelor’s vs. High school	−0.238	0.036
VSA → IR	High school vs. Undergraduate	0.240	0.026
PA → IR	Bachelor’s vs. Master’s+	−0.348	0.020
PA → IR	High school vs. Master’s+	−0.561	0.000
PA → IR	High school vs. Undergraduate	−0.392	0.000
VSA → IR	Exposure heritage destination promotions No vs. Not sure	−0.313	0.011
VSA → IR	Exposure heritage destination promotions No vs. Yes	0.179	0.013
VSA → IR	Exposure heritage destination promotions Not sure vs. Yes	0.492	0.001
NI → DM	Exposure heritage destination promotions No vs. Not sure	−0.277	0.018
NI → DM	Exposure heritage destination promotions Not sure vs. Yes	0.321	0.009
PA → EE	Exposure heritage destination promotions Not sure vs. Yes	0.381	0.031
DM → IR	Exposure heritage destination promotions Not sure vs. Yes	−0.209	0.038
NI → IR	Exposure heritage destination promotions No vs. Yes	0.128	0.048
NI → IR	Exposure heritage destination promotions Not sure vs. Yes	0.263	0.031
EE → IR	Exposure heritage destination promotions No vs. Yes	−0.263	0.001
NI → EE	Cultural-travel frequency High vs. Low	−0.156	0.005
PA → DM	Cultural-travel frequency High vs. Low	0.210	0.010
VSA → DM	Cultural-travel frequency High vs. Low	−0.165	0.032
DM → IR	Interest in cultural preservation/sustainability High vs. Low	−0.213	0.014
PA → IR	Interest in cultural preservation/sustainability High vs. Low	0.235	0.014
PA → IR	Interest in cultural preservation/sustainability High vs. Moderate	0.261	0.017
VSA → DM	Interest in cultural preservation/sustainability Low vs. Moderate	−0.263	0.007
VSA → EE	Interest in cultural preservation/sustainability Low vs. Moderate	0.317	0.003
PA → DM	Interest in cultural preservation/sustainability Low vs. Moderate	0.187	0.049
VSA → DM	Prior experience with heritage tourism No vs. Not sure	−0.209	0.038
VSA → DM	Prior experience with heritage tourism Not sure vs. Yes	0.196	0.042
PA → EE	Prior experience with heritage tourism Not sure vs. Yes	0.311	0.002
PA → IR	Prior experience with heritage tourism Not sure vs. Yes	0.275	0.018
VSA → EE	Prior experience with heritage tourism Not sure vs. Yes	−0.189	0.042
PA → EE	Technology use for tourism planning High vs. Low	0.404	0.000
PA → EE	Technology use for tourism planning Low vs. Moderate	−0.479	0.000
VSA → EE	Technology use for tourism planning High vs. Low	−0.311	0.001
VSA → EE	Technology use for tourism planning Low vs. Moderate	0.356	0.003
PA → IR	Technology use for tourism planning High vs. Low	0.264	0.021
EE → IR	Technology use for tourism planning High vs. Low	−0.237	0.027
NI → IR	Technology use for tourism planning High vs. Low	0.206	0.027
NI → IR	Technology use for tourism planning High vs. Moderate	0.363	0.000

Note: This table reports only the statistically significant differences in structural path coefficients (Δβ) between groups.

; any paths not listed were non-significant.

5. Discussion

The direct-effects model accounts for which organismic and communicative factors best predict Intention to Recommend (IR) in digital heritage advertising. There were three structural paths that were significant: visual sensory appeal (VSA → IR), emotional engagement (EE → IR), and destination memory (DM → IR). Narrative engagement and perceived authenticity did not have significant direct predictions of IR when estimated as a set of predictors. The trend provides pale hints about operations of Stimulus–Organism–Response (S–O–R) processes in heritage communication and agrees with recent allegations of the dominance of visual, affective, mnemonic, and narrative mechanisms.

5.1. Direct Relationships

5.1.1. Visual Design as a Proximal Driver of Advocacy

The significant positive influence of VSA on IR (β = 0.334, p < 0.001) indicates that visual design is a good and direct predictor of advocacy intent. This is in line with the underlying S–O–R paradigm, with sensorially rich, aesthetically homogenous stimuli eliciting approach tendencies without requiring prolonged contemplation. Visual design therefore is not only an antecedent to attention but is itself an influence.

This finding is consistent with research in cultural tourism where eye tracking identified compositional balance, color contrast, and ornamentation cues to be determining gaze patterns and intention to behave. Likewise, experimental research on heritage VR and AR environments holds that aesthetic design and visually prominent cues enhance visit attitudes when interface quality is high. Neuromarketing data support this view: aesthetic coherence is found to increase processing fluency and pleasure responses, both related to positive judgments.

For heritage representation, the implication is self-evident: high-fidelity visualization, proportionate proportions, and focal salience cannot be utilized as superficial styling but as inevitable drivers of influence. Advertising campaigns to incorporate visually optimized content—photography, graphics, or digital narratives—are set to create advocacy. Relative to more muted signs of narrative or authenticity, visual composition is overt, cross-cultural, and less dependent on knowledge of or context about culture and is therefore especially well-suited to attention-limited digital spaces.

5.1.2. Emotion and Memory as Organismic Levers of Intention

EE and DM were both strong positive predictors of IR, with DM having a stronger coefficient (β = 0.281, p < 0.001) than EE (β = 0.134, p < 0.001). The two-way effect is in line with enduring evidence in tourism and psychology. Emotion is a primary determinant of heightened encoding and retrieval processes, which, in turn, influence destination memory. Memorable Tourism Experience (MTE) studies always place affect and memory as antecedents to loyalty, recommendation, and electronic word-of-mouth.

The more robust DM → IR process highlights the point that, although emotional arousal is necessary, it is the transfer of that arousal into stable accessible memory that best facilitates advocacy. In line with [51], who showed that affect intensity alone is not able to ensure loyalty unless combined with schemas of memory, emotional involvement is the spark but memory consolidation drives advocacy over time. For heritage marketing, the operating effect is two-way: (a) create experiences that elicit authentic emotional resonance—via evocative visualizations, symbolic hints, or story-telling snippets—and (b) make certain that those instances are designed to maximize memorability. Unforgettable moments, narrative anchors, and temporal cues can serve as mnemonic devices that burn the experience into long-term memory. When later deciding whether or not to recommend, it is those affectively charged, vividly remembered shards that are most likely to resurface.

5.1.3. Why Narrative Immersion and Authenticity Were Not Directly Predictive

To our surprise, neither NI nor PA was significantly directly predictive of IR in terms of VSA, EE, and DM. This counterintuitive result is consistent with theory-based work. Narrative immersion works indirectly: in engaging a reader or viewer in the world of a narrative, it creates emotion, significance, and presence that indirectly influence judgments. Meta-analyses indicate that immersion effects are not to be counted upon once affective resonance and clarity have been controlled for. Narrative design would therefore be most important as a facilitator of emotional engagement but not an independent cause.

Similarly, authenticity has long been a thorn in the side of tourism scholarship. Though perceived authenticity has been shown to enhance satisfaction, attachment, and memorability, its role in affecting recommendation is context-dependent. It has been emphasized that authenticity speaks differently to diverse expectations among audiences and cultural contexts; what is “authentic” to one tourist might be disappointing or, worse, irrelevant to another. In this model, after emotion and memory were added, the incremental impact of authenticity on IR went down, demonstrating that it indirectly influences advocacy by bringing emotional strength and memorability. This is a more precise theoretical reading: NI and PA are not to be jettisoned but relocated as antecedents or moderators within more general affective–mnemonic processes, not as blanket direct predictors of campaign activity.

These results advance S–O–R theory in heritage communication in three key respects. They first again confirm the existence of a direct stimulus-level pathway from visual design to advocacy, underscoring perceptual cues’ persuasive efficacy regardless of following higher-order judgments. They secondly again confirm emotional engagement and memory prioritization as organismic processes that make stimuli behaviorally effective, aligned with MTE perspectives and neurocognitive theories of affect–memory interaction. Third, they qualify the generally expected primacy of authenticity and immersion, placing their effects within mediated as opposed to immediate channels.

Complementarily, these findings reconcile conflicting findings in authenticity and immersion studies. They imply that outcomes are less reliant on intensity of immersion or “look and feel” authenticity but rather rely on the extent to which design solutions are supportive of emotional engagement and memory formation. Through heightening the theoretical boundary conditions of what most influentially impacts recommendation, this research disseminates a more careful comprehension of persuasion processes in digital heritage promotion.

While the research opportunistically drew on a Greek, bilingual (Greek/English) sample and cited heritage promotions typically rather than a single site, the interpretive paradigms of the respondents remain predominantly Hellenic/European. Estimated path strengths can thus be interpreted as locally anchored and the underlying S–O–R mechanism theoretically transportable. Cross-cultural work suggests that aesthetic appeal by vision, story conventions, and authenticity standards are culture-coded; what is “balanced composition,” an effective story arc, or plausible provenance can vary by regime of heritage (tangible vs. intangible), religious–secular domain, and media infrastructures. To evaluate scope conditions, future research will need to employ multi-country sampling with back-translation and expert evaluation and formally test configural/metric/scalar measurement invariance before making structure path comparisons. Moderation-mediator designs can examine if cultural orientation (e.g., collectivism–individualism), religiosity, or usage patterns on platforms moderate the roles of authenticity and story. Replications across regions and diasporic subgroups—as well as across campaign formats (short-form video, carousels, AR overlays)—would determine which effects act universally and which culture-specifically, thus securing external validity without stretching inferences beyond a single-nation case study.

5.2. Mediation Analysis

Mediation analysis proves the psychological processes by which communicative signals in online heritage environments exert their influence on Intention to Recommend (IR). Emotional Engagement (EE) and Destination Memory (DM) were always the leading mediators to describe how meaning-based and design aspects are received by the audience prior to their emergence as advocacy behavior. Collectively, these findings point to a twofold organismic process: emotion is the short-term affective link that engages people with the heritage message, while memory consolidation is the longer-term cognitive anchor that is the basis of recommendation behavior.

5.2.1. Partial Mediation of Visual Sensory Appeal

Visual Sensory Appeal (VSA) had both direct and indirect influences on IR, with significant mediation through both EE and DM. That is, VSA → EE → IR (β = 0.038, p = 0.003) and VSA → DM → IR (β = 0.057, p < 0.001) both played a role, with the DM path having greater explanatory power. This twin pattern indicates that evocative images serve a double role: they are immediate stimuli that instigate approach behavior immediately and also instigate internal states that heighten emotional resonance and memorability.

Previous research in visual persuasion and consumer psychology supports this result. Ref. [2] illustrated that aesthetic appeal stimulates brain areas related to reward and emotional processing, offering a neuroscientific explanation for visual persuasive influence. In heritage settings, it has validated that compositional balance and ornamental richness had a direct impact on tourists’ behavioral intentions in the Historic Centre of Macau. Likewise, studies of virtual and augmented reality tourism experiences indicate that visually coherent design facilitates both affective engagement and place experience memorability.

Practically, what these results highlight is that heritage campaigns need to center on visual fidelity and consistency. Noble photography, proportionally balanced composition, and focal salience are not issues of aesthetic finish but are strategic commitments of advocacy. By enhancing immediate affect and resilient memory, visual design is a central mover in making exposure into recommendation.

5.2.2. Full Mediation of Narrative Immersion

Narrative Immersion (NI) did not have a direct effect on IR but exerted indirect and powerful effects through EE and DM. NI → EE → IR (β = 0.014, p = 0.004) and NI → DM → IR (β = 0.066, p < 0.001) affirm that narratives indirectly impact advocacy through emotional connection and hard-wired representations of experience and place. This phenomenon fits with narrative transportation theory, which posits that stories convince not through explicit guidance of audience members but through the implantation of people into a virtual world that makes room for affect and imagery.

The benefit of the DM pathway aligns with studies that suggest that stories are more plausible when they produce persistent cognitive residues. Research has illustrated how stories produce self-brand connections through the insertion of brand information into autobiographical memory frameworks. In tourism, storytelling has been proved to increase memorability because it allows for the provision of a temporal and symbolic framework of interpreting heritage sites.

In practice, what it means is that immersion should not be sought for its own sake—sensory presence alone might not be enough to compel advocacy. Heritage campaigns, in contrast, need to have stories that capture people’s hearts and incorporate specific details that get etched into memory. These involve situating grounding experiences in symbolic events, incorporating personalized testimonies, or linking heritage stories to broader cultural values that can be remembered and shared.

5.2.3. Full Mediation of Authenticity

Perceived Authenticity (PA) had no direct influence on IR but had very strong indirect effects via EE (β = 0.055, p = 0.001) and, importantly, via DM (β = 0.104, p < 0.001). Thus, authenticity has influence on advocacy essentially when inscribed as long-lasting memory traces. This profile is consistent with studies interpreting authenticity as increasing satisfaction, attachment, and recall more than behavior per se and with appraisals that vary by context: effects become more pronounced where cues conform to visitors’ cultural schemas and diminish where a mismatch occurs. Here, authenticity serves as a contextual judgment that inputs on the basis of the subject’s “feel” of experiences and endures more than as a driving variable to commitment via recommendation. In S–O–R terms, PA (and NI) operate via an organismic state, with memory the overall path to IR, consistent with data that memories of experiences lead to fidelity and advocacy. In practice, authenticity cues need to be written into memory—scaffolded, affectively engaging text; bare descriptions come to have persuasive influence where narrative and design ease encoding supportive of memory.

5.3. Multi-Group Analysis (MGA) Results

MGA demonstrates that the same stimuli take on varying organismic pathways within segments of an audience—a fundamental S–O–R assumption. In line with S–O–R, there were group differences in paths. Women showed stronger VSA—direct and via memory—and stronger authenticity-to-emotion/recommendation (PA → EE; VSA → IR), and men were stronger in affective transfer and authenticity-to-memory (VSA → EE; EE → IR; NI → IR; PA → DM). Overall, 18–29-year-olds preferred narrative and memory (higher NI → IR; DM → IR), 25–34-year-olds had the highest PA → IR and stronger NI → EE than 35–44-year-olds, and 45–54-year-olds showed the strongest visual direct path (VSA → IR). In terms of education, participants with a bachelor’s+ drew on authenticity-based internalization (PA → DM; PA → EE; NI → DM), whereas high-school-educated responders relied on visual routes (VSA → DM/EE/IR); direct PA → IR increased with education. Exposure made a difference: when “Not sure,” VSA → IR, NI → DM, and PA → EE became stronger, whereas among exposed subjects, DM → IR and EE → IR prevailed and NI → IR weakened. Experience with high cultural-travel frequency weakened NI → EE and VSA → DM but strengthened PA → DM. High preservation interest decreased DM → IR but increased PA → IR; low interest strengthened visual/memory routes (VSA → EE/DM; PA → DM). For experience (where prior heritage exposure status was unclear), the VSA → DM and PA → EE/IR pathways were more pronounced, and experience strengthened the VSA → EE. Tech-active planners exhibited higher PA → EE, PA → IR, and NI → IR and lower VSA → EE and EE → IR, reflecting a stronger direct transfer of narrative/authenticity, whereas low tech users relied more on visual affect and emotion-to-action translation.

Theoretically, MGA accounts for why group effects of “immersion” and authenticity seem to be at odds: their impact is segment-constrained and is achieved primarily through emotion and memory. In application, segmentation is relevant: stress-balanced, high-fidelity imagery to female, mature age, less-educated, low-interest, low-tech, and uncertain-experience segments (Macau-style proportionality, focal salience); promote authenticity (embodied in mnemonic scaffolds) to male, infrequent traveler, more-educated, and high-interest/highly tech-active segments; and design story formats to seed salient anchors for young age groups. Follow-up research should pre-register MGA contrasts, adjust for multiple testing, and test moderated mediation (e.g., S–O–R with MTE/memory as conditional mediators), preferably in longitudinal or two-wave designs (cf. Cao et al. [33]), along with eye tracking and affect measures to track pathway dynamics from exposure through long-term memory and advocacy.

6. Practical Implications

The research aimed to reveal the impact of communicable design cues such as visual sensory appeal, narrative interest, and perceived authenticity on recommendation intention by means of affective engagement and destination memory. The findings confirm a sequential pattern of influence: visuals influenced advocacy directly, while narrative and authenticity were mainly engaged via emotion and in particular memory. Multi-group analyses also indicate that these paths vary significantly by gender, age, education, prior exposure and experience, interest in sustainability, frequency of cultural travel, and technology use. These findings correspond to the following practical implications for policy, management, and education.

6.1. Policymakers and Destination Governance

Experience-quality measures need to be mapped against the mechanisms that have been revealed in the study. Stakeholders, excluding arrival records, need to track affective engagement metrics (e.g., dwell time on interpretive media, sentiment in visitor feedback) and mnemonic results (e.g., aided recall of core heritage significations, follow-on eWOM/recommendation rates). Funding and procurement assessment needs to reward projects that incorporate high-fidelity visual communication and interpretation for memory (tidy narrative structures, symbolic anchors, then-and-now framings). Authenticity standards must be incorporated into permits and grants through the development of co-created on-site and virtual stories with community custodians and authenticating claims of authenticity using contextual information and provenance. Multi-sensory, inclusive design—tactile models, media captions, audio description, and thoughtful soundscapes—must be taken up to make emotional and memory paths accessible to older adults and those with mobility or sensory impairment. Finally, segment-sensitive stewardship should inform demand management: since visually oriented advocacy is strongest among given segments (e.g., age 45–54, lower formal education, little recent exposure), visually dominant campaigns can redistribute visitation to less visited places, and authenticity-rich, narrative content can increase visitation at sensitive flagship destinations.

6.2. Destination Managers and Heritage Marketers

Heritage marketers and destination managers need to design content to do two things at the same time: (1) encourage front-end involvement with compositional balance, clear foci, and material specificity and (2) help long-term recall with distinctive narrative moments (rituals, craft sequences, disputed histories), considered spatial sequencing, and temporal signposts. Because memory drives recommendation, include “sticky” cues—signature viewpoints, symbolic objects, brief narrative arcs—and repeat them at locations with wayfinding and micro-interpretation.

One strategy is to adopt segment-specific creative tactics in accordance with the MGA. For women, highlight proven authenticity among sophisticated visual content in an attempt to enhance direct and memory-based persuasion; for men, highlight affect-rich narrative and behind-the-scenes proofs that crystallize in memory. For 18–29-year-olds, introduce serial micro-stories, co-created UGC assignments, AR notifications, and post-noticing rehearsal queries; for 25–34-year-olds, emphasize provenance and conservation efforts via narrated stories that continually generate emotions and verbal share invitations; for 45–54-year-olds, offer high-impact imagery (panoramas, close-ups with texture) with abbreviated copy. In terms of education, stay visually oriented and jargon-free for lower formal education readers; provide more curatorial context setting and research linkage for bachelor’s/master’s-level segments.

Prioritize aligning content against prior exposure. When unknown or exposed, lead with high-contrast visual and linear storytelling to sow emotion and memory. When already exposed, deploy memory-activation assets (nostalgia cues, “remember this” reminders, reawakening user memories) and explicit recommendation signals, since emotional affinity and destination memory convert more powerfully into advocacy here.

In addition, focus should be given to tailoring to experience and orientation. For return cultural visitors, reduce high-gloss images and apply authenticity-to-memory strategies (time-lapse conservation, expert interviews, craft demonstrations). For intermittent visitors, preserve strong visual scaffolds that support encoding. For sustainability-oriented visitors, highlight authenticity and stewardship (transparency in conservation activity, benefit sharing with the community).

It is important to account for technology usage in planning. Tech-planners are more sensitive to story and authenticity and less to mere visual presence—so provide longer narratives, provenance chronologies, and interactively mapped content. Low tech users are more sensitive to imagery and emotion-to-action path maps, such as short, image-driven assets with clear, easy-to-send prompts.

6.3. Site Operations and Interpretation

Implement the S–O–R framework on site by choreographing visitor flow through story inspirations to high-salience viewing points or experiences. Employ AR “then-and-now” overlays, specially crafted soundscapes (e.g., liturgy, sounds of a market, workshop din), and short hands-on exercises to elicit affect and increase encoding. Offer memory aids, digital postcards, stamp trails, and collectable micro-stories to facilitate rehearsal and subsequent eWOM. In brief, evidentiary tokens of truthfulness (dates, materials, interventions, voices of the community) with one powerful image or artifact detail to connect appraisal and feeling. Include accessibility at the beginning, seating at story nodes, readable type, flat signage, and multi-modal interpretation (text, sound, touch) so emotional and mnemonic routes are accessible to all. Position content against previous exposure: for a new/exposed audience, employ visual salience and clear narrative to create emotion and memory; with repeat viewers, position memory-reactivation assets (nostalgia triggers, “remember this” language, reactivated user content) and overt share requests first, since the impact of emotion/memory on advocacy equivalence is greater. Differentiate by experience and orientation: returning cultural visitors register authenticity more easily than narrative or visual excess—use authenticity-to-memory tools (conservation time-lapses, craft demonstrations, curator interviews); sustainability-oriented visitors are most receptive to credible authenticity—highlight transparent conservation processes, community benefit sharing, and protection of intangible heritage.

6.4. Education, Capacity Building, and Partnerships

We suggest capacity development for staff and partners (creators, guides, agencies) to design and assess content in emotion- and memory-based metrics. Integrate light A/B testing into production processes (visual arrangement options; presence/absence of authenticity markers) and post-consumption recall and intention-to-share as key performance measures. Co-create with school and community organizations to facilitate authenticity and diversification of narrative repertoires and deliberately incorporate intangible heritage and under-represented voices. We also call for the practice of moral social media that draws on perceived “social return” without coercion—edit joint moments, generate open opt-in anchors, and safeguard privacy and cultural respect. At the site, apply the same S–O–R logic by structuring flows through brief story nodes to high-salience interaction; apply AR overlays, soft soundscapes, and brief hands-on exercises to induce affect and enhance encoding. Interleave brief markers of authenticity (dates, materials, interventions, voices of the community) with a single suggestive image or artifact detail to bridge appraisal and feeling. Then plan for accessibility from the beginning—seating at points of significance, flat-wayfinding, readable typography, and multi-modal interpretation—so the feeling in the memory loop is accessible to everyone who comes through and the potential advocacy pool is broadened.

6.5. Monitoring and Continuous Improvement

We suggest a closed-loop analytics approach connecting inputs (visual, narrative, authenticity cues) to organismic proxies (engagement, sentiment) and to downstream behavioral outcomes (saves, shares, reviews). Segment-level tracking should inform iterative creative optimization of the pathways most robust for each audience, e.g., visual-driven creative for 45–54-year-old adults and authenticity-driven creative for sustainability-focused visitors. We also note the limited generalizability of cross-sectional, self-report data. New assets, before scale-up, ought to be tested in small-batch studies (e.g., eye tracking, speeded recognition/recall tests) to ensure that intended emotions in memory mechanisms are engaged in the target segment. This mechanism-based testing, being rigorous, connects design decisions to advocacy outcomes and the long-term engagement of heritage sites.

7. Conclusions, Limitations, and Future Directions

This research aimed to clarify how designable communicative cues in digital heritage marketing—visual sensory appeal, engagement in storytelling, and perceived authenticity—are manifested in recommendation intention through organismic states of affective engagement and destination memory. Based on an S–O–R grounded PLS-SEM framework, the following three findings are pivotal. First, visual design has a robust direct impact on advocacy (VSA → IR), establishing that high-fidelity, visually consistent images are able to cause approach behavior without the necessity of high-level appraisal. Second, emotion and memory are both predictors of recommendation, with destination memory more strongly predicting it (DM → IR > EE → IR), suggesting that that which is retained in memory best predicts advocacy. Third, narrative immersion and authenticity are not direct causes of recommendation once visual and organismic factors are controlled; rather, both function indirectly and collectively through emotion and memory. Mediation tests showed partial mediation for VSA (visuals work directly and through EE/DM) and full mediation for NI and PA (their influence on IR is mediated through EE and particularly DM). Multi-group tests also exhibited systematic heterogeneity by gender, age, education, exposure, frequency of travel, sustainability orientation, previous heritage experience, and use of technology, suggesting segment-sensitive approaches.

This pattern collectively enhances S–O–R accounts of heritage communication: design decisions are worth their own while (visuals on advocacy), but the most generalizable story for narrative and authenticity markers is emotion to memory to recommendation. This reconciles conflicting findings in immersive technology and authenticity research by demonstrating that results are less a product of hardware intensity or authenticity “claims” per se and more a product of the content’s ability to engage emotions and remain in strong memory.

Some fruitful avenues can expand and develop the current contribution yet further. In the first place, the model’s stimulus–organism–response reasoning and the highlighted emotion to memory to recommendation sequence suggest causal research that manipulates communicative cues and observes short- and longer-term effects. Since the sampling frame consisted mainly of Greek participants who were bilingual, salience of visual balance, narrative conventions, and authenticity norms probably reflects local cultural repertoires and platform ecologies. Instead of being a limitation, it provides a baseline against which cross-cultural extensions can be made: follow-up studies can employ multi-country sampling, back-translation, and expert panels and formally test measurement invariance (configural/metric/scalar) prior to comparing path strengths. Adding cultural value orientations (e.g., collectivism–individualism), religiosity, and media-use patterns as moderators will draw scope conditions of the emotion → memory mechanism across heritage regimes (tangible versus intangible), sacred–secular settings, and diasporic publics. Controlled tests (lab, online A/B testing, and in situ field experiments) can rigidly manipulate visual structure (e.g., proportional balance, focal salience), narrative organization (e.g., transportation devices, pacing), and authenticity framings (object-based vs. existential cues) and introduce multi-sensory layers—i.e., soundscapes and spatialized audio—to test the hypothesis that sonic affordances facilitate emotional resonance and mnemonic consolidation in heritage environments. Second, consolidation mechanisms are best characterized with temporal designs. Experience-sampling or panel studies around bursts of campaign activity can index rehearsal and decay effects (e.g., recap reels, “memory hooks”), demystifying when initial affect crystallizes into destination memory and when reminders optimally trade off memory for advocacy. Staggered follow-ups (e.g., 24 h/1 week/1 month) would follow the emotion to memory path and determine optimal “re-engagement” windows. Third, avoiding single-method self-reports will make inferences tighter. Multi-method measurement—attention eye tracking, arousal psychophysiology (EDA/HR), appraisal implicit association tasks, and recognition/recall memory tasks—can be combined with objective digital traces (saves, shares, CTR, referral codes, eWOM volume) to calibrate psychological states against observable behavior. Hybrid designs combining platform analytics into experimental manipulations can determine if the same cues producing felt emotion and memory also produce real-world advocacy. In addition, theoretical insight into memory and authenticity can be obtained through head-to-head construct testing. In one design, contrast destination memory with other mnemonic constructs (e.g., MTE, autobiographical memory—rehearsal/impact) and define authenticity as both stimulus (designable cue) and appraisal (visitor judgment). This method specifies construct boundaries, tests for equivalence, and establishes which operationalization most fully predicts downstream effects in digital heritage communication. Moreover, in response to interpretable heterogeneity identified by multi-group analyses, subsequent research can establish formal measurement invariance and estimate moderated mediation to determine if certain segments evidence different trajectories. Gender, age groups, education, previous experience/exposure, interest in cultural preservation, frequency of travel, and use of technology were all found to be moderators. Dichotomously rated gender (male/female) produced no non-binary cases and low external validity with gender-non-conformist samples. Future studies should adopt inclusive items on the measure of gender (non-binary, self-describe, prefer not to answer; multiple selection), explore continuous measures of identification with gender, and predict MGA or mixture models on extended categories with sufficient sample sizes.

Mixture models or fsQCA can identify equifinal patterns (e.g., “visual-dominant,” “authenticity–memory,” “narrative–tech-savvy”) that lead to similar advocacy through varied combinations of cues and organismic states. In addition to intention to recommend, incorporate responsible visitation (off-season alternatives, code-of-conduct acceptance), donation/volunteering inclination, and learning transfer (maintaining heritage knowledge). Because authenticity and story may influence ethical positions as much as advocacy, exemplifying such conduct will bring the mechanism discussed here into alignment with stewardship objectives. Lastly, for better reproducibility and cumulative findings, subsequent studies can resort to pre-registration, open materials, and multiple-site replications (by heritage categories and cultures). Various studies by differential heritage type (tangible vs. intangible), visit phase (pre-trip vs. on-site vs. post-trip), and cultural audiences (domestic vs. international) will determine when the visual direct effect predominates and when authenticity or narrative more strongly flows through emotion and memory.

Together, these directions make a single model into a more sustained path: experiments that ground causality; longitudinal traces that follow slow memory making; multi-method triangulation that allows inner states and outer behavior to converge; comparative modeling that refines our imagination; enlarged outcomes that pair stewardship to success; and studies of complex affect, where ambivalence and wonder both reside. If we tread this path, heritage communication is no longer a megaphone but rather a bridge—taking people across generations with not only strong but enduring and inclusive messages as well, so that which inspires us today may judiciously be remembered tomorrow.