Touchpoints' role in the museum context: from affect to customer loyalty

Over time museums have undergone a significant transformation, becoming experience-centred institutions that prioritize visitors' needs for education, entertainment, reflection and knowledge sharing (e.g. Traboulsi et al., 2018; Belenioti and Vassiliadis, 2017). Meanwhile, they have encountered increasing competition from the highly dynamic leisure market, making it essential to create and maintain a loyal customer base (Falk and Dierking, 2016). It has become evident that the customer base represents one of the most valuable resources for museums and the COVID-19 pandemic has made this even clearer (Giulierini, 2021). Thus, the adoption of a market-oriented approach aimed at superior value creation has gained increasing relevance to create a strong relationship between the museum and its public. The goal is promoting not just a visit but a positive experience (Vesci et al., 2020), starting to involve customers before the visit itself and continuing after it, thus fostering customer loyalty.

Given these premises, understanding customer loyalty intentions in the museum setting is key for the success of museums (Harrison and Shaw, 2004). The present work addresses this point by focusing on the relationship between the affective visitor response to the touchpoints of the museum and the visitor loyalty intentions. It is known that multiple interactions occur, directly and indirectly, between a brand and its customers. Since the outcomes stemming from these interactions play a pivotal role, museums are required to properly leverage the touchpoints that affect visitor experiences and, ultimately, loyalty intentions (Addis, 2023). Touchpoints vary in terms of control, as described by Lemon and Verhoef (2016), and it is important to consider all the different types of touchpoints: museum-owned touchpoints (under the museum's full control), partner-owned touchpoints (less controllable by the museum since they depend on its partners) and social/external touchpoints (under the control of the visitors). By reviewing literature on touchpoints, a lack of research emerges in the museum setting (De Keyser et al., 2020), with only a few studies addressing the role of multiple touchpoints with a comprehensive perspective (e.g. Chen et al., 2021). More importantly, no study focuses on the link between the visitor affective response to the museum's touchpoints (touchpoint positivity) and visitor loyalty intentions (Luo et al., 2024). The present study aims to fill this gap, offering valuable academic and managerial implications.

Specifically, the study advances the museum literature and practice as follows. First, it offers a fresh perspective on visitors' behavioural loyalty intentions, going beyond the experience economy approach (e.g. Vesci et al., 2020) and the brand experience concept (e.g. Piancatelli et al., 2021; Recuero-Virto et al., 2017), which have been extensively employed in museum research.

Second, this study considers the relative impact of affective response to multiple touchpoints by adopting the relational cohesion theory. Specifically, as proposed by Baxendale et al. (2015), different types of touchpoints that encompass the concepts of channels and media are considered, providing a deeper understanding of the more fruitful set of touchpoints that can lead to empowered affective responses with positive effects on word-of-mouth (WOM) and the decision to visit the museum more than once. As museums are increasingly embracing digital transformation (Massi et al., 2021) and applying digital technologies and social media to their daily work (Palumbo et al., 2022; Karayilanoğlu, 2019), understanding the role of each touchpoint can support the search for a proper balance between traditional and new touchpoints within the visitor journey. In this perspective, the study also contributes to the body of museum literature that, in contrast, in recent years has mainly focused on the exclusive role of digital on-site touchpoints (Carù et al., 2021).

Third, as far as implications for practice are concerned, this research provides insights to help museum managers to design and optimize the customer journey of their visitors, eliciting revisit intention and advocacy. The latter is key as it can stimulate the visits from many people positively influenced by the experience of friends and family members. Understanding the role of touchpoints in influencing visitor loyalty intentions can offer valuable perspectives to managers for enhancing museums' capabilities in touchpoint journey design, touchpoint prioritization, touchpoint journey monitoring and touchpoint adaptation (Homburg et al., 2017).

Attracting new visitors and fostering visitors' loyalty is critical for museums to achieve sustainable profitability and long-term growth (e.g. Giachino, 2022; Piancatelli et al., 2021). It is therefore not surprising that visitors' behavioural intentions have become a topic increasingly investigated in the museum marketing literature, according to two main research streams.

The most established stream of studies built on the experiential marketing perspective. For example, a recent study by Vesci et al. (2020) focused on the model of the four realms of experience of Pine and Gilmore (1999) – aesthetic, learning, entertainment and escapism – to investigate the relation between the perceived quality of museum experiences and visitors' WOM behavioural intentions. The second, more recent, stream of studies relies on the brand experience perspective. For instance, Recuero-Virto et al. (2017) investigated the influence of a museum's brand image and satisfaction on visitors' loyalty, and Piancatelli et al. (2021) analysed the effects of museum atmosphere on brand perceptions and visitor behavioural intentions.

In line with the entire tourism sector, museums are a service-intensive industry that focuses on visitor experiences not only during their stay but also before and after the visit (Sheng and Chen, 2011). Indeed, the museum experience is the result of a series of interactions between the organization and the customer across multiple channels (Klaus and Maklan, 2013) and can be defined as the customer’ subjective responses to the encounters with the organization, encompassing the communication encounter, the service encounter and the consumption encounter (Lemke et al., 2011). Experiences arise from the interactions occurring across various touchpoints during the whole customer journey, which covers the pre-service encounter stage, the service encounter stage and the post-service encounter stage (e.g. Lemon and Verhoef, 2016). Touchpoints differ in terms of control, nature and stage of the customer journey. Touchpoint control involves who is in charge of the considered point of interaction: the museum may control a given touchpoint or not. Building on the touchpoint framework developed by Lemon and Verhoef (2016), the present study considers museum-owned touchpoints, partner-owned touchpoints and social/external touchpoints. Museum-owned touchpoints are designed and managed by the museum and under the museum's control (e.g. advertising, websites and apps). Partner-owned touchpoints are jointly designed, managed, or controlled by the museum and one or more of its partners, such as marketing agencies, distribution partners or communication partners. Therefore, partner-owned touchpoints are within the museum sphere of influence. It is important to consider that the distinction between brand-owned and partner-owned touchpoints can become less clear depending on the degree of autonomy attained from the partner. Moreover, partners might also have an impact on certain touchpoints that are owned by the museum. Finally, social/external touchpoints acknowledge the significant roles played by others in shaping the customer experience. Throughout the customer journey, customers are exposed to external touchpoints, such as other customers, independent information sources or peer influences, that have the potential to impact the overall process. The museum lacks substantial influence over these touchpoints.

The nature of a touchpoint delineates the way in which the touchpoint represents the museum. A touchpoint can therefore be human (such as frontline employees), physical (e.g. physical evidence and self-check-in kiosks), digital (such as chatbots and websites) or a blend of these (Manthiou and Klaus, 2022).

Table 1 summarizes empirical research on multiple touchpoints and their related impact on consumer attitudes and behaviours and compares the present study with previous research.

Behavioural loyalty intention refers to a customer's anticipation of the probability of engaging in a specific action (Lam and Hsu, 2006). Service literature has recognized five advantageous behavioural intentions (Zeithaml et al., 1996): saying positive things about the service provider, suggesting the service provider to fellow customers, remaining loyal to the service provider, increasing expenditure with the service provider and being willing to pay higher prices. Nevertheless, the two main indicators of loyalty are repurchasing the service and recommending it to others (Donovan and Rossiter, 1982, among others).

In the museum context, Harrison and Shaw (2004) contended that individuals typically do not express a strong inclination to revisit museums. Their primary argument stems from the notion that this inclination is linked to the inherent characteristics of the museum experience, which does not readily lend itself to frequent revisits in the short term. Thus, while visitors may leave the museum feeling satisfied, they are unlikely to return unless significant changes are implemented to the museum's offerings. Hence, the authors affirm that engaging in a positive WOM may be the most appropriate indicator of favourable behavioural loyalty intentions in the museum setting.

Nevertheless, the present study conceptualizes visitors' loyalty as a dimension that comprises both recommendations intentions and revisit intentions. The latter is commonly used in other infrequently purchased service contexts, including hospitality and tourism destinations (e.g. Chi and Qu, 2008; Yoon and Uysal, 2005). Bulatova et al. (2020) state that visitors whose cognitive process is properly stimulated often return, and Harrison and Shaw themselves (2004) acknowledge that very satisfied visitors may express an intention to return in the near future. In addition, museums are increasingly hosting or organizing events, temporary exhibitions or revising exhibition itineraries to offer additional reasons to revisit the museum.

The present study develops a theoretical framework based on the theory of relational cohesion (Lawler et al., 2000). The theory of relational cohesion is centred on the assumption that emotions trigger a cognitive process in which the emotions are associated with the source of emotions (Hernàndez-Ortega et al., 2022). This brings to the development of subjective perceptions of cohesive relations with the source of emotions. Therefore, a psychological mechanism is generated: from the exchange between parties to the generation of emotions, influencing individuals' behavioural outcomes, namely the propensity to remain in a relationship (Lawler et al., 2000). The above positive exchanges produce positive feedback, encouraging customers to engage again in the positive interactions with the related source. Hence, this leads to higher loyalty intentions towards the service provider, namely intention to repurchase and to spread positive WOM. Traditionally, the theory of relational cohesion has been applied to understand and explain the formation and the consequences of interpersonal relationships. More recently, it has been used to also understand the interactions between subjects and technological service agents.

The present work aims to extend the theory of relational cohesion to interactions between consumers and service encounters, namely touchpoints, in a museum context. In this respect, visitors within their customer journey encounter different touchpoints, for instance, brochures, staff, totems, websites, etc. The present study considers all the categories of touchpoints throughout the pre-visit stage and the service encounter stage (in the museum) and addresses the relationship between affective response to museum touchpoints and visitors' loyalty intentions. Affective response is modelled as touchpoint positivity, which is defined as the valence of the customer's response to each touchpoint (Kahn and Isen, 1993). While affective response can be understood multi-dimensionally (Chitturi et al., 2008), qualitatively different emotions can be associated with the unidimensional construct of affective valence or positivity (Kahneman and Krueger, 2006). In service literature, positivity has been previously employed and related to different outcomes including loyalty intentions (e.g. Ieva and Ziliani, 2018). Affective response has been shown to influence spending and repeat purchase intentions (Arnold and Reynolds, 2009; Liu, 2006): if affective responses are positive, they lead to higher loyalty intentions towards the source of these positive responses. Therefore, positive emotions elicited by the museum touchpoints could reinforce the relationship with the museum, fostering two types of loyalty intentions towards the museum: visitor intention to revisit the museum and the intention to recommend the museum to friends and relatives. This relationship might vary according to the touchpoint considered. Building on the above, the conceptual framework reported in Figure 1 is proposed. We formulate the following hypotheses that will be tested at the museum touchpoint level:

To provide theoretical and managerial implications, the role of touchpoints will also be addressed in the discussion in light of their control nature, with reference to the three main abovementioned categories: brand-owned, partner-owned and social/external. As in previous academic studies, positivity is declined at the touchpoint level. A list of touchpoints emerged with reference to museums by considering and integrating lists of touchpoints employed in literature (e.g. Romaniuk et al., 2013; Baxendale et al., 2015; Ieva and Ziliani, 2018): audio guides; bookshop; brochures and flyers; interactive on-site devices; mobile app; museum map, totems and museum signage; newspapers and magazines; online infomediaries; outdoor ads and paid advertising; road signs and museum banner; social media; staff; website.

The study aims to identify the relationship between the affective customer response to the touchpoints of a museum and loyalty intentions. Data were collected by means of a cross-sectional survey on museum visitors thanks to the cooperation with a small museum based in a medium-sized city in a European country (in the country of the reference about 90% of existing museums are categorized as small museums). The museum serving as the setting of the study is an ethnographic museum housing a permanent collection dedicated to world cultures, such as Chinese art and African art. In addition to its core displays, the museum curates a series of temporary exhibitions on a scheduled calendar and organizes special events tailored to its broader cultural programming. Visitors were asked to answer the survey questions with reference to their last visit to the museum. The survey was presented as a part of a university research project aimed at better understanding their experience with the museum, ensuring anonymity of the answers. No incentives were offered to the respondents. The survey was administered online to a convenience sample of museum visitors that were contacted via email, social media or in the museum at the end of the visit. A filter question at the beginning of the survey was ensuring to allow responses only from subjects that had visited the museum in the past. Respondents were asked to share the approximate date of their last visit so that it could be employed to control for recall bias: the number of days elapsing between the date of their last visit and the date of response was computed and included in the statistical models as a control variable. Consumers could be biased in recalling past purchase behaviour. However, studies on touchpoint exposure (Romaniuk et al., 2013; Herhausen et al., 2019) employ surveys as a valid option to assess consumer exposure to marketing stimuli. Respondents also had to report whether they visited the museum to attend a specific event, as the visit experience during events might be slightly different from the usual one. The data collection was completed with a final sample including 244 subjects over 18 years old who had visited the museum in the last 4 years. A list of 13 touchpoints was identified with reference to museums by considering and integrating lists of touchpoints employed in previous studies across different industries (e.g. Romaniuk et al., 2013; Baxendale et al., 2015; Ieva and Ziliani, 2018). Respondents had to recall their encounters with the following abovementioned touchpoints: audio guides; bookshop; brochures and flyers; interactive on-site devices; mobile app; museum map, totems and museum signage; newspapers and magazines; online infomediaries; outdoor ads and paid advertising; road signs and museum banner; social media; staff; website. The touchpoint list was randomized per respondent to avoid any order bias in the responses.

Positivity was measured, following Baxendale et al. (2015), by means of the single Likert-scale item, “How did it make you feel about the museum?” on a five-point scale anchored by “very positive” and “very negative.” This variable was then recentred around zero, and if the participant did not report any interactions with a touchpoint, positivity was imputed as zero, following Baxendale et al. (2015). Therefore, only experienced touchpoints were actually rated as positive or negative to ensure that positivity scores are reliable and comparable across different subjects. Respondents were also asked to indicate the frequency of encounter with each touchpoint over the previous three months on a single Likert-scale item (seven-point scale) anchored by “never” and “very often.” Furthermore, to reduce omitted-variable bias and avoid overestimating the effects of touchpoint positivity, we measured other variables serving as controls, such as age and gender, in addition to the number of days elapsing from the last visit. Intention to recommend the museum was measured by means of a scale from Vesci et al. (2020), adapted by the authors to the context of a museum from Lee et al. (2008) and Lockwood and Pyun (2019). Intention to revisit the museum was measured by means of a scale adapted from Hwang and Lyu (2015).

As far as analytic strategy, two ordinary least squares (OLS) regression models were run. Model 1 and Model 2 included, respectively, intention to recommend the museum and intention to revisit the museum as outcome variables. The set of independent variables in both models was as follows: positivity of each touchpoint, age, gender, the number of days elapsing from the last visit (to control for recall bias) and whether the visit was in the context of an event (to control for the different circumstances occurring during an event). Both models were run with a backward stepwise selection procedure: at each stage, the model was evaluating the significance of coefficients and removing the variable displaying the highest p-value. The backward stepwise selection procedure allows to include in the final model only the set of touchpoints (and eventual control variables) that was found to be significantly related to the outcome variable. The procedure ended when all the variables included in the model were displaying a p-value lower than 0.10. This procedure was implemented to cope with the high number of touchpoints involved in the estimation. Standardized coefficients were computed and compared to highlight the relative importance of each touchpoint in terms of positivity. Finally, to also control for frequency of exposure to touchpoints, a robustness check was conducted to understand whether results change when considering frequency. The analyses were conducted using IBM SPSS.

The sample involved respondents with an average age of 43 years old, ranging from 18 to 83 years old; 53% of them were female and 64% held a university degree. More than 50% of the recalled experiences occurred within the last year and 28% of respondents visited the museum, as far as the recalled experience is concerned, to attend a special event. Table 2 displays, for each touchpoint, the values in terms of positivity and frequency of exposure. The top three touchpoints in terms of positivity are social media, staff and website, which are also the most frequently encountered by museum visitors.

Before the regression models, a confirmatory factor analysis on the two outcomes was run. Cronbach's alpha was measured to assess reliability and all measures displayed values above 0.80. The average variance extracted (AVE) values for the two constructs were above the 0.50 threshold. All the standardized loadings displayed values higher than 0.70. Results confirmed discriminant validity: the correlation estimate between the two constructs was below 0.80 and significantly different from 1, and each construct's square root of AVE was higher than their corresponding correlation, complying with Fornell and Larcker's criterion (Fornell and Larcker, 1981). Table 3 displays the results from the confirmatory factor analysis and the related items and factor loadings.

Finally, common method bias was tested on the self-reported variables by employing Harman's single-factor approach: the first factor explained a portion of variance smaller than the critical value of 50% (43%), thus confirming that our model was free from common method bias.

The first OLS model with intention to revisit the museum as an outcome was run. Variance inflation factor values were below 3; thus, multicollinearity was not a concern. Durbin–Watson statistic is equal to 2, thus signalling no autocorrelation of residuals. The ANOVA statistic is significant, showing that at least one variable in the model is significant. The backward procedure ended up with a model displaying an adjusted R² of 0.24 and a set of six variables significantly related to intention to revisit at the 0.10 level. Table 4 displays the final variables with related coefficients: standardized coefficients express the importance of each touchpoint's positivity in being related to the intention to revisit. Results show that the most important touchpoint explaining intention to revisit is newspapers and magazines, followed by the staff and the website. Therefore, H1 is supported: positivity, as far as certain touchpoints are concerned, displays a positive relationship with intention to revisit the museum. Positive affect driven by encountering some touchpoints is a significant driver of behavioural loyalty intentions.

The second OLS model with intention to recommend the museum as outcome was run. Values were below 3, thus multicollinearity was not a concern. The Durbin–Watson statistic is close to 2 (2.3), thus signalling no autocorrelation of residuals. The ANOVA statistic is significant, showing that at least one variable in the model is significant. The backward procedure ended up with a model displaying an adjusted R² of 0.34 and a set of six variables significantly related to intention to recommend at the 0.10 level. Table 5 displays the final variables with related coefficients. Results show that the most important touchpoints explaining intention to recommend the museum are staff and social media, followed by newspaper and magazines. Therefore, H2 is supported: positivity, as far as certain touchpoints are concerned, displays a positive relationship with the intention to recommend the museum. The positive affective response resulting from encountering certain touchpoints is a significant driver of positive WOM.

Given that previous studies have also shown that frequency of exposure to touchpoints can play either a positive or negative role (e.g. Baxendale et al., 2015), we run a robustness check by including touchpoint frequency in both Model 1 and Model 2. The goal of the robustness check is to understand whether results change when frequency of exposure to touchpoints is taken into account. To avoid the inclusion of a very high number of variables compared to the number of cases – which might cause statistical power issues, overfitting and multicollinearity issues – a specific approach was adopted for including frequency of exposure: an exploratory factor analysis on frequency of exposure variables with varimax rotation was run, leading to identifying two factors accounting for 71% of the variance, as only two eigenvalues displayed a value higher than 1 and the explained variance was higher than the commonly employed threshold of 60%. Varimax rotation was chosen to maximize the variance of each factor while reducing at minimum the loadings of variables that do not contribute to the factor. All touchpoint frequency variables, apart from staff, loaded high on Factor 1, while staff was particularly high on Factor 2: Table 6 displays factor loadings. Therefore, we could interpret Factor 1 as a general frequency of exposure to the museum touchpoints excluding staff, while Factor 2 is a more specific exposure to staff. The main goal of the exploratory factor analysis was to summarize frequency of exposure, so the portion of accounted variance of the two factors is satisfactory.

The two frequency factors were log-transformed, as suggested by Baxendale et al. (2015), to account for communication wear-out through over-exposure, which leads to diminishing returns, and we included both log-transformed factors in the two models, Model 1 and Model 2, that were rerun with the same backward selection process. Table 7 displays results from Model 1 after the variable selection process including the two frequency factors: the set of touchpoints related to intention to revisit the museum is the same as before, and the two frequency factors have been removed by the procedure, as they were not significantly related to intention to revisit.

Table 8 displays results from Model 2 after the variable selection process, including the two frequency factors: the set of touchpoints related to intention to recommend the museum is the same as in the main analysis, apart from brochures and flyers and the website, for which there is weak support of a significant relationship with intention to recommend (p < 0.10). Interestingly, Factor 2 that loads high and positively with staff displays a negative and (barely) significant relationship with intention to recommend, while the affective response to the staff is positively related to intention to recommend: this might suggest that staff plays a positive role, but there is a threshold of exposure beyond which it can backfire.

To summarize, both robustness checks confirm the role of the identified touchpoints in the main analysis. Only with reference to explanatory variables of intention to recommend, slight variations occur with weak support provided to two additional touchpoints.

Museums have evolved into multifaceted institutions, with a role extending beyond culture, art conservation, and heritage to encompass a diverse array of objectives within society. At the same time, visitors differ from those in the past: they now engage with their experiences through technology, seek to share their emotions and are increasingly drawn to opportunities facilitated by innovative tools. In this context, seamlessly blending the physical and digital domains is an essential objective for museums (e.g. Giulierini, 2021).

The results of this study reveal the importance of measuring the role of touchpoints in influencing visitor loyalty intentions. The present work provides theoretical implications with reference to the theory of relational cohesion and practical implications for the design, monitoring, prioritization and adaptation of touchpoints in the museum context.

First, the present work provides support for the theory of relational cohesion in a museum context (Lawler et al., 2000) by emphasizing the role of affect in eliciting loyalty intentions through different types of touchpoints. The emotional side of the museum experience should then be taken into account in the majority of the interactions with visitors. Moreover, the affective visitor response to both physical and digital touchpoints (as well as to both on-site and off-site touchpoints) is positively related to loyalty intentions: prioritization efforts should consider the need for the museum to embrace an omnichannel perspective across touchpoints – such as the museum's website, museum's staff and social media – in order to manage the visitor experience in an effective way.

Second, the positive experience with three touchpoints helps explain the intention to revisit the museum: newspapers and magazines, staff and website. All these touchpoints but one are under the control of the museum. The most important touchpoint is a partner-owned touchpoint, namely newspapers and magazines: this should guide the museum to properly manage media relations (press kits, press releases, etc.), especially on the occasion of events, considering that a museum visit motivated by an event stimulates (in itself) the intention to return.

Third, the positive experience with four touchpoints – two of which are under the museum's control – helps explain the intention to recommend the museum: staff, social media, newspaper and magazines and interactive on-site devices. Staff and social media are the most important touchpoints whose experience is positively related to the intention to recommend the museum to others. Manthiou and Klaus (2022) highlight that human representatives are suitable for tasks requiring soft or non-routine skills: humans can play a key role in the complex and multidimensional museum experience by using communication, listening and negotiation skills. This should guide the museum to invest in providing staff with the appropriate training for the museum experience delivery (communication, listening, problem solving, empathy and adaptability) and in emphasizing their key role through internal communication. WOM plays a key role in cultural contexts (Harrison and Shaw, 2004). Therefore, an effective social media strategy is key to stimulate WOM and peer-to-peer feedback. Social media can leverage satisfactory experiences with the museum and influence other visitors to plan their visit or to improve their awareness about the museum (Piancatelli et al., 2021). In that way, the museum can manage and monitor WOM communication and peer-to-peer suggestions.

The study entails several managerial implications. First, findings, consistently with studies in other contexts and sectors, highlight the need for allowing museum visitors to move across on-site and off-site, digital and physical touchpoints in a seamless way. This might lead to improved loyalty interactions.

Second, results point to the need for prioritizing investments in those “museum-owned touchpoints” that are positively related to customer loyalty intentions: website, staff and interactive on-site devices. Designing journeys mainly around these touchpoints can support the museum in maximizing investment return in such a resource-constrained context.

Third, given the important role of affect in stimulating loyalty intentions, it is essential to rely on staff to pay attention to and manage visitors' emotions, especially during the visit. Providing staff with training on how to elicit emotional and affective responses in the museum – by involving visitors, adopting experiential discourse, and creating opportunities for social interaction – can lead to positive WOM and increased revisit intentions. Such training should aim to enhance their experiential intelligence, flexibility and communication skills.

Finally, the results underscore the importance of maintaining consistent quality standards across all relevant touchpoints, including those that lie within the museum's sphere of influence but are not under its direct control. For example, newspapers and magazines – whose coverage can significantly shape revisit and recommendation intentions – should be regarded as long-term strategic partners in designing an effective and seamless visitor experience. Governmental bodies and cultural institutions are encouraged to revise existing frameworks in European countries – such as the Uniform Quality Levels for Museums in Italy, the Museumsbund Quality Standards in Germany or individual cultural service charters issued by museums – to incorporate new key performance indicators. These should reflect the critical role of visitor emotions and offer updated guidelines for comprehensive and effective touchpoint management.

The present study is not without limitations. First, data have been collected via survey where customers were asked to recall their previous experience with the museum. Therefore, the presence of recall bias cannot be excluded, even though efforts have been made to minimize it by including a relevant control variable (number of days elapsing since the last visit). The employment of a survey, as in many studies, could also lead to self-selection bias with an over-representation of respondents with positive (or negative) experiences. Future research could address this issue by using different methodologies (e.g. experiments) to reduce self-selection bias, which is common in surveys. Second, touchpoints may vary in the ways they elicit an effective response among consumers over time. For instance, certain touchpoints might be very impactful at the time of the encounter, but their significance could be reduced rapidly afterwards. Vice versa, other touchpoints could display a more lasting affective role. This heterogeneity in terms of the timing of the emotional response has not been captured due to our methodological approach based on a cross-sectional survey. Future research could adopt a longitudinal design to monitor whether the emotional impact of certain touchpoints varies over time. Third, the employed survey is based on one small museum only, thus limiting the degree of external validity of the study. Fourth, affective response to each touchpoint has been measured with a single-item measure, as in previous studies, with a self-reported questionnaire. Future studies could improve this point by adopting more complex measures in this respect. Future research could also measure the impact of different touchpoints on other outcomes, such as customer engagement and willingness to pay, that represent interesting indicators for museums. Finally, response to touchpoints has been measured without considering differences among visitor segments based on personal and behavioural characteristics (e.g. age, motivations and familiarity with the museum). Future studies could address heterogeneity of visitor response to touchpoints by adopting a sub-group analysis.

The authors would like to express their sincere gratitude to Chiara Allegri, for her invaluable support and collaboration throughout the course of this study and to Elena La Barbera and Tiziana Battiata for their support in data collection.