Corporate Twitter use and stakeholder engagement: An empirical analysis of the Spanish hotel industry

This study provides a general overview of the way in which hotel chains use Twitter as a communication tool to engage with their stakeholders. A sample of 109 of the largest Spanish hotel chains was examined, obtaining the result that the majority of them maintain an official corporate Twitter account. In addition, it was found that there is a significant positive relationship between a hotel’s size and its Twitter activity and a significant negative relationship between the audience and the stakeholder engagement. Moreover, the findings of the paper show that particular media and content types generate higher stakeholder engagement. Given that this is the first study of its kind, the paper offers both theoretical implications for the social media literature and practical implications for the hospitality industry.


Introduction
The trend of social media has become unstoppable. According to Statista (2015a), there are 1.96 billion social network users worldwide. Given that figure, nowadays, an increasing number of companies are using social media as a tool to communicate with their stakeholders. Using social media platforms can help companies to set up a two-way communication protocol with their stakeholders (Bonsón and Flores, 2011;Hsu, 2012;Lee et al., 2012) and this not only generates stakeholder engagement and loyalty (Wang and Fesenmaier, 2004) but also creates a better image of the companies, which helps to improve their online reputation (De Moya and Jain, 2013). Several studies point out the importance of social media to the topic of customer relationship, loyalty management and public engagement (Brodie et al., 2011;Lee et al., 2013). With the rapid development of social media, scholars also analyse how professionals can take advantage of this online tool to help them gain a better understanding of their customers' needs and the factors that drive stakeholders to engage with companies. However, little is known about how Twitter is being used in the hospitality sector to engage with stakeholders. Previous studies focus on the general usage of Twitter rather than its application in the tourism sector; therefore, it is worthwhile extending the earlier Web 2.0 research to explore this phenomenon.
The main objective of this paper is to provide an initial assessment of communication and stakeholder engagement practices using Twitter in hotel chains. The paper aims to answer the following questions. (1) To what extent do hotel chains use Twitter as a communication channel to engage with their stakeholders? (2) Which factors influence the Twitter activity of hotel chains? (3) Which factors influence the stakeholder engagement on Twitter?
In this study, we focus on the social network Twitter, given that it is the largest micro-blog site and the second-largest social media platform that is used in business (Statista, 2015b). In addition, it focuses on information and opinion sharing (Kwak, Changhun and Moon, 2010) and on prompting conversation rather than just self-presentation (Kietzman et al., 2011). It is not only a social media platform on which its users can connect with their family, friends or co-workers; it also offers an opportunity for companies to interact with different types of individuals and allows them to broadcast information. It is recognised by strategic communicators for its ability to reach a large number of stakeholders, allowing organisations to make better use of social media for the purposes of public relations, advertising and marketing campaigns (Stelzner, 2012).
An increasing number of companies are tending to use social media as a new and creative way to communicate with their stakeholders and to achieve stronger engagement (Lai and Li, 2005;Mangold and Faulds, 2009;Men and Tsai, 2013). According to Stelzner (2012), Facebook and Twitter are the top two choices for marketers and nearly all marketers with more than three years of experience in social media use Twitter (91%). Consultants emphasize the importance of building a social media strategy to raise virtual communities and strengthen engagement (Li and Bernoff, 2008;Solis and Breakenridge, 2009). Academics claim that online tools can be seen as a means to maintain and manage customer relationships (Boyd and Ellison, 2007;Chan and Guillet, 2011;Garau-Vadell and Orfila-Sintes, 2008;Schmidt et al., 2008) and that using social media platforms helps organisations to disseminate and exchange information rapidly and to obtain real-time feedback and engagement in conversation (Lovejoy et al., 2012).
According to Crotts (1999), the hospitality industry is a service-oriented industry, in which the majority of goods and services are experiential, which makes reviews from customers very important for potential customers. Thus, the rapid development of new technologies, such as Web 2.0 and social media, has led hotels and their website users to a new era, in which the integration of these new technologies enables them not only to disseminate information (one-way communication) but also to interact with each other (two-way communication) more easily (Park and Gretezl, 2007). Therefore, the tourism industry may take advantage of the benefits of online social media tools to build a better customer relationship by starting a conversation, listening and replying to its customers and by doing so generating a better engagement with them (Chan and Guillet, 2011;Drury, 2008, Leung andBai, 2013).
According to Chan andGuillet (2011) andHolzner (2009), Web 2.0 and social media platforms are the most appropriate tools for the tourism industry to communicate and interact with its users and develop effective relationships with them. The research of the European Travel Commission (2010) also recommends that hotels should start taking action to interact online with their customers. Similarly, Schmidt et al. (2008) and Wilson (2005) express the importance of adopting interactive networking and becoming part of the "new marketing media".
Recently, an increasing number of hotels have been using social media platforms as their marketing tools, and according to the study by Leung and Bai (2013), 75% of hotels worldwide use social media for marketing purposes.
Regarding the tourism industry, some researchers indicate that there is a need for studies that focus on how companies are using social media as a new communication tool to engage with their stakeholders (Leung and Bai, 2013) and that there is a lack of empirical studies to help professionals to integrate academic research with the industry's needs.
For the purposes of this study, we analysed the Twitter practices of 109 Spanish hotel chains (78 out of 109 have an official Twitter account). This sample was used given that Spain is considered to be one of the leading countries in tourism. In addition, Spain is ranked number two both in international tourist arrivals in Europe and in international tourism receipts (UWTO Tourism Highlight, 2013).
To find out whether particular media or content types influence stakeholder engagement, 3705 tweets were analysed in total. Our findings suggest that hotels are adopting social media as a new communication channel to engage with their stakeholders. Furthermore, there is a significant relation between the hotel size and the Twitter activity; however, the size of the hotel does not influence stakeholders' engagement. Moreover, it was found that hotels with smaller audiences tend to have higher stakeholder engagement than channels with a high number of followers. The results of the paper also show that, by using different media and content types of tweets, they generate different levels of stakeholder engagement.
We believe that our paper has implications for both research and practice. This is the first study to explore Twitter as a communication and engagement platform in hotel chains. In addition, the findings of this paper provide hotels with some practical information about the kind of content or media type that could generate more stakeholder engagement.

Literature Review, Theory and Research Questions
The rapid increase in social media usage in general has attracted the attention of academics and a large number of studies have been conducted on that topic. In our study, we focus on the previous research on social media conducted in the hotel industry. The study by  offers an overview of all the investigations related to social media published in academic journals from 2007 to 2011 in the tourism and hospitality areas. In general, they highlight the strategic importance of social media for the hospitality industry and the need to develop metrics to measure the effectiveness of social media communications. Kwok and Yu (2013) argue that there is a lack of studies focusing on stakeholder engagement, and Kumar et al. (2010) point out the importance of metrics to analyse this engagement.
The most widely used social media platforms in the hospitality industry are Facebook and Twitter (Leung and Bai, 2013). According to Leung and Bai (2013), Facebook is the platform that accounts for the highest number of studies conducted. Twitter itself is analysed in the hospitality industry together with Facebook and other social networks by Chan and Guillet (2011), Leung and Bai (2013) and . Leung and Bai (2013) apply the motivation, opportunity and ability (MOA) theory and the concept of involvement to social media pages. They find that customers' motivation and opportunity have positive relations with their social media involvement, which in turn positively affects their intention to revisit a hotel's social media page. Another study  analyses the marketing effectiveness of social networks in the hotel industry, proposing a theoretical model integrating the attitudetoward-the-ad model with the concept of attitude toward social media pages. Their findings indicate that customers' attitude toward social media pages is influenced by their social media experience, which also influences their attitude toward hotel brands, booking intentions and electronic word of mouth.
The dearth of academic attention paid to Twitter usage in the hotel industry in spite of its great interactive capability and the lack of studies focusing on stakeholder engagement and applying appropriate metrics were the two main reasons for conducting research about Twitter activities and stakeholder engagement in the hotel industry.

Stakeholder Theory
Twitter allows two-way communication, which leads to interaction and engagement with stakeholders. Therefore, the stakeholder theory was applied as the background for our study. According to Freeman (1984), stakeholders can be defined as "any group of individuals which can affect or be affected by the achievement of the organisation's objective". This theory stresses that considering the interests of all stakeholders, not only shareholders, could generate a higher performance level for the company. Therefore, a company should seek to strike a balance between the interests of its diverse stakeholders, and communication through social media might be an effective tool to satisfy their information needs and minimize the information asymmetry. According to Freeman (1984), companies try to achieve greater transparency to gain the approval of their diverse stakeholders.
Social media sites provide companies in the tourism and hospitality industry with an opportunity to engage with potential customers and help them to increase their online presence as well . According to Wang and Fesemaier (2004), companies can use social media as a means to manage customer relations and engage with customers through social interaction. By applying this theory, we aimed to respond to the question of whether the stakeholder engagement on Twitter in Spanish hotel chains might be explained by stakeholder theory (more concretely, this is addressed in RQ3).

Research Questions
The main aim of our study was to provide a general overview of the Twitter usage in hotel chains. Our next objectives were to find out how active they are, how large their audience is and what kinds of content, media type and interaction, are prevalent. As this is an exploratory study, six research questions were formulated to identify the factors associated with the activity and stakeholder engagement on this platform. (2013), Gallego et al. (2009), Larrán and Giner (2002), Mallin and Ow-Yong (2012) and Utama (2012) argues that the size of the company is considered to be one of the most frequently used variables to explain the level of information disclosure. Furthermore, academics suggest that if a company reaches a certain size, it tends to formulate a web development strategy, which also includes a social media strategy (Bonsón et al., 2008(Bonsón et al., , 2014Rodríguez et al., 2012). Therefore, it can be stated that large companies disclose more information on the Internet than small ones. Based on that, our first research question was created:

RQ 1 : Does a hotel's size influence its activity on Twitter?
Some previous studies suggest a negative relationship between the size of a company and its stakeholder engagement (Men and Tsai, 2013). Thus, we aimed to explore this phenomenon on Twitter and formulated our second research question:

RQ 2 : Does a hotel's size influence its stakeholder engagement on Twitter?
According to Deegan and Samkin (2009), the loyalty of stakeholders (based on good relationships and reflecting good stakeholder engagement) can have an important impact on the company's performance. Indeed, the first step in stakeholder engagement is to initiate a dialogue (e.g. posting a tweet). Stakeholder theory points out that greater transparency (tweeting activity) leads to higher stakeholder approval (represented by stakeholder engagement). Thus, we aimed to find out whether there is a statistically significant correlation between the activity on the Twitter channel (measured by the number of tweets) and the stakeholder engagement. Based on that, our third research question was formulated:

RQ 3 : Does the activity on Twitter influence the stakeholder engagement?
A large number of followers does not necessarily lead to higher stakeholder engagement (Men and Tsai, 2013). Greater engagement in smaller hotel chains can be explained by the argument that in smaller social network communities, followers feel closer to the company as they see that their opinion matters, which encourages them to interact more with the company. Brodie et al.
(2011) also find greater user engagement in small online companies because of the sense of belonging to the group, generating a feeling of being part of a community. Hence, to examine this phenomenon, our fourth research question was formulated:

RQ 4 : Does the audience influence the stakeholder engagement?
The disclosure literature highlights the importance of the medium through which the information is presented. Academics (Cho et al., 2009;Davison, 2007;Graves et al., 1996) suggest that text disclosure alone is no longer adequate to articulate information and stress the importance of visual imagery, such as graphics, photos and videos, claiming that these are more powerful communication tools that affect the followers' perception of the richness and potency of a given message. Hence, we intended to explore whether different media types, depending on their degree of richness, provoked different levels of stakeholder engagement.

RQ 5 : Does the media type influence the stakeholder engagement?
In addition, we aimed to ascertain whether proxies for stakeholder engagement, such as the number of favourites, replies and retweets, depend on particular video content based on the six categories defined in the methodology part. Moreover, we tried to identify a content type that might lead to a higher level of stakeholder engagement. Hence, our last research question was: RQ 6 : Does the content type influence the stakeholder engagement?

Sampling and Data Collection
The sample of this study is taken from the list published yearly by Hosteltur. It consists of the 109 largest Spanish hotel chains ranked by the number of guest rooms. The Twitter account of each hotel chain was collected by following the icon link on the hotel official website or by searching the Twitter platform to identify the verified hotel chain account. Consequently, 78 verified Twitter accounts were obtained for the 109 hotel groups. The data collection took place in October 2013.
A manual collection process was adopted, starting with the identification of Twitter channels for Spanish hotel chains. The hotel size was measured by the number of rooms of the whole group. The activity was measured by the total number of tweets posted by the hotel on its official Twitter account. The number of followers was used as a proxy for the popularity/audience of the Twitter account. An adaptation of the Facebook metrics proposed by Bonsón and Ratkai (2013) was used to measure the engagement of stakeholders, as shown in Table 1. It considers the following interactions: replies vs. comments, favourites vs. likes and retweets vs. shares. This paper includes specific terminology defined by the Twitter platform. Favouriting a specific tweet indicates that you liked it. A follower is another Twitter user who has followed you to receive your tweets in his or her home stream. A response to another user's tweet that begins with the @username of the person to whom

Content Analysis
This study also aimed to answer the questions concerning whether particular content and media types lead to higher stakeholder engagement (RQ5 and RQ6). The usage of media types on Twitter was classified and coded according to four categories: web links (1), video (2), photo (3) and text (4). Regarding the content analysis, the content types were identified and coded as shown in table 2. Therefore, a necessary step in this process was to conduct a content analysis of the tweets.
Content analysis offers a systematic, complex and objective way to compare content for large samples. In the literature, it is defined as a multipurpose research method to investigate a broad spectrum of problems by systematically and objectively identifying the special characteristics of messages (Holsti, 1969). It is considered to be a standard method for systematically comparing the content of communication. Therefore, it is used in a number of previous studies related to communication on social networks (Bonsón et al., 2014;Smith et al., 2012).
The whole process of the content analysis conducted in this study consists of four steps: (1) sampling, (2) coding, (3) analysing the content and (4) consolidating the results.
Sampling: To keep the sampling scope manageable but able to reflect a representative sample of hotels' tweets, a restricted number of tweets was established. Based on this, we analysed 50 tweets per account. If the number of tweets per account did not exceed 50, we analysed all of them; otherwise, we analysed the 50 most recent ones. Altogether we analysed 3705 tweets.
Coding: Before the coding was established, operational definitions and categories were developed for the content of the tweets and an initial research framework for content analysis was created. Tweets' content was classified by the authors according to the following categories (Table 2).
To test the classification scheme and the clarity of the descriptions, two independent research assistants cooperated on the research. A detailed explanation of the content analysis framework and tweet categories was given to the independent coders. Consequently, they were asked, together with one of the authors, to classify an initial portion of the sample to approve the suggested classification scheme. During this stage, it was possible to identify the areas of confusion and the necessary changes were made to the classification descriptions. These changes were then incorporated into the content analysis framework and the final coding was established.
Analysing the content: After establishing the coding, every tweet was carefully analysed and the category frequencies were tabulated,

Content type Tweet content Code
Marketing Promotion/ publicity of hotel itself. 1

Reply to Stakeholder
Tweets directly answering to stakeholders' comments and questions 2 Retweet Information from stakeholders retweeted by company 3

General Information
Travel information, News, tourism information that are not related to hotel itself but city or region. 4

Social and Environmental Concern
Social, Society, and Environmental concern 5 Employment Employment concern 6 Consolidating the results: As the final step, the results obtained throughout the content analysis were consolidated and the statistical differences were assessed.

Descriptive Statistics
Among the 109 Spanish hotels that were listed in Hosteltur in 2012, 71.55% of them (78 out of 109) use Twitter as a communication channel. The maximum number of tweets, belonging to the most active Twitter account, was found to be 11,128, and the minimum was 0 tweets. On the other hand, the account with the largest audience had 112,584 followers, and the account with the smallest audience had only 18 followers. Thus, it might be seen that there are substantial differences between hotel corporate Twitter accounts regarding the number of total tweets and the audience due to the high standard deviations in the case of both variables (Table 3).
In Table 4, the most popular content type among the hotels in our sample can be observed. The results show that hotels tend to use their Twitter account as another marketing tool as 42% of the tweets are related to marketing or promotional issues. The second purpose is to interact with stakeholders by retweeting their posts (22%). In addition, we can see that Twitter provides hotels with opportunities to engage in direct communication with their stakeholders by replying to their comments or questions (18%). Approximately 16% of the tweets were dedicated to general information, while the other content types, such as environmental, social or employment (CSR content), appeared to be scarce. Table 5 shows that the most frequently used media type was the website link (55.76%). After  the website link, the analysis found that hotels still tend to stick to the traditional way (plain text) of communicating with their stakeholders, rather than employing other visual tools, like photos (11.79%) or videos (2.86%). To examine the stakeholders' engagement, the metrics proposed by Bonsón and Ratkai (2013) were adopted. The results are provided in table 6. On average, 11% of the tweets were favourited (P1), 8% (C1) were replied to and 18% were retweeted (V1) by stakeholders. The average numbers of favourited tweets and replies were quite similar, specifically 0.13 for P2 and 0.12 for C2, and the average number of retweets per tweet was 0.26 (V2). Dividing it by the number of fans and multiplying it by 1000, following the methodology of Bonsón and Ratkai (2013), we obtained the average number of favourites/replies/retweets per tweet per 1000 fans. This result shows that among these 3 ways to interact with hotels, stakeholders tend to choose "retweets" more often than favourites or replies as a means to engage with the hotel. By summing P3 (0.21), C3 (0.16) and V3 (0.43), we obtained the average index of stakeholder engagement E of 0.59.

Statistical Analysis
Additionally, we aimed to identify the factors associated with both the channel activity and the stakeholder engagement on this platform. First, applying univariable statistics, we examined whether there was any significant relationship between the hotel size and the channel activity by adopting the Spearman coefficient (Table 7). As we can see, a significant relationship between the hotel size and its activity was found.
However, based on the results provided in table 8, examining the relationship between stakeholder engagement and the hotel size or activity, no positive significant relationship was found. Regarding the audience and stakeholder engagement, a significant negative relationship was detected. This means that the hotel size or activity is not associated with the stakeholder engagement level and that there is a negative relationship between the audience and the To understand better the relationship between the media type and the reaction of the stakeholders, we summarise the descriptive statistics in Table 9.
It can be seen that the engagement of stakeholders varies among different media types. According to the results, photos, as a particular media type, generate more retweets (0.48) and favourites (0.28) than other media types do. On the other hand, hotels receive To confirm that there is a significant relationship between media types and stakeholder engagement, the Kruskal-Wallis test was applied. The results presented in Table 10 show that there is a significant relationship between media types and stakeholder engagement. Hence, using a particular media type leads to increased stakeholder engagement (measured by favourites, replies and retweets). We found that photos and text lead to higher stakeholder engagement than other media types.
In this study, we also aimed to analyse whether different content types generate different levels of stakeholder engagement and whether there is a significant relationship between them. Table 11 shows that each tweet generates an average of 0.15 favourites, 0.13 replies and 0.28 retweets. As discussed earlier, stakeholders tend to engage with hotels by using the retweet function as it offers a fast way to express their thoughts about or agreement with the published content. This table also provides interesting insights regarding the content type related to employment, which seems to generate more engagement in the concept of favourites than other content types. Our study also shows that stakeholders of large hotel chains tend to retweet content that contains social and environmental (CSR) concerns.
As shown in Table 12, the Kruskal-Wallis test was applied again to identify whether there is a significant relationship between content type and stakeholder engagement. The results confirm that different content types influence stakeholder engagement (measured by favourites, replies and retweets).

Multivariate Statistics
To confirm the results of the Kruskal-Wallis tests suggesting that the media and content types are associated with stakeholder engagement, multivariate statistics applying the generalised linear model (GLM) were obtained. A GLM is a flexible generalisation of an ordinary linear regression that allows for response variables that do not have a normal distribution.
Thus, it was considered appropriate for our sample as the distributions of the dependent variables representing the stakeholder engagement (favourites, replies and retweets) were non-normal. The research questions were tested at the multivariate level with two independent variables: media type and content type. The results of the statistical analysis are shown in Table 13.
The findings of the GLM confirm the assumption that particular content and media types can generate higher stakeholder engagement. More specifically, photos and text are considered to be the most engaging media types, while tweets related to employment or social and environmental issues and replies to customers are the most engaging content types.

Discussion and Conclusions
This is the first study to explore communication and stakeholder engagement practices using Twitter in hotel chains. The main objective of this paper was to provide an initial assessment of such practices. The study aimed to answer the following questions. (1) To what extent do hotel chains use Twitter as a communication channel to engage with their stakeholders? (2) Which factors influence the Twitter activity of hotel chains? (3) Which factors influence the stakeholder engagement on Twitter?
The results show that 72% (78 out of 109) of the largest hotel chains have an official Twitter account. However, the channels differ in their activity (measured by the number of tweets) and their audience (measured by the number of followers). Regarding the content type, the results show that the most common content type of the tweets is marketing (42%). Considering the lower costs of promoting the products and services through this platform (by posting special offers, news, events, pictures, etc.) in comparison with other ways of advertising, it is understandable that hotel chains tend to use this cost-effective tool and post a huge amount of tweets with a promotional character. The results of our study also show that the most frequently used media type is the website link (55.76%), which might be due to the fact that Twitter only allows posts of up to 140 characters. Therefore, a website link is posted instead and navigates the stakeholder to the web page on which the target information is placed. This link could be, for example, the hotel's booking page or another social media site. Another finding that emerged from this study is that among the 3 ways to interact with hotels on Twitter (favourite, reply and re-tweet), stakeholders tend to choose "retweets" more often than favourites or replies as a means to engage with the hotel. This might be due to the simplicity of clicking the retweet button, which is remarkably convenient for stakeholders to express their opinions quickly without writing a comment. In addition, the retweet button enables greater visibility than the favourite button as the content is visible on the wall of the user to all his or her followers.
In addition, we aimed to answer six research questions related to Twitter activity and stakeholder engagement on this platform. A significant relationship was found between the hotel size and the Twitter activity. Hence, this finding suggests that bigger hotels have a more active presence on Twitter. This might stem from the fact that big hotels have more resources, which allows them to put more effort into social media site maintenance and develop better social media strategies. The results of our study comply with previous research confirming that bigger companies tend to have a greater social media presence than smaller ones (Bonsón and Bednárová, 2013;Gallego et al., 2009;Larrán and Giner, 2002;Mallin and Ow-Yong, 2012;Utama, 2012). Furthermore, we examined which factors influence the stakeholder engagement on Twitter by applying the metrics proposed by Bonsón and Ratkai (2013). It was found that neither the size of the hotel nor its Twitter activity has a positive influence on the stakeholder engagement on this platform. The latter suggests that stakeholders' propensity to engage with hotels via Twitter is in contradiction to the stakeholder theory, which suggests that greater transparency (activity) leads to higher stakeholder approval (engagement).
Regarding the audience measured by the number of followers, we found a significant negative impact on stakeholder engagement. However, this finding is in agreement with some previous studies claiming that a smaller audience feels closer to the community as the audience members are able to see that their opinion matters, which encourages them to interact more with the hotel (Men and Tsai, 2013).
Additionally, we tried to identify which media and content types lead to higher stakeholder engagement (considering the number of favourites, replies and retweets). It was found that photos, as a particular media type, generate more favourites and retweets than other media types. The reason behind this might be that, compared with reading a large amount of text or watching a video, stakeholders receive the full information from only viewing the photo (with a brief description if available), which is remarkably time saving and therefore more likely to generate some kind of response (such as favourite or retweet). These results are in compliance with the literature on media richness conducted by Cho et al. (2009), Davison (2007) and Graves et al. (1996). On the other hand, the media typeplain text, which usually leads to open dialogue in which the hotel replies to the stakeholders' comments, tends to receive more replies from stakeholders.
In terms of which content type generates higher stakeholder engagement, we found that social and environmental content generates the most retweets; responses to customers tend to be the most answered; and employment content is the most favourited. This phenomenon might be explained by the serious problem of unemployment that Spain is currently facing. Hence, based on these insights, we conclude that providing more information about CSR via a hotel's Twitter account is beneficial for stakeholder engagement on this platform.

Implications
Since there is no previous in-depth study exploring the communication and stakeholder engagement practices using Twitter in hotel chains, our study provides a new contribution to the debate about current communication practices through social media in the tourism sector. In addition, the study suggests some practical implications.
Potentially interesting practical implications for marketers or social media community managers might stem from our findings indicating that different content and media types generate different levels of engagement. Another finding of the present study is that the size of the hotel does not influence the stakeholder engagement; therefore, it offers some implications for smaller hotels or hotel groups indicating that the effectiveness of social media usage is not limited to big hotel chains. On the contrary, smaller-sized hotels should take advantage of social media as they represent low-cost solutions and their stakeholders may feel more engaged with the hotel when they see that the hotel is willing to listen to them and interact with them. On the other hand, the pending challenge for big hotel chains is to determine how they can make their stakeholders feel more special and willing to interact with the hotel. Additionally, our study shows that other content related to environmental, social and employment issues, leads to higher engagement of the audience and therefore might positively influence the hotel's awareness and online presence.

Limitations and Future Research
Before closing, a number of limitations and future research suggestions should be acknowledged. As our data did not have a normal distribution, nonparametric alternatives such as Mann-Whitney, Kruskal-Wallis, and Generalised Linear Model were adopted. Nonparametric tests are usually less powerful than corresponding tests designed for use on data that come from a specific distribution. Another limitation stems from the content analysis which is, in general, exposed to increased error as it is difficult to automate or computerize it. The content analysis is limited by availability of data and there is a risk that observed trends may not express an accurate reflection of reality.
As our study is limited to Spanish hotel chains, future studies could adopt our approach and apply it to a worldwide context, which might arguably improve the generalizability of the results. In addition, it might be of great interest to conduct research on the reasons that lead users to engage with the Twitter pages of hotel chains. Another interesting stream of research might be to propose a Twitter engagement model for hotel chains to explain the relationship between antecedent factors, the engagement that occurs and the attitudinal and behavioural effects on users.