The Factors that A ﬀ ect Usage Intentions and Travel Intentions of Travel-Related WeChat O ﬃ cial Accounts

: WeChat O ﬃ cial Accounts (WOAs)—a feature of WeChat—are thriving. However, studies on travel-related WOAs remain scarce and lack depth. This study aimed to identify the factors that inﬂuence usage intentions and travel intentions in regard to travel-related WOAs. Based on the Uses and Gratiﬁcations (U&G) theory, this study developed a theoretical model and examined the model through a web-based questionnaire. The questionnaire was distributed through the online survey tool Sojump. After analyzing the obtained data via structural equation modeling, the results revealed that: (1) social, entertainment, and information motivations are the primary factors inﬂuencing usage intentions; (2) information and entertainment motivations are important factors inﬂuencing travel intentions; and (3) usage intentions positively a ﬀ ect travel intentions and play a mediating role. The ﬁndings carry theoretical and practical signiﬁcance regarding how travel-related WOAs can improve users’ usage and travel intentions, as well as value for the actual management and marketing of WOAs.


Introduction
Since it was first launched by Tencent in January 2011, WeChat has accumulated over one billion active users and has developed into one of the most widely used smartphone applications [1]. The WeChat Official Accounts (WOAs) were established on 23 August 2012, as a platform designed primarily to allow enterprises, media companies, government departments, and individuals to share, communicate, and access information through features such as push publishing, information queries, online interactions [2]. According to the WeChat 2018 Influence Report and WeChat Economic Data Report, the number of WeChat Official Accounts (WOAs) exceeded 10 million by the end of 2017, among which active accounts reached 3.5 million, making WOAs important platforms for public discourse and important channels for the dissemination [3]. As of 2020, the number of WeChat official accounts had risen to 20 million [4].
With the upgrading of tourism infrastructure, people travel more frequently. China's tourism industry has expanded enormously and developed an extensive market base, becoming the foremost market and the largest innovation testbed in global tourism [5]. In 2019, people traveling within China reached 6.01 billion and tourism revenue amounted to about 5.73 trillion yuan (amount to approximately 0.81 trillion dollars), for annual growth rates of 8.4% and 11.7% respectively [6]. An increasing number of tourist destinations are establishing their own WOAs in attempts to enhance promotion; expand the coverage, forms, and scope of their services; and consolidate brand image [7]. From a sustainability perspective, mobile technology in general is expected to contribute to the development of sustainable tourism [8]. Information and Communication Technologies (ICT) and applications (apps) for tourists which is invisible; extrinsic motivation is the pull factor, mainly the characteristics and attraction of the destination. The causality of motivation for behavior is confirmed [45][46][47], and recent studies also show that travel motivation has a positive impact on travel intentions, which themselves lead to travel behavior [48][49][50][51].
Travel intentions, an important concept in the study of tourism, are the psychological process through which travelers select their destinations during travel-related decision-making. In their attempt to predict travelers' behaviors, Woodside and Lysonski examined situational factors and consequently opined that travel intentions reflect travelers' strong desire to visit a certain location at a certain time [52].
Previous research has found that entertainment motivation provides psychological benefits of travel which, in turn, influences travel intentions. A study by Mannell and Iso-Ahola posited that the psychological benefits produced by leisure and tourist experiences originate from the interplay between two motivational forces: the urge to escape from routine and stressful environments and the urge to seek recreational opportunities [53]. Studies have also found that information and social motivation affect traveler satisfaction which, in turn, influences travel intentions. For instance, Ross and Iso-Ahola identified a considerable similarity between the dimensions of motivation and traveler satisfaction, and suggested that knowledge-seeking, social interaction, and escapism are important shared motivation and satisfaction factors [54]. The senior tourism market has received increased attention. Using Taiwanese seniors as the study sample, the authors identified that novelty-seeking not only can be stimulated by affect, but also aroused travel intention [55]. The effects of social motivations, entertainment motivations, and information-seeking on travel intentions have also been highlighted in more recent studies [56,57]. For example, Martin et al. (2019) found that utilitarian/hedonic motivations of consumers influenced brand interaction and behavioral intention of consumers.
Considering the above and the feature of travel-related WOAs, information-seeking, self-service, entertainment, and social motivations may be the primary motivations behind travel intentions. Therefore, we propose the following hypotheses: Hypothesis 5 (H5). Information motivation positively affects travel intentions regarding travel-related WOAs.

Usage Intentions and Travel Intentions
Previous studies have shown that attitude and behavioral intentions are positively correlated [58][59][60][61]. Verhagen et al. clearly show that utilitarian, social, and hedonic motivations have a significant impact on customer behavioral intention [62]. Moreover, Charfi and Cui et al. show the mediating effects of usage intentions between utilitarian and hedonic motivations and travel intentions [63,64]. A study by Leung and Bai found that travelers' active engagement with hotel social media pages influences their revisit intentions [65]. Further, Chen et al. revealed that the novelty, understandability, and interestingness of travel blog contents influence behavioral intentions through blog-usage intentions [66].
Analysis of the above literature shows that usage intentions may positively affect travel intentions and play a mediating role. Thus, we develop the following hypothesis: Based on the above literature review and the development of hypotheses, the theoretical model is shown in Figure 1. The theoretical model comprised six variables: information motivation, self-service motivation, entertainment motivation, social motivation, usage intentions, and travel intentions. According to the U&G theory, the four types of motivations mentioned above correspond to four types of user gratification, namely: information gratification, self-service gratification, entertainment gratification, and social gratification.
Analysis of the above literature shows that usage intentions may positively affect travel intentions and play a mediating role. Thus, we develop the following hypothesis: Hypothesis 9 (H9). Usage intentions for WOAs positively affect travel intentions regarding travel-related WOAs.
Based on the above literature review and the development of hypotheses, the theoretical model is shown in Figure 1. The theoretical model comprised six variables: information motivation, self-service motivation, entertainment motivation, social motivation, usage intentions, and travel intentions. According to the U&G theory, the four types of motivations mentioned above correspond to four types of user gratification, namely: information gratification, self-service gratification, entertainment gratification, and social gratification.

Research Setting
There are abundant tourism resources in China's Fujian Province, with three UNESCO World Heritage sites (Wuyi Mountain, Fujian Earth Building and Danxia) two UNESCO Global Geoparks (Taining and Ningde), 10 national 5A tourist attractions, 10 national geological parks, 15 national nature reserves, 29 national forest parks, 137 national key cultural relic protection units, and 4 national historical and cultural cities [67]. It is one of the most important tourism provinces in China.
In recent years, the tourism market in Fujian has shown year-on-year growth. In 2018, Fujian achieved total tourism revenue of 663.46 billion yuan (amount to approximately 94.61 billion dollars), an increase of 30.5%. It received 9.01 million inbound tourists, an increase of 16.2% over the previous year. Revenue reached $9.09 billion yuan (amount to approximately 1.27 billion dollars). The number of domestic tourists received was 451.39 million, an increase of 20.3%; domestic tourism income was 603.30 billion yuan, up 32.0% [68].
This study selected the "Fujian Tourism" WOA (fjta12301) for this research, based on the following rationale: first, Fujian is a core area of China's "21st Century Maritime Silk Road" initiative (a Chinese strategic initiative to increase investment and foster collaboration) and enjoys an abundance of tourism resources; second, it is China's first national ecological civilization pilot zone (that is, under a plan to explore new practices and conduct reforms to preserve the ecology and

Research Setting
There are abundant tourism resources in China's Fujian Province, with three UNESCO World Heritage sites (Wuyi Mountain, Fujian Earth Building and Danxia) two UNESCO Global Geoparks (Taining and Ningde), 10 national 5A tourist attractions, 10 national geological parks, 15 national nature reserves, 29 national forest parks, 137 national key cultural relic protection units, and 4 national historical and cultural cities [67]. It is one of the most important tourism provinces in China.
In recent years, the tourism market in Fujian has shown year-on-year growth. In 2018, Fujian achieved total tourism revenue of 663.46 billion yuan (amount to approximately 94.61 billion dollars), an increase of 30.5%. It received 9.01 million inbound tourists, an increase of 16.2% over the previous year. Revenue reached $9.09 billion yuan (amount to approximately 1.27 billion dollars). The number of domestic tourists received was 451.39 million, an increase of 20.3%; domestic tourism income was 603.30 billion yuan, up 32.0% [68].
This study selected the "Fujian Tourism" WOA (fjta12301) for this research, based on the following rationale: first, Fujian is a core area of China's "21st Century Maritime Silk Road" initiative (a Chinese strategic initiative to increase investment and foster collaboration) and enjoys an abundance of tourism resources; second, it is China's first national ecological civilization pilot zone (that is, under a plan to explore new practices and conduct reforms to preserve the ecology and protect the environment), with "Refreshing Fujian" being the slogan of Fujian tourism [69]; third, the "Fujian Tourism" WOA is not only an important online platform for external officially regarding Fujian-related tourism information, but is also a brand new official interactive platform.

Questionnaire Design
A questionnaire survey was used to gather the necessary data for model validation. Relevant previous studies were examined in order to design and adapt model dimensions and items for the present study. Appendix A shows the sources of the items used. The initial questionnaire was modified based on two experts' opinions (that is, those of a full professor and associate professor in the tourism studies field) before testing. Then, five master's students majoring in tourism studies were recruited to read the questionnaire and identify any ambiguity. The results showed that WOA users would have a good understanding of the questionnaire instructions and items. The official questionnaire was then finalized.
The questionnaire was divided into three parts. The first section comprised the questionnaire instructions. The second section concerned the various variables of the research framework, including motivation and gratification variables (information, self-service, entertainment, and sociability) and usage and travel intentions associated with the travel-related WOAs; each item was scored using a seven-point Likert scale, with "1" representing "strongly disagree" and "7" representing "strongly agree." Finally, the third section concerned the respondents' demographic information (gender, age, education level, marital status, and monthly personal income), their daily screen time on WeChat, their history of WeChat usage, and their number of WeChat friends (please refer to Appendix A for the questionnaire items and their sources).

Data Collection
The questionnaire was published on Sojump (www.Sojump.com), the most popular online survey tool in China [70]. Many recent research studies have employed Sojump to solicit survey responses [56,71,72].
The interviewees who were addressed were randomly selected from the members who had used the "Fujian Tourism" WOA (fjta12301). They were of different ages, identities, occupations, and regions. Respondents provided their answers via computer or cellphone.
Completed respondents were collected via Sojump for statistical analysis, with a total of 741 retrieved from November 6 to 23, 2018. After excluding respondents-with repetitive IP addresses and those completed in less than 120 seconds or in more than 1800 seconds, a total of 729 valid questionnaires were eventually obtained.

Data Processing and Analytical Methods
Validity testing of the questionnaire was initially conducted using Cronbach's α. After verifying that the questionnaire possessed sufficient validity, exploratory factor analysis (EFA) was then performed to construct the dimensions of the motivation and gratification variables. Afterward, Amos 24.0 software (IBM, Somers, NY, USA, version 24.0) was employed to perform confirmatory factor analysis (CFA) and to determine the goodness-of-fit, convergent validity, and discriminant validity of the measurement model. Finally, structural equation modeling was conducted, and bootstrapping was used to test the mediation effects of the model. Table 1 shows the demographic data of the surveyed sample. Among the 729 respondents, 59% were females; the majority were unmarried (51.9%) and had a high level of education (62.3% had a bachelor's degree). In addition, 74.3% of the respondents spent over two hours daily on WeChat, and 77% had been using WeChat for over two years; 39.8% had a monthly income that exceeded 5000 yuan (equivalent to approximately $728), and 24.7% had over 200 WeChat friends. The results of this demographic data are consistent with the results of the 44th China Internet Network Development Report of the China Internet Network Information Center (CNNIC), as well as with the entire Internet user sample database. The report shows that as of June 2019, the largest group of Internet users in China were students, accounting for 26.0%; regarding age structure, as of June 2019, Internet users aged 10-39 accounted for 65.1% of all internet users, among which those aged 20-29 accounted for 24.6%. Income structure shows that by June 2019, netizens with a monthly income of 2001-5000 yuan accounted for more than one-third of the total, that is, 33.4%. All this is consistent with our demographic data [73].

Analysis of the Measurement Model
First, reliability and validity testing were conducted on the questionnaire, yielding a Cronbach's α value of 0.862. According to Kline, a reliability coefficient of over 0.8 is very satisfactory [74]. The 12 items regarding motivation and gratification then underwent EFA; analytical results revealed that the factors influencing usage and travel intentions could be classified into four dimensions, namely, information motivation, self-service motivation, entertainment motivation, and social motivation. Then, Sustainability 2020, 12, 6108 8 of 17 the remaining five items, which involved usage and travel intentions, also underwent factor analysis. The results showed that cross-loading occurred for the item "I will recommend this travel-related WOAs to my family or friends." Hair et al. suggested that any item that has a significant correlation with two or more factors should be eliminated, as it may imply a poor item design [75]. Thus, this item was eliminated, and the finalized questionnaire was used as the official version for CFA validation.
During validation, Amos version 24.0 was employed to test the goodness-of-fit, convergent validity, and discriminant validity of the model. Table 2 illustrates the goodness-of-fit results (χ2/df = 3.280; GFI = 0.951; RMSEA = 0.056; CFI= 0.941; IFI= 0.942; TLI= 0.920). During the CFA, composite reliability (CR) was used to determine the internal consistency of the model; all CRs were consequently found to exceed 0.6, which indicated satisfactory reliability [72]. The convergent validity was measured using factor loadings and average variance extracted (AVE) [75,76]. All factor loadings were consequently found to exceed 0.5, which implied a relatively high level of convergent validity. AVE reflects the average variance of item loading on a construct. Each of the constructs had an AVE value of 0.5, which indicated satisfactory convergent validity. Table 3 summarizes the statistical outcomes of the CFA. In terms of discriminant validity, Fornell and Larcker suggested that the AVE of a dimension should be larger than the square of the correlation between any two constructs [77]. The results for the present data illustrated that for all constructs the AVE values (the diagonal elements in Table 4) exceeded the square of the correlation between any two constructs (the off-diagonal elements in Table 4); hence, discriminant validity was supported. Based on the above, we determined that the measurement model possessed sufficient goodness-of-fit and satisfactory reliability, convergent validity, and discriminant validity. The diagonal values are the average variance explained (square root of AVE); the off-diagonal values are the correlation coefficients with other constructs. ** significantly correlated at the 01 level (two-tailed); * significantly correlated at the 0.05 level (two-tailed).

Analysis of the Structural Model
The structural model was validated using Amos version 24.0. The goodness-of-fit indices (χ2/df = 3.280; RMSEA = 0.056; GFI = 0.951; CFI = 0.941; IFI = 0.942; TLI = 0.920) implied a good model fit. The structural model was then evaluated to validate the hypothesized relationships. The results showed that the standardized path coefficients were statistically significant. Information motivation (B = 0.190, p < 0.01), entertainment motivation (B = 0.242, p < 0.01), and social motivation (B = 0.406, p < 0.001) were found to have significantly positive effects on usage intentions, which indicated that they were important predictors of usage intentions. Thus, H1, H3, and H4 were statistically validated. Similarly, information motivation (B = 0.241, p < 0.01) and entertainment motivation (B = 0.217, p < 0.01) also showed significant positive effects on travel intentions, which identified them as important predictors of travel intentions. Thus, H5 and H7 were statistically validated. Finally, usage intentions (B = 0.262, p < 0.01) significantly influenced travel intentions, which implied that the former is a factor that influences the latter. H9, therefore, was statistically supported. Figure 2 and Table 5 present the results of the hypothesis testing and path coefficients.

Analysis of the Structural Model
The structural model was validated using Amos version 24.0. The goodness-of-fit indices (χ2/df = 3.280; RMSEA = 0.056; GFI = 0.951; CFI = 0.941; IFI = 0.942; TLI = 0.920) implied a good model fit. The structural model was then evaluated to validate the hypothesized relationships. The results showed that the standardized path coefficients were statistically significant. Information motivation (B = 0.190, p < 0.01), entertainment motivation (B = 0.242, p < 0.01), and social motivation (B = 0.406, p < 0.001) were found to have significantly positive effects on usage intentions, which indicated that they were important predictors of usage intentions. Thus, H1, H3, and H4 were statistically validated. Similarly, information motivation (B = 0.241, p < 0.01) and entertainment motivation (B = 0.217, p < 0.01) also showed significant positive effects on travel intentions, which identified them as important predictors of travel intentions. Thus, H5 and H7 were statistically validated. Finally, usage intentions (B = 0.262, p < 0.01) significantly influenced travel intentions, which implied that the former is a factor that influences the latter. H9, therefore, was statistically supported. Figure 2 and Table 5 present the results of the hypothesis testing and path coefficients.

Testing for Mediation Effect
The mediation effect of the model was tested through bootstrapping. As the calculated bootstrap confidence interval (CI) did not include zero, we concluded that there was an indirect effect [77,78]. The 95% bias-corrected CI and 95% percentile CI were calculated using Amos 24.0 and a bootstrapping approach with 2000 iterations (Table 6).  Table 6 shows that the total effect of information on travel intentions was 0.291, which indicated that the path model had a significant total effect; meanwhile, the direct effect of information on travel intentions was 0.241, which indicated a significant direct effect; finally, the indirect effect of information on travel intentions through usage intentions was 0.050, which implied a significant indirect effect. Thus, usage intentions partially mediate the effect of information on travel intentions.
The total effect of entertainment on travel intentions was 0.487, which indicated that the total effect of the path model was significant. Meanwhile, the direct effect of entertainment on travel intentions was 0.049, which implied an insignificant direct effect; finally, the indirect effect of entertainment on travel intentions through usage intentions was 0.151, denoting a significant indirect effect. Thus, usage intentions fully mediate the effect of entertainment on travel intentions.
Given that the total effect of social interaction on travel intentions was 0.155, the total effect of the path model was insignificant.
In summary, usage intentions mediate the model comprising factors influencing WOAs users' travel intentions. More specifically, they partially mediate the relationship between information motivation and travel intentions while fully mediating that between entertainment motivation and travel intentions.

Discussion
This is a preliminary study that aimed to examine the effects of information, self-service, entertainment, and social motivations have on usage intentions for travel-related WOAs and, consequently, on users' travel intentions. This study also sought to evaluate the relationship between users' usage intentions and travel intentions in relation to travel-related WOAs. The results validated the positive effects of information, entertainment, and social motivations on usage intentions regarding travel-related WOAs, the positive effects of information and entertainment motivations on WOA users' travel intentions, and the positive effects of users' usage intentions on their travel intentions.

Factor Influencing WOAs Users' Usage Intentions
Analysis of the direct effect of psychological gratification types (information, entertainment, and social interaction) on usage intentions showed that social motivation (0.406) is the main factor influencing usage intentions. This verified H4, and past research has also pointed out that one of the main motivations for people to use social media is sociality [19,35,78]. In regard to social interaction, WOAs allows users to easily connect with strangers and make new friends. The significant positive influence of sociability on user intentions identified in the present study accords with the findings of Chang and Zhu (2011), who indicated that social interactions are crucial motivating factors behind WOA users' usage behaviors [14]. Thus, actively encouraging WOA users to interact with their friends is an effective method of advancing the development of WOAs.
Information (0.190) and entertainment (0.242) are secondary influencing factors regarding usage intentions, which validates the hypothesis H1 and H3. These findings are consistent with the previous research [79][80][81][82]. Entertainment has a positive effect on users' intention [83], but due to platform characteristics of WOAs, the main function is a content push, which is a passively received communication mode. Interaction is poor, and it is not as good as immersive media such as WeChat, Weibo, or online games.

Effect on Travel Intentions
Regarding the direct effect of psychological gratification types (information and entertainment) on travel intentions, information (0.241) was found to be the main factor influencing travel intentions, followed by entertainment (0.217); social gratifications do not influence travel intentions. The positive effect of information on users' travel intentions verifies hypothesis H5 and is consistent with previous research [66,84,85]. Wen and Huang (2019) analyzed the motivations of 348 visitors and found that unique seeking information predicted travel intentions [50].
Entertainment motivation (0.217) exerts a positive influence on travel intentions. H7 was validated, in line with previous research by Chiu and Yang [86]. Usage intentions influence travel intentions (0.262), so H9 was also validated. Thus, entertainment ultimately influences users' travel intentions, but this influence is fully mediated by usage intentions, matching the findings of Huang (2008) [28]. Unlike previous findings by Aluri et al. (2016) demonstrating that sociality affects behavior intentions [87], this result shows that sociality does not affect the user's travel intentions, which does not support H8. Nevertheless, travel intentions can be effectively reinforced by enhancing the amusement and entertainment facilitated by WOAs.

Self-Service Motivation
Self-service factors have no significant effect on either usage intentions or travel intentions; that is, H2 and H6 are not supported. Generally speaking, all travel-related WOAs now possess a reservation feature. WOAs also offer other self-service features, such as audio services, navigation maps, and membership offers [88]. However, the present study found that these features do not influence either usage intentions or travel intentions. It is recommended, therefore, that WOAs reduce the focus on these features. The reason for these features' lack of influence, as inferred by the author, is that these self-services can all be substituted by specialized platforms or media. Use of navigation maps, for example, may be undermined by users' preference for more dedicated applications such as Gaode Maps, Baidu Maps, and Google Maps.
These findings underline the dominant role of social interaction for WOAs users' usage intentions, the dominant role of information for WOA users' travel intentions, and the mediating effect of usage intentions on travel intentions.
Our study aims to equip marketing analysts and executives with insights that can be implemented in the future. According to the results, the travel-related WOAs will be designed to be more entertaining. Reviews, photos, and videos shared via WOAs or online platforms motivate users to visit a place, attend an event, or buy a product. The real-time access to relevant tips and guides, traveling instructions, or inspirational photos/videos has ultimately changed the way users plan an activity, which almost is to increase tourists [89]. Furthermore, big data from the WOAs and mobile device signaling elucidates wider patterns of tourist movement, as applied to forecast travel demands and sustainable management of a destination [90].

Conclusions and Limitations
There are three implications of the research ranging from theory to practice for sharpening the skills of tourism managers.

Improve the Social Design of Interfaces in Order to Enhance Usage Intentions
This study has shown that social gratification constitutes the foremost factor influencing WOA users' usage intentions and serves as a means by which travel-related WOAs can stay connected with users (who can meet new friends through traveling and interact through the WOAs). Regardless of how such relationships are formed, travel-related WOAs should preferably establish features for interacting and socializing in order to better enhance usage intentions. For instance, they could make announcements through notifications and promote a transition from one-way communication with users to many-way interactions involving relevant parties.

Enhancing Travel Intentions by Enriching the Content Published
This study found that information is the primary factor influencing users' travel intentions, and that they can also be enhanced by a high level of entertainment. Hence, WOA managers should publish professional information that is useful for users, such as information that can help travelers make informed travel-related decisions, enhancing their usage intentions. WOAs can also enrich the content published via various other means. For instance, travel information could be presented in the form of building an information recommender system for users according to the similarities in groups of people [91].

Reducing Focus on Self-Service Features
Travel-related WOAs offer self-services including ticket reservations (entry tickets and other reservations), navigation maps, online shopping, online customer services, and member centers [63]. Despite their appeal to users, however, this study proved that they fail to enhance users' usage and travel intentions. Thus, travel-related WOAs should remove some or all of these features and focus on the advantages of the WOA design, which relate to information, entertainment, and social gratification. Efforts should be made to enhance features such as information regarding tourist attractions, characteristic activities and attractions, recommended itineraries, nearby services, city guides, audio guides, and reviews and comments, as well as opportunities for interactions among travelers, as such features gratify users' information, entertainment, and social needs.
As with all studies, the present study has some limitations. First, as it focused on travel-related WOAs in China, the results might be limited to the Chinese population. Caution should be exercised when extrapolating the results to other countries and other social media; and future studies should