Keywords

1 Introduction

Customer engagement (CE) can bring ample benefits to tourism firms (see a systematic review of So et al. [1]). Among others, CE can improve value creation, customer trust, brand loyalty, and relationship quality [1, 2]. In the social media domain, the enhancement of CE depends on the quality of social media content shared by tourism firms [2]. This suggests that knowledge of content marketing is imperative. Content marketing assists firms in formulating and distributing content that can satisfy their customers, which eventually leads to an increase in CE on social media [3]. One of the constructs that can influence CE on social media is nonverbal information. In interactive technology, nonverbal information may include pictures, videos, and sounds [4].

Content marketing scholars have examined how nonverbal information can impact CE [3, 5]. Despite the attempts, some limitations are still present. To begin with, the focus was merely on photos and videos. Research has neglected the audio aspect (e.g., sound) of nonverbal information. In fact, in marketing in general, studies about how sound and music on social media play a role in shaping consumer behaviour remain unavailable. Knowledge about this matter is still waiting to be discovered. Given such a situation and the ubiquity of sound-based social media (e.g., TikTok), there is a need to explore the role of sound in content marketing. Another literature gap worth investigating is the fact that the object of content marketing research was never explicitly tourism small and medium enterprises (SMEs). Compared to their larger enterprises counterpart, tourism SMEs have a limited understanding of deploying social media as a marketing tool [6]. This situation highlights the necessity to study content marketing strategies and CE on social media in the context of tourism SMEs to enhance theory and practice. In addition, content marketing literature only concentrated either on Facebook or Instagram (see e.g., [3, 5]). To the best of our knowledge, content marketing studies involving TikTok remain absent. Because CE differs across social media platforms [7] and TikTok's global monthly active users have been growing exponentially in recent years (i.e., from 689 million in 2020 to more than 1 billion in 2021 [8]), content marketing scholars need to scrutinize TikTok.

Based on the literature lacunae above, this study aims to explore the role of sound in influencing CE (i.e., views and shares) in the context of TikTok and tourism SMEs. Theoretically, this study enriches tourism CE and content marketing literature. Practically, insights from this research can help tourism SMEs strengthen their content marketing strategies on TikTok.

2 Literature Review

2.1 Sound

In this explorative study on TikTok, we define sound as all the audio formats present on marketer-generated content (MGC; e.g., monologues, music, songs, movie dialogues, and other formats). Based on its sources, we categorize sound into two: cover and original. Cover sound originates from other sources (e.g., Justin Bieber songs and sounds that other TikTok users create), while original sound is created by firms themselves.

Extant social media and content marketing scholars drew on media richness theory [9] when studying the effects of nonverbal information on CE. The theory contends that rich media formats can enhance understanding and stimulate senses. Communication media containing various elements (e.g., videos) are rich, whereas those with fewer attributes (e.g., photos) are low in richness [3]. In this study, we classify cover sound as high richness media because the sound consists of sound effects, music with various tones, or noises. Meanwhile, we consider original sound as low richness media because it usually involves simple sound, such as people speaking.

Findings on the topic of nonverbal information from the lens of media richness theory remain inconclusive. For instance, while rich media (i.e., videos) are significantly positive for likes, they are insignificant for comments [10, 11]. Also, while Moran et al. [5] discovered that all rich media formats positively influence CE, Cvijikj and Michahelles [12] found varying effects. Considering such contrasting evidence, there is no guarantee that cover sound can enhance CE albeit the use of it is pervasive on TikTok.

2.2 Marketer-Generated Content

As mentioned above, sound materializes along with MGC on TikTok. Thus, it is important to analyze the interaction between sound and MGC in impacting CE. MGC in this research corresponds to firms’ messages created and distributed through social media [13], which we divide into three categories: emotional, informational, and transactional. Emotional content relates to affect-laden MGC aimed to elicit sensory and emotional reactions [3]. Informational content is those social media posts conveying information (either related or unrelated to firms’ products and services) in a non-promotional fashion. Transactional social media content involves those messages containing giveaways, promotions, donations, sales, and other monetary benefits [7].

2.3 Customer Engagement

CE refers to “a consumer's positively valenced brand-related cognitive, emotional and behavioral activity during or related to focal consumer/brand interactions” [14]. In this study, the spotlight is the behavioural facet of CE consisting of views and shares. TikTok is still in secrecy about how its algorithm delivers content to its users [15], and thus how the counts of views can be increased remains questionable. Due to this reason, it is substantial to inspect factors driving views on TikTok. Additionally, we concentrate on shares because the more people share content, the more the content reaches current and potential customers, making marketing efforts more efficient [16].

3 Methods

Following extant relevant studies [3, 5, 7], we conducted content analysis on TikTok. We collected data from all the posted content of seven travel guide services for a highly visited mountain in Indonesia, Mount Bromo. The services are Bromo Project (@bromoproject.id), Yuk Bromo (@yukbromo), Hai Bromo (@haibromo), Bromo Creative.com (@bromocreative.com), Tripjawaindah.com (@bromopedia), Bromo Hore (@bromo_horee), and Bromo Full Senyum (@bromo250k). We used these services as our sample because they are highly active on TikTok. Two research assistants coded the data, and they must adhere to the coding instructions we have created. Particularly for the three MGC constructs, we only considered the texts written on the TikTok videos. This decision manifested as a result of our observation prior to the data collection. We noticed that all the travel guide services shared entertaining videos on their accounts. Therefore, if we code the types of videos, our study will yield less meaningful results (i.e., we only compare the informational and transactional posts). We then further observed the videos and found that the services mostly delivered written messages on their videos. Because the texts are in the forms of emotional, informational, and transactional, we captured these types of texts as MGC. Besides MGC and sound, we also recorded branded hashtags (no branded hashtag as the baseline) and days of posting (weekend as the baseline) as our control variables. Also, while all other constructs are mutually exclusive, MGC is non-mutually exclusive. This means that one TikTok video can contain one or more MGC types (e.g., a TikTok video can be emotional and transactional simultaneously). After removing outliers (i.e., extreme high views and shares), our final dataset comprised 660 TikTok posts, 4,092,289 views, and 10,920 shares. These data were then analyzed using negative binomial regression (NBR). According to research [17], NBR is more efficient and appropriate than other regression methods when (1) the dependent variables are count data with positive-only integers (e.g., views and shares); and (2) the Poisson assumption of equidispersion is violated (i.e., the variance values of the dependent variables are large). Our data fit these criteria.

4 Results

Our models explaining views (LR χ2 (9, N = 660) = 32.418, p < 0.01) and shares (LR χ2 (9, N = 660) = 20.508, p < 0.05) are significant as a whole. The results show that cover sound has no direct effect either on views (0.143, p > 0.05) and shares (0.324, p > 0.05). Additionally, we discover that cover sound has no interaction effects with any MGC in affecting views (emotional: −0.244, p > 0.05; informational: −0.234, p > 0.05; transactional: −0.164, p > 0.05) and shares (emotional: −0.207, p > 0.05; informational: −0.1, p > 0.05; transactional:  −0.888, p > 0.05). Individually, emotional content has significant and positive effects on views (0.44, p < 0.05), while informational (0.317, p > 0.05) and transactional (0.527, p > 0.05) posts have no significant effects on views. Also, transactional social media posts have significant and positive effects on shares (0.898, p < 0.05), while informational (−0.18, p > 0.05) and emotional (0.203, p > 0.05) posts have no significant effects on shares.

5 Implications

Our explorative paper contributes to tourism and content marketing literature in three ways. First, we explore the role of sound in driving CE. We demonstrate that although cover sound is ubiquitous on TikTok, it has no direct and interaction effects on CE. This endeavour expands content marketing literature as existing studies only concentrated on the visual aspects of nonverbal information (e.g., photos and videos). Second, while extant tourism CE studies merely examined government-controlled or large tourism firms, we explicitly and particularly analyzed factors influencing CE in the context of tourism SMEs. Third, we enrich content marketing literature by discussing TikTok. This is substantial, considering social media is highly contextual [3], and prior studies [5, 7] primarily scrutinized Instagram or Facebook. Practically, our findings can guide tourism SMEs in optimising their content marketing strategies on TikTok. We suggest tourism SMEs to focus less on sound but more on MGC types. To increase views, tourism SMEs need to share more emotional content; and to improve shares, they need to distribute more transactional posts.