探討社群媒體行銷在旅遊業的影響 - 政大學術集成

全文

(1)國立政治大學資訊管理學系. 碩士學位論文指導教授:張欣綠博士. 立. 政治大. ‧ 國. 學. 探討社群媒體行銷在旅遊業的影響. ‧. in Tourism. n. al. er. io. sit. y. Nat. Examining the Effect of Social Media Marketing. Ch. engchi. i n U. v. 研究生：吳岱諭中華民國一○三年七月.

(2) ACKNOWLEDGEMENT My deepest gratitude goes first and foremost to my advisor, Dr. Hsin-Lu Chang. Thank you so much for your constant guidance and support on this thesis as well as your careful reading of the manuscript. You have walked me through all the stages of the thesis and helped me work out my questions during the difficult time of the thesis. Without your patient instruction, this thesis could not have reached its present form. I would also like to express my sincere thanks to the case company, a well-known travel agency in Taiwan. Concrete advice and assistance offered by the managers make a valuable contribution to this study. I am grateful to my partner, Yu-Hsian Kung. I am thankful that you have always helped me. I'm really glad you're my partner and friend. Additionally, members in EBLAB are so nice and friendly. Thank you for your kind response to my problems in. 立. the past two years.. 政治大. ‧ 國. 學. I also owe my special gratitude to my friends and classmates who gave me their help and time in listening to me and providing suggestions for my thesis. Additionally,. ‧. thank you for sharing your life stories with me and for your friendship.. Nat. y. sit. n. al. er. io. Last but not least, to my dear family, thank you for encouraging me when I was feeling stressed out and upset. I am especially grateful to my parents for their support through these years. I always know that they believe in me and want the best for me.. Ch. engchi. i n U. v. Dai Yu Wu 07.2014.

(3) 探討社群媒體行銷在旅遊業的影響摘要近年來，透過社群媒體向消費者宣傳商品與理念的旅遊服務供應商日益增多，但卻有不少企業苦於衡量其中的效果。雖然文獻已提出社群媒體的行銷機會與發展性，然而實際對商品銷售進行分析的文獻仍有限，因此本研究試圖驗證社群媒體行銷活動是否會正向影響旅遊商品的銷售。此外考量到不同的商品特質可能會對媒體行銷效果帶來不同程度的影響，因此本研究將藉由以下五種旅遊商品特性，包括旅遊結構、旅客參與度、旅遊範圍、商品價格與天數，來探討社群媒體行銷與商品銷售的關係。本研究將一間國內知名旅行社於2012至2013年期間在 Facebook粉絲專頁上所行銷過的相關商品作為實驗組，並找出同時期販售但未在 Facebook上行銷的類似商品作為控制組，進行差異中之差異法(Difference in Difference)，以探討在Facebook行銷前後，實驗組與對照組商品銷售的差異。研究結果顯示Facebook行銷活動對旅遊商品的購買有正向影響，此外透過社群媒體宣傳較低結構的旅遊、較為放鬆的行程、範圍較小的目的地或相對低價的商品更有機會提升銷售，而天數為唯一無法調節Facebook行銷活動與銷售關係之商品特性。. 立. 政治大. ‧. ‧ 國. 學. 關鍵詞：社群媒體行銷、Facebook 行銷活動、旅遊商品分類、旅遊產業. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v.

(4) Examining the Effect of Social Media Marketing in Tourism ABSTRACT A growing number of tourism industry suppliers are using social media to promote their services and reach target customers despite some doubt regarding the effectiveness of these tools. Nevertheless, few researchers have examined the effects of social media marketing, and especially its effects on sales. Therefore, our study aims to assess the impact of social media on the purchase of tourism products. Additionally, to explore whether the influence of social media marketing changes among different types of tourism products, we develop a typology of tourism products that classifies tourism along the following five dimensions: (1) the structure of tourism, (2) the involvement of tourists, (3) the scope of tourism, (4) the price of products, and (5) the length of a tour. All tours with Facebook campaigns conducted by our case company from February 1, 2012 to November 30, 2013 are selected as our target tourism products. Moreover, we choose certain products without Facebook. 立. 政治大. ‧. ‧ 國. 學. n. al. er. io. sit. y. Nat. campaigns that were available for purchase during the same time period as a control group. We obtain the sales data from the case company and calculate the sales of each product before and after the Facebook campaigns. We then apply a difference-in-difference approach, comparing the average changes in sales performance of the treatment group with those of the control group. The results show that Facebook campaign activities have a positive impact on purchases of tourism products. Furthermore, sales are more likely to increase when a travel agency promotes products that are less structured, limited in scope, relatively lower-priced, or require less tourist involvement.. Ch. engchi. i n U. v. Keywords: Social Media Marketing, Facebook Campaigns, Tourism Product Classification, Tourism Industry.

(5) CONTENT. CONTENT ..................................................................................................................... i TABLES ...................................................................................................................... iii FIGURES ..................................................................................................................... iv CHAPTER 1: INTRODUCTION ............................................................................... 1 CHAPTER 2: LITERATURE REVIEW................................................................... 4 2.1. Social media activities in tourism ............................................................................ 4. 政治大 CHAPTER 3: RESEARCH FRAMEWORK .......................................................... 13 立 Media effects with different types of products ........................................................ 7. 3.1. Traditional EC marketing versus Traditional EC and Social media marketing ..... 13. 3.2. Effect of social media marketing on tourism products .......................................... 14. ‧ 國. Involvement of Tourists ............................................................................ 15. sit. y. Nat. 3.2.2. Structure of Tourism ................................................................................. 14. ‧. 3.2.1. 學. 2.2. 3.2.4. Price of Products ....................................................................................... 16. 3.2.5. Length of Tour .......................................................................................... 17. n. al. er. Scope of Tourism ...................................................................................... 16. io. 3.2.3. Ch. engchi. i n U. v. CHAPTER 4: RESEARCH METHODOLOGY .................................................... 18 4.1. Case Background ................................................................................................... 18. 4.2. Data Collection ...................................................................................................... 19. 4.3. Sample Design and Evaluation .............................................................................. 22. CHAPTER 5: RESULTS AND DISCUSSION ....................................................... 24 5.1. Data Analysis .................................................................................................. 24. 5.2. Findings.................................................................................................................. 32. 5.3. Discussion .............................................................................................................. 34. CHAPTER 6: CONCLUSION.................................................................................. 36. i.

(6) 6.1. Summary ................................................................................................................ 36. 6.2. Limitations and Implications for Future Research ................................................. 37. REFERENCE ............................................................................................................. 38. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. ii. i n U. v.

(7) TABLES Table 2-1: Explanations of the Five Usages of Social Media ........................................ 5 Table 2-2: Social Media Usages in Different Tourism Cases ........................................ 7 Table 2-3: Five Dimensions of Tourism Products ....................................................... 11 Table 4-1: Classification of Company A’s Facebook Campaigns ................................ 18 Table 4-2: Measures of Tourism Product Classification and Sales .............................. 20 Table 4-3: Means for Key Variables ............................................................................ 22 Table 5-1: Average Sales Before and After Facebook Campaign Activities Conducted in the Treatment Group ........................................................................ 24 Table 5-2: Testing of Hypothesis 1 .............................................................................. 25 Table 5-3: Testing of Hypothesis 2 .............................................................................. 26 Table 5-4: Testing of Hypothesis 3 .............................................................................. 27 Table 5-5: Testing of Hypothesis 4 .............................................................................. 28. 立. 政治大. ‧. ‧ 國. 學. Table 5-6: Testing of Hypothesis 5 .............................................................................. 30 Table 5-7: Testing of Hypothesis 6 .............................................................................. 31. n. er. io. sit. y. Nat. al. Ch. engchi. iii. i n U. v.

(8) FIGURES Figure 3-1: Research Framework................................................................................. 13 Figure 5-1-a.................................................................................................................. 26 Figure 5-1-b ................................................................................................................. 27 Figure 5-2-a.................................................................................................................. 28 Figure 5-2-b ................................................................................................................. 28 Figure 5-3-a.................................................................................................................. 29 Figure 5-3-b ................................................................................................................. 29 Figure 5-4-a.................................................................................................................. 30 Figure 5-4-b ................................................................................................................. 30 Figure 5-5-a.................................................................................................................. 31 Figure 5-5-b ................................................................................................................. 32. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. iv. i n U. v.

(9) CHAPTER 1: INTRODUCTION Web 2.0 and social media heavily influence online activities. Nevertheless, there appears to be confusion regarding the two terms, which are often used interchangeably. The term “Web 2.0,” used to represent the second generation of the World Wide Web, can be thought of as the technical infrastructure that facilitates consumer-generated content (Berthon, Pitt, Plangger, & Shapiro, 2012). Indeed, Web 2.0 technologies help to create easy-to-use platforms to allow people to create their own content and that enables a shift in focus from companies to consumers. How does the term “social media” differ from Web 2.0? Social media are perceived as a Web 2.0 development, which is a set of Internet-based applications that build on the ideological and technological foundations of Web 2.0 and that allow the creation and exchange of content (Kaplan & Haenlein, 2010). In other words, social media not only allow users to generate content but also encourage users to share the content among their networks. Kaplan et al. classified social media into six different types: collaborative projects, blogs, content communities, social networking sites, virtual game worlds, and virtual social worlds.. 立. 政治大. ‧. ‧ 國. 學. In this paper, we will use social media as a term based on Web 2.0 applications. n. al. er. io. sit. y. Nat. that satisfy the above-mentioned definitions. Today, the growing popularity of social media has forged connections among Internet users, and the media have become increasingly important sources of information for consumers. According to Market Force (Olenski, 2012), 81 percent of US respondents indicated that Internet posts from their friends directly influenced their purchase decisions.. Ch. engchi. i n U. v. Tourism has long been one of the most important components of the electronic commerce (e-commerce or EC) world, and its influence has profoundly changed the structure of the industry (Milano, Baggio, & Piattelli, 2011). In tourism, social media have dramatically changed how consumers plan and buy travel-related products (Buhalis & Law, 2008). A study conducted by the Opinion Research Corporation indicated that 82 percent of respondents expressed that they had checked online reviews, blogs, and other online customer feedback before purchasing a travel-related product, accounting for the largest portion of all products and services (eMarketer, 2008). These results show that social media serve as an interactive platform to help consumers obtain product information or recommendations from other people, especially for travel products. Similarly, companies are likely to benefit from using social media as a marketing channel to approach potential consumers. For example, comScore conducted an 1.

(10) experiment with Starbucks, following the in-store purchasing behavior of Starbucks’ Facebook fans. The result was that those who were exposed to Starbucks’ messages visited Starbucks and purchased 38 percent more often than those who were not exposed to such messages (Koetsier, 2012). In the tourism industry, research from Funsherpa Company illustrates the influence of social media on US travelers, with 52 percent of travelers having changed their plans after researching their trip on social media sites and with half of travel companies claiming that direct bookings have been generated from social media (Bennett, 2012). As social media have appeared to gradually change existing marketing practices on a fundamental level, companies cannot ignore this new marketing channel. Today, businesses continue to turn to social media to promote their services and to reach customers in their own way, but businesses appear to have different views on the effect of social media marketing. According to the Social Media Marketing Industry Report in 2013, 97 percent of marketers indicated that they are participating in social media marketing, but only approximately one in four marketers claimed that they are able to measure the return from their social media activities. Additionally,. 立. 政治大. ‧ 國. 學. ‧. regarding the effectiveness of Facebook marketing, 37 percent of marketers agreed that their Facebook efforts are effective, whereas the remainder are uncertain or have opposite opinions (Stelzner, 2013). Another report surveyed by American Express. n. al. er. io. sit. y. Nat. Business Travel (2010) indicated that half of corporate travel professionals use social media to communicate with travelers and to stay on top of the latest travel industry information; nevertheless, for one-fifth of travel professionals, the greatest barrier to the use of social media is an unclear business case or return on investment (ROI). From the above-mentioned data, we know that an increasing number of companies are attempting to use social media as important marketing tools despite widespread doubt regarding the effectiveness of such tools. Such uncertainty is well recognized as a large barrier to developing social media marketing.. Ch. engchi. i n U. v. Social media marketing has been attracting more attention from both companies and researchers. A number of studies have discussed opportunities related to social media marketing and provided strategies and principles for utilizing online social networking applications for managers (Berthon et al., 2012; Hsu, 2012; Kaplan et al., 2010); however, few researchers have examined the effects of social media marketing, especially the effects on sales. Moreover, although some studies have focused on the important influence of social media usage such as online reviews or word of mouth in the travel planning process (Sparks & Browning, 2011; Chen, Shang & Li, 2013), few studies have considered the effect of social media on travel-related purchasing decisions. 2.

(11) Therefore, our study aims to assess the effects of social media marketing in the tourism industry by examining the influence of social media on the purchase of tourism products. We are also interested in exploring whether the influence of social media marketing changes among different types of tourism products. The results can help companies in the tourism industry to ensure the financial outcome of social media marketing and further enhance their understanding of which types of tourism products are suitable for selling through social media. Our study collects data from a tourism agency, which has already used social networking sites such as Facebook and blogs to promote their business and to reach customers in recent years. The following questions will be investigated: (1) Do social media marketing activities have a positive effect on the purchase of tourism products and even increase sales? (2) Do different types of tourism products provided by marketers in their social networking websites have different levels of influence on sales?. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 3. i n U. v.

(12) CHAPTER 2: LITERATURE REVIEW With marketing strategies and techniques used in Web 2.0 applications potentially available to businesses, a wide range of social media activities continue to appear and develop on the Internet. Our study aims to assess the effects of social media marketing in tourism; hence, the first section will review the literature that elaborates the functionalities and characteristics of social media that can support different activities for businesses, and especially for tourism industry suppliers (e.g., hotels, transportation providers, attractions, travel agencies). Meanwhile, given the influence of product types on media marketing, we will also review relevant literature in the second section to identify important tourism products that should be considered in our study.. 政治大 2.1 Social media activities in tourism 立. ‧. ‧ 國. 學. The original Web 1.0 allowed only one-way communication through static webpages; hence, consumers could only view and could not contribute to the content of webpages. In a sense, website publishers communicated with users as if through lecture (Thackeray, Neiger, Hanson, & McKenzie, 2008). In contrast, today’s social. sit. y. Nat. media not only facilitate new firm-customer interactions but also expose these interactions to other users (Gallaugher & Ransbotham, 2010).. n. al. er. io. According to Culnan et al. (2010), social media can support multiple activities that include branding (including advertising, public relations, and content delivery), sales (including calls to action, e.g., links to purchase items), customer service and support, and product development. Furthermore, companies potentially gain business value from these activities when customers engage with them on a regular basis, co-creating content and sharing power. Kietzmann et al. (2011) proposed a framework that defines seven social media functionalities, including identity, conversations, sharing, presence, relationships, reputation, and groups. Different social media activities can be defined by the extent to which they focus on some or all of these functionalities.. Ch. engchi. i n U. v. Companies launch social media activities for different purposes, which can be divided into diverse categories. According to Bernoff and Li (2008), there are five primary ways for companies to use social media: research and development (listening), marketing (talking), sales (energizing), customer support (supporting) and operations (managing). Furthermore, social media metrics have been proposed to measure performance in the above categories. Li and Soonius (2012) noted that 4.

(13) companies conduct Facebook campaigns for similar reasons, and they further distinguish the following twelve campaign categories: news, history, corporate social responsibility, contests, winners, albums, promotions, recruitment, co-creation, likes, polls, and questions. The two studies above are summarized in Table 2-1 with respect to the five categories of using social media. Table 2-1: Explanations of the Five Usages of Social Media Study. Bernoff and Li (2008). Li and Soonius (2012). Gaining insights from customers and using that input in the innovation process. Engaging customers to co-create products and services or simply improve existing ones (e.g., using questions, polls, co-creation). products or services. (e.g., using photos, contests, likes, news, history). Category. customers and using them to products and services influence others conducting a promotion). sit. Enabling customers to help Providing easier and quicker one another solve problems customer support. io. Providing employees with tools to enable them to assist one another in finding more effective ways to conduct business. al. n. Managing. (e.g.,. y. Nat. Supporting. enthusiastic Directly or indirectly selling. ‧. Energizing. 學. Identifying. Ch. er. Talking. 治政大 fans aware of Using conversations with Making 立 to promote companies’ brands and activities customers ‧ 國. Listening. Serving as an internal platform for sharing knowledge or attracting new employees (e.g., recruiting). engchi. i n U. v. Tourism industry suppliers have implemented social media activities to reach some or all of the above objectives. For example, Malaysia's low-cost airline AirAsia conducted the Friendsy campaign to increase brand awareness of Sydney as a new Australian hub. The company developed a Facebook app showing a virtual plane that AirAsia’s fans could fill with 302 of their Facebook friends. The winner of the campaign would have an opportunity to take their friends from Sydney to Kuala Lumpur, where they would then spend three nights with accommodations paid for by AirAsia, before returning home on their private plane. As a result, the AirAsia Facebook fan base grew by 30 percent, and the Friendsy competition reached 2,291,483 people on Facebook. The campaign helped achieve an average 82.5 percent 5.

(14) load factor on all routes from Sydney, exceeding the initial goal of AirAsia (Facebook Studio, 2013). Obermutten is another noted example of the use of social media marketing. This village, located in Switzerland's remote mountain region of Graubunden, uses Facebook to increase awareness of this place that has fewer than 80 residents. Obermutten began a campaign that allows anyone who likes the village's Facebook page to have their pictures posted on the village bulletin board. Because generating the connection between the online world and this remote village attracts attention from users all over the world, the Facebook page quickly reached more 40,000 fans. Additionally, this activity also draws traffic to the Graubunden tourism website, with visits increasing by 250 percent. An effort that cost 10,000 Swiss francs resulted in the media equivalent of $2.4 million, with approximately 60 million media impressions. Obermutten subsequently expanded this successful activity to attempt to promote the village as a long-term and enduring attraction. Therefore, Obermutten established a new landmark called the Obermutten International Museum of Friendship (OIMOF) using co-creation with its fans. All presents from fans around the world are stored in OIMOF and are also photographed, described and featured in the Obermutten Facebook photo gallery (Advertising Age, 2012).. 立. 政治大. ‧. ‧ 國. 學. n. al. er. io. sit. y. Nat. Some activities can engage fans and provide tourists with a new way to travel. For example, Gettysburg National Military Park uses Facebook to connect followers to some of the battlefield's lesser-known sites. Moreover, the park launched the “52 Footsteps” Facebook Challenge, which outlines a weekly story that can lead visitors to specific spots on the battlefield. Upon finding a spot, visitors are encouraged to take a photo and upload it to the park's Facebook page (Rowland & Rowland, 2013).. Ch. engchi. i n U. v. A company may also use official social media platforms to provide travelers with the latest news or inside stories about the company. As an example, in the Marriott on the Move blog (www.blogs.marriott.com), J.W. Bill Marriott, the chairman and CEO of an American diversified hospitality company, Marriott International, Inc., regularly updates articles. He replies to the comments posted by visitors and adds a personal touch, such as news about his family, to the blog (Thevenot, 2007). An increasing number of travel providers use social media to address customer problems and to manage their companies. JetBlue, for instance, has hired six staffers to monitor missives from the company's 1.4 million Twitter followers. In one incident in which a JetBlue passenger tweeted about not being able to obtain a seat assignment next to his child, the airline quickly responded while the passenger was still in the boarding area (Laura, n.d.). Sometimes social media are also used to facilitate 6.

(15) interactions between consumers and even to enable customers to help one another solve problems. For example, KLM Royal Dutch Airlines has established “Club China” on its website for frequent travelers to China to exchange itineraries, travel tips, and business contacts (Winer, 2009). The above-mentioned examples show that various social media activities are conducted by travel providers to serve different objectives. We summarize these six cases in Table 2-2 based on the categorization made in prior literature (Bernoff & Li, 2008; Li & Soonius, 2012). Table2-2: Social Media Usages in Different Tourism Cases Gettysburg. Case AirAsia. Obermutten. Category. 立√. Listening. Supporting. JetBlue. Royal Dutch Airline. √. √. √. √. √. ‧. Managing. Military Inc. 治政 Park 大. √. ‧ 國. Energizing. √. International,. 學. Talking. National. KLM. Marriott. √. Nat. y. √. sit. n. al. er. io. In our study, we focus on the marketing effects of social media in the tourism industry. According to Bernoff and Li (2008), increased sales can be used as a metric to evaluate the effect of social media applications with respect to talking and energizing. These usages also allow users to take part in conversations and create their own content, such as reviews, questions, compliments or suggestions, which are known as User Generated Content (UGC); hence, these campaigns can be characterized as firm-initiated but customer-implemented (Godes & Mayzlin, 2009). We will identify these types of social media activities from all the campaigns conducted on the Facebook fan page of our case company and further explore their influences by examining changes in the sales of its tourism products. Moreover, we will explore whether these influences change with the promotion of different types of tourism products.. Ch. engchi. i n U. v. 2.2 Media effects with different types of products The suitability of the media for marketing to consumers greatly depends on the characteristics of the marketed products and services (Peterson et al., 1997). Legg and Baker (1987) proposed that the effectiveness of service advertising depends on both 7.

(16) the service type and the media forms that are selected. Therefore, it is necessary to clearly consider product and service characteristics when evaluating the effects of media. Some early research suggested that different categories of products were likely to affect the marketing performance of traditional media. A study by Bowen and Chaffee (1974) categorized products into two types: high-involvement and low-involvement. Print advertisements for these two types of products were selected and classified as either informational or non-informational. The results of the study showed that informational print advertisements were much more effective for high-involvement products than for low-involvement products in both evaluation of marketing content and willingness to buy the advertised product. In contrast, non-informational print advertisements had no differential effect between the two types of products.. 政治大. Preston (1970) proposed that the content of various media had inherent differences that affected how people responded. The results of the study showed that television advertisements may be suitable for low-involvement products because of the concentration of advertisements that distinguish advertised items only as brands of a given product. In contrast, magazine advertisements could elicit a relatively high-involvement response because these advertisements differentiate advertised. 立. ‧. ‧ 國. 學. items based on distinct product attributes.. Nat. y. sit. n. al. er. io. Stafford and Day (1995) examined two media (print and radio) and two message appeals (rational and emotional) to determine whether certain media or message strategies were more suitable than others for two types of retail services, classified as experiential and utilitarian services. Full-service restaurants and hotels were identified as examples of experiential services; photograph processing and automobile repair were examples of utilitarian services. The results of the study indicated that a rational appeal generated more positive attitudes toward the advertisement than an emotional appeal. Additionally, experiential retail services may benefit more from radio and print advertising than utilitarian services.. Ch. engchi. i n U. v. Since the mid-1990s, the Internet has become increasingly popular. Some researchers began to compare it with traditional media with respect to the marketing effects based on types of products. For example, Yoon and Kim (2001) aimed to investigate the attitudes of respondents toward online advertising in comparison with traditional media advertisements; furthermore, they attempted to link product characteristics to choice of media. Their results suggested that the Internet is a medium that is better suited to high-involvement products. The classification system proposed by Peterson et al. (1997) categorized 8.

(17) products and services on the Internet along three dimensions: cost and frequency of purchase (low outlay and frequently purchased versus high outlay and infrequently purchased), value proposition (tangible or physical versus intangible or information), and the degree of differentiation (high versus low). Building on the above-mentioned classification criteria, the results of a study by Phau and Poon (2000) indicated that the different types of products and services would significantly influence the consumer choice between a conventional retail channel and an Internet channel. Generally, products and services that have a low outlay, are frequently purchased, have intangible value proposition, and that are relatively highly differentiated are more likely to be purchased via the Internet. A study by Dijkstra et al. (2005) explored the effects of advertising campaigns for books and wine via television, print, and Internet banners. For the book advertisement, the participants who were exposed to print only had a higher level of advertisement recall than the Internet banner-only participants. For the wine advertisement, however, the advertisement recall of the television-only condition was significantly higher than that of the print-only condition.. 立. 政治大. ‧ 國. 學. ‧. The use of online video games with subtle or overt commercial messages related to a company, product, or service is perceived as a new advertising channel for. n. al. er. io. sit. y. Nat. marketers in today’s increasingly diverse media environment. An existing study (Choi & Lee, 2012) explored whether the use of an animated spokes-character in an online video game affects the persuasiveness of an advertisement and examined how hedonic and utilitarian aspects of products moderate the magnitude of such effects. Tissue is selected as an example of a utilitarian product that is purchased to fulfill functional and practical needs. In contrast, coffee is chosen as an example of a hedonic product that is primarily used to satisfy experiential needs and fun. The results show that the influences of the spokes-character on brand evaluation and purchase intentions are more obvious for utilitarian products than for hedonic products.. Ch. engchi. i n U. v. Based on the previous research, there is ample evidence that different types of products or services influence the effects of media marketing. Therefore, we will consider the suitability of media based on the categories of advertised products in our study. Today, the growing number of companies turning to social media to promote their products and services has increased the importance of linking social media marketing and product types. However, few academic studies have examined this topic; consequently, we will investigate the effects of product types on social media activities in tourism. Ideally, tourism products are produced to meet market demands in a 9.

(18) cost-efficient manner and are based on the wise use of the cultural and natural resources of the destination (Smith, 1994). Lew (1987) proposed a classification of ideographic tourist attractions that includes three categories: general environments, specific features and inclusive environments. General environments are broad in scope and often large in scale. These environments generally require little or no tourist involvement. Specific features are smaller in scale and often have clear connections to tourism. Inclusive environments are the principal attractions that draw tourists to a destination and cause tourists to become completely engrossed in the experience. According to Foo et al. (2004), the researchers Yiannakis and Gibson (1992) designed the Tourist Role Preference Scale (TRPS), in which the positioning of each tourist role reflects the optimal destination characteristics and relative behavior with respect to all others. The TRPS includes the following three dimensions:. The strangeness–familiarity dimension indicates the level of novelty and strangeness that a tourist prefers in a travel environment. The structure–independence (high structure–low structure) dimension. ‧. 3.. 學. 2.. ‧ 國. 1.. 政治大 The stimulation–tranquility dimension explains an individual’s need for 立 active stimulation from his/her surrounding travel environment. indicates the level of organization and planning undertaken to travel.. Nat. y. sit. n. al. er. io. Orth and Turecková (2002) proposed five factors in destination choices, including push factors, pull factors, exogenous factors, total travel cost and available time, to identify travel styles. Push factors refer to psychological motives, such as the need to escape, relax or belong. Pull factors refer to man-made, natural and socio-cultural attractions. Exogenous factors refer to the political and social environment, health threats, natural disasters, and terrorism. Total travel costs include the cost of tips and on-site costs. Available time refers to the amount of travel and vacation time.. Ch. engchi. i n U. v. To identify recognizable product characteristics, Buckley (2007) examined price, duration, group size, skill requirements and remoteness for a set of commercial adventure tours involving various activities and locations. The results of the study indicated that most adventure tours can be distinguished broadly along two product dimensions: price and tour duration.. 10.

(19) Table 2-3: Five Dimensions of Tourism Products Proposed Dimensions in. Yiannakis and. Lew (1987). Gibson (1992). this Study. Orth and Turecková (2002). (1) Structure. structure– independence dimension. (2) Tourist Involvement. stimulation– tranquility dimension. push factors. strangeness– familiarity dimension. pull factors. (3) Scope. (4) Price. total travel cost price available time. ‧. (5) Length. skill requirements. 政治大. ‧ 國. 立. (2007). 學. general environments; specific features; inclusive environments. Buckley. Nat. y. duration. sit. n. al. er. io. According to the studies listed in Table 2-3, we develop a typology of tourism products that classifies tourism along the following five dimensions in our study:. i n U. v. (1) Structure of tourism (structure versus independence). Ch. engchi. Highly structured travel refers to trips that are to a large extent arranged by travel agencies. This type of tour generally contains transportation, food, accommodation and entertainment, which are advertised and sold together at an inclusive price. In contrast, independent travel does not include a set of scheduled sightseeing activities, and hence, tourists have more flexibility in making their own travel arrangements. (2) Involvement of tourists (stimulation versus tranquility) Stimulation indicates a desire for high levels of interaction with surrounding environments that can be satisfied by participating in adventure activities, water sports (e.g., sailing, scuba diving, whitewater rafting) or winter activities (e.g., skiing, snowboarding). In contrast, a more tranquil journey involves little interaction with the local environment. Tourists merely want to relax and escape to a vacation destination. They enjoy the change of pace and want to take it easy during their stay (Dolnicar & Leisch, 2003). 11.

(20) (3) Scope of tourism (broad versus limited) The conditions of specific features and the inclusive environment proposed by Lew (1987) are relatively similar in our study; thus, we combine them into a category called specific tourism, which means that tourism is limited in scope and there is often a specific theme to satisfy the needs of tourists. Additionally, the listing of specific attractions by name is most often used in studies of small areas, such as cities (Lew, 1987). Many of the attractions visited and events attended by tourists are in urban areas (Silberberg, 1995). For example, tourists can visit museums and locations of cultural and historical significance. They can also enjoy the urban environment along with entertainment, nightlife and sporting events (Edwards, et al., 2008). In our study, island destinations can also be regarded as specific tourism. An island destination is developed based not only on its sea-based amenities but also the local history, natural attractions and expressive culture (Cameron & Gatewood, 2008). When tourists visit only one city or island during a trip, they can spend more time in an area. In contrast, tourists are less likely to visit local places for an extended amount of time when they visit many cities or regions during a trip (i.e., the tourism is broad in scope).. 立. 政治大. ‧ 國. 學. (4) Price of products (high versus low). ‧. er. io. sit. y. Nat. The cost of tourism to visitors includes the cost of transportation to and from the destination and the cost of ground content, such as accommodation, tour services, food and beverages, and entertainment (Dwyer et al., 2000). In our study, tourism products were classified as either higher- or lower-priced based on their selling price.. al. v i n C h in a tourist’sUdecision-making process, and Length of stay is one of the key elements engchi n. (5) Length of tour (long versus short). promotional campaigns must therefore be adjusted to the tourist’s decisions with regard to the length of time spent at a destination (Martínez-Garcia & Raya, 2008). In this study, tourism products were classified as either longer or shorter trips based on duration of travel. The tourism products that have been advertised on social media for our case company will be classified along the above-mentioned five dimensions to enable us to examine the effects of their social media marketing under the influences of different product types. Furthermore, the results are likely to identify which types of tourism products are more or less suitable for using social media marketing in terms of generating more sales.. 12.

(21) CHAPTER 3: RESEARCH FRAMEWORK According to the existing literature, we develop our research framework as shown below (Figure 3-1). We seek to explore the effect of social media marketing and tourism product types on sales performance. Therefore, in this framework, we will examine the relationship between these effects and sales and answer the following questions: (1) can social media marketing facilitate the sale of tourism products, and (2) which types of tourism products are more suitable to conduct social media marketing activities for increasing sales, whereas which types of products are less suitable? We will discuss in detail how to construct each research hypothesis in the following sections.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. i n U. Figure 3-1: Research Framework. engchi. v. 3.1 Traditional EC marketing versus Traditional EC and Social media marketing Traditional e-commerce is enabled by Web 1.0, featuring one-way communication. Social commerce is enabled by Web 2.0, a platform that allows bidirectional communication, making consumers become better informed than ever before (Baghdadi, 2013). According to Stephen and Toubia (2010), social media allow people to actively participate in the marketing and selling of products and services in online communities. Moreover, Kim and Hawamdeh (2011) stated that companies utilized social media that help them improve their marketing strategies, increase website traffic and improve sales. That is, traditional EC marketing has some weaknesses that can be complemented by social media, and thus, companies that incorporate social media marketing into traditional EC marketing can generate better 13.

(22) sales. Some companies can indeed increase sales using social media marketing activities. Dell, for example, stated that its use of Twitter had generated $1 million in incremental revenue resulting from sales alerts (Kaplan et al., 2010). The Royal Opera House in London made 30 percent of its sales through its Facebook fan page (Ang, 2011). Old Spice, an American brand of male grooming products, was almost certainly in the decline phase of its product lifecycle; however, social media marketing activities revived the brand. Sales of Old Spice increased by 55 percent between April and June 2010 through posting on YouTube with millions of downloads, successful humorous tweets on Twitter, and the creation of a massively popular fan page on Facebook (Berthon, Pitt, Plangger, & Shapiro, 2012). Therefore, we expect that tourism agencies can increase the sales of tourism products with the assistance of social media marketing. Our hypothesis is as follows:. 立. 政治大. ‧. ‧ 國. 學. H1. Marketing tourism products in dual channels (i.e., e-commerce and social media) is likely to achieve better sales performance than simply marketing in traditional e-commerce.. 3.2 Effect of social media marketing on tourism products. al. n. tour, which will be discussed in the following section. 3.2.1 Structure of Tourism. Ch. engchi. er. io. sit. y. Nat. To realize the relationship between social media marketing and sales performance, we propose five dimensions to classify tourism products: the structure of tourism, the involvement of tourists, the scope of tourism, the price of products, and the length of a. i n U. v. Based on Chapter 2, the structure of tourism indicates the level of planning that is performed by travel agencies. Accordingly, highly structured travel refers to trips that are to a large extent arranged by travel agencies, such as package tours. In contrast, less structured travel refers to trips that are primarily planned by tourists themselves, such as independent travel, flight and hotel packages. Shih (2011) indicated that Facebook enables companies to develop online relationships with customers and even to influence conversations about their products. Moreover, given the flexibility to start marketing campaigns at any time, companies can lunch timely promotions. For travel agencies, after promoting tourism products to their followers on Facebook, these potential consumers may be interested in certain products or special offers even if they did not originally have any travel plan. In such 14.

(23) a case, Facebook can be a preferable channel for companies to market independent travel compared with package tours because independent travel typically requires more conversations between companies and tourists than package tours do, and Facebook can thus provide a more flexible channel for companies to discuss with customers in terms of travel dates and travel plans compared with traditional websites. Furthermore, according to Lo et al. (2011), people who like to share travel photos on social networking sites appear to prefer independent travel rather than package tours. Such individuals are more likely to become acquainted with locals and even attempt to learn some of the local language. Therefore, we expect a stronger effect of using social media marketing activities to promote independent tours compared with marketing package tours. Our hypothesis is as follows: H2. Incorporating social media into the marketing of independent travel is more likely to increase the sales than in the marketing of package tours.. 立. 政治大. 3.2.2 Involvement of Tourists. ‧ 國. 學. ‧. According to Chapter 2, a high-involvement tourist refers to an individual’s need for active stimulation from his/her surrounding vacation environment. Their desires can be satisfied by physical or adventure activities that require more interactions with the. sit. y. Nat. surrounding environment. In contrast, those who prefer tranquil environments tend to have little interaction with local tourist attractions.. n. al. er. io. One of the main characteristics of social media is that these media gather engaged and active participants of a given company (Kaplan et al., 2010). Bartram (2001) stated that increased exposure of adventure activities in the media may indeed stimulate involvement in an activity. Additionally, those who primarily use social media to share their trips tend to be quite young and appear to want a deeper, more authentic travel experience (Lo et al., 2011). According to Foo et al. (2004), younger tourists may prefer higher levels of stimulation, whereas older tourists prefer more tranquil environments. Therefore, using social media to promote tourism products that include more exciting activities may target active users and increase their intention to engage in these activities and to subsequently purchase the products. Thus, our hypothesis is formed as follows:. Ch. engchi. i n U. v. H3. Incorporating social media into marketing products that require greater tourist. involvement is more likely to increase sales than in marketing products that require less tourist involvement.. 15.

(24) 3.2.3 Scope of Tourism According to Chapter 2, tourists may have fewer opportunities to gain a deeper understanding of local destinations when tourism is broad in scope. In contrast, specific tourism is relatively limited in scope, such as visiting only one city or island during a trip. Moreover, there is often a specific theme to satisfy the needs of tourists. According to Lew (2008), specialty tourism products that are different from mass tourism destinations may have less information available and may be treated with greater caution by consumers. However, advances in communication technology and social media have led to rapid growth in special-interest travel destinations and experiences as well as travel to more remote corners of the globe. Therefore, tourism businesses can also participate in online social media to build a web presence and reputation for these products. For example, providing special interest blogs and tailoring the content may be an effective approach to target specific interest groups (Schmallegger & Carson, 2008). Consequently, we hypothesize as follows:. 立. 政治大. ‧ 國. 學. H4. Incorporating social media into marketing tourism that is limited in scope is more likely to increase sales than in marketing tourism that is broad in scope.. ‧. 3.2.4 Price of Products. Nat. y. sit. n. al. er. io. When the price of a product is relatively high, customers tend to search for more information (Beatty & Smith, 1987). Because product price is a stimulus for customer thinking, high-priced products lead customers to devote more thought to their purchase intentions (Wathieu & Bertini, 2007). Additionally, product price is related to perceived risk (Wang & Chang, 2013). That is, higher risk is involved when selecting products with higher prices.. Ch. engchi. i n U. v. According to the research findings of Wang et al. (2013), if a product’s price is relatively high, then the information and recommendations provided by strong-tie contacts from Facebook have a more significant effect on purchase intentions than the information and recommendations provided by weak-tie sources from Facebook. However, this effect is not found for low-priced products. Accordingly, we expect that interpersonal influence and communication on social media are more important when consumers consider buying relatively high-priced tourism products. Thus, we develop the following hypothesis: H5. Incorporating social media into the marketing of higher-priced tourism products is more likely to increase sales than in the marketing of lower-priced tourism 16.

(25) products. 3.2.5 Length of Tour In general, the risk of travel increases with the length of stay. Consequently, the anticipation of a longer trip stimulates more information search (Fodness & Murray, 1999). According to Tsiotsou and Ratten (2010), Web 2.0 technologies can foster the sharing of information and lower product uncertainty. Additionally, Litvin et al. (2008) described online interpersonal influence or eWOM as a potentially cost-effective means of marketing tourism, especially for high-risk tourism products. Therefore, we believe that social media supported by Web 2.0 technologies can be an effective tool to reduce the perceived risk of a longer trip and thus to promote better marketing effects. Our hypothesis is as follows:. 治政 H6. Incorporating social media into the marketing of大 longer trips is more likely to 立 of shorter trips. increase sales than in the marketing ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 17. i n U. v.

(26) CHAPTER 4: RESEARCH METHODOLOGY 4.1 Case Background Tourism has grown and developed into one of the largest industries in the world because of the increased interest in leisure activities. For outbound tourism from Taiwan, the official tourism bureau indicated that the total number of overseas trips in 2012 was over 10.2 million. Approximately 20.6 percent of people traveled abroad from Taiwan at least once in 2012. Moreover, approximately 85 percent of those individuals sought the assistance of travel agencies, whereas less than 5 percent entrusted their domestic travel to travel agencies. In view of the importance of international travel for travel agencies, we will focus on this type of travel product offered by our case company (called Company A).. 政治大 Company A was founded in 1978 and is one of the best-known travel agencies in 立 Taiwan. Company A has three subsidiary companies, one of which is responsible for. ‧ 國. 學. ‧. managing e-commerce and online marketing. The company sells various tourism products, including domestic travel, international travel, personal travel services, corporate travel services, and inbound tourism, on its official website. When it. n. al. er. io. sit. y. Nat. appeared that business would most likely benefit from the advantages of Web 2.0 applications, Company A began to use them in interactions with consumers. Company A has had its own Facebook business fan page since September 2009, and the page currently has more than 266,000 Facebook fans. Based on prior literature on the five uses of social media (Bernoff & Li, 2008; Li & Soonius, 2012), we classified Company A’s Facebook campaigns conducted from February 1, 2012 to November 30, 2013 and identified those that would be measured in this study (Table 4-1).. Ch. engchi. i n U. v. Table 4-1: Classification of Company A’s Facebook Campaigns Five Usages of Social Media. Listening. Descriptions of Company A’s Facebook Campaigns The company uses a “review app” that allows consumers to rate and share opinions of its products and services. There are currently more than 4,000 reviews on the Facebook page, and anyone can respond to these reviews.. 18. Objects of Observation in this Study.

(27) The page includes discussion of daily life and shares moods with customers while uploading cheerful photographs. The fan page uses calming landscape images, inspiring words, sentimental quotes, questions or contests to engage customers. Moreover, the page may provide links to related products for customers.. √. Energizing. The company posts special offers and links directly on Facebook.. √. Supporting. When customers raise questions about their products or services, the company can address customers’ doubts quickly.. Managing. Company A does not use Facebook as an internal tool for sharing knowledge or recruiting new employees.. 立. 政治大. 學. ‧ 國. Talking. 4.2 Data Collection. ‧. n. al. er. io. sit. y. Nat. First, we identified all the international tours that have been marketed using any of the Facebook campaigns listed in Table 4-1. All the tours with campaigns conducted from February 1, 2012 to November 30, 2013 were selected as our target tourism products. Because we wanted to inspect the changes in the sales of tourism products before and after the Facebook campaigns, we excluded the products that were only made available a few days before the marketing campaigns were conducted or that would be removed from the online store soon after the campaigns concluded (i.e., there might be no orders for the products before or after the Facebook campaigns). We obtained the sales data of all our targeted products for 2012-2013 from Company A.. Ch. engchi. i n U. v. The manager of the case company indicated that the sales price of each product changes over time, and hence the company uses the number of customers to evaluate annual sales performance. Therefore, according to the trading volume of each target product, we calculated the average monthly number of customers to measure sales performance before and after the Facebook campaigns. Additionally, because some products might be the subjects of more than one campaign, for each target product, we calculated the sales (i.e., the average monthly number of customers) before and after the first campaign activity began. However, if the length of time between the first campaign and any subsequent ones to market the same product exceeded four months, we also calculated the sales performance before and after the later campaign to 19.

(28) examine its influence on the target product. Essentially, we calculated the sales result of each target product based on its sales start and end dates. That is, for each target product, sales performance pre-Facebook campaign was measured from its sales start date to the time when its first Facebook campaign was conducted. Sales performance post-Facebook campaign was calculated from the first Facebook campaign to its sales end date. However, there are long and short periods for products sold in the store. Hence, if the sale's start date was earlier than three months before the first Facebook campaign, the average monthly number of customers was calculated in the three months before the first campaign. Additionally, if the sale’s end date was later than three months after the first Facebook campaign concluded, the sales result was calculated in the first three months after the campaign was conducted.. 政治大. To examine the effects of Facebook marketing activities based on different types of tourism products, the measurements of tourism product classification, including the structure of tourism (ST), the involvement of tourists (IN), the scope of tourism (SC), the price of products (PR) and the length of tour (LE), are defined in Table 4-2. Another point to note is how to determine the price and the length of a product. Although package tours and independent group trips had fixed departure and return. 立. ‧. ‧ 國. 學. n. al. er. io. sit. y. Nat. dates, consumers who bought flight and hotel packages could decide their length of stay at tourism destinations for themselves. Therefore, for each flight and hotel package, the number of travel days was decided by calculating the average length of tour in the sales data. Additionally, travel agencies sell the same product at different prices that vary according to the moment, the place of purchase and other factors; hence, the price of each tourism product was decided by calculating the average sales price based on sales data.. Ch. engchi. i n U. v. Table 4-2: Measures of Tourism Product Classification and Sales Components. Structure of tourism (adapted from Yiannakis and Gibson, 1992). Items. Measures of Tourism Product Classification. ST. The structure of tourism is measured by the level of planning that is performed by travel agencies. Accordingly, a tourism product is less structured when it includes only air tickets and accommodations and not a set of scheduled sightseeing activities (e.g., independent travel, flight and hotel packages). In contrast, a package tour is labeled a highly structured tourism product. 20.

(29) The involvement of tourists is measured by their need for active stimulation from the surrounding vacation environment. Therefore, a tourism product requires greater tourist involvement Involvement of tourists (adapted from Yiannakis and Gibson, 1992). when it includes a least one adventure activity (e.g., whitewater rafting, scuba diving, visiting a theme park or amusement park) to satisfy the need for stimulation. In contrast, a tourism product requires less tourist involvement when it is more tranquil and does not include any adventure activity. Tourists merely want to sightsee or shop for a relaxing vacation.. IN. Scope of tourism (adapted from Lew, 1987). 政治大. A tour is broad in scope when tourists travel to many cities or regions during a trip. In contrast, a tourism product is limited in scope when tourists travel to one city or island during a trip and often when there is a specific theme to satisfy the needs of tourists.. 立 SC. ‧ 國. 學. ‧. A tourism product is labeled a higher-priced. io. y. n. al. sit. PR. er. Nat. Price of products. Length of tour. Sales performance. product when its price is greater than the mean sales price of all the samples plus 0.5 times standard deviation. In contrast, it is labeled a lower-priced product when its price is less than the mean sales price of all the samples minus 0.5 times standard deviation.. Ch. engchi. i n U. v. LE. A tourism product is labeled a longer trip when its length is greater than the mean length of all the samples plus 0.5 times standard deviation. In contrast, it is considered a shorter trip when its length is less than the mean length of all the samples minus 0.5 times standard deviation.. Sales. The average monthly number of customers (i.e., the average monthly trading volume of each product) is used as a measure of the sales performance of each target product.. 21.

(30) 4.3 Sample Design and Evaluation Based on the sale start and end dates of each target product, we also selected a product from those without Facebook campaigns conducted during the same time period. Moreover, to eliminate some of the effect of selection bias, we chose those that matched the target products in term of the types or geographical relations of tourism destinations as a control group. The treatment group contained 188 samples, as did the control group. We then calculated the sales (average monthly number of customers) of the control group based on the dates and times when Facebook campaign activities were conducted for their paired samples in the treatment group. Comparisons of sales in the treatment and control groups before and after the Facebook campaigns offer a simple method for evaluating the effects of Facebook marketing.. 政治大. Table 4-3 presents the means for several key variables in our data set. The means are presented separately for stores in the treatment and control groups, along with t statistics for the null hypothesis that the means are equal in the two groups. The results suggest that the distribution of product types in the control group is similar to that in the treatment group, which meets the threshold for further research. Additionally, the average sales before the Facebook campaigns in the treatment and. 立. ‧. ‧ 國. 學. n. al. er. io. sit. y. Nat. control groups are not significantly different; however, the average sales after Facebook campaigns in the treatment group are significantly greater than those in the control group. That is, Facebook campaigns are likely to impact the sales of tourism products.. Ch. i n U. v. Table 4-3: Means for Key Variables Variable. i e nSamples g c hin:. Treatment Group. Control Group. t*. a. Independent travel. 37.77% (0.49). 38.3% (0.49). -0.106. b. Package tour. 62.23% (0.49). 61.7% (0.49). 0.106. c. East Asia. 60.64% (0.49). 59.57% (0.49). 0.210. d. Southeast Asia. 15.96% (0.37). 18.09% (0.39). -0.548. 1. Distribution of product types (percentage):. 22.

(31) e. Europe. 10.64% (0.31). 11.70% (0.32). -0.327. f. America. 7.45% (0.26). 3.72% (0.19). 1.573. g. Oceania. 4.26% (0.20). 5.85% (0.24). -0.705. h. South Africa, Middle East, South Asia. 1.06% (0.10). 1.06% (0.10). 0.000. 36346.78 (32486). 35606.94 (28072). 0.236. 5.54 (2.62). 5.59 (2.38). -0.206. i. Price j. Tour length 2. Sales (means):. 立. treatment Facebook. 33.36 (36.96). -0.800. 3.099***. ‧. campaigns. 49.15 (59.30). 學. After group’s. 政治 37.32大(44.83). 33.89 (37.97). ‧ 國. Before treatment group’s Facebook campaigns. Nat. n. al. er. io. sit. y. Note: See text for definitions. Standard deviations are given in parentheses. *Test of equality of means in treatment and control groups. *p< 0.1 **p<0.05 ***p<0.01. Ch. engchi. 23. i n U. v.

(32) CHAPTER 5: RESULTS AND DISCUSSION 5.1 Data Analysis To determine the impact of Facebook campaign activities on product sales, we employ a difference-in-difference approach, comparing the average change in sales before and after Facebook activities for the treatment group with the average changes in sales over the same time period for the control group. This method is adopted because it is intended to eliminate some of the effect of selection bias. The following Table 5-1 summarizes the changes in average sales in our data. We present the data by different sample group in columns (i) and (ii). We also show the differences in average sales between the treatment group and control group in column (iii). Row 3 of the table presents the changes in average sales over time. These entries are simply the differences between the averages before and after the Facebook activities (i.e., row 2 minus row 1). We find that the relative gain (the “difference in difference” of the changes in sales) is 19.22, with a t statistic for the null hypothesis that the changes in mean sales are equal in the two groups (t = 4.79, p = 0.00). The results in Table 5-1 suggest that Facebook campaigns conducted by the. 立. 政治大. 19.22. increase. in. ‧. ‧ 國. 學. a. sales. (the average. monthly. sit. y. Nat. case company led to number of customers).. n. al. er. io. Table 5-1: Average Sales Before and After Facebook Campaign Activities Conducted in the Treatment Group. Ch. e nSamples gchi. i n U. v. Variable. Treatment Group (i). Control Group (ii). Difference, Treatment-Control (iii). 1. Average sales before Facebook campaigns. 33.89. 37.32. -3.43. 2. Average sales after Facebook campaigns. 49.15. 33.36. 15.79. 3. Changes sales. 15.26. -3.96. 19.22. in. mean. 24.

(33) The same result can be obtained by using the following regression model:. Sales = α+ β1 Post + β2 Face + β3 Post * Face + ε Where Sales is the average monthly number of customers per product, Post is a dummy variable equal to one if the period is after the Facebook campaign and zero otherwise, Face is an indicator variable taking the value of one if the tourism product has been marketed by Facebook campaign activities between February 2012 and November 2013 and zero otherwise, Post * Face is an interaction term between Post and Face, and ε is an error term. The estimate of coefficient β3 is directly comparable to the simple difference-in-difference of sales changes in column (iii), row 3 of Table 5-1.. 政治大. The coefficients of the key variables are presented in Table 5-2. The coefficient of Post*Face (β3 = 0.181) is found to be significant at the 0.01 level and thus provides support for H1. That is, Facebook campaign activities have a significantly positive impact on sales of tourism products.. 立. ‧ 國. 學. Table 5-2: Testing of Hypothesis 1. Β -0.037. -0.728. y. 0.181. 2.887 ***. io. 752. al. n. N. -0.842. mean of sales. 38.43. SD of sales. 46.00. 2. R. Note: *p< 0.1. sit. Post*Face. Nat. Face. -0.043. Ch. 0.019 **p<0.05. t. er. Post. ‧. Variable. engchi. i n U. v. ***p<0.01. While the comparisons in Table 5-1 make no allowance for examining the moderating effects of product type on the influence of Facebook campaigns, these effects can be incorporated in the regression model. First, according to Table 4-2 regarding the measures of tourism product classification in Chapter 4, three categorical variables for tourism product types, including the structure of tourism (ST), the involvement of tourists (IN), and the scope of tourism (SC), are created. For each sample, if it is labeled less structured, requiring greater tourist involvement or limited in scope, it takes the value of one; if not, it takes the value of zero. Additionally, two ratio variables, including the price of products (PR) and the length of tour (LE), are also created to divide the samples into two groups. 25.

(34) We then conducted a linear regression analysis such as the one described above; however, this time, we selected part of the data set by specifying a rule based on values of new variables, as shown in Table 5-3. The coefficients β3 in both clusters of structure of tourism are significant at the 0.05 level; however, the coefficient of less structured tourism products (β3 = 0.215) is greater than that of highly structured tourism products (β3 = 0.167). Then, we used Pseudo-F statistics to compare β3 between two clusters. For p = 0.01, the result of the analysis (f-value = 45.8) is greater than F (0.99, 1, 749) = 6.669, and hence, the difference in coefficients between the two groups is significant at the 0.01 level. That is, H2 is supported, and the result provides evidence that treatment has a stronger effect when Facebook campaigns are conducted to market less structured tourism products (e.g., independent tours) than those that are highly structured (e.g., package tours). Figure 5-1 shows the interaction patterns for the two groups (structure versus independence) in the treatment and control groups. Table 5-3: Testing of Hypothesis 2. 立. Independence (ST=1). -0.037 (-0.572). -0.057 (-0.682). n. al. mean SD. 466. C h41.59 engchi 49.62. 2. R. Note: *p< 0.1. sit. 0.167 (2.103)**. er. io. n. -0.153 (-1.841)*. y. 0.017 (0.266). Nat. Post*Face. Structure (ST=0). ‧. Face. ‧ 國. Post. Structure of Tourism. 學. Variable. 政治大. 0.026 **p<0.05. i n U. v. 0.215 (2.111)** 286 33.28 38.93 0.021. ***p<0.01 Treatment Group. 60 50 40 sales 30 20 10 0. 54.94 39.48 39.61 24.68. before. after Time. Figure 5-1-a 26. structure independence.

(35) Control Group 40 38 36 sales 34 32 30 28. 37.77 36.6. 34.08 32.2. before. structure independence. after Time. Figure 5-1-b In Table 5-4, the coefficient β3 for tourism products requiring greater tourist involvement is 0.167, whereas the coefficient β3 for those that require less tourist involvement is 0.188, and its treatment effect is significant at the 0.05 level. Additionally, for p = 0.01, f-value = 19.37 is greater than F (0.99, 1, 749) = 6.669, indicating that the difference in coefficients between the two groups is significant at the 0.01 level. Therefore, H3 is not supported but is significant in the opposite direction. That is, treatment has a greater effect when Company A promotes products that are more relaxing and less stimulating on its Facebook fan page. Figure 5-2 shows the comparison between the two clusters (tranquility versus stimulation) in the. 立. 政治大. ‧. ‧ 國. 學. treatment and control groups.. io. Tranquility (IN=0). al. n. Post. y. sit. Involvement of Tourists. Face Post*Face n. er. Nat. Variable. Table 5-4: Testing of Hypothesis 3. -0.038 n C h(-0.611) U engchi. Stimulation (IN=1). iv. -0.056 (-0.606). -0.047 (-0.753). -0.021 (-0.234). 0.188 (2.492)**. 0.167 (1.479). 514. 238. mean. 36.15. 43.35. SD. 46.52. 44.55. R2. 0.021. 0.016. Note: *p< 0.1. **p<0.05. ***p<0.01. 27.

(36) Treatment Group 60. 52.6 47.49. 40.56 40 sales. Tranquility Stimulation. 30.69 20 0 before. Time. after. Figure 5-2-a Control Group 50. 42.46. 40 sales. 37.46. 30 20 10. 立. 治31.53 Tranquility 35.03 政大Stimulation. ‧ 國. before. 學. 0 after Time. Figure 5-2-b. ‧. n. al. er. io. sit. y. Nat. In Table 5-5, the coefficient β3 of tourism products that are limited in scope (β3 = 0.232) is significant at the 0.01 level and greater than the coefficient β3 of those that are broad in scope (β3 = 0.081). Moreover, for p = 0.01, f-value = 147.26 is greater than F (0.99, 1, 749) = 6.669, showing that the coefficients differ significantly at the 0.01 level. Therefore, H4 is supported. That is, treatment has a stronger effect when Facebook campaigns are conducted to market city tours abroad or island destinations than travel that is broad in scope (i.e., tourists travel to many cities, regions or even countries during a trip). Figure 5-3 presents the changes in the average sales of the two groups (broad versus limited) in the treatment and control groups.. Ch. engchi. i n U. v. Table 5-5: Testing of Hypothesis 4 Variable. Scope of Tourism Broad (SC=0). Limited (SC=1). Post. -0.012 (-0.128). -0.056 (-0.914). Face. -0.094 (-1.078). -0.011 (-0.180). Post*Face. 0.081 (0.734). 0.232 (3.048) ***. 28.

(37) n. 268. 484. mean. 38.16. 38.58. SD. 42.61. 47.82. R2. 0.006. 0.04. Note. *p< 0.1. **p<0.05. ***p<0.01 Treatment Group. 60 50 40 sales 30 20 10 0. 55.19 34.50 39.63. broad. 32.93. limited after 政Time 治大 Figure 5-3-a. before. 40.92. 39.92. 40. 30.21. 30. ‧. sales. Control Group 50. 學. ‧ 國. 立. 35.59. broad. 20. Nat. limited. sit. y. 10. io. before. n. al. Ch. Time. after. Figure 5-3-b. engchi. er. 0. i n U. v. To obtain a better understanding of the influence of price on the purchase of products, samples that are larger than 0.5 times standard deviation from the mean are included in the analysis. That is, we use the mean sales price of all the samples (mean = 35,977) plus or minus 0.5 times standard deviation (SD = 30,301) as cut-off points, as shown in Table 5-6. As a result, the coefficients β3 in lower-priced products (β3 = 0.203) are significant at the 0.1 level and greater than that in higher-priced products (β3 = 0.066). Then, for p = 0.01, f-value = 64.12 is larger than F (0.99, 1, 373) = 6.703. That is, the difference in coefficients between the two groups is significant at the 0.01 level. Therefore, H5 is not supported, and the prices of products posted on Facebook are negatively related to sales performance. This result shows that the effect of Facebook campaigns on sales is stronger when Company A promotes products with relatively lower prices. Figure 5-4 shows the interaction patterns for the two clusters (low versus high) in the treatment and control groups. 29.

(38) Table 5-6: Testing of Hypothesis 5 Price of Products. Variables. Low (PR<20,827). High (PR>51,128). Post. -0.073 (-0.822). 0.014 (0.114). Face. -0.174 (-1.936)*. 0.044 (0.354). Post*Face. 0.203 (1.861)*. 0.066 (0.439). n. 246. 130. mean. 34.67. 20.52. SD. 39.01. 17.21. 2. R. 50 38.81. 40 26. 30. ‧. sales. Treatment Group. 0.011. 學. ‧ 國. Note. *p< 0.1. 0.019 治政大 **p<0.05 ***p<0.01 立. low. 23.59. 20. Nat. 20.44. high. sit. y. 10 0. n. al. Ch. time. after. er. io. before. i n U. Figure 5-4-a. engchi. Control Group. v. 50 40 sales. 39.58 33.93. 30 20. 18.92. 19.14. low high. 10 0 before. time. after. Figure 5-4-b Using the mean length of all the samples (mean = 5.6) plus or minus 0.5 times standard deviation (SD = 2.49) as cut-off points, samples that are larger than 6.85 or less than 4.36 are listed in the analysis, as shown in Table 5-7. Compared with the coefficient of longer trips (β3 = 0.06), that of shorter tours (β3 = 0.162) is larger 30.

(39) because they are significantly different from each other at the 0.01 level (for p = 0.01, f-value = 49.54 > F (0.99, 1, 407) = 6.698). However, the coefficients of interaction term in both the clusters of tours’ lengths are not significant on sales. That is, the change in mean sales between the treatment and control groups are not significantly different for both clusters. Therefore, H6 is not supported. This result indicates that although the treatment effect of shorter trips is much greater than that of longer tours, the moderating effect of the length of stay is not strong enough to affect the relation between Facebook campaign and sales performance. Additionally, Figure 5-5 provides the interaction patterns for the two groups (long versus short) in the treatment and control groups. Table 5-7: Testing of Hypothesis 6 Length of Tour. Variable. 立. Note. *p< 0.1. y. sit. 36.53. al. **p<0.05. 39.20. er. R. (0.429). 256. n. 2. 0.060. (1.460). io. SD. 0.162. 0.016. Ch. ***p<0.01. engchi. i n U. v. Treatment Group 50 39.11. 40 sales. 30 20 10. 0.118 (1.038). ‧. mean. 0.023 (0.207). (-0.550) -0.163 (-1.846)*. Nat. n. Long (LE>6.85). 學. Post*Face. ‧ 國. Post Face. 政治大 -0.051. Short (LE<4.36). 28.75 23.48 20.48. long. 0 before. after time. Figure 5-5-a. 31. short. 154 19.47 16.29 0.028.