An Analysis of Search Engine Use for Travel Planning

Transcription

1 An Analysis of Search Engine Use for Travel Planning Daniel R. Fesenmaier Professor and Director National Laboratory for Tourism & ecommerce School of Tourism and Hospitality Management Temple University, Philadelphia, PA 19122, USA Telephone: Zheng Xiang Assistant Professor School of Merchandising and Hospitality Management University of North Texas, Denton, TX USA Telephone: Bing Pan Assistant Professor Department of Hospitality and Tourism Management School of Business and Economics College of Charleston, Charleston, SC , USA Telephone: Rob Law Professor School of Hotel and Tourism Management Hong Kong Polytechnic University, Kowloon, Hong Kong Telephone:

2 An Analysis of Search Engine Use for Travel Planning ABSTRACT Search engines have become one of the primary tools for travel planning and, as such, have become an important part of the Internet marketing strategy of destination organizations. Recently, it has been demonstrated that because of the dynamic relationships among the search engine providers, the tourism industry and travellers, it is essential that destination marketing organizations have a substantial understanding of how search engines are used within the travel planning process. This study proposed a three stage model that provides a framework for examining how online travellers use search engines and how different aspects of the travel planning process shape this use. The model was evaluated based upon a national survey of American online travel planners and the findings provide significant insight into the role of search engines for travel planning. The implications of the study for search engine marketing are discussed as well several challenges for future research. Keywords: Search engine marketing, search engine, Internet, travel planning, destination marketing. 2 P age

3 An Analysis of Search Engine Use for Travel Planning INTRODUCTION Search engine marketing is emerging as one of the most important components of an overall Internet marketing program. With the growth of information on all facets of the tourism experience, general search engines such as Google and Yahoo! have become the Hubble of the Internet galaxy, enabling travellers to navigate through this space so as to find information that might be useful in the travel planning process (Xiang, Wöber and Fesenmaier, 2008). Indeed, recent studies by the Travel Industry Association of America (TIA, 2008; 2009) and others have shown that the huge majority of U.S. travellers use search engines for vacation planning. Additionally, the general search engines such as Google have developed a variety of systems that enable destination marketing organizations (DMOs) to identify the specific interests as well as the keywords used by potential visitors. As a result, search engines have been recognized as the first step in a travel planning process, and therefore, are seen as a critical starting point with which DMOs can communicate with existing and potential visitors. The research on search engines and information search has largely been based in computer sciences, information sciences and consumer behavior. This research, for example, has documented the use of keywords (Jansen and Spink, 2006; Jansen, Booth and Spink, 2008), while Marchinoni (1997) and others (Knight and Spink, 2008; Pirolli and Card, 1999), have examined the factors affecting nature and structure of navigation through the Internet while others have identified the strengths and weaknesses of various search engines (Ding and Marchionini, 1996; Dong and Su, 1997; Su, 1998). Recently a number of studies on search have been conducted within the tourism context (Döring, 2008); for example, Pan and Fesenmaier 3 P age

4 (2006) and Xiang et al. (2008) found that people follow a reasonably structured process as they use the Internet for travel planning. In addition, Wöber (2006), Xiang et al. (2008); Xiang, Gretzel and Fesenmaier (2008) as well as Xiang and Pan (2009) concluded that search engines provide a simplified view of tourism space (i.e., those websites that are available through a search engine), and this view changes daily. As such, they argue that search engine marketing is a complex and dynamic process which requires a substantial understanding of how potential visitors use search engines as part of the travel planning process (Moran and Hunt, 2005; Murphy and Kielgast, 2008; Sen, 2005). The goal of this study was to build upon this literature to develop a general framework that links together the stages of search, i.e., the antecedents, the search process, and the outcomes of search, in travellers use of a search engine for travel planning purposes. It is argued that this framework provides as essential foundation for the development of an effective search engine marketing program. This paper is organized into several sections. Following the Introduction, Related Literature is critically reviewed to provide the foundation for the current study. Then, a General Framework is proposed to describe the basic search process and its constructs. In the Methodology Section, research design involving a national survey of U.S. travellers is described as well as the basic data analysis methods. Research Results are then summarized with respect to the testing of the conceptual framework. Finally, the implications for both theory and practice as well as limitations and future research directions are discussed. RELATED LITERATURE Due to the growing importance of online information search in travel, the use of search engines in the travel planning context has become an increasingly important topic in tourism. This section synthesizes the recent literature on consumers and travellers use of search engines 4 P age

5 and then proposes a general framework that describes the various aspects of this process. Studies on the use of search engines generally fall into three categories with the focus on the process of search, the nature of user queries, and the interaction between the user and the search interface, respectively. From the process standpoint, using a search engine can be understood as consisting of three distinct steps: 1) query formulation, wherein the user enters a query into the search engine interface (Levene 2006; Slone, 2002); 2) search results generation, wherein the search engine retrieves a number of search results that match the query and then displays them in a pre-defined format; and, 3) search results evaluation, wherein the user evaluates the search results and then navigates back and forth between the search engine interface and the web pages linked to those results (Jansen, Spink and Saracevic, 2000; Jansen and Spink 2005; Jansen and Pooch 2001; Marchionini, 1997; Su, 2002). With respect to user queries, studies by Jansen and his colleagues, and more recently by Xiang et al. (2008) indicate that users questions tend to be short, consisting of less than four keywords. Studies by Pan et al. (2007) and Hwang, Xiang, Gretzel, & Fesenmaier (2009) also indicate that searchers usually focus on cities as the geographical boundary instead of states or countries. For example, a search for a city combined with a specific hotel or a hotel brand is one of most common queries that occurs when travellers perform accommodation-related searches. In addition, they conclude that travellers often combine their searches for accommodations with other aspects of the trip, including dining, attractions, destinations, or transportation; additionally, many travellers engage in a switching behaviour that varies between broad and focused search strategies (Hwang et al., 2009). Several studies have focused on users interaction with the search engine interface, indicating that the rank of web pages significantly influences Internet navigation. For example, 5 P age

6 the majority of search engine users do not look beyond the first three pages of search results (Henzinger, 2007; Pan et al., 2007), which implies that only a relatively small number of search results are relevant from the marketing perspective. Studies also show that users trust more organic listings, which, in turn, have a higher conversion rate (Jansen and Spink, 2006). Recently, Pan et al. (2007) found that the order of search results presented by Google, etc. dramatically affects selection of the respective link; specifically, the subjects were significantly more likely to select the first and second suggested links, that the those links presented 8th 10th were also likely to be chosen, and those links presented 3rd 7th were very unlikely to be chosen, due to the scrolling effects. Additionally, Kim and Fesenmaier (2008) suggest that the use of search engines has a significant effect on impression formation, and consequently on the selection and overall evaluation of the website of the tourist firm. While the literature in both computer/information sciences and the tourism field has defined a new research area in travellers use of new technology, they are limited in a number of ways. First, most studies have narrowly focused on the search process, i.e., the interaction, user queries, and search strategies. A number of potentially important questions remain to be answered. For instance, what factors condition the search process? What factors influence a traveller s search strategies (e.g., choice of query terms) and in what way? How does the search process impact on the traveller s mental state and behaviour with respect to attitude change and learning? Second, many of existing studies are primarily descriptive in nature. Particularly, many of them utilized search engine server transaction logs with the aim to describe user behaviour. While this approach is, of course, useful, it does not allow the researchers gain insights into the big picture with necessary explanatory power. As such, it is argued that a more general 6 P age

7 framework is needed to provide a comprehensive understanding of the use of search engines for travel. A GENERAL FRAMEWORK OF SEARCH ENGINE USE FOR TRAVEL PLANNING Building upon this research, a general framework is proposed to describe search engine use in relation to the travel planning process. This framework views the use of a search engine for travel planning as a system with its antecedents as well as certain outcomes and feedbacks. According to the literature in tourism and computer/information sciences, it is postulated that there are three general components or stages of search engine use as related to the travel planning process: 1. The pre-search conditions which is a form of knowledge representation that reflects the integration of the information search process used to plan a trip and the perceived usefulness of various travel planning tools available on the Internet; 2. The search process which includes the frames used to evaluate search results; and, 3. The evaluation of the overall search process which culminates in attitude formation toward search engines and online travel planning. As shown in Figure 1, the first stage is important in that it sets the foundation for the search strategy used by the traveller. Davis (1989), Davis and Venkatesh (1996), Johnson-Laird (1983) and Slone (2002), among others (Brandt and Uden, 2003; Hendry and Efthimiadia, 2008; Stibel, 2005), indicate that one s perceptions of the task as well as the functions of the tools used to support this task directly affect the extent and nature of use of information technologies such as search engines. Thus, within the context of online travel planning, it is argued that these presearch conditions reflect travellers use of (or preference for) various types of information as well as the perceived usefulness of the various travel tools (i.e., types of websites) available on the Internet to find this information. The second stage of the search process describes the basic strategies travellers use to navigate through the Internet to find relevant information so that the 7 P age

8 various travel decisions (i.e., destination, accommodations, attractions, routing, etc) can be made. As such, these strategies act as frames within which the information accessed through use of search engines is evaluated (Dholakia and Bagozzi, 2001; Fesenmaier and Jeng, 2002). The third stage, then, focuses on the overall evaluation of search engines within the travel planning process. Importantly, this stage of search engine use not only results in some sort of overall evaluation (i.e., satisfied vs. not satisfied), but also attitude formation toward search engine use for travel planning (Pan and Fesenmaier, 2006; Su, 2002). Finally, it is important to note that this third stage sets the stage for future use of search engine for trip planning, and therefore it is linked to Stage 1 in the overall process. The following sections elaborate on the dimensions or aspects of each of these components as well as some of the key relationships. Insert Figure 1 here Pre-Search Conditions As can be seen in Figure 1, it is hypothesized that two constructs can be used to define the pre-search conditions (Stage 1) which determine (to a large extent) whether or not the traveller regularly uses a search engine (i.e., Google or Yahoo!) as part of the online travel planning process. These constructs are defined by the various types of information used (Fodness and Murray, 1998; Vogt and Fesenmaier, 1998; Vogt, Fesenmaier and Mackay, 1993) and the perceived usefulness of the various tools/websites on the Internet for trip planning (Brandt and Uden, 2003; Hendry and Efthimiadia, 2008; Jansen and Molina 2006; Su, 2002). In this study, information search tasks were measured by asking American online traveller to identify the types of information (out of a list of 15) they typically use when planning a pleasure trip; these included information about a particular destination or a number of alternative destinations, 8 P age

9 information about attractions, accommodations or car rental information, and/or choosing routes. The perceived usefulness of various travel tools/websites (i.e., travel agency sites, travel company sites, travel magazines, virtual communities, etc.), on the other hand, focuses on the ability of the Internet to support the travel planning effort and, as such, was measured by asking travellers to rate the perceived usefulness (i.e., not at useful - essential) of the various types of online tools/websites for travel planning. Thus, one might expect that travellers with different information needs and their assessment of the various Internet travel planning tools to differ significantly in their use of search engines. Recent research by Jansen et al. (2008) found, for example, that online user queries are mostly informational (81%), followed by navigational tasks (10%) and transactional tasks (9%) and is consistent with research on the information search process of travellers by Vogt, MacKay and Fesenmaier (1993). It is further hypothesized that the interaction of these two constructs (i.e., search for travel information and use of various types of websites) enables one to gain the experience, knowledge, and understanding necessary to evaluate the relative usefulness of search engines in helping find information necessary to plan a trip (Fisher et al., 2008). Thus, it is expected that, for example, those persons that actively seek information about a destination and consider search engines very useful/essential would also consider search engines essential for the specific task of finding information about a particular or alternative destinations. On the other hand, if a traveller has already chosen a destination (and does not need any additional information), the traverller might not consider search engines particularly useful for this aspect of trip planning. Search Process The search strategies used by the online travel planner (the second stage of the model) can be conceptualized as a series of four frames within which the search engine results are 9 P age

10 evaluated (Dholakia and Bagozzi, 2001; Fesenmaier and Jeng, 2001; Pan and Fesenmaier, 2006). Within the context of tourism planning, it is argued that the first task (i.e., deciding on destination, accommodations, transportation, etc.) within the travel planning process sets the stage for the strategies used in evaluating search engine results (Fesenmaier and Jeng, 2001; Pan and Fesenmaier, 2006). Also following Jeng and Fesenmaier (2002), Pan and Fesenmaier (2000) and Hwang and Fesenmaier (2009), it is argued that the degree of specificity (i.e., general vs. specific) of the planning task also affects the way one evaluates search engine results. For example, it might be expected that a person having already decided to visit a particular destination might use an online travel agency in order to arrange the flight or use a company hotel website to book accommodations (i.e., a very specific task). However, if a person is undecided about a specific destination and was seeking information about alternative destinations, that person might use a travel community website or an online magazine to learn about potentially interesting places to visit. Further, it is argued that based upon these considerations, an online search engine user (i.e., travel planner) will differ significantly in terms the keywords entered into the search engine as well as the number of search results considered (Jansen and Pooch, 2001). Recent studies by Pan et al. (2006), Pan, Litvin, and O Donnell (2007) and Xiang et al. (2008) indicate that search strategies (i.e., keyword and number of search results) differ significantly depending upon the nature of task (i.e., facet of the trip being considered and level of specificity) and user goals. For example, one might expect that when the traveller has already decided the destination and wishes to seek information about places to visit or, perhaps, a hotel at the destination, he/she would use the name of the destination as the keyword within a search engine; and, it might be expected that the number searches he/she would evaluate would be limited. However, if the person perceives 10 P age

11 greater flexibility in the place to visit or the activity(ies) in which to participate, it might be expected that he/she searches much further among the search engine results. Overall Evaluation The third stage of a search engine process (i.e., the evaluation stage) can be defined using four constructs. First, it is hypothesized that the online travel planner forms an overall evaluation of search engines and their use for travel planning, and that this evaluation can be reflected in a general satisfaction measure (Pan and Fesenmaier, 2006; Su, 2002). Also, it is hypothesized that the experience is translated into attitudes towards search engines in terms of trust, ease of use, and their efficacy in supporting the travel planning process (Gefen, Karahanna and Straub, 2003; Gretzel et al., 2006; Nakamura et al, 2007; Pan and Fesenmaier, 2006; Su, 2002). Thus, it is expected that a person who is very satisfied with her/his use of search engines would consider search engines as trustworthy, easy to use and the results they provide enable the traveller to make better decisions. On the other hand, dissatisfied users might see search engines as untrustworthy, difficult to use, and/or not very effective in helping them plan a trip. Finally, it is hypothesized that these evaluations (and resulting attitudes) shape (or reshape) the mental model that travellers have of the Internet (and travel planning tools) as well the use of search engines. METHODOLOGY A national survey was conducted from January 5 th 15 th 2009 on Americans who had travelled and used the Internet during calendar year A quota sampling procedure was used whereby 10,000 members of the online panel of Americans maintained by Survey Sampling International (SSI) were invited to participate in the survey, and the first 1,000 respondents (at least) meeting specific conditions were selected. The conditions for participating in this study 11 P age

12 were that the respondents had to have taken at least one business or pleasure trip (that was at least 50 miles one way from home, or included an overnight stay) and they had used the Internet for travel planning including getting information about destinations, checking prices or schedules on the Internet during calendar year The respective panel members were invited once by SSI to complete the survey with the promise of a small monetary incentive which is established by SSI for all panel members. This effort resulted in a total of 2,508 respondents (representing a 25.1% response rate) completing the initial filtering questions; 1,288 respondents actually met all conditions of the study and completed the survey. The online survey was organized into five sections. The first section included a series of filtering questions relevant to all respondents and focused on their use of the Internet, their skills and knowledge of the Internet, travel and involvement in the travel planning process. The second section of the survey focused entirely on various aspects of online travel planning including the types of information sought, their attitudes towards various types of websites such as travel agencies, general travel sites and their use of search engines for travel planning. The third section of the survey included questions regarding the respondent s attitudes toward paid listings and other forms of search engine advertising. The next section included a series of questions related to overall satisfaction with the online travel planning process, the use of search engines as well as their attitudes towards search engines in terms of their ability to support the travel planning process. The last section of the survey included a series of questions regarding age, gender, marital status, income, and race of the respondent. Frequency and Chi square analyses were conducted to assess the nature and extent to which online American travellers include search engines as part of their travel planning effort and to assess a set of key relationships within the proposed framework. Chi square analysis was 12 P age

13 considered appropriate in this exploratory analysis as most of the responses to the questions were discrete, and some questions solicited unstructured responses. Specifically, the analyses focused on testing the following: 1) the relationships among information search activities, evaluation of travel websites, and use of search engines for travel planning (Stage 1); that is, this analysis sought to identify the differences between search engine users (SETs) and non-users (NSETs) in terms of their online information search activities and the types of websites they use; 2) the relationships between the frames travellers establish (i.e., first planning task and specificity of search) and the nature of their search query used and their evaluation of SERPs (Stage 2); 3) the relationships between overall evaluation (i.e., satisfaction with the use of search engines as part of the travel planning process) and their attitudes toward search engines including trust, perceived ease of use, and the quality of the trip plan (Stage 3); and, 4) the relationships between the attitudes toward search engines (i.e., trust, perceived ease of use and trip plan quality) and information search activities and evaluation of travel websites. RESULTS This section reports the results of analyses of a random survey 1,288 American travellers who used the Internet for travel planning. Almost all (96.3%) respondents accessed the Internet from home, while 32.6 percent of the respondents reported that they also had access from work; in addition, most respondents (74.7%) reported going online from home several times a day and an additional 18 percent indicated they went online (from home) about once a day. Interestingly, many of the respondents rated themselves somewhat to very skilled at using the Internet, and agreed that they were relatively successful in finding what they want on the Internet. Further, about a third of the respondents perceive themselves as leaders or innovators in the use of Internet technologies. Questions were also asked regarding their involvement in the travel 13 P age

14 planning process. The results indicate that most respondents enjoy travel planning very much; about 50 percent of the respondents indicated that they tend to devote a lot of time to planning a trip and almost 75 percent of the respondents indicated that they become very involved in the travel planning process, and therefore, are unlikely to leave trip planning to others. The respondents were equally distributed in terms of gender (52% women and 48% men); 59 percent reported themselves as married, 20 percent are single or never married, 13 percent are divorced or separated and 7 percent reported living with a partner. About 38 percent of the respondents indicated that they graduated from high school and 28 percent graduated from college; 33 percent earned less than $40,000; an additional 33 percent earned between $30,000 and $60,000 while 12.4 percent of the respondents indicated they earned $100,000 or more during Search Engine Use in Travel Planning The analyses also indicate that a huge majority of online American travellers (86%) use general search engines to assist in their travel planning efforts; indeed, 44 percent of the respondents indicated that they used 2 3 different search engines while 15 percent indicated that they regularly used 4 or more different search engines to assist in travel planning. Also, a majority (65%) indicated that they use search engines most of the time or almost always (25%). Interestingly, however, most respondents rated general search engines only somewhat useful for travel planning, and very few considered them an essential travel planning tool; indeed, search engines were seen as essential for finding maps and/or driving by only 35 percent of the respondents, and the next highest activity was for seeking information about a particular destination at 23 percent. Frequency analyses were conducted to better understand the variability in the ten components that comprise the three stages of proposed model. As shown in Table 1, the range in 14 P age

15 response to the two constructs is believed to explain search engine use (the types of online planning and the perceived usefulness of various types of websites in assisting the planning process) suggests that while they are very important planning tasks, many aspects of travel planning are somehow done offline; also, it seems that, consistent with previous findings (Vogt and Fesenmaier, 1995), much of the information search process is highly directed and functional. Specifically, most respondents indicated that they search for information about a particular destination (76%), searched for hotel prices or places to stay (69%), or searched for airline fares and/or schedules (62%). Interestingly, relatively few persons searched for 800 numbers (8%), information about stores (24%) or cruises (23%). However, it appears that there substantial agreement in the degree to which online travel agencies, company websites, destination sites and search engines in the travel planning process. Specifically, 65 percent of the respondents indicated that general search engines were very useful or essential to the travel planning process, 62 percent similarly rated company websites and 57 percent of respondents also indicated that online travel agencies such as Expedia and Orbitz were very useful - essential. Insert Table 1 here Frequency analyses were also conducted to understand the search strategies used by online American travellers (see Table 2). As discussed previously, online search strategy was characterized by four constructs: 1) the first task conducted in the travel planning process; 2) the level of detail (general vs. specific) that is used to frame the search process; 3) the nature of keywords used to initiate the search process; and, 4) the depth to which search engine results are evaluated. The results as shown in Table 2 indicate that online travel search generally starts with one of three tasks: Accommodations (29%), destination selection (29%), or transportation (23%). 15 P age

16 Very few respondents indicated that they started the online travel planning process by considering shopping alternatives (0.7%), attractions (7.0%), or routes (9.4%). Interestingly, however, the results also indicate that most travellers have some sort of general frame of reference (58.5%) which guides their planning process whereas 24 percent of the respondents reported that they usually start planning their trip with a specific of where and when they will travel. It appears that, in large part, this frame is defined by a destination as the majority (55%) of respondents indicated reported that they use all (or part) of the name of the destination to initiate the search process. This finding contrasts sharply with only 10 percent of the respondents indicating that they use the name of online travel company or the 13 percent that searched for an online travel agency. Last, the results indicate that many (31%) American travellers use only the top three search results; and, an additional 20 percent typically use the results reported in the top page (i.e., the top 10 results). This later finding is consistent with the literature which indicates that many people (50% of the respondents) limit their selection to the first page of search results and an additional 18 percent limit their search to 2 pages (Pan et al., 2008). Insert Table 2 here The final frequency analyses focused attention on the travellers evaluations of this tool as part of the travel planning process (see Table 3). Specifically, the respondent was asked to rate their overall satisfaction with general search engines when planning their pleasure travel; then, they were asked to assess their attitudes toward general search engines in terms of trust, ease of use, and plan quality. The results of these analyses seem to indicate that online American travellers are fairly satisfied with the results provided by search engines as 86.2 percent of the respondents indicated that they were somewhat or extreme satisfied. 16 P age

17 Insert Table 3 here Model Evaluation The second phase of the analysis used a series of post hoc Chi-square tests to examine the hypothesized relationships between the respective constructs described in Figure 1. Again, it is argued that Chi-square analysis is appropriate in this exploratory study as most of the response formats used are discrete and nonmonotonic, and therefore, do not lend themselves to conventional parametric analyses such as structural equation modeling (SEM). Unfortunately, however, the pair wise tests do not provide an overall idea of goodness-of-fit. The following provides a brief summary description of the results of these analyses organized by the three stages of the search engine planning model. Pre-Search Conditions Two sets of analyses were conducted to assess the differences between those respondents using search engines for travel planning (labeled SETs) and those typically not using search engines (NSETs) as part of the online travel planning process. As discussed previously, it is argued that a search engine use can be determined by the types of online information activities in which the respondent participates as well as evaluations of the usefulness of other types of online planning tools. The results of the Chi Square analyses as shown in Table 4 indicate that there are consistently significant (α = 0.05) differences between SETs (search engine users for travel planning) and NSETs (those respondents indicating that they do not regularly use search engines for travel planning) in terms of extent (depth and breadth) of online search activities as well as use of the various online travel planning tools. For example, over 80 percent of the SETs 17 P age

18 indicated that they normally search for information about the destination vs. only 56 percent of NSETs; 72 percent of SETs searched for hotel prices or places to stay as compared to 49 percent of NSETs; and, 59 percent of SETs search for and/or printed out maps as compared to 35 percent of NSETs. Consistent with these findings, SETS tend to use many more of the online travel planning tools including online travel agencies (SETs = 69% vs. 42 for NSETs), company websites (64 % vs. 51%), destination websites (47% vs. 33%) and community websites (25% vs. 8%). Insert Table 4 here Search Process Six sets of analyses were conducted to assess the relationships within the constructs included in this stage of the search engine use within the travel planning process. First, it was hypothesized that the perceived usefulness of search engines for the various travel planning tasks should relate directly to the first task of the travel planning process and the degree of specificity within which online travel planning occurs (see Table 5). The results of Chi square analyses indicate that (as discussed previously) while there are general patterns of behavior, there are consistently significant (α = 0.05) differences among SETs. That is, the results indicate that destination and accommodation selections are consistently the most popular starting points for trip planning with transportation the third most popular. However, the results of this analysis indicate the travel task addressed first differs substantially, depending upon the perceived usefulness of search engines. For example, 36 percent of those persons who rated search engines essential for finding information about particular destinations initiated travel planning through the destination perspective; this compares to 32 percent for those considering attractions and/or 18 P age

19 accommodations as the starting point, and 17 percent of those starting with the route. Also, 31 percent of those rating search engines essential for finding airline fares and schedules started first transportation issues; this compared to 29 percent of the respondents who started with accommodation and 24 percent started with destination selection. Insert Table 5 here Analyses were also conducted assessing the relationships between the perceived usefulness of search engines for the various travel-related tasks and the frame (general vs. specific) which directs the search/travel planning process (see Table 6). The results, again, indicate that while there are some strong trends in the responses, there are significant (α = 0.05) differences in the nature of travel planning frame, depending upon perceptions of the usefulness of search engines for travel planning. For example, 44 percent of those travellers rating search engines essential for finding information about a particular travel destination usually have a general idea about the trip, and 30 percent of these respondents indicated that usually have a specific problem in mind. This finding contrasts sharply with those persons rating search engines essential to find out about potential destinations to visit in that 55 percent of these travellers start the planning process with a general frame. Indeed, about 60 percent of those persons considering search engines essential for finding things to do at a destination and websites that offer free travel brochures start planning with a general idea as compared to the 20 percent that start with specific travel-related tasks. Insert Table 6 here 19 P age

20 A Chi-square analysis was conducted to examine the relationship between the frame and the problems/aspect of travel planning started first (see Table 7). The results indicate significant (α = 0.05), but not substantially different, results. For example, those travellers starting with attractions (65%) or destination planning (62%) are more likely to take a general perspective than those focusing on accommodations (53%); concomitantly, those focusing on accommodations (29%) are much more likely to have a very specific problem when initiating their travel planning efforts. Insert Table 7 here Two additional analyses in the second phase of travel planning focused on the relationship between the planning task, search frame, and the keyword they first used to initiate the information search process. Again, Chi square analyses were conducted and the results presented in Tables 8 and 9 indicate that there are significant (α = 0.05) and meaningful relationships between the keywords that one uses and the frames defining the search process. Table 8 shows that most people use the name of the destination as the keyword used first but the percentage varies substantially by the initial task within the trip planning effort. In this study, 34 percent of those using a destination as a keyword consider the destination as the starting point; however, 30 percent of these online travel planners consider accommodations as the initial planning task, and 18 percent indicated that transportation was their starting point. This finding contrasts sharply with those entering the name of an online travel agency where transportation related issues was the most popular task (36%), accommodations was the second most popular (29%) and destination was the third most frequent starting task (24%). Lastly, 40 percent of those persons knowing the website address focused first on destination and 25 percent started 20 P age

21 with transportation-related planning. Insert Table 8 here This analysis also considered the relationship between travel planning frame (as defined by specific vs. general problem) and the keywords used to initiate the search process (Table 9). As discussed previously, most respondents use the name of the destination to start the search process; however, the results indicate that the types of keywords used differ significantly (α = 0.05) depending upon the search frame. For example, 82 percent of those who searched for an online travel agency had a general idea in terms of travel planning effort, as compared to only 35 percent of those entering the name of a slogan or advertising, or 59 percent of those entering a name of a destination. Interestingly, almost 30 percent of those searching a travel-related company such as a hotel or airline indicated that they had a specific problem when starting the travel planning process; this compares to 24 percent of those searching for a destination and 22 percent that are searching for an online travel agency. Insert Table 9 here The last analysis considered within this stage of the model focused on assessing the relationship between the types of keywords first entered into a search engine and the number of search results evaluated. The results are presented in Table 10 and indicate that significant (α = 0.01) differences exist among the respondents. For example, it appears that those respondents entering a company slogan or advertising or a travel-related company are much more likely to limit their use of only the first 3 results (45% and 38%, respectively), as compared to those entering (as a keyword) the name of a destination (25%) or an online travel agency (26%). 21 P age

22 Interestingly, those entering the name of a travel-related company are also likely to limit their search to the first page (66%); alternatively, those persons entering the name of an online travel agency (34.3%), a destination (25.2%), or a travel-related company (20.7%) are much more likely to extend their use of search results beyond two pages. Insert Table 10 here Overall Evaluation of Search Engine Use The analyses conducted in this stage of the study focused on the evaluation of the results and attitude formation. The first analysis examined the relationship between the length of search results considered and overall satisfaction with the use of search engine for travel planning. Interestingly, the results of this analysis showed no meaningfully significant (α = 0.05) differences. That is, those persons who tend to select the first search result are equally likely to be extremely satisfied (44%) as those who use up to 3 pages of search results (40%). However, a series of analyses between satisfaction and the twelve items used to measure attitudes toward search engines indicate that those persons that are extremely satisfied with search engines were significantly (α = 0.01) more likely to trust the results of search engines, consider them easy to use, and agree that their use improves the quality of the travel plan. In particular, 41 percent of those extremely satisfied agreed that search engines are reliable (vs. 9% for those somewhat satisfied); 61 percent of those extremely satisfied felt that search engines make travel planning easier to complete as compared to 24% of those somewhat satisfied with search engines; finally, 59 percent of those satisfied strongly agreed that the use of search engines enable them to make better travel decisions vs. 22 percent of those only somewhat satisfied. 22 P age

23 Insert Table 11 here Closing the Circle: Linking Post-Search Evaluation and Pre-Search Conditions A final set of analyses were conducted to examine the relationships between Overall Evaluation and the two components of Stage 1 (the Pre-Search Conditions): online information search and the perceived usefulness of online travel planning tools. The results indicate that those extremely satisfied with general search engines are more likely to search for information about a particular destination (83%) as well as information about hotel prices or places to stay (72%) and information about things to do at the destination (61%); beyond this, there appears to be no correlation between information search strategies for travel planning and satisfaction with general search engines. Interestingly, and perhaps as expected, the extremely satisfied SETs search engine travellers consistently perceive other types of online planning tools to be significantly (α = 0.05) more useful in finding online travel information than those who were not satisfied. For example, 29 percent of the extremely satisfied SETs found online travel agency sites essential (and, an additional 41% very useful) vs. zero (0) percent for those that were extremely unsatisfied. These differences were consistent for most of the tools/websites considered. However, the reverse was true for those using travel guidebook sites such as Fodors or Lonely Planet; in this case, extremely unsatisfied SETS were more likely to consider this tools essential and/or very useful. Also, there was no significant (α = 0.05) correlation between search engine satisfaction and the perceived usefulness of newspaper and/or magazine sites and social networking sites. CONCLUSION AND IMPLICATIONS Search engines have become an essential tool in the travel planning process and, thus, 23 P age

24 they serve as one of the most important spaces on the Internet within which online destination marketing can be conducted. A number of recent studies have confirmed the importance of search engines as well have considered various facets of an information search process, an online travel planning process, and various factors affecting the use of search engines. Indeed, the literature clearly documents three important stages of the travel planning process as related to the use of search engines: 1) the pre-search conditions that drive the search process; 2) the actual search process including search strategies and the frames used to evaluate search results; and, 3) the evaluation of the overall search process which culminates in attitude formation toward search engines. It is argued that the first stage (pre-search conditions) is important in that it sets the foundation for the search strategy used by a traveller in terms of types of information obtained as well the Internet tools used to search for this information. The second stage, on the other hand, describes the basic strategies travellers use to navigate through the Internet in search of relevant information so that the various travel decisions (i.e., destination, accommodations, attractions, routing, etc) can be made; as such, these strategies act as frames within which the information accessed is evaluated. The third stage, then, focuses on the overall evaluation of the use of search engines within the travel planning process and sets the stage for future use of search engines for travel planning. The results of this study based upon a survey of online American travellers confirm that most Americans use general search engines quite extensively for travel planning along with many other online tools. The results also show that there are strong and consistent relationships between the respective constructs in the proposed model in that those using search engines tend to be very active and involved travel planners; also, they tend to use a variety of different online tools websites within the overall travel planning effort. Importantly, the search frame (as 24 P age

25 defined by the first task of the online planning process and the level of specificity) sets the stage for the keywords used to start the search effort, which in turn, affects the depth of search. Lastly, the results clearly indicate that most online American travellers differ substantially in terms of their satisfaction with the results of general search engines such as Google and Yahoo! and that this satisfaction is reflected in the extent to which search engines can be trusted, are perceived to provide useful results and the extent to which they enable the traveller to make good travel decisions. This study contributes to the existing knowledge about the use of search engines in online travel planning in several ways. First and foremost, this study offers a theoretical framework to explain the antecedents, the process, as well as the outcomes in the use of search engines for travel planning by building upon the classic literature on travel information search and the emergent research stream in understanding of online search behavior for travel. As such, this study can be used as the stepping stone for future research endeavors to explore and empirically test many of the important relationships in online search behavior within a travel planning context. Second, this study examined some of the key relationships in the proposed theoretical framework based upon large-scale empirical data. Different from conventional approach to studying search engine use, which usually relies on server transaction log data, the research design allows the researcher to gain insights into what consumers expect, how they perform, as well as what they learn through the use of search engines for travel. While there are some limitations (as addressed below), this study has certainly established some basic understandings about factors that drive the search process, which further leads to a comprehensive understanding of online consumers attitudinal and behavioral aspects involved in search engine use. The findings also have several implications for the design of an effective program in 25 P age

26 search engine marketing. First, it is clear from this and past research that search engines often represent the starting (or entry) point for the travel planning process; as such, the design and placement of search engine results is a critical part of the communication process and therefore, considerable effort should be invested to make this a positive persuasive experience. Second, search engine travellers (SETs) are substantially different from the general online traveling population in that they are much more active and involved in the travel planning process. This suggests that destination marketing organizations can, and should, employ a variety of strategies with which to engage the travel planning including the use of Web 2.0 (i.e., social communities and consumer generated content) to attract and involve potential visitors. Third, the finding that many people use destination-oriented keywords regardless of the specific planning problem suggests that destination websites should incorporate a range of functions beyond simply providing information about a destination; these functions might include links to travel agencies, accommodations and transportation providers. However, the differences found in this study suggest that it is important to develop smarter search systems that recognize the goals (and search strategy) of the travel planner. Finally, it is an important finding (and, perhaps not unexpected) that SETs are very positive about the role that search engines play in the overall travel planning process. From this, it seems that DMOs should consider ways to build on this goodwill by co-developing new tools with the search engines to better support tourism products. These tools may include destination-specific recommendation engines, virtual communities, and interactive interpretive systems. While these findings strongly support the proposed model, it is clear that there are a number of limitations to this study. First, it appears that important constructs are missing including variables describing the user (i.e., Internet expertise) and the nature of the specific trip 26 P age

27 being planned. Beyond inclusion of these constructs, an important challenge is to identify order or causal linkages between the respective constructs. That is, the issue of structure was assumed and not specifically addressed in this analysis, but future studies should examine the role of each construct in shaping or being shaped. For example, one might argue that the attitudes identified in Stage 3 should be included in Stage 1 as they form the basis of one s mental model of the system as well as the online travel planning process. Third, this study used information that was generalized across all trips taken throughout a year; one might expect that while certain aspects of the model are relatively stable across this time period, the role of other aspects especially related to the trip differ substantially. Finally, this exploratory model used constructs that were measured in a variety of formats which limited the use of a variety of statistical analyses and therefore, our ability to assess the relative importance of each component of the model. With these limitations, it is clear that substantial research should be conducted to better identify constructs and their role(s) within the model. With this said, however, it is argued that the results of this study provide a substantial foundation for developing a comprehensive understanding of the role of search engines within the online travel planning process. 27 P age