Activity-based segmentation of the outbound leisure tourism market of Saudi Arabia

Abstract

Using an array of multivariate statistical methods (principal components, cluster, discriminant, and analysis of variance), this study empirically develops market segments for the outbound leisure travel market of Saudi Arabia based on respondents’ stated preferences for common vacation activities. Analysis of 455 responses to a structured survey reveals three main travel segments—conservatives, fun seekers, and variety seekers. Conservatives (older, married, and male respondents) have a profound dislike for entertainment-oriented activities; fun seekers (young, single, and female) prefer shopping and leisure activities, and variety seekers (middle aged, single, and female) like all vacation activities. Implications for destination marketers and tourism segmentation research are discussed.

Keywords

Activity-based segmentation Saudi Arabia segmentation tourism vacation

Introduction

The Kingdom of Saudi Arabia is an important tourism-generating market and the largest in the Middle East, accounting for around 40% of all outbound tourists from the region (WTO, 2003). Approximately 10 million Saudi residents travel abroad each year (Colliers International, 2011), 49.7% of these travel for leisure, 20.9% for business, 21.5% for visiting friends and relatives, and 7.9% for other purposes including medical, education, and training purposes. Total spending by outbound tourists from Saudi Arabia amounted to an estimated $10 billion in 2011 (Business Monitor International, 2012), with the biggest portion going to shopping (37.9%) and leisure activities (27.2%). These numbers are expected to increase in the future (to approximately 12.73 million by 2015) (Business Monitor International, 2012) due to several factors, including a large, young, and increasingly affluent population with limited opportunities for out-of-home leisure and entertainment within their country.

Recent data from Saudi Arabia’s Tourism Information and Research Center show that the majority of Saudi tourists visited neighboring Gulf Cooperation Council (GCC) countries (29.5%) and other Middle East countries (31.8%). However, a substantial portion visited other countries in South Asia (21.8%), Europe, and the United States (16.9%). Tourism organizations in these destinations need to understand the activities that Saudi tourists deem necessary as part of a vacation, if they are to succeed in developing effective marketing programs to target them. In particular, it is important to understand the segments of consumers that exist in this market, so that destinations can decide which of them to target based on what they have to offer. Market segmentation is a tool for developing this understanding. It is also a powerful tool that can be used to identify tourism market opportunities and aid in tourism product development (Middleton and Clarke, 2001). Accordingly, many tourism segmentation studies have been reported in the literature (e.g. Graham and Wall, 1978; Formica and Uysal, 1996, 1998; Johns and Gyimothy, 2002; Kim et al., 2003; Masiero and Nicolau, 2012; Mehmetoglu et al., 2010; Sirakaya et al., 2003; Thurau et al., 2007).

The purpose of this study is to contribute to this literature by undertaking a similar segmentation study of the outbound leisure tourism market in Saudi Arabia. The goal is to develop profiles of outbound tourists from the country based on the extent to which they see various vacation activities as a necessary part of a successful vacation. Specifically, the study seeks answers to the following research questions:

Can the outbound Saudi leisure travel market be segmented on the basis of activities that tourists prefer to engage in?

Will activity-based segmentation of this market lead to identifiable, measurable, actionable, and stable market segments?

Can these activity-based segments be linked to other travel-related variables, such as destination and travel mode preferences?

The study is useful in two respects. First, it will provide marketers in destinations frequented by tourists from Saudi Arabia with insights that will help them better plan and develop marketing and promotional strategies aimed at this market. Many previous studies (Lang et al., 1993) have noted the usefulness of activity-based segmentation in providing actionable guidelines for destination managers. Second, although previous research has sought to understand important aspects of the behavior of Saudi vacation tourists (e.g. Bogari et al., 2004; Seddon and Khoja, 2003), we are not aware of any studies that have segmented the market for Saudi tourists to international destinations, and certainly not any that have done so using activity-based segmentation. We are also not aware of similar studies using data from other Arab world contexts. Thus, our study will add to the literature on international tourism by profiling tourists from an important source market for global tourism. Furthermore, the study adds to the growing literature on activity-based segmentation in tourism and its usefulness in understanding tourist behavior. In particular, it validates the usefulness of activity-based segmentation. Moreover, we extend the literature on activity-based tourism segmentation by showing that destination preferences and attitudes toward alternative travel arrangements can be predicted on the basis of segments developed using the procedure.

In the next section, we review the tourism segmentation literature. This is followed by a discussion of our research methods. Analysis and results of our empirical study are next reported, followed by theoretical and practical implications of our results. The final section presents the study’s limitations and offers suggestions for future research.

Tourism segmentation studies

As in traditional market segmentation, tourism segmentation studies encompass researchers’ attempts to group tourists from particular countries (or tourists to particular destinations) into homogeneous categories based on similarity with respect to a particular variable or set of variables. At the heart of any segmentation effort is the segmentation basis or variable set to be used in classifying consumers. Not surprisingly, numerous dimensions and variables have been used by researchers in tourism segmentation. These have included travel motivation (Andreu et al., 2005; Bieger and Laesser, 2002; Cha et al., 1995; Kau and Lim, 2005; Park and Yoon, 2009), benefits sought in the destination (Jang et al., 2002; Sarigollu and Huang, 2005; Shoemaker, 1994), travel mode (Mehmetoglu, 2006), visit frequency (Tsiotsou, 2006) Internet use (Bonn et al., 1999), expenditure amounts and patterns (Spotts and Mahoney, 1991; Thrane and Farstad, 2012), traveler psychographics (Silverberg et al., 1996), and preferences for tourism activities, among others.

An important distinction in tourism research has been that between push and pull factors as determinants of people’s travel decisions (Crompton, 1979). Push factors are individual-related sociopsychological motivation factors that predispose the individual to want to travel. They are answers to the question ‘why does a person choose to travel on vacation?’ Numerous travel motivation factors have been proposed in the literature, including escape from a perceived mundane environment, exploration and evaluation of self, relaxation, prestige, regression, enhancement of kinship relationships, and facilitation of social interaction (Crompton, 1979; see also Dann, 1981 and Iso-Ahola, 1987 for earlier discussions of tourism motivation factors). Pull factors, on the other hand, are characteristics of a given destination (such as infrastructure or cultural attractions) that draw people to it.

Many tourism segmentation studies (such as Andreu et al., 2005; Bieger and Laesser, 2002; Cha et al., 1995; Kau and Lim, 2005; Park and Yoon, 2009) have used travel motivations as input. However, it has been suggested that developing tourist segments on the basis of motivational factors alone can be problematic because motives are simply not stable. Consumer travel motivations differ between individuals, and even from one decision-making context to another for the same individual. Accordingly, activity-based segmentation has been offered as an alternative to developing stable tourist segments across different national and cultural contexts (e.g. Beritelli and Boksberger, 2005; Choi and Tsang, 2000; Choi et al., 2011; Hsieh et al., 1992; Kim and Jogaratnam, 2003; Lang et al., 1993; McKercher, et al., 2002; Sung et al., 2000; Yang, 2009).

Hsieh et al. (1992) identify five activity-based segments in a study of Hong Kong international pleasure travelers, namely ‘visiting friends and relatives’, ‘outdoor sports’, ‘sightseeing’, ‘full-house activity’, and ‘entertainment’. Focusing on the same Hong Kong market, Choi and Tsang (2000) identify four activity-based clusters of travelers (sightseeing, outdoor sports, entertainment and outdoor activities, and friends/relatives visiting), while McKercher et al. (2002) identify six cultural tourism market segments. Lang et al.’s (1993) study of Japanese female overseas travelers identifies five main segments—‘outdoor sports’, ‘sightseeing’, ‘life-seeing’, ‘activity combo’, and ‘naturalist’. In a study of visitors to New Brunswick from other Canadian provinces (specifically Quebec, Ontario, Nova Scotia, and Prince Edward Island), Choi et al. (2011) identify three activity-based segments that they name ‘outdoor lovers’, ‘active explorers’, and ‘cultural shoppers’. Sung et al. (2000) focus specifically on adventure tourists. They factor analyze 48 adventure travel activities to derive six segments—‘soft nature’, ‘risk equipped’, ‘question marks’, ‘hard challenge’, ‘rugged nature’, and ‘winter snow’. Using a cluster analysis of 39 sport and 35 non-sport activities, Beritelli and Boksberger (2005) find six clusters of Swiss vacation travelers, which are named as ‘family/partner holiday’, ‘hanging around’, ‘active relaxation’, ‘destination orientation’, and ‘beach holiday’. Similarly, Kim and Jogaratnam (2003) factor analyze 16 vacation activity preferences of US and international students and identify four factors that they labeled ‘cultural’, ‘sports’, ‘leisure’, and ‘touring’. Yang (2009) cluster analyze 45 activities to derive three segments of visitors to rural Michigan—outdoor tourists, cultural tourists, and general tourists.

In general, two broad segmentation approaches can be identified in the literature—a priori and a posteriori approaches (Mazanec, 1992), although see Dolnicar (2004) for a more nuanced categorization. In a priori segmentation, the grouping criteria are defined by the researcher in advance, and the predetermined segments are described and further profiled using some selected descriptors. For instance, age may be used a priori to segment a particular travel market into young, middle-aged, and older travelers who are then profiled using additional variables. In a posteriori segmentation on the other hand, statistical approaches such as factor and cluster analyses are applied on attitudinal and/or behavioral data to derive clusters (or segments) of tourists such that each cluster contains tourists who are similar to each other with respect to the relevant attitudinal or behavioral variables. The resulting segments are then profiled based on similarity with respect to a further set of variables, typically sociodemographic variables. Using the a posteriori approach in activity-based segmentation, tourist segments are developed based on respondents’ preferences for various vacation activities. Typically, a large set of initial activities is subjected to factor analysis to derive broad activity dimensions that are then used as input into a clustering algorithm. The present study employs such a factor–cluster segmentation approach applied on a list of tourism activities that respondents see as necessary to participate in during their ideal vacation. The overall hypothesis is that this approach will provide meaningful and actionable segments and lead to a better understanding of travel preferences of Saudi vacation tourists. Details of the study approach, including specific activities included and the factor–cluster techniques used in the data analysis are described in the method section. But first we briefly review the literature on vacation activities.

Tourism activities

At the heart of activity-based tourism segmentation studies are tourism activities. In these studies, tourists from a particular country, to a particular destination, or in a particular category (e.g. adventure tourists) are grouped into clusters based on specific activities that they prefer to engage in during a vacation. The logic is that tourism activities are central to understanding the tourism decision-making process (Choi and Tsang, 2000) because activities not only translate directly to behavior but can also be treated as the outcome of traveler preferences (Jang et al., 2005). An important criterion affecting a traveler’s destination choice is the type of activities to be engaged in while visiting the destination. Accordingly, it has been argued that activities can be used to connect tourists’ motivations (push factors) to destinations (pull factors), since motivations can be viewed as tourists’ expectations of activities and destinations as staging places of these activities (Moscardo et al., 1996).

Many tourism activities have been used in previous tourism segmentation studies, and it will not be possible to list them all here. For instance, Beritelli and Boksberger (2005) use a set of 74 activities (39 sport and 35 non–sport-related activities) to classify the general Swiss travel population into six clusters. As is to be expected, the specific activities included in any study often depend on the focus of the study. For example, nature-based tourism studies typically include a large proportion of nature-based activities like camping, cycling, fishing, hunting, and so on; leisure-based studies emphasize sunbathing, lying in the beach, recreation activities, and so on; and culture-based studies will emphasize activities relating to visiting cultural attractions, museums, and the like. Certain activities (e.g. shopping and visiting museums), however, tend to transcend most studies.

Since most studies typically include a large number of tourist activities—16 is the minimum we encountered in the studies reviewed in this article—researchers frequently seek to identify smaller sets of activity dimensions. For instance, Mehmetoglu (2007) derived 5 activity dimensions—visiting historic/cultural activities, challenging nature-based activities, relaxing nature-based activities, and pleasure-based activities—from the 17 original activities; Choi and Tsang (2000) derived 4 dimensions—sightseeing activities, outdoor sport activities, entertainment activities, and friends/relatives visiting activities—from 33 activities; Madrigal and Kahle (1994) identified 5 dimensions—cultural activities, outdoor activities, sport activities, visiting ancestral homeland activities, and visiting friends/relatives activities—from their original 18 activities, and Kim and Jogaratnam (2003) identified 4 activity dimensions—cultural activities, sport activities, leisure activities, and touring activities—from 16 activities.

Method

Data for our empirical study were collected as part of a larger survey of vacationing preferences and behavior of consumers in Saudi Arabia using a structured self-administered questionnaire. The questionnaire addressed several issues, including respondents’ general vacation preferences (domestic versus foreign), factors influencing choice of vacation destination, attitudes toward domestic vacations, and preferred activities during a vacation. Respondents also provided sociodemographic information. In the part of the questionnaire of interest to this article, that is, preferred activities during a vacation, respondents were given a list of possible vacation activities and asked to indicate how necessary each activity would be for them during a perfect vacation. A 5-point response scale was used with anchors only at the end points (1 = Not at all necessary; 5 = Very necessary).

The list of activities was obtained from a review of the activity-based segmentation literature described earlier and feedback from a pilot test of an initial draft questionnaire. A final list of 17 vacation activities was included in the section of interest to the present article. The list of factors, along with cited sources, is shown in Appendix 1.

Data collection

Mall-intercept sampling was used in the data collection. Cooperation was solicited from the management of four large shopping malls in four Saudi cities for research assistants to mount questionnaire completion booths during one weekend. The assistants then intercepted shoppers in the mall and personally solicited their cooperation in filling out the questionnaires at the booth. To increase cooperation, respondents had the opportunity to enter a draw to win one of three gift certificates worth up to Saudi riyal 200, equivalent to about US$53. Admittedly, an ideal sampling method would have been to intercept outbound travelers in international airports. However, we used mall intercepts because the objective of the larger study was to examine consumers’ vacation plans in advance of the upcoming travel season for that year, with a specific focus on their plans for domestic versus international travel. The data were collected in March 2011, shortly before the annual summer vacation for the majority of consumers. Convenience sampling is widely used in tourism studies. In a review of articles published in five primary hospitality management journals, Baloglu and Assante (1999) found that the majority of the articles (between 64 and 84.8% for different journals), used non-probability sampling. A total of 455 completed questionnaires are used for analysis in this article. Males constitute around 60% of this sample. It is almost evenly split between married and single respondents, with the majority in the age-groups 18–25 years (35%) and 26–30 years (25%).

Analysis

Similar to the approach in previous tourism segmentation studies (e.g. Cha et al., 1995), we used a multistage approach in the data analysis. First, we conducted exploratory principal components analysis on the 17 vacation activities to reduce them to a smaller set of dimensions for input into a cluster analysis algorithm. Second, we clustered respondents based on their scores on the new activity dimensions that were extracted from the principal components analysis. Third, we validated the cluster solution through discriminant analysis, which also allowed us to confirm the cluster designations and ascertain the role of each activity dimension in discriminating among clusters. Fourth, we tested for statistical significance of differences in cluster means on the activity dimensions using analysis of variance (ANOVA) procedures. Fifth, we developed sociodemographic profiles of the clusters by examining sociodemographic differences among cluster members. We used a series of cross-tabulations between cluster membership on one hand and each of six sociodemographic variables—gender, age, marital status, income, education, and number of children. In all analyses, we also tested for statistical significance of differences in cluster membership across the sociodemographic variables using χ ² analyses. We also tested for differences in vacation destination and other travel-related preferences among the clusters.

The data used for analysis in this article are part of data from a larger study that also sought to examine how to promote domestic tourism among residents in Saudi Arabia. The study was conducted just before the annual summer vacation, and a filter question asked respondents to indicate where they planned to spend that year’s vacation—at a domestic or international destination. The data used in this article are for those who reported that they planned to spend their vacation at an international destination.

Results

Descriptive statistics

Prior to presenting results of the principal components, cluster, discriminant, ANOVA, and χ ² analyses, we present descriptive statistics (means and SDs) of the 17 original vacation activities used in the questionnaire (Table 1). The variables are listed in descending order of means, and the table also includes results of statistical tests for differences between each mean and the scale midpoint of 3. At the 5% level of significance, except those for exploring the outdoors and going on safari, the means of all activities are significantly different from the scale midpoint (above the midpoint for variables with positive t values and below for those with negative values). In general, relaxation is the vacation activity considered most necessary by all respondents. This is followed by visiting beaches, lazing around, visiting amusement parks and recreation centers, and shopping. The three least necessary activities, are attending sporting events, music concerts, and visiting nightclubs.

Table 1.

Descriptive statistics for study variables.

	Mean	SD	t	Significance
1. Just relaxing	4.51	0.84	38.09	0.000
2. Visiting beaches	4.09	1.16	19.89	0.000
3. Lazing around	4.05	1.10	20.10	0.000
4. Visiting amusement parks and so on	4.07	1.11	20.21	0.000
5. Shopping	3.98	1.15	17.92	0.000
6. Sleeping	3.96	1.14	16.83	0.000
7. Visiting historical places	3.89	1.22	15.55	0.000
8. Visiting museums	3.74	1.22	12.87	0.000
9. Attending cultural festivals	3.49	1.26	8.19	0.000
10. Watching movies	3.47	1.45	6.91	0.000
11. Camping	3.38	1.33	6.06	0.000
12. Exploring the outdoors	3.13	1.49	1.91	0.057
13. Going on safari	3.04	1.40	0.64	0.521
14. Diving	2.73	1.51	−3.74	0.000
15. Attending sporting events	2.66	1.39	−5.13	0.000
16. Attending music concerts	2.63	1.56	−5.02	0.000
17. Visiting nightclubs	2.26	1.54	−10.13	0.000

Principal components

In the principal components analysis, we analyzed the correlation matrix, used the varimax method to rotate factors, and retained factors with eigenvalues greater than one. Five factors were extracted that explained 67% of variance in the variables. Based on the patterns of factor loadings (Table 2), we labeled these as outdoor adventure, knowledge-seeking, entertainment, leisure, and relaxation. Cronbach’s α reliabilities for the dimensions are also shown in Table 2.

Table 2.

Factor loadings and scale reliabilities for vacation activity dimensions.

Dimension/Variables	Loading	Eigenvalue	Percentage of variance	Cronbach’s α
Outdoor adventure		3.44	22.9	0.72
1. Camping	0.662
2. Attending sporting events	0.615
3. Going on safari	0.670
4. Diving	0.744
5. Exploring the outdoors	0.618
Knowledge-seeking		2.22	14.8	0.70
6. Visiting museums	0.825
7. Visiting historical places	0.811
8. Attending cultural events	0.691
Entertainment				0.72
9. Visiting nightclubs	0.702	1.42	9.5
10. Watching movies	0.747
11. Attending music concerts	0.825
Leisure		1.06	7.1	0.67
12. Shopping	0.696
13. Visiting beaches	0.686
14. Visiting amusement parks/recreation centers	0.710
Relaxation		1.02	6.8	0.65
15. Lazing around	0.640
16. Sleeping	0.598
17. Just relaxing	0.780

Clusters

Following standard practice as recommended in multivariate statistics texts, we conducted the cluster analysis in two stages. We first run hierarchical cluster analysis using the five activity dimensions as input, inspected the agglomeration schedule and associated dendrogram, and found that a three-cluster solution is a reasonable approximation to the data. We then run nonhierarchical clustering using the K-means procedure in Statistical Package for Social Sciences and specified a three-cluster solution. Mean scores of the three clusters on the vacation activity dimensions are shown in Table 3 along with the results of one-way ANOVA tests for statistical significance of differences in the means. Pairwise differences were assessed using the Bonferroni procedure.

Table 3.

Cluster means on vacation activity dimensions.

		Activity dimension
Cluster	N (%)	Outdoor adventure	Knowledge seeking	Entertainment	Leisure	Relaxation
I—Conservative	186 (41)	2.4^a	3.7^a	1.7^a	3.6^a	4.4^a
II—Variety seekers	152 (34)	3.8^b	4.4^b	3.5^b	4.5^b	4.8^b
III—Fun seekers	113 (25)	2.8^c	2.8^c	3.7^b	4.2^c	4.2^c
Total sample	451 (100)	3.0	3.7	2.8	4.0	4.5
ANOVA statistics
F		139.96	160.22	342.06	58.27	20.60
p		0.000	0.000	0.000	0.000	0.000

ANOVA: analysis of variance.

Note: Scores are on a 5-point scale with higher scores indicating greater preference for the particular activity. Within each activity dimension, means with different superscripts are statistically different; those with the same are not. For example, for the ‘leisure’ dimension, means for conservatives, variety seekers, and fun seekers are all statistically different from each other; for the ‘entertainment’ dimension, means for variety seekers and fun-seekers are not different from each other; however, each is different from the mean for conservatives.

The mean scores of Cluster I are close to the sample average on the knowledge-seeking and relaxation dimensions, slightly below average on the leisure and outdoor adventure dimensions, but substantially below average on the entertainment dimension. Thus, this cluster is distinguished not by an above-average liking for any particular activity but rather a strong dislike for entertainment activities. Accordingly, we named this the ‘conservative’ cluster. It constitutes 41% of the sample. In sharp contrast to this, Cluster III (25% of the sample) has above-average scores on the leisure and entertainment dimensions and slightly below-average scores on the remaining dimensions. Accordingly, we named this the fun seekers because it appears to be a cluster that seeks fun activities during their vacation. Finally, the above-average scores of Cluster II on all five activity dimensions led us to label it as variety-seeker cluster. This is made up of 34% of the sample.

The ANOVA results show statistically significant differences in means of all five activity dimensions across the three clusters. However, in relative terms, the F values show that the differences are larger for the entertainment dimension, followed by the knowledge-seeking, outdoor adventure, and leisure dimensions. The relaxation dimension shows the smallest difference among clusters. In terms of pairwise differences, all three groups differ from each other for all activity dimensions except entertainment. For this dimension, the conservative cluster differs significantly from the remaining two, but these do not differ from each other.

Cluster validation

We run discriminant analysis to validate the cluster solution and identify the activity dimensions that discriminate among the clusters. In this analysis, we used respondents’ scores on the activity dimensions as independent variables and membership in the three-cluster solution as dependent variable. The results are shown in Tables 4 to 6.

Table 4.

Unstandardized (standardized) discriminant function coefficients.

Vacation activity dimension	Function 1	Function 2
1. Outdoor adventure	0.572 (0.442)	0.434 (0.336)
2. Knowledge seeking	−0.092 (−0.069)	1.084 (0.811)
3. Entertainment	1.118 (0.821)	−0.422 (−0.310)
4. Leisure	0.357 (0.263)	0.185 (0.136)
5. Relaxation	0.013 (0.011)	0.451 (0.363)
Constant	−5.990	−6.923
Eigenvalue	1.992	1.034
Percentage of variance	65.8	34.2
Canonical correlation	0.816	0.713
Wilks’ λ	0.164	0.492
χ ²	798.2	313.8
df	10	4
p	0.000

Table 5.

Discriminant functions at group centroids.

Cluster	Function 1	Function 2
I: Conservatives	−1.669	0.098
II: Variety seekers	1.308	1.080
III: Fun seekers	1.007	−1.584

Table 6.

Classification of results.

		Predicted group membership^a
Actual group	No. of Cases	I	II	III
I: Conservatives	186 (41%)	185 (100.0%)	0 (0.0%)	0 (0.0%)
II: Variety seekers	152 (34%)	6 (4.0%)	140 (94.0%)	3 (2.0%)
III: Fun seekers	113 (25%)	4 (3.5%)	6 (5.3%)	103 (91.2%)

^aPercentage of grouped cases correctly classified; hit ratio = 95.7%.

Table 4 shows the discriminant function coefficients and associated model statistics. Given that there are three clusters, two discriminant functions were extracted. Both functions have eigenvalues greater than one and have statistically significant discriminating ability as reflected in the associated Wilk’s λ, χ ², and p values. Function 1 accounts for almost 66% of the variance in the vacation activity dimensions, while Function 2 accounts for the remaining 34% variance. The function coefficients (both unstandardized and standardized) show that Function 1 is defined primarily by the entertainment activity dimension, with the adventure and leisure dimensions making modest contributions, while Function 2 is defined primarily by the knowledge-seeking dimension, with adventure and relaxation also making modest contributions. Thus, Function 1 is largely an entertainment-driven function, while Function 2 is mainly knowledge driven.

Table 5 shows the cluster centroids on the discriminant functions. Cluster I has a relatively high negative score on Function 1 (the entertainment-dominant function), but a low positive score on Function 2. Thus, this cluster is defined more by a dislike for entertainment activities than by a liking for any particular activity. This reaffirms our naming it the conservative cluster. In contrast, Cluster III has a positive score on Function 1 and high negative score on Function 2, indicating a general liking for entertainment-related activities and a relative dislike for the knowledge-seeking activities that dominate Function 2. This reaffirms our naming this cluster the fun seekers. Finally, Cluster II has high positive scores on both functions, reaffirming its designation as variety seekers.

The classification statistics (confusion matrix) in Table 6 show that the two discriminant functions are able to correctly classify almost 96% of the respondents into their respective cluster, indicating a high reliability of the clustering solution.

Cluster sociodemographic profiles

Results of χ ² analyses of sociodemographic differences in clusters are shown in Table 7. In our analyses, we designated the respective sociodemographic variable as independent variable and cluster membership as dependent variable. This allowed for determination of the conditional probability of cluster membership, given the respective sociodemographic variable.¹

Table 7.

Cluster sociodemographic profiles.

	Conservatives	Variety seekers	Fun seekers	Total
Gender (χ ² = 7.83; p = 0.020)
Male	46.0%	30.9%	23.2%	100%
Female	32.6%	38.4%	29.1%	100%
Age (χ ² = 33.11; p = 0.000)
Up to 25 years	26.2%	36.9%	36.9%	100%
26–35 years	43.4%	34.9%	21.7%	100%
36–45 years	53.7%	30.5%	15.9%	100%
Over 45 years	70.8%	12.5%	16.7%	100%
Marital status (χ ² = 31.48; p = 0.000)
Married	51.7%	26.6%	21.7%	100%
Single	25.3%	44.0%	30.8%	100%
Income (χ ² = 18.23; p = 0.108)
<SR 1000	21.6%	43.2%	35.1%	100%
SR 1000–SR 4999	35.8%	37.9%	26.3%	100%
SR 5000–SR 9999	40.0%	37.3%	22.7%	100%
SR 10,000–SR 14,999	43.1%	30.4%	26.5%	100%
SR 15,000–SR 20,000	57.1%	19.0%	23.8%	100%
>SR 20,000	53.1%	21.9%	25.0%	100%
Education (χ ² = 10.12; p = 0.120)
High school and below	39.3%	32.8%	27.9%	100%
Diploma	29.8%	39.5%	30.7%	100%
University degree	45.7%	31.6%	22.7%	100%
Other	57.1%	21.4%	21.4%	100%

SR: Saudi riyal; US 1.00 = SR 3.75.

Note: Percentages in boldface and italics indicate cells for which observed frequencies are greater than expected under the respective null hypothesis.

The results in Table 7 show statistically significant differences in probabilities of cluster membership across gender, age, and marital status. Specifically, males are more likely to be in the conservative cluster than in the other two (46% conservatives vs. 31 and 23% for variety seekers and fun seekers, respectively). In contrast, females are more likely to be variety seekers (38% of females are in this cluster) than fun seekers or conservatives. On age differences, older respondents are significantly more likely to be conservatives, while younger respondents are more likely to be fun seekers or variety seekers. As shown in Table 7, 71% of all respondents over 45 years fall into the conservative cluster, compared to only 26% of those below 25 years. In contrast, 37% of respondents under age 25 fall into fun and variety seekers compared with only 12–18% of respondents above 45 years. Indeed, there is an increasing monotonic relationship between age-group and probability of membership in the conservative cluster and a corresponding decreasing monotonic relationship between age-group and probability of membership in the fun- and variety-seeking clusters. Finally, significant differences exist also for marital status, with married respondents more likely to be in the conservative cluster and non-married respondents more likely to be in the remaining two clusters.

Segment differences in travel preferences

In addition to sociodemographically profiling the segments, we also examined differences in their travel preferences. In the questionnaire, respondents indicated where they generally prefer to spend their annual vacation and what their plans were for this year’s vacation. For both questions, we elicited dichotomous responses with response options in-Kingdom and out of Kingdom. Regarding the plans for this year’s vacation, we also asked respondents to indicate in an open-ended format, which specific country they planned to travel to, if out of Kingdom. We examined segment differences in travel destination preferences through cross-tabulation and χ ² analysis involving the dichotomous responses and cluster membership. The results (Table 8) show statistically significant differences among the clusters. Relative to the other two groups, the conservative segment generally prefer to spend their annual vacation locally in-Kingdom, while the fun-seeking segment generally prefer to go out of Kingdom. Plans for the current year’s annual vacation mirror these general preferences.

Table 8.

Segment differences in travel destination preferences.

	General vacation preferences		Plan for this year’s vacation
Segment	In-Kingdom	Out of Kingdom	In-Kingdom	Out of Kingdom
Conservatives (n = 183)	39.9%	60.1%	38.3%	61.7%
Variety seeking (n = 152)	25.7%	74.3%	20.9%	79.1%
Fun seeking (n = 110)	14.5%	85.5%	17.4%	82.6%
χ ²	22.63		19.43
p	0.000		0.000

In the questionnaire, respondents also indicated their level of agreement with a set of travel-related statements drawn from Lehto et al. (2002) and reflective of preferences for travel modes and attitudes toward travel. We also examined differences among the segments in their responses to these statements using one-way ANOVA and the Bonferroni procedure for pairwise differences. The results (Table 9) also show significant segment differences at p = 0.05, particularly between the conservative segment on one hand and the variety-seeking and fun-seeking segments on the other.

Table 9.

Segment differences in travel preferences.

	Means			ANOVA statistics
	Conservative	Variety seeking	Fun seeking	F	p
I usually like to travel on all-inclusive holiday packages	3.15^a	3.49^b	3.42^b	3.10	.046
A vacation should be spent exploring new things and learning about other people and their cultures	4.16^a	4.61^b	4.04^c	14.72	.000
I enjoy making my own arrangement for my holidays	4.18^a	4.36^a,b	4.08^a,c	2.91	.055
I like to travel during holidays just to get away from home	2.94^a	3.37^b	3.59^b	8.33	.000
I generally like to travel to places I have not been to before	4.11^a	4.54^b	4.35^b	7.58	.001
I generally like to travel to places I can talk about when I get home	3.32^a	3.96^b	3.93^b	13.04	.000
I generally like to travel to places that my friends have not been to before	2.46^a	3.20^b	3.25^b	16.77	.000
It is prestigious to spend the annual vacation outside the Kingdom	1.86^a	2.52^b	2.71^b	15.26	.000

ANOVA: analysis of variance.

Note: Responses were on a 5-point scale ranging from 1 = strongly disagree to 5 = strongly agree. Means with the same superscripts are not significantly different from each other at p = 0.05; those with different superscripts are. For example, for item 1, the mean score for conservatives is significantly different from the means for both variety seekers and fun seekers; however, the means for the latter two are not significantly different.

For five of the seven statements, the latter two segments have similar preferences; they like to travel to get away from home (confirming their preferences for out-of-Kingdom vacations), they like to travel to new destinations, to places their peers have not been to before that they can talk about when they return home, and they believe (more than the conservatives) that it is prestigious to spend a vacation out of Kingdom. However, in terms of preferences for vacation arrangements and learning about other people and cultures, the pattern of differences is somewhat different. Variety-seeking respondents have a higher preference for making their own holiday arrangements than the remaining two segments and have a significantly greater preference for traveling to destinations with different people and cultures.

Discussion

The present study has identified three main segments of outbound Saudi vacation tourists—conservatives, fun seekers, and variety seekers—based on the activities they prefer to engage in during the ideal vacation. The segments exhibit different sociodemographic profiles, as well as differences in vacation travel preferences and attitudes. In terms of vacation activities, the conservatives are unique in their vehement dislike for entertainment-oriented activities (nightclubs, music concerts, and movies), although they have a slight inclination toward liking knowledge-seeking activities. This is the largest segment, comprising 41% of the population, and is dominated by relatively older, married, and male respondents. They are more likely than the other segments to prefer domestic vacation destinations and to stick with the familiar. Fun seekers emphasize shopping and fun-related activities like visiting amusement parks and beaches. This segment comprises about 25% of the market and is dominated by young, single, and female respondents. They are the most likely segment to prefer international vacation destinations and, similar to the variety seekers, to prefer travel to new destinations, people, and cultures. Finally, the variety seekers (34% of the population) have above-average ratings for the necessity of all the included vacation activities as part of a vacation. During their vacations, they like to do as much as they can afford in terms of time and money. They like to shop, engage in leisure activities (such as visit beaches and amusement parks), and won’t mind visiting historical and cultural attractions. This segment is also made up of majority female, single respondents but with relatively higher age categories than the fun-seeker segment. They are more likely to prefer international destinations than the conservatives but less likely than the fun seekers; in essence, appearing to balance between the two destinations. Similar to the fun seekers, they prefer to travel to new destinations, people, and cultures.

As discussed earlier, segmentation results like these provide powerful input to marketing strategy planning by government and private tourism and hospitality organizations in destinations targeting leisure tourists from Saudi Arabia. The pattern of sociodemographic differences in profiles of these segments suggests that destination marketers need to understand that the Saudi outbound tourist market is not one homogeneous market that can be reached with one offering and communication. Rather destinations need to develop their profiles and communications based on who they intend to target in this market. For instance, destinations targeting families need to be wary of and strive to accommodate the different activities that married male and married female respondents deem necessary as part of an ideal vacation. In particular, since relatively older, married males tend to be more conservative while their female counterparts prefer fun and leisure-oriented activities, these destinations need to develop attractions that offer the possibility to engage in both leisure- and fun-oriented activities without overly emphasizing any entertainment-oriented activities. Vacation and/or tour packages that combine historical and cultural attractions, beach and amusement park visitation, as well as ample opportunities for high-quality shopping, will particularly appeal to this group of tourists.

On the other hand, destinations and tour organizations targeting singles, particularly single males, will need to emphasize the variety of activities they have to offer. But, depending on who they are targeting, even these will need to be circumspect in how they communicate. For middle-aged to older consumers, communications should de-emphasize entertainment-related activities; while for younger to middle-aged singles, communications could include these activities. But even marketers still need to be cautious in how they communicate this information. Saudi society is still a conservative society, as evidenced by the relative size and composition of the conservative segment. It is the largest segment, and although dominated by older, married males, it is important to note that other sociodemographic segments are also well represented in this segment. For example, roughly 25% of respondents below 25 years and 25% of non-married respondents are classified in this category. They do not particularly consider entertainment activities as a necessary part of a perfect vacation. Destination marketers, who think that the majority of Saudis are eagerly waiting to get out of their entertainment-deprived country during vacations to engage in activities that they otherwise do not have an opportunity to engage in, need to rethink their positions. Conservatism is alive and well in the country and permeates different sociodemographic segments. Thus, it is important for these marketers to understand the need to always take into account Saudi culture and traditions both in the design of destination offers and in communication with this market.

Within the context of the specific results of the present study, it is also important for destination marketers targeting Saudi tourists to note which vacation activities are particularly important to leisure tourists from Saudi Arabia as a whole. As identified in Table 1, these are relaxation, amusement and fun activities, and shopping. Destinations that have these activities to offer need to emphasize them in their marketing communication, while those that do not possess them need to determine the feasibility or otherwise of developing them and the eventual likelihood of successfully targeting Saudi tourists.

The present study contributes to the academic literature as well. The most obvious is that it further demonstrates the relevance, benefits, and practicality of using preferred vacation activities as a basis for segmentation. The benefits of activity-based tourism segmentation has earlier been identified (Lang et al., 1993) as including better understanding of visitors’ travel choices and patterns, better destination product development that is focused on bundled activities, and effectiveness in marketing strategy development as this is done on the basis of a bundle of activities rather than individual activities (Lang et al., 1993). With clearly distinct segments identified, sociodemographically profiled, and linked to destination and travel preferences, this study has provided a better understanding of Saudi vacation tourists and their travel preferences than any purely demographic segmentation would have provided.

The activity dimensions and segments identified in the present study are quite similar to those found in some previous studies. To be sure, specific vacation activities included in this and previous studies necessarily tend to be context specific. For example, Yang (2009) included activities—like deer hunting, turkey hunting, snowmobiling, casino gaming, and so on—that are common in the Michigan destinations that are the subject of his study. Nevertheless, certain other activities—such as sightseeing, shopping, and visiting museums and amusement parks—transcend several studies, resulting in fairly comparable activity dimensions—and in some instances, segments—across studies. For instance, the activity dimensions extracted from our analysis of ideal vacation activities are similar to those found by Kozak (2002) and validated by Jonsson and Devonish (2008). The segments identified in our study can also be found in most of the activity-based segmentation studies reviewed earlier. In particular, the entertainment and outdoor adventure segments appear to endure in a lot of studies conducted in different national contexts. This indicates the possibility that there exist tourist segments that are global and that could be attracted by destinations offering activities desired by those segments.

Limitations and suggestions for future research

Two caveats need to be taken into account when interpreting the results as well as improved upon in future studies. First, we asked respondents to indicate what activities they consider necessary during the ideal vacation; we did not actually observe what activities they were engaged in. There is a possibility that the responses could be affected by social desirability bias. For instance, the relatively low-stated importance of entertainment-oriented activities could be attributed to this. In the religiously conservative Saudi context, people may be apprehensive about admitting that they go on vacation to indulge in entertainment activities like watching movies and attending concerts—activities that are otherwise not allowed in Saudi Arabia. Future studies could attempt to survey Saudi tourists at particular destinations so that researchers can observe what activities they actually engage in. Related to this is the timing and mode of data collection. As indicated in the method section, the data were collected through mall intercepts shortly before the travel season began because a key objective of the larger study was to examine respondents’ vacation plans for that year. It is possible that responses were affected by the fact that respondents were still in the planning phase of their vacations. Again a survey of tourists in the actual destination setting could help overcome this potential bias.

Footnotes

Appendix

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Note

References

Andreu

Kozak

Avci

. (2005) Market segmentation by motivations to travel: British tourists visiting Turkey. Journal of Travel and Tourism Marketing 19(1): 1–14.

Baloglu

Assante

(1999) A content analysis of subject areas and research methods used in five hospitality management journals. Journal of Hospitality and Tourism Research 23(1): 53–70.

Beritelli

Boksberger

(2005) Activity-based market segmentation: a behavioural approach. Tourism 53(3): 259–266.

Bieger

Laesser

(2002) Market segmentation by motivation: the case of Switzerland. Journal of Travel Research 41(1): 68–76.

Bogari

Crowther

Marr

(2004) Motivation for domestic tourism: a case study of the Kingdom of Saudi Arabia. Tourism Analysis 8: 137–141.

Bonn

Furr

Susskind

(1999) Predicting a behavioral profile for pleasure travelers on the basis of Internet use segmentation. Journal of Travel Research 37(4): 333–340.

Business Monitor International (BMI) (2012) Saudi Arabia Tourism Report Q1 2012, London, UK: BMI.

Cha

McCleary

Uysal

(1995) Travel motivations of Japanese overseas travelers: a factor-cluster approach. Journal of Travel Research 34(1): 33–39.

Choi

Tsang

CKL

(2000) Activity based segmentation on pleasure travel market of Hong Kong private housing residents. Journal of Travel and Tourism Marketing 8(2): 75–97.

10.

Choi

HSC

Murraya

Kwana

(2011) Activity-based segmentation of Canadian domestic pleasure travelers to New Brunswick. International Journal of Hospitality and Tourism Administration 12(3): 202–224.

11.

Colliers International (2011) Saudi Arabia tourism opinion piece. Available at: http://www.colliers.com/~/media/files/emea/uk/research/destination-consulting/saudi-arabia-tourism-overview.pdf. Accessed 31 January, 2014.

12.

Crompton

(1979) Motivations of pleasure travel. Annals of Tourism Research 6: 408–424.

13.

Dann

(1981) Tourist motivation: an appraisal. Annals of Tourism Research 8: 187–219.

14.

Dolnicar

(2004) Beyond ‘commonsense segmentation’: a systematics of segmentation approaches in tourism. Journal of Travel Research 42: 244–250.

15.

Formica

Uysal

(1996) A market segmentation of festival visitors: Umbria jazz festival in Italy. Festival Management and Event Tourism 3(1): 175–182.

16.

Formica

Uysal

(1998) Market segmentation of an international cultural-historical event in Italy. Journal of Travel Research 36(4): 16–24.

17.

Graham

JEJ

Wall

(1978) American visitors to Canada: a study in market segmentation. Journal of Travel Research 16(3): 21–24.

18.

Hsieh

O’Leary

Morrison

(1992) Segmenting the international travel market by activity. Tourism Management 13(2): 209–223.

19.

Iso-Ahola

(1987) Towards a social psychology theory of tourist motivation. Annals of Tourism Research 9: 256–262.

20.

Jang

Morrison

O’Leary

(2002) Benefit segmentation of Japanese pleasure travelers to the USA and Canada: selecting target markets based on the profitability and risk of individual market segments. Tourism Management 23(4): 67–378.

21.

Jang

Cai

Morrison

. (2005) The effects of travel activities and seasons on expenditure. International Journal of Tourism Research 7(6): 335–346.

22.

Johns

Gyimothy

(2002) Market segmentation and the prediction of tourist behavior: the case of Bornholm, Denmark. Journal of Travel Research 40(3): 316–327.

23.

Jonsson

Devonish

(2008) Does nationality, gender, and age affect travel motivation? A case of visitors to the Caribbean island of Barbados. Journal of Travel and Tourism Marketing 25 (3/4): 398–408.

24.

Kau

Lim

(2005) Clustering of Chinese tourists to Singapore: an analysis of their motivations, values, and satisfaction. International Journal of Tourism Research 7 (4–5): 231–248.

25.

Kim

Jogaratnam

(2003) Activity preferences of Asian international and domestic American university students: an alternate basis for segmentation. Journal of Vacation Marketing 9(3): 260–270.

26.

Kim

Wei

Ruys

(2003) Segmenting the market of West Australian senior tourist using an artificial neural network. Tourism Management 24(1): 25–34.

27.

Kozak

(2002) Comparative analysis of tourist motivations by nationality and destinations. Tourism Management 23(3): 221–232.

28.

Lang

O'Leary

Morrison

(1993) Activity segmentation of Japanese female overseas travelers. Journal of Travel and Tourism Marketing 2(4): 1–22.

29.

Lehto

O’Leary

Morrison

(2002) Do psychographics influence vacation destination choices? A comparison of British travellers to North America, Asia, and Oceania. Journal of Vacation Marketing 8(2): 109–125.

30.

Madrigal

Kahle

(1994) Predicting vacation activity preferences on the basis of value-system segmentation. Journal of Travel Research 32: 22–28.

31.

Masiero

Nicolau

(2012) Tourism market segmentation based on price sensitivity: finding similar price preferences on tourism activities. Journal of Travel Research 51(4): 426–435.

32.

Mazanec

(1992) Classifying tourists into segments: a neural network approach. Journal of Travel and Tourism Marketing 1(1): 2–6.

33.

McKercher

PSY

du Cros

. (2002) Activities-based segmentation of the cultural tourism market. Journal of Travel and Tourism Marketing 12(1): 23–45.

34.

Mehmetoglu

(2006) Segmenting the nature-based tourists based on travel mode choice. Journal of Hospitality and Leisure Marketing 14(4): 47–67.

35.

Mehmetoglu

(2007) Typologising nature-based tourists by activity: theoretical and practical implications. Tourism Management 28: 651–660.

36.

Mehmetoglu

Hines

Graumann

. (2010) The relationship between personal values and tourism behaviour: a segmentation approach. Journal of Vacation Marketing 16(1): 17–27.

37.

Middleton

Clarke

(2001) Marketing in Travel and Tourism. Woburn, MA: Butterworth-Heinemann.

38.

Moscardo

Morrison

Pearce

. (1996) Understanding vacation destination choice through travel motivation and activities. Journal of Vacation Marketing 2(2): 109–122.

39.

Park

Yoon

(2009) Segmentation by motivation in rural tourism: a Korean case study. Tourism Management 30(1): 99–108.

40.

Sarigollu

Huang

(2005) Benefits segmentation of visitors to Latin America. Journal of Travel Research 43(3): 277–293.

41.

Seddon

Khoja

(2003) Saudi Arabian tourism patterns and attitudes. Annals of Tourism Research 30(4): 957–959.

42.

Shoemaker

(1994) Segmenting the U.S. travel market according to benefits realized. Journal of Travel Research 32(3): 8–21.

43.

Silverberg

Backman

(1996) A preliminary investigation into the psychographics of nature-based travelers to the southeastern United States. Journal of Travel Research 35(2): 2–21.

44.

Sirakaya

Uysal

Yoshioka

(2003) Segmenting the Japanese tour market to Turkey. Journal of Travel Research 14(1): 293–304.

45.

Spotts

Mahoney

(1991) Segmenting visitors to a destination region based on the volume of their expenditures. Journal of Travel Research 32(4): 24–31.

46.

Sung

Morrison

O’Leary

(2000) Segmenting the adventure travel market by activities: from the North American industry providers’ perspective. Journal of Travel and Tourism Marketing 9(4): 1–20.

47.

Thrane

Farstad

(2012) Nationality as a segmentation criterion in tourism research: the case of international tourists’ expenditures while on trips in Norway. Tourism Economics 18(1): 203–217(15).

48.

Thurau

Carver

Mangun

. (2007) A market segmentation analysis of cruise ship tourists visiting the Panama Canal watershed: opportunities for ecotourism development. Journal of Ecotourism 6(1): 1–18.

49.

Tsiotsou

(2006) Using visit frequency to segment ski resorts customers. Journal of Vacation Marketing 12(1): 15–26.

50.

WTO (2003) Outbound tourism from Saudi Arabia. Madrid, Spain: World Tourism Organization.

51.

Yang

(2009) Activity-based market segmentation in a rural tourism destination: a case study of west-central Michigan. PhD Dissertation, Michigan State University, East Lansing, MI, USA