“Destination image and its consequences in the perspective of four-stage loyalty model (an empirical evidence from visitors of Tarakan City, Indonesia)

The study was conducted to test the postulate of four-stage loyalty model through the concept of destination image to represent cognitive component, satisfaction for affective aspect, and intention loyalty as the representative of conative component. The sample in this research was 102 tourists who visited Tarakan City, Indonesia in November 2016. The structural equation model showed that destination image was satisfaction antecedent, and satisfaction was the antecedent of intention loyalty. The impact of destination image towards intention loyalty was fully mediated by satisfaction. Theoretically, this research result had the implication in developing loyalty theory, while the managerial implication was beneficial to increase visitors loyalty to Tarakan City as a tourism destination.


INTRODUCTION
During the last six decades, tourism had experienced sustainable expansion and diversification thus it served as one of the largest economic sectors with fastest growth in the world.In 2014, the number of international tourist (overnight visitors) arrivals reached 1,138 million arrivals or increased by 4,000% from 25 million tourist arrivals in the 1950's and increased by 51 million arrivals or 4.7% by 2013 (UNWTO, 2015).With the ever increasing growth, tourism became a highly competitive industry for every tourism destination marketer.Morgan et al. (2002) mentioned that 70% of all tourists only visited the main tourism destinations in ten countries known as "best niche players", and more than 70% revisited the same tourism destinations (Chen & Tsai, 2007;Richards, 2006;Chaminuka et al., 2011).Hence, a successful tourism destination marketing strategy will not only focus on winning new tourists, but it should also be able to build tourist loyalty they already won (Palmer, 1994).Loyalty was a fundamental strategic component for the company (Bigne et al., 2001), and it was proven to be related with profitabilities (Hallowell, 1996).Loyalty could be predicted from the intention loyalty as conative component, which, according to the postulate of four-stage loyalty model (Oliver, 1997) served as the antecedent of affective component, and affective component was the antecedent of cognitive component.
In the field of tourism, satisfaction was a concept from affective component and mainstream concept tested as the antecedent of inten-tion loyalty (Kozak, 2001 Cho et al., 2014).Although the majority of empirical findings showed that satisfaction had significant direct impact on intention loyalty, the study of Bigne et al. (2001) towards the visitors of two tourism destinations in Valencia, Spain, namely Peniscola and Torrevieja with two concepts to represent conative loyalty, namely intention to revisit and intention to recommend instead found that satisfaction had positive and significant impact on intention to recommend on both samples, but the indicator of intention to revisit showed a different result.On Torrevieja sample, Bigne et al. (2001) confirmed that satisfaction had positive and significant impact on intention to revisit, but it was insignificant on Peniscola sample.Inconsistent result was also shown by the research result by Mechinda et al. (2009), towards Domestic and International tourists who visited Chiangmai City, namely that satisfaction had positive and significant impact on attitudinal loyalty on Domestic tourists' sample, but it was insignificant on International tourists' sample.The aforementioned findings by Bigne et al. (2001) and Mechinda et al. (2009) showed that the difference in research setting and unit will result in differences of satisfaction impact towards intention loyalty; hence satisfaction could not be concluded as antecedent of tourist loyalty.
Aside from contradictive findings regarding satisfaction direct impact towards intention loyalty on particular research setting and unit, the postulate of four-stage loyalty model (Oliver, 1997) stated that loyalty was formed through four stages with the pattern of cognitive → affective → conative → action (p.394) and was not supported by empirical model developed by Tang (2014).Tang (2014) stated that the impact of satisfaction on behavioral intention was smaller compared to the impact of tourist attraction towards behavioral intention.Hence, the role of satisfaction to represent affective component as the antecedent of intention loyalty would need a retest in tourist behavioral study.
Tourist attraction was one of destination image dimensions (Byon & Zhang, 2010), which recently attracted the interest of researchers in the field of tourism.Some empirical studies reported that destination image had significant direct impact on satisfaction (Loureiro & Gonzalez, 2008 Chen and Tsai (2007) found that direct impact of destination image on satisfaction was insignificant, also Prayag and Ryan (2011) also found that destination image was mediated by satisfaction towards intention loyalty.Thus, the role of destination image as the antecedent of satisfaction and intention loyalty was not yet generalizable; hence it will still need to be tested again in empirical study.Related to that, thus the structural equation model developed in this research was intended to know the impact of destination image on satisfaction, impact of destination image on loyalty, and impact of satisfaction on intention loyalty.The test result of inter-concept relations was expected to be beneficial as the development of tourist loyalty theory in the perspective of four-stage loyalty model (Oliver, 1997, p. 394) and tourism destination marketing strategy, particularly Tarakan City.

Four-stage loyalty model
To learn about post-purchase consumer behavior, Oliver (1997) introduced loyalty theory by the label of four-stage loyalty model.Oliver (1997) admitted that four-stage loyalty model, which em-ployed the pattern of cognitive → affective → conative → action was developed from attribute-based models of attitude initiated by Fishbein and Ajzen (1975).According to Oliver (1997), behavioral components which consisted of cognitive (knowledge), affective (emotional or feelings aspect), and conative (intention or commitment) were basically inconsonant dimensions of loyalty.Due to that, said components will become consonants through cognitive sense, affective means, conative sense; and behavioral ways sequentially.Additionally, Oliver (1997) explained that the three stages of attitudinal loyalty (cognitive, affective and conative) involved "sustainers" to strengthen loyalty level (e.g., value based on functional aspect for cognition, satisfaction for affect, and commitment for conation).The relations and occurrence sequence of said three components could be seen in Figure 1.
Cognitive loyalty was the available information base to attract consumers' attention to one of the brands.As the first stage in the model, cognitive loyalty was deemed as loyalty in cognitive sense, or consideration.Due to that, this stage was based on functional characteristic, cost and primary advantage.
Second stage would be affective loyalty, namely loyalty based on affect (feelings).Commitment in this stage was known as affective loyalty.In this case, loyalty was planted inside the consumers' mind as an affect (feelings) and no longer as mere cognition (consideration).Affective loyalty occurred during the consumption hence it involved the preference, fulfilling experiences, also preferences; and this stage also had the potential of causing dissatisfaction.
Conative loyalty as the third stage of loyalty shaping appeared as conative (behavioral intention) and directly affected by affective change towards brand.Conation was an intention or commitment to behave towards the destination in particular way containing statement of commitment to purchase, which later would turn into action loyalty, namely consumers' real action or behavior of repeat purchase.Measurement of repeat-purchase patterns should be reserved for situations where only purchase data (e.g., scanner data) are available (Oliver, 1997, p. 397).For this reason, the fourth stage (action loyalty) was not tested in this study.

Destination image
In the last two decades, various studies showed that image was a vital concept to understand the process taken by the tourists in selecting the tourism destination (Baloglu & McCleary, 1999).A favorable image was viewed as a critical aspect of a company's ability to maintain its market position (Faullant et al., 2008).Image was a behavioral construct consisted of individual knowledge, mental representation (belief), feelings, and global impressions regarding object or destination (Baloglu & McCleary, 1999).
In the field of tourism, the concept of destina- Source: Oliver (1997, p. 394).
assessing the destination's image was crucial to the design of effective place marketing strategies (Faullant et al., 2008).(Oliver, 1980).Stedman (2002) described place satisfaction as a multidimensional summary judgment of the perceived quality of a setting.It was viewed as the utilitarian value of a place to meet certain basic needs ranging from sociability to services to physical characteristics (Stedman, 2002).
According to Gallarza et al. (2013), tourists' satisfaction with their visit to particular tourism destination could be measured by three indicators, namely: My choice to purchase the trip to "X" City was a wise one, I did the right thing when I purchased the trip to "X" City, and This experience that I received in "X" City is exactly what I needed.
The strong relations between customer satisfaction and loyalty had caused the effort to maximize visitors' satisfaction becoming one of the main purposes of tourism destination administrators (Yuksel, 2010).
Past research, adopting the proposed conceptual sequence of loyalty (Oliver, 1997) H2: Satisfaction had direct and significant effect on intention loyalty.

Intention loyalty
The efforts to increase company competitiveness and profitabilities brought the term 'loyalty' into business activities.Loyalty was defined as ''a deeply held commitment to re-buy or re-patronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior'' (Oliver, 1997, p. 34).However, loyalty measurement in tourism context based on the definition above was a difficult issue because tourism products purchase was a rare kind of purchase (Opperman, 1999 H3: Destination image had direct and significant influence on loyalty intention.

Sample and data collection
The

Data analysis
To achieve the purpose of this research, the PLS-SEM approach was used to test the hypotheses, and processed with SmartPLS Professional 3.0 software (Ringle et al., 2015).PLS-SEM could be utilized for research which aimed to predict target variables and using latent variable values in the subsequent analysis (Hair et al., 2014, p. 19).
PLS was an iterative combination from the analysis of main components and regressions to explain construct variants in the model (Chin, 1998)

Measurement model
Variable measurements of this research used the reflective approach.Destination image variable was measured in multidimensions, while satisfaction and intention loyalty were measured by unidimension.The result of measurement model evaluation (

Structural model
The evaluation result of structural model (Figure 2) showed that two out of three developed paths in the structural model had significant coefficients, while the coefficient of "ecotourism destination image to intention" path was insignificant.The explicable change variation was the change occurred on each predictor that had a high predictive power, because R 2 is greater than 0.20 (Hair et al., 2014, p. 175).Intervariable relations in the model also had their predictive relevance, as the Q 2 value of every endogenous latent variable was marked positive (Hair et al., 2014, p. 184).If the value of VA F was less than 20% it could be concluded that there was (almost) no mediation.On the contrary, when VA F had a very great result, namely above 80%, there was a full mediation happened.The situation where VA F was greater than 20% and lesser than 80% could be categorized as partial mediation (Hair et al., 2014, p. 225).Table 2 showed that destination image indirect impact was 0.435 and significant at p < 0.000.With VA F value of 94.77% it could be concluded that the impact of destination image towards intention loyalty was fully mediated by satisfaction.

DISCUSSION
The first purpose of this research was identifying tourists' image towards a city as a tourism destination, while the second purpose was testing the postulate of four-stage loyalty model (Oliver, 1997) through the concept of destination image, satisfaction, and intention loyalty.
The result of this research showed that destination image was formed through four dimensions, namely: dimensions of infrastructure, attraction, value for money, and enjoyment.The test result of the first hypothesis showed that destination image had positive and significant direct impact towards satisfaction.Hence, the research result was a support towards findings from the preceding empirical study (Loureiro & Gonzalez, 2008  The test result of the second hypothesis confirmed that satisfaction had positive and significant direct impact on intention loyalty, supporting the result of preceding study (Kozak, 2001 The pattern found on the tourists' loyalty model based on the findings of this research was destination image (cognitive) → satisfaction (affective) → loyalty intention (conative).Hence, the result of this study was one of the empirical evidence confirming the postulate of four-stage loyalty model (Oliver, 1997), namely conative component as the consequence of affective component, and affective component was the antecedent of cognitive component.

Theoretical implication
Loyalty model pattern based on this research result was: destination image (cognitive) → satisfaction (affective) → intention loyalty (conative), an empirical evidence from the field of tourism which confirmed three out of four stages in four-stage loyalty model (Oliver, 1997).Therefore, the role of a concept representing affective component to mediate the impact of a concept representing cognitive component towards a concept representing conative component was vital in the development of tourists' loyalty theory.

Managerial implication
To position Tarakan City as a tourism destination and creating tourists's loyalty, the Local Government must be able to create image for the tourists.The more appropriate Tarakan City as tourism destination, the higher the tourists' satisfaction with their visit to Tarakan would be.Furthermore, the satisfied tourists would recommend other people to visit Tarakan, telling positive things about Tarakan City to others, and returning for another visit.
Tourists's image towards Tarakan City as a tourism destination could be shaped by maintaining its uniqueness and improving the quality of various tourism attractions and locations currently existed in Tarakan City, not burdening the business lines involved in tourism industry by local tax and retribution to keep it from impacting tariff setting principles to maintain the image of Tarakan City as a "good value for money" tourism destination.The capability to maintain security stability and improvement of infrastructure quality would also establish visitors' convenience and the impression of Tarakan City as a good place for enjoyment.

Table 1 .
Measurement model evaluations Note: a not valid; LV: latent variable; AVE: average variance extracted; Sqrt: square root.

Table 2 .
Effects of destination image on intention loyalty (Chen & Tsai, 2007ho et al., 2014), but it contradicted the findings byTang (2014).The third hypothesis, namely 'destination image which directly affects intention loyalty in positive and significant manner' is not supported by the data.Thus, the findings support the research result by Prayag and Ryan (2011) who stated that the impact of destination image on intention loyalty was mediated by satisfaction.On the contrary, the research findings were contradictory with the preceding research result which stated that intention loyalty was directly predicted by destination image(Chen & Tsai, 2007; Bigne et al., 2001; Loureiro & Gonzales, 2008; Qu et al., 2011; Faullant et al., 2008; Assaker & Hallack, 2013; Byon & Zhang, 2010).