Abstract
This research explores the role of dual-processing systems in evaluating complex information on online travel agency (OTA) websites, explicitly focusing on the transformative effects of multiple cues considering information sources. Two experimental studies were conducted to explore the processing modes of information using priming in assessing the impact of a controllable cue (brand in study 1 and scarcity in study 2 and an uncontrollable cue (customer ratings). The findings indicate that customer ratings function as a primary cue. With positive priming producing heuristics, the value of a brand offsets the negative impact of low ratings, while it is disregarded with high ratings. The perceived value of scarcity aligns consistently with customer ratings due to its inherent uncertainty with irrelevant priming triggering a systematic mode. This research advances the understanding of customers’ evaluation of complex information and suggests practical guidelines for implementing optimal marketing strategies in OTA settings.
Introduction
The online travel agency (OTA) is a competitive environment in which hotel brands compete for the customer’s booking decisions. The vast majority of consumers use OTAs to research and compare hotel options well in advance of the purchase decision (Besiada, 2023). Online travel agencies (OTAs) present customers with a plethora of information, such as brand, price, marketing messages, and online reviews and ratings (Guo & Li, 2022). This information is vital to assist customers in making informed purchase decisions by describing the travel products comprehensively. However, excessive information can overwhelm customers and lead to information overload (Chen et al., 2009; Guo & Li, 2022). Although many studies have demonstrated the impact of information on OTAs on decision-making, the integrated effect of multiple cues is still indeterminate. Hotels have more control over information presented on their own websites than they do on OTAs; therefore, the effect of multiple cues is heightened in the competitive OTA environment.
It is well documented that hotel reviews and ratings have a strong impact on consumer decision-making (e.g., Wen et al., 2021). Research suggests that while customer reviews and ratings critically impact hotel booking behaviors, the influence of negative reviews outweighs that of positive reviews (Kim et al., 2021). Despite the substantial negative impacts, a hotel is restricted from manipulating or eliminating negative customer-generated information (Alvarez & Campo, 2011). Hotels can improve their service offerings to make ratings more favorable or respond online to enhance perceptions of trust (Sparks et al., 2016), but the ratings themselves are uncontrollable.
External cues transmitted through social media provide information that is present in the hotel booking environment, although it may be irrelevant to the purchase decision. People are exposed to pop-up messages, Instagram posts, and other cues that are out of the hotel’s control. Research suggests that social media posts can influence OTA behaviors and interact with the relevant information (Cain et al., 2024; Tanford et al., 2020). Along with these uncontrollable cues, operators have factors that they can control. For example, a well-known brand name creates certain expectations about the hotel’s quality and reputation. (Akdeniz et al., 2013; El-Said, 2020). Marketing messages, such as scarcity cues, can be used to create a sense of urgency and desirability of the product (Huang et al., 2020).
These cues and others co-exist in the OTA booking environment. In order to predict and manage their effects, it is important to know how they interact, and which cues prevail in the hotel judgment. The dual-processing theory explains how individuals make decisions utilizing two modes of processing, systematic and heuristic (Chen & Chaiken, 1999). A heuristic processing mode enables individuals to make intuitive decisions using selective information with minimal effort and energy, while a systematic mode requires a high degree of mental effort and time to make logical decisions considering all available information (Tversky & Kahneman, 1974). Priming effects make certain information more accessible and increase its impact on decision-making (Janiszewski & Wyer, 2014). This research proposes that external cues transmitted through social media create priming effects that activate a systematic or heuristic mode and influence the weight placed on multiple cues during the hotel evaluation process.
This research report two experiments that examine the effects of one controllable cue, one uncontrollable cue and one external priming stimulus on hotel evaluations and booking intentions. Although hotel choice is the ultimate goal, evaluations are at least as important since the typical consumer spends more than 5 hr over several weeks evaluating travel options on OTAs prior to booking (Besiada, 2023). In both studies, a social media post serves as the external priming stimulus and customer rating is the uncontrollable cue. In study 1, positive or negative affective priming is used in conjunction with hotel brand as the controllable cue. In study 2, relevant or irrelevant procedural priming is used in conjunction with scarcity as the uncertain controllable cue. Dual processing and priming principles provide a theoretical foundation to predict and explain the relative impact of multiple cues in the online hotel booking context.
Although research on ratings and reviews is prevalent, less is known about how they interact with other cues present in the OTA environment. The research contributes to knowledge by examining three distinct types of cues that coexist and providing a theoretical foundation to explain their relative impact. The findings are important to operators who must compete for business on OTA sites where many factors, such as ratings, are out of their control. By understanding how to leverage controllable factors while managing those that are uncontrollable, operators can gain a competitive advantage.
Literature Review
Dual-Processing Theory
Dual-processing theory is a widely recognized psychological concept that postulates how humans process information through two distinct mechanisms (Chen & Chaiken, 1999; Kahneman, 2011). The theory suggests that people can process information analytically using systematic or heuristic processing to make decisions. Individuals endeavor to reach logical conclusions using “rule-based reasoning” but often rely on “belief-based reasoning” due to their finite cognitive resources (Evans & Frankish, 2009, p. 41).
The systematic processing mode involves a controlled operation relying on explicit processes to generate reasonable attitudes and behaviors (Chen & Chaiken, 1999; Evans & Frankish, 2009). This mode involves a rigorous approach to scrutinizing information through a cognitive and deliberate analytical process, incorporating all aspects of the relevant content (Chartrand, 2005). Thus, individuals employ the systematic process when dealing with content-related cues, requiring a high level of cognitive effort, time, and energy (Tan et al., 2018; Tversky & Kahneman, 1974). However, systematic processing may be slow and less effective, as the human brain has limited mental capacity to process the vast amount of data (Evans & Frankish, 2009; Kahneman, 2011).
The heuristic processing mode simplifies judgments and generates intuitive conclusions through an implicit process, requiring minimal mental effort and time (Chen & Chaiken, 1999; Tversky & Kahneman, 1974). This mode involves selective attention to specific information, disregarding all available sources, thereby overcoming limited mental capacity and reducing cognitive overload (Chen et al., 2009; Kahneman, 2011). Heuristic processing can lead to biased and inaccurate conclusions by relying only on selective information (Evans & Frankish, 2009; Kahneman, 2011).
Priming
People are constantly exposed to peripheral information, referring to the background details in the environment that are not the center of attention. Despite its insignificance, peripheral information can influence the level of distraction, emotion, or cognitive load, which in turn affect the decision-making process (Zhang et al., 2016). For example, people often make impulsive purchases at a grocery store when hungry, and French wine sales increase with playing French music (Dijksterhuis, 2004; North et al., 1997). Even though customers are not consciously aware of the French music in the background, its presence subconsciously makes French-related information more accessible thereby influencing decisions to buy French wine. This illustrates how individuals can have multiple attitudes toward the same object based on their mental modes (Tan et al., 2018; Wood, 2000). According to dual-processing theory, two different modes of processing information depend on situational cues or circumstances.
The influence of peripheral information can be explored through experimental designs that incorporate priming techniques (Cain et al., 2024; Tanford et al., 2020). Priming is a mental framework that triggers specific ideas and increases the accessibility of the associated thoughts by processing stimuli prior to the main task (Janiszewski & Wyer, 2014; Minton et al., 2017). Peripheral information projected in priming stimuli leads to changes in attitudes and behaviors toward the primary task. Previous research demonstrated that priming using irrelevant social media posts can arouse certain feelings and, in turn, impact donation intentions (Tanford et al., 2020). Another study revealed that a priming task measuring individuals’ risk aversion increases sensitivity to risks, affecting hotel booking behaviors (Cain et al., 2024). Previous studies suggest that priming using peripheral information can modify decision-making, even when not directly associated with the hotel choice.
As a priming technique, affective priming is a psychological framework to lead to intended moods in response to “affect-loaded stimuli” (Minton et al., 2017, p. 311). People in a positive mood will likely leverage a heuristic mode as it signals a sense of security in decision-making, implying a decreased need for deliberate processing (Bohner et al., 1995). Conversely, negative emotion is sensitive to situations involving problems and risks, necessitating a thorough and analytical approach through a systematic mode (Bohner et al., 1995; Tan et al., 2018).
Procedural priming employs cognitive actions, tasks, or procedures to modify behaviors and actions in response to previously presented stimuli (Janiszewski & Wyer, 2014). By priming a process through a particular task, procedural priming facilitates the execution of prime-activated behaviors on the primary decision (Minton et al., 2017). A task with relevant information for the primary judgment is likely to engage a systematic process, whereas a task using irrelevant information encourages the use of a heuristic process (e.g., Dijksterhuis, 2004).
Multiple Cues in the Cue Assessment System
Online Ratings
Online reviews and ratings provide a simple and concise evaluation of a hotel based on the collective opinions of other travelers (de Langhe et al., 2016; Gavilan et al., 2018; Zhao et al., 2015). Generated by fellow customers rather than crafted marketing messages, reviews and ratings yield a powerful impact on the booking decisions due to their inherent credibility (Casaló et al., 2015; Chakraborty, 2019; Zhao et al., 2015). Other customers’ evaluations, reflected in the rating score, critically influence the decision-making process for choosing a hotel. That is, customer ratings displayed on the OTA environment are readily interpretable (Casaló et al., 2015; Wen et al., 2021). Cue diagnosticity theory emphasizes that highly diagnostic information is prioritized in decision-making, highlighting the importance of information adequacy (Akdeniz et al., 2013). Thus, customer ratings are prioritized over other available information in hotel booking decisions. The underlying mechanism behind this prioritization is signaling theory, which suggests that partial information signals the quality of the product as a whole (Spence, 1978). Consequently, salient information from customer ratings can override other less diagnostic cues, becoming the sole reference as a primary cue for hotel booking decisions.
Several studies suggest that while customer reviews and ratings consistently affect hotel booking behaviors, the influence of negative reviews outweighs that of positive reviews (Cain et al., 2024; El-Said, 2020; Miyazaki et al., 2005). Similarly, positive ratings increase travelers’ hotel choices, whereas unfavorable ratings can deter bookings (Gavilan et al., 2018). Hotels are constrained from directly manipulating or removing unfavorable scores to mitigate risk. Given the limited control over this customer-generated information, the customer rating is considered an uncontrollable cue (Alvarez & Campo, 2011; Browning et al., 2013). While hotels can manage unfavorable ratings by improving their services or responding to negative opinions on OTAs, such efforts constitute a long-term strategy, and the actual ratings themselves remain beyond direct control (Park & Allen, 2013).
Hotel Brand
In addition to uncontrollable cues, such as ratings, customers also encounter controllable cues that hotels can actively control and maintain in the OTA settings (de Langhe et al., 2016). Despite lacking control over uncontrollable cues, hotels can achieve favorable outcomes by strategically leveraging controllable cues. Brand knowledge is considered highly credible and diagnostic as a critical factor in evaluating product quality (Akdeniz et al., 2013). Brand plays an essential role in shaping customers’ perceptions of a product or service based on accumulated knowledge and experiences in the customers’ minds that help specify the brand through past and present marketing activities (Keller & Lehmann, 2006). In the OTA context of searching for hotels, customers with higher brand knowledge are likely to spend less time searching and clicking on fewer options to book hotels (Lee & Cranage, 2010).
However, the impact of a brand is influenced by the value of uncontrollable cues or vice versa in hotel booking situations where multiple cues are involved. A previous study demonstrated that the effect of a brand on hotel booking behaviors can be altered by the presence of online reviews and pricing (Wen et al., 2021). Furthermore, a brand can also modify the influence of uncontrollable cues, as indicated by a study illustrating that negative reviews are mitigated by the hotel’s established reputation and perceived credibility (Vermeulen & Seegers, 2009). Dual processing theory provides a framework to reconcile the role of hotel brand.
In Study 1, affective priming is manipulated in the form of a social media post when evaluating a hotel with a well-known brand or an independent hotel. Within the affective priming framework, research suggests that positive priming activates a heuristic mode to process information, whereas negative priming activates a systematic mode (Bohner et al., 1995; Tan et al., 2018).
Since people tend to be cognitive misers, they often rely on salient signals and include additional information in the decision process only when necessary. This principle explains that individuals tend to simplify the judgment process to optimize mental effort (Fiske & Taylor, 1991). It involves a heuristic process in which selective information is utilized to make effective decisions (Chen & Chaiken, 1999; Tversky & Kahneman, 1974). Previous studies have demonstrated that the effect of multiple cues depends on a primary cue when integrating information (Vermeulen & Seegers, 2009; Wen et al., 2021). In other words, a salient cue, such as customer ratings, determines the extent to which additional information is involved in decision-making. Additional information can be incorporated into the decision, particularly when the value of a primary cue is unfavorable. Information integration theory suggests that all available information can be accumulated and reflected in the decision-making process, with each piece of information contributing to the final judgment (Anderson, 2014). Therefore, it is hypothesized that hotel brand will offset the impact of a low rating when using a heuristic process with positive priming. When a rating is high, however, the decision will not include the value of the hotel brand, as the high rating itself serves as a satisfactory signal.
When incorporating negative priming to activate a systematic process, all available information is utilized for making analytical decisions in evaluating the value of a target product (Bohner et al., 1995; Chen & Chaiken, 1999; Evans & Frankish, 2009). In this process, the negative value of a primary cue sends a strong signal for the overall assessment, and this negative impact is challenging to overcome with other available information (Wen et al., 2021). Consequently, the hotel brand has limited influence in mitigating the negative impact of a low rating. On the contrary, a high rating is deemed sufficient to assess the value of a hotel, leaving little room to reflect the value of the hotel brand. As a result, the evaluation of a hotel is primarily determined by customer ratings, with minimal consideration for the brand.
Therefore, this study proposes a 3-way interaction among three factors: an uncontrollable cue, a controllable cue, and the mode of information processing. A customer rating serves as an uncontrollable cue, considered a primary cue, a hotel brand is utilized as a controllable cue, and priming activates processing mode. In this regard, this study proposes the following hypotheses:
Scarcity Message
Hotels leverage the scarcity principle to entice travelers into a prompt booking decision by displaying the limited number of remaining rooms (Huang et al., 2020; Kim et al., 2020). When the availability of a scarce product decreases, customers perceive it as more valuable because the opportunity to acquire it declines (Huang et al., 2020; Lynn, 1991). Consequently, the desire for the product increases as customers place a higher value on the product. By emphasizing the limited availability and implying the risk of losing the opportunity to secure the desired product, hotels can increase customers’ booking intentions (Huang et al., 2020; Kim et al., 2020). However, the marketing strategy using the scarcity principle also has negative impacts, leading to anxiety and cognitive dissonance due to the impulsivity of the decision (Kim et al., 2020; Li et al., 2021).
While a controllable cue—exemplified by a hotel brand—tends to have a clear value proposition, the question arises as to whether the cue assessment mechanism remains consistent when there is uncertainty about the value it attains. For example, a scarcity message gives rise to both positive and negative evaluations due to inherent uncertainty (Kim et al., 2020; Li et al., 2021). Although a scarcity message can offer valuable information for obtaining a limited-supply item, it can also pose potential risks and result in urgent decisions that may later be regretted. In this case, a heuristic process may not accurately assess the value of an uncertain cue because it selectively prioritizes salient information (Chen & Chaiken, 1999; Kahneman, 2011). Consequently, the value of the scarcity message can potentially be disregarded during the decision-making process due to its uncertainty.
On the other hand, the value of an uncertain cue can be determined by employing a normative rationale through a systematic process (Evans & Frankish, 2009). Systematic processing can produce an accurate judgment to clarify the uncertainty when purchase decisions involve possible risk (Huang et al., 2020; Zhang et al., 2014). The value of uncertain information is expected to depend on salient information, resulting in consistent evaluation among available information (Connelly et al., 2011). Hence, the value of an uncertain cue can be determined by the level of the primary cue through a systematic process.
In study 2, relevant or irrelevant procedural priming is implemented in conjunction with a scarcity message. Within the procedural priming framework, a task unrelated to the current decision can activate a heuristic mode, whereas a task containing information related to the decision operates a systematic mode (e.g., Dijksterhuis, 2004). Therefore, this study anticipates that an uncontrollable cue alters the impact of an uncertain controllable cue, and the dual-processing system changes the evaluations of the multiple cues. A customer rating serves as the uncontrollable cue, a scarcity message is used as the uncertain controllable cue, and priming is employed to trigger the dual-processing system. Therefore, a 3-way interaction between priming, scarcity and customer rating is hypothesized.
Method
This research comprised one pretest and two main studies aimed at verifying the proposed hypotheses. The pretest was conducted to determine appropriate levels of stimuli for main studies. Both studies focused on how travelers assess the value of multiple cues available on an online travel agency (OTA) website and how the evaluation changes depending on the processing mode. Study 1 tested hypotheses 1 and 2 to investigate the combined effect of uncontrollable and controllable cues through the application of affective priming. Study 2 examined hypotheses 3 and 4 to understand the effects of multiple cues, including an uncertain controllable cue, with procedural priming.
Priming was implemented to activate the dual-processing system. using social media content. Considering the substantial amount of time individuals spend on social media, the information derived from the channel is regarded as peripheral signal that disrupts primary tasks such as hotel reservations (e.g., Cain et al., 2024). In study 1, affective priming was applied using text-based social media posts. These posts were manipulated to elicit polarized emotions using negative or positive words, consequently activating the dual-processing system (See Appendix A). Procedural priming was employed for study 2, manipulating the relevance of the priming stimuli in the context of a hotel booking decision (See Appendix B). The response to the priming task operates a certain mode of processing depending on its relevance.
An online survey using Qualtrics panel data was utilized, with data collection taking place in February 2020. Participants were required to have the following criteria: having OTA experience within the last 12 months, being aged 18 or older, and residing in the U.S. A total of 459 responses for Study 1 and 450 responses for Study 2 were initially collected. To ensure data quality, responses identified as outliers, straight-liners, and those that failed attention check questions were eliminated. Consequently, 430 responses for study 1 and 418 responses for study 2 were used for the main analysis. The demographic characteristics of the sample are outlined to offer background information. Random assignment in the experimental design minimizes their influence on the research outcomes, ensuring the internal validity of the primary focus. The demographic information of all participants is presented in Table 1.
Demographic Characteristics.
Pretest
Stimuli for study 1 (hotel brand, customer rating, and affective priming) and study 2 (the level of scarcity, customer rating, and procedural priming) were selected through pretest with a total of 132 participants. For study 1, the overall brand value was measured using a 7-point Likert scale to choose a hotel brand. Among four well-known brands (e.g., Hilton and Marriott) and six independent brands with fictitious names, Hilton was selected for the well-known brand (M = 6.10), and Rockwellton (M = 3.23) was chosen for the independent brand (F1, 63 = 191.88, p < .001). For customer ratings in study 1, hotel credibility was assessed on a 7-point Likert scale across eleven options. Two levels of online ratings, 9.2/10 (M = 6.51) for high and 6.4/10 (M = 4.84) for low, were selected (F1, 56 = 163.01, p < .001). To select affective priming stimuli, ten positive social media posts including positive words (e.g., joy, happy, and grateful) and ten negative posts showing negative words (e.g., sad, annoy, and disappointed) were examined on a 7-point bipolar emotion scale (1: negative, 7: positive). A set of three social media posts for each condition (Mpositive > 6.14, Mnegative < 2.42) was selected, resulting in significantly different emotion levels at p < .001.
To determine an appropriate number of remaining rooms indicating the scarcity in study 2, six options were tested using a 7-point bipolar scale (1: high, 7: low). The effective level of scarcity was identified when a hotel displayed three remaining rooms (M = 3.77), significantly differentiating it from a hotel without a scarcity message (M = 4.39, F1, 63 = 7.46, p = .008). In study 2, two levels of customer ratings—specifically 4.5/5 (Mhigh = 5.94) and 3.0/5 (Mlow = 4.83)—were selected from a set of seven options (F1, 67 = 95.22, p < .001). To select procedural priming stimuli, a set of 16 social media posts was examined on a 7-point bipolar scale (1: irrelevant, 7: relevant) based on their relevance to hotel booking behaviors. Three irrelevant posts (M < 2.08) and three hotel-related posts (M > 5.68) were selected, reaching a significant level of p < .001.
Study 1
Design and Stimuli
The study employed 2 customer rating (low vs. high) × 2 hotel brand (independent vs. well-known) × 3 priming (no priming vs. positive vs. negative) factorial design. Customer rating indicates customers’ overall score out of 10 as an uncontrollable cue, which was 9.2 for high and 6.4 for low. Hotel brand refers to the value attached to a brand name as a controllable cue. The well-known brand was Hilton, and the independent brand was Rockwellton. Priming is utilized to manipulate a mode of processing information, simulating Facebook posts. Positive priming is employed to evoke pleasant feelings, triggering a heuristic processing mode, whereas negative priming aims to induce unpleasant feelings, activating a systematic processing mode. No priming is included as a control group. Examples of the stimuli are displayed in Appendix A.
Procedure and Measures
After consent and screening questions, participants were randomly assigned to one of twelve conditions, each representing various combinations of independent variables, priming, hotel brand, and customer ratings. In each condition, social media posts as a priming stimulus were initially presented to activate specific emotions that engage an associated processing mode, followed by a primary stimulus simulating an OTA website. Subsequently, participants were asked to rate three items of overall hotel evaluations (a positive opinion about this hotel; booking this hotel is a good idea; this hotel is appealing) and three items of booking intentions (how likely are you to book this hotel; the probability you will book this hotel; I would consider booking this hotel) on a 7-point Likert scale (Casaló et al., 2015). Finally, manipulation checks and demographic questions concluded the survey.
Results
Manipulation Checks
A series of one-way analysis of variances was conducted to check the effectiveness of manipulations. The well-known hotel brand (M = 6.20) had significantly higher brand awareness (1: low, 7: high) than the independent brand (M = 3.54, F1, 428 = 285.63, p < .001). The effect of customer rating was tested with the degree of the rating scores (1: low, 7: high) illustrating a significant difference between a high (M = 6.15) and low (M = 5.09) rating (F1, 428 = 86.81, p < .001). Furthermore, priming stimuli were examined with the valence of the social media posts (1: negative, 7: positive), indicating significant mean differences by conditions (F2, 427 = 136.52, p < .001). A Bonferroni post-hoc test revealed that three conditions (Mnegative = 2.83, Mnone = 5.06, Mpositive = 5.86) were significantly different from each other at p < .05. Therefore, the manipulations were effective.
Hypotheses Testing
To examine the effect of Hotel Brand (HB), Customer Rating (CR), and Priming (PM), a three-way analysis of variance (ANOVA) was conducted on the average of the three items of overall hotel evaluation (α = .920). A three-way ANOVA was utilized to examine the main effects and interactions of all combinations of the independent variables. The results revealed a significant main effect of hotel brand (F1, 418 = 23.52, p < .001, η p 2 = .053), customer rating (F1, 418 = 18.42, p < .001, η p 2 = .042), a two-way HB × CR interaction (F1, 418 = 7.18, p = .008, η p 2 = .017), and a three-way HB × CR × PM interaction (F1, 418 = 3.15, p = .044, η p 2 = .015). The three-way interaction indicated how the effect of hotel brand on hotel evaluations changes depending on the simultaneous presence of customer rating and priming. Table 2 presents the results of the PM × HB × CR interaction.
Three-way Interaction Results (Study 1).
Follow-up analyses were conducted to test the proposed hypotheses by focusing on the impact of the highest-order interaction (See Table 2). Two-way ANOVAs were performed to break down the HB × CR interaction at each level of priming. Significant two-way HB × CR interactions were found at no PM (F1, 136 = 9.80, p = .002, η p 2 = .067) and positive PM (F1, 142 = 4.20, p = .042, η p 2 = .029). To seek the source of the interactions, one-way ANOVAs were further conducted at each level of customer rating. With positive priming, a well-known hotel brand significantly increased the overall evaluation of a low-rated hotel (M = 5.90) compared to an independent hotel brand (M = 4.87), as H1a proposed. However, the effect of the hotel brand was not observed when customer ratings were high, as there was no statistically significant difference between hotel brands (5.72 vs. 5.93), thus supporting H1b. Figure 1 displays the interaction, illustrating the significant impact of hotel brand for low customer rating and the attenuated impact of hotel brand for high customer rating.
The effect of hotel brand and customer ratings with positive priming.
In contrast, the assessment of multiple cues differs in negative priming. Under negative priming, the HB x CR interaction was not observed (F1, 140 = .186, p = .667), while a simple main effect of customer rating was found (F1, 140 = 9.421, p = .011, η p 2 = .045). The overall evaluation of a high-rated hotel (M = 5.85) was significantly greater than a low-rated hotel (M = 5.34) regardless of hotel brand, as H2 proposed. The results indicated that the combined effect of customer rating and hotel brand varies with pre-processed information (priming), and that the role of hotel brand changes depending on the rating score.
Additionally, this study employed PROCESS Model 12 (Hayes, 2018) to examine the impact of multiple cues on customers’ booking intentions, mediated by overall evaluations as an integrated model. The conceptual model is depicted in Figure 2. The results revealed a direct effect of priming (B = .257, SE = .113, t = 2.268, p = .02) and overall evaluations (B = .985, SE = .040, t = 24.625, p < .001) on booking intentions. The analysis also included the indirect effect of hotel brand on booking intentions through overall evaluations, conditional on customer rating and priming (point estimate = −.758, 95% bootstrap CI = −1.424, −.069). The results indicated that hotel brand increases overall evaluations of a hotel, which in turn increases booking intentions when customer rating is low with positive priming.
Integrated model.
Study 2
Design and Stimuli
Study 2 used a 2: customer rating (low vs. high) × 2: scarcity message (no scarcity vs. scarcity) × 3: priming (no priming vs. relevant vs. irrelevant) factorial design. Customer rating refers to an uncontrollable cue, which was operationalized graphically with 4.5 (high) or 3 (low) filled in circles. A different rating format was used since review sites use various graphical and numerical systems. A scarcity message indicates the number of remaining rooms (three rooms left) as an uncertain cue. Priming involves manipulating the information processing mode by leveraging Instagram posts. Relevant priming showed hotel-related information to activate a systematic mode, whereas irrelevant priming used a post unrelated to hotels to operate a heuristic mode. A no-priming condition is included as a control group. Examples of the stimuli are displayed in Appendix B.
Procedure
Random assignment was applied throughout twelve conditions, representing all possible combinations of priming, scarcity, and customer ratings. The participants were first exposed to procedural priming stimuli and subsequently tasked with recalling five words from the social media posts in an open-ended form. Next, the participants were presented with the main stimuli and asked to rate three items of overall hotel evaluations and three items of booking intentions (Casaló et al., 2015). The survey concluded with manipulation checks and demographic questions.
Results
Manipulation Checks
One-way analysis of variances was performed to verify the manipulation of the stimuli. To measure the perceived scarcity of the hotel deal, participants rated the sufficiency of rooms employing a bipolar scale (1: insufficient, 7: sufficient). The result revealed that a hotel with a scarcity message (M = 5.04) has significantly lower sufficiency than one without the message (M = 5.36, F1, 416 = 5.78, p = .017). The degree of rating scores (1: low, 7: high) showed a significant difference between high ratings (M = 5.87) and low ratings (M = 4.44, F1, 416 = 140.51, p < .001). The effect of priming was examined with the degree of relevance to hotel booking (1: not at all relevant, 7: extremely relevant). Relevant priming (M = 5.15) was perceived as more related to the hotel than irrelevant priming (M = 3.73) and no priming (M = 4.47). The result of a Bonferroni post-hoc test indicated that each condition’s relevance was significantly different at p < .05 (F2, 415 = 23.36, p < .001). Therefore, the manipulations for scarcity, customer ratings, and priming stimuli were effective.
Hypotheses Testing
A three-way analysis of variance (ANOVA) was conducted to examine the effects of scarcity (SC), customer ratings (CR), and priming (PM) on overall hotel evaluations (α = .911). There were significant main effects of SC (F1, 406 = 4.07, p < .001, η p 2 = .010), CR (F1, 406 = 40.74, p < .001, η p 2 = .091), a two-way SC × CR interaction (F1, 406 = 6.85, p = .009, η p 2 = .017), and a three-way SC × CR × PM interaction (F1, 406 = 7.39, p < .001, η p 2 = .035). The three-way interaction was broken down to identify the source of the interaction at each level of priming (See Table 3).
Three-way Interaction Results (Study 2).
With irrelevant priming, customer rating significantly influenced overall evaluations (F1, 135 = 18.16, p < .001, η p 2 = .119), revealing higher evaluations for a hotel with high ratings (M = 5.76) compared to low ratings (M = 4.90), regardless of the presence of scarcity. As H3 proposed, customer ratings dominated the assessment of multiple cues, thereby mitigating the effect of a scarcity message when the pre-processed information was irrelevant to hotels. Similarly, when priming was not presented, overall evaluations were significantly influenced by customer rating (F1, 134 = 35.505, p < .001, η p 2 = .129, Mhigh = 5.68, Mlow = 4.66).
With relevant priming, there was a significant SC × CR interaction (F1, 137 = 21.32, p < .001, η p 2 = .135). A follow-up analysis was further conducted to specify the simple effect of scarcity at each level of customer rating. As illustrated in Figure 3, the results indicated overall evaluations of a high-rated hotel were significantly higher with scarcity (M = 5.80) than without scarcity (M = 5.04). On the other hands, scarcity decreased overall evaluations when customer rating was low (M = 4.30) compared to no scarcity (M = 5.53). The results of the three-way interaction support H4, elucidating how the impact of scarcity changes depending on the level of customer rating when relevant priming is present.
The effect of scarcity and customer ratings with relevant priming.
Additional analysis using PROCESS Model 12 (Hayes, 2018) was conducted to demonstrate the influence of multiple cues and mediating effect of overall evaluations on booking intentions (see Figure 2). The results showed a direct effect of overall evaluations (B = .889, SE = .037, t = 23.696, p = .001) and direct interaction effects between customer rating and priming (B = .600, SE = .169, t = 3.555, p = .001) on booking intentions. In this model, there was a conditional indirect effect of scarcity and customer rating for irrelevant priming (point estimate = .492, 95% bootstrap CI = .065, .926) and relevant priming (point estimate = 1.185, 95% bootstrap CI = .509, 1.871). The results revealed that scarcity decreases overall evaluations of a hotel, which affects booking intentions when customer rating is low, regardless of priming.
Discussion
In today’s hotel booking landscape, understanding how customers access multiple cues is crucial. The research findings highlight that the impact of multiple cues varies based on peripheral information, the value of a primary cue, and cue characteristics. This insight sets the stage for a comprehensive exploration of customer perspectives in the hospitality industry. Table 4 presents the results of the hypotheses testing.
Hypothesis Support.
A heuristic process enables customers to make decisions by relying on selective information (Chen & Chaiken, 1999). In study 1, polarized emotions (positive vs. negative) using affective priming are presumed to activate different modes of processing information (Bohner et al., 1995; Tan et al., 2018). With positive priming, postulated to operate a heuristic mode, customer ratings are selected for hotel booking judgments and play a primary role as a salient cue that sends a strong signal based on source credibility (Chakraborty, 2019; Zhao et al., 2015). As H1b confirmed, positive customer ratings alone suffice for evaluating a hotel product without extra information, supporting Spence’s (1978) signaling theory. However, the secondary cue is incorporated into the evaluation process when the primary cue is unsatisfactory (Anderson, 2014), aligned with the proposition of H1a. The finding indicates that the value of hotel brand is included in the evaluation of a hotel only when customer ratings are not sufficient to compete the judgment.
Similarly, Study 2 emphasizes the importance of customer ratings as a primary cue. In study 2, procedural priming was used as peripheral information, presumed to trigger different modes of information processing depending on the relevance to a hotel. An interesting finding is that the mechanism for determining the value of uncertain information is distinct depending on the peripheral information processed earlier. Previous research suggests that online reviews and ratings are considered highly diagnostic, followed by brand and price (Wen et al., 2021). Brand information is integrated into the decision process to support the weak value of the primary cue based on its transparency and highly diagnostic nature (Akdeniz et al., 2013). Conversely, uncertain cues like scarcity messages pose a challenge in determining their value under heuristic process, resulting in the absence of an integrated information mechanism. Consequently, the impact of a scarcity message is disregarded when priming information is irrelevant; thus, the hotel evaluations explicitly rely on customer ratings, as H3 confirmed. Under heuristics, judgments may omit uncertain information because an optimal solution is unavailable (Neth & Gigerenzer, 2015).
According to the dual-processing theory, individuals engaged in systematic processing tend to make rational assessments by computing the values of all available information (Chen & Chaiken, 1999). However, the effect of the brand was not observed with negative priming, which is presumed to activate the systematic process, as H2 proposed. In other words, while all available information is integrated into the rational assessment, the prominent signal of customer ratings outweighs the value of the brand. This phenomenon is frequently observed in hospitality research (de Langhe et al., 2016; Wen et al., 2021). For example, Wen et al. (2021) found brand familiarity to be a low-scope cue to online reviews. The findings suggest that the positive value of a rating alone is considered enough information for hotel booking decisions without needing additional details. Conversely, a low rating is a clear deterrent, discouraging potential purchases. Previous research emphasizes that the impact of negative ratings outweighs that of positive ratings, as individuals tend to avoid potential risks in travel (Casaló et al., 2015). Thus, the impact of a brand in counterbalancing the negative impact of low customer ratings is ineffective. The results indicate that the value of a secondary cue cannot overcome the unfavorable value of a primary cue unless the minimum threshold of this cue is met when the judgment is made using rational assessments.
The findings of Study 2 suggest that peripheral information related to hotels leads to a diverse assessment of uncertain cues depending on the value of the primary cue. In this study, irrelevant priming is intended to induce a systematic mode; thus, hotel-related peripheral information is presumed to facilitate logical thinking processes. A scarcity message is intended to improve the perceived value of a product and encourage customers to purchase by producing a sense of urgency rooted in limited availability (Huang et al., 2020; Lynn, 1991). However, a scarcity message can be perceived as doubtful due to its inherent uncertainty, producing cognitive dissonance and post-purchase regret (Kim et al., 2020; Li et al., 2021). Consequently, people are likely to seek additional information to ensure the assessment of a scarcity message. The findings of H4 indicate that a scarcity message reduces the perceived value of a hotel deal when the customer rating is low, while it increases the value of the high-rated hotel. The findings also align with cue consistency principles, wherein the evaluation of an uncertain cue can be considered with the value of a primary cue (Miyazaki et al., 2005). These outcomes add to Noone’s and Lin’s (2020) findings that price promotions with limited-time and limited quantity scarcity are effective when booking lead time is long.
Theoretical Implications
This research contributes to the existing knowledge of cue assessment by utilizing priming frameworks to alter thinking processes when multiple cues are presented in the context of OTAs. The findings of this research help us understand the decision-making processes within the online hotel booking environment by incorporating various mechanisms.
When salient information such as customer ratings is available, it plays a primary role in determining the use of other available information. The cue assessment mechanism varies based on the value of the primary cue. As a result, it adds to signaling theory as it is applied when the value of the primary cue is favorable (Spence, 1978). In this research, the strong signal from the primary cue, accounting for partial information, represents the overall value of an entity.
This study also contributes to information integration theory (Anderson, 2014) as it is utilized in decision-making by reflecting all information describing the travel product when the signal from a primary cue is weak. In this research, other available information is considered to complete the decision process by overcoming the negative impact of the primary cue. In addition, this study extends on the existing choice overload literature (Chen et al., 2009; Guo & Li, 2022) which focused mainly on choice size, the amount of information presented to customers and the format in which the information was presented. This research adds also to previous research by determining the integrated effect of multiple cues heightened in the competitive OTA environment.
This study enhances our understanding of cognitive processing mechanisms, by applying the dual-processing system to the hospitality industry. It serves as a significant framework for comprehending the diverse responses to available information in the online travel environment. This research highlights the benefits of an experimental method using priming to elicit distinct judgments toward identical information through dual processing. Depending on priming, various cue assessment mechanisms are accepted to understand the value of information, resulting in diverse evaluations. Thus, positive priming and irrelevant priming aim to activate the heuristic process, leading to simpler decisions that rely on a primary cue and only expand the consideration range when necessary. Conversely, negative priming and relevant priming are intended to trigger the systematic process, leading to a logical process aimed at avoiding possible errors and biases arising from the heuristic process (Chen & Chaiken, 1999). For instance, the effect of scarcity is mitigated with positive and irrelevant priming, while it is considered a good deal with the support of good ratings under negative and relevant priming. This research illustrates the benefits of both processes in comprehending the impact of multiple cues within a hotel booking scenario.
This study adds to previous literature by specifying that uncertain cues, such as scarcity messages, are challenging to ascertain their value due to the inherent risks of the deal (Huang et al., 2020; Kim et al., 2020; Lee & Cranage, 2010). Priming intended to activate a heuristic mode leads to the disregard of the significance of such information due to the lack of clear, intuitive benefits. This distinguishes it from a transparent cue such as a brand, which can counterbalance the adverse effects of other cues. Conversely, priming intended to operate in a systematic mode aligns with the value of a primary cue when assessing an uncertain cue. The study’s findings demonstrate the significance of cue consistency (Miyazaki et al., 2005), highlighting the reliance of an uncertain cue on salient information within the context of multiple cues.
Practical Implications
This study provides practical guidelines for effectively utilizing complex information to reduce customers’ choice overload phenomena (Chen et al., 2009; Guo & Li, 2022) and to manage cues co-existing in the OTA environment. While the source of information is a critical factor in promoting products on OTAs (Alvarez & Campo, 2011), the significance of the relationships between multiple cues is often overlooked. Despite the powerful impact of customer ratings and reviews, hotels have limited control over them (Gavilan et al., 2018). Instead, hotels should optimize information that they can control, such as their brands and marketing strategies, to increase bookings. On the other hand, when customer-generated information is unfavorable, hotels can take prompt actions with quick diagnostic information to overcome the negative effect (Wen et al., 2021). For example, hotels can emphasize the value of their brands or offer bundled packages or items with transparent pricing (e.g., Tanford et al., 2012).
In addition, specifying the value of a cue is critical for hotels when they seek to utilize information involving uncertainty, such as a scarcity message and cancellation policy, which can be interpreted as either beneficial or risky (e.g., Kim et al., 2020; Kim et al., 2023). Linking this research with Noone and Lin’s (2020) findings, it is suggested that scarcity messages are most effective early in the booking window with a cue to trigger rational thinking as long as ratings are high. Conversely, when ratings are low, scarcity messages can backfire and decrease hotel evaluations. In a case where ratings suddenly decline or negative reviews appear prominently, hotels should consider retracting the scarcity marketing or implementing a cue to activate the heuristic system. Therefore, managers should carefully review all other available information when building marketing strategies that involve inherent uncertainty. Monitoring a salient cue, such as customer ratings throughout the booking window, is important. To achieve optimal results, it is essential to align marketing messages with other relevant information.
The findings highlight the potential of priming to manipulate responses and show differences in attitudes and behaviors toward complex information describing a hotel product. Although the information from social media is unrelated to booking decisions, hotels can utilize peripheral information to derive desirable outcomes. Therefore, this study pioneered the use of a priming framework for hotel marketing procedures. The mode of processing information can be applied through emotional changes and simple behavioral tasks. For example, hotels can use emotional pictures and texts that evoke positive and/or happy feelings to prime customers into a heuristic mode requiring minimal mental effort to make a decision (e.g., Cain et al., 2024). These steps may be desirable when the goal is to motivate customers to make impromptu buying decisions, such as reacting to last minute discount offers. Furthermore, distracting customers from hotel evaluation situations by using irrelevant information or tasks, such as an advertisement or a short online survey that requires clicks, can activate the heuristic process (e.g., Tanford et al., 2020).
In contrast, as previous research demonstrated, negative emotions and relevant tasks can activate a systematic mode requiring maximum mental efforts to make a decision in the hotel booking process (Tan et al., 2018). For example, hotels may induce negative emotions in customers by highlighting the adverse effects of environmental issues, which may emphasize the hotel’s commitment to addressing such concerns. By using priming techniques, hotels can influence customers’ cognitive processes and enhance customer evaluations in online bookings, possibly resulting in a competitive advantage.
Limitations and Future Research
The research has several limitations that need to be addressed in future studies. While this research focuses the impact of multiple cues, it assessed a limited number of these cues to represent both controllable and uncontrollable cues. As the value and attributes of information can alter customer judgment, a replicated study using other information, such as OTA loyalty rewards and cancellation policy, will expand our understanding of the relationship among multiple cues.
This research utilized a between-subjects design to simulate online booking environment in a laboratory setting. Since multiple cues can affect each other’s value, information on alternative options may also impact decision-making. Future research can address this limitation by exhibiting a list of options as a form of within-subjects design, allowing individuals to compare all options.
Another limitation of the study is the challenge in determining the processing mode in which people are operating. Future research using automatic cognition can expand the applications of the dual-processing system (Miles et al., 2019). Methods to examine automatic cognition, such as the Brief Implicit Association Test and the Affect Misattribution Procedure, will help understand how the two modes of processing operate in the complex decision process of online hotel booking.
In today’s era of information overload in OTAs, understanding customers’ assessment and prioritization of multiple cues as well as their relationships is imperative. This research discovers that the influence of these cues varies depending on information processing methods, the importance placed on primary cues, and the specific characteristics of each cue. Therefore, implementing the actionable insights from this research can significantly improve hotels’ performance in the competitive online hotel booking market.
Footnotes
Appendix
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Harrah College of Hospitality, University of Nevada, Las Vegas.
