Based on data from Airbnb, this study employs the Heuristic-Systematic Model as its theoretical framework. Utilizing facial recognition technology, we extract characteristic variables and construct a multiple linear regression model for empirical testing, aiming to explore the impact of heuristic cues on consumer purchase decisions in sharing-based short-term rental platforms. The main findings are as follows: the attractiveness score of the host’s profile photo, stereotypical impressions, and word-of-mouth reputation all positively influence consumer purchase decisions, while the blurriness of the host’s profile photo negatively affects them. These results provide valuable insights for sharing economy platforms and hosts to optimize information presentation and enhance transaction conversion rates.
Against the backdrop of rapid advancements in internet technology and profound shifts in consumer attitudes, the sharing economy―aimed at enabling efficient and rapid matching of supply and demand―has demonstrated vigorous growth. The concept of the sharing economy was first introduced by American sociology professors Felson & Spaeth (1978) in a paper. In the context of the internet, the sharing economy refers to a model where internet platforms facilitate the more efficient allocation of idle resources from suppliers. Through specific transaction mechanisms, resource providers transfer the usage rights of these idle assets to obtain expected economic benefits, while resource seekers gain access to desired services or products at a higher cost-performance ratio.
As a key component of the sharing economy ecosystem, the shared accommodation sector has reshaped the traditional lodging market through innovative service models, offering consumers diversified, convenient, and cost-effective accommodation solutions. Representative platform enterprises such as Airbnb have leveraged a disintermediated peer-to-peer (P2P) transaction model to integrate global accommodation resources. Their core technological infrastructure serves as a digital bridge enabling direct interactions between supply and demand, allowing users to complete full-process services―including property search, online booking, and electronic payment―seamlessly via smart devices.
The transaction process within the shared accommodation industry is also a process of information exchange between transacting parties. Information serves as one of the key bases for users to make purchase decisions after clarifying their needs (Deaton & Muellbauer, 1980). The Heuristic-Systematic Model (HSM), proposed by Chaiken (1980), is one of the important theoretical frameworks in the field of social cognition and is widely used to analyze the strategies individuals employ when receiving and processing persuasive information. This model categorizes two distinct modes of information processing during individual decision-making: “heuristic processing” and “systematic processing”. Among these, “heuristic processing” is particularly crucial for capturing consumer attention in online platform environments. Although some studies have explored how various information cues on shared short-term rental platforms influence consumer purchase decisions, there remains a lack of clear classification and quantitative characterization of these cues, making it difficult to gain a deeper understanding of the specific mechanisms through which information cues operate in the decision-making process.
Therefore, this study adopts the HSM as the theoretical framework to extract heuristic cues from the information presented on shared short-term rental platforms and systematically investigates their impact on consumer purchase decisions. It aims to address the following research questions: 1) How can heuristic cues on shared short-term rental platforms be categorized? 2) What features can be used to quantify these heuristic cues? 3) How do different heuristic cues vary in their influence on consumer purchase decisions in shared short-term rental platforms?
Heuristic information processing relies on readily accessible informational cues and simple decision rules (Tam & Ho, 2005). In this mode of processing, individuals do not need to expend significant cognitive resources, relying instead on superficial characteristics of the information―such as the credibility of the source and the manner of presentation―and often make decisions based on intuition. In the context of shared short-term rentals, a property’s homepage generally contains two main types of information: one is information self-presented by the host, and the other is evaluative information provided by other parties regarding the property or the host. Among these, visual elements such as images that offer an immediate impression to consumers, as well as processed secondary information such as word-of-mouth ratings and trust verifications, serve as heuristic cues that further influence consumers’ booking behavior.
Current research exhibits the following limitations: First, most studies focus on a single type of cue―such as host photos or ratings―to explore their impact mechanisms, with few adopting the perspective of heuristic information cues to examine how effortlessly processed information characteristics influence purchase decisions. Second, existing studies on host photos in short-term rental platforms primarily use surveys or experiments to investigate the effect of a particular photo attribute on user behavior. To date, no study has conducted comprehensive facial analysis based on a large dataset of objective and authentic host photos from short-term rental platforms to examine how host photo characteristics and facial features affect user trust. Therefore, this study will focus on host profile photos, property word-of-mouth information, and host reputation data on sharing-based short-term rental platforms to investigate the impact of heuristic cues on consumer purchase decisions.
In the field of consumer behavior, physical appearance plays a pivotal role. For instance, the appearance of individuals in advertisements exerts a positive influence on consumers’ attitudes toward both the advertisement and the product (Li et al., 2016). The physical attractiveness of service personnel can affect customers’ perceptions of their credibility, professional image, and likability, thereby influencing perceived service quality (Li et al., 2016). Similarly, salespeople with higher levels of physical attractiveness are more likely to gain customer trust (Ahearne, Gruen, &Jarvis, 1999). Cyr et al. (2007) suggested that seller photos in online environments enhance consumer trust by increasing perceived credibility. Ying & Rao (2019) conducted research on how consumers’ trust perception based on host photos influences their choice preferences, confirming the significant impact of attractiveness features. Therefore, this study proposes the following hypothesis:
H1: The attractiveness score of the host’s photo significantly positively influences consumer purchase decisions on shared short-term rental platforms.
As a non-verbal form of communication, smiling can lead others to evaluate an individual more leniently and perceive them as trustworthy (LaFrance & Hecht, 1995) or kind and friendly (Hess et al., 2002). Scharlemann et al. (2001) found that individuals who smile are more likely to gain the trust of other participants. On shared short-term rental platforms, consumers typically cannot interact with hosts face-to-face. Therefore, the smile intensity in a host’s profile photo becomes an important cue for consumers to assess the host’s attitude and warmth. In the context of online short-term rentals, Fagerstrom et al. (2017) employed an experimental simulation method and found that hosts who smile in their photos tend to receive more bookings. Accordingly, this study proposes the following hypothesis.
H2: The smile intensity in the host’s profile photo has a significantly positive influence on consumer purchase decisions on shared short-term rental platforms.
Stereotype refers to a generalized and relatively stable perception or cognition that people hold toward a certain category of individuals or objects (Macrae, Stangor, & Hewstone, 1996; Fiske, Rosenblum, & Travis, 2009). The Stereotype Content Model (SCM) reveals that the content of stereotypes focuses on evaluations along two primary dimensions: competence and warmth, and many evaluations of individuals or groups are based on this framework. Fiske (2018) conducted cross-cultural studies on individuals and groups across different countries, and their results demonstrated the universal applicability of the SCM across cultural contexts.
On Airbnb, the personal photos provided by hosts often reveal information such as gender and age. Since people generally hold stereotypes toward unfamiliar individuals or groups, the information users infer from host photos―such as gender and age―can influence their perception of the host’s trustworthiness, thereby affecting their booking decisions. According to the SCM: Older adults are stereotypically perceived as high in warmth but low in competence; University students (young people) are stereotypically perceived as high in both warmth and competence; Men are often stereotyped as low in warmth but high in competence; Women are often stereotyped as high in both warmth and competence. Therefore, this study proposes the following hypotheses:
H3: Stereotypes significantly influence consumer purchase decisions on shared short-term rental platforms.
H3a: Consumers are more inclined to book properties offered by female hosts.
H3b: Consumers are more inclined to book properties offered by younger hosts.
Within trust theory, trust is regarded as a form of dependency relationship, the establishment of which relies on the transparency and reliability of information. Multiple studies have indicated that image quality is an important factor affecting perceived trust. High-quality images provide richer and more authentic visual information, which helps enhance the sense of trust. Conversely, blurry or low-quality images may be perceived as indicative of missing or concealed information, thereby evoking distrust. Ambiguous risk information can influence decision-making behavior; factors such as information content, trust in the source, and aversion to ambiguity all play roles in the decision-making process (Melkonyan, 2011). Some scholars have experimentally quantified the relationship between the degree of photo blurriness and trust level (Hao, 2020). The results show that as the level of blurriness in photos increases, consumer trust gradually decreases. Therefore, this study proposes the following hypothesis:
H4: The blurriness of the host’s personal photo has a significantly negative influence on consumer purchase decisions.
Numerous studies have indicated that reputation systems on e-commerce platforms can effectively foster the formation of consumer trust. For example, Resnick et al. (2000) investigated e-commerce reputation on eBay and found that sellers with better reputations received more orders, as consumers tend to trust those with favorable reputations. Some studies have pointed out that the “Super host” badge can significantly enhance consumers’ trust in hosts and subsequently influence their booking decisions (Li et al., 2023). This is because the “Super host badge serves as an official endorsement by the platform of the host’s service quality, providing consumers with a reliable reference. Ratings and acceptance rates are important indicators of host service quality. High ratings typically indicate that the host’s service has been recognized by the majority of consumers, while a high acceptance rate reflects strong service efficiency and awareness. Together, these indicators form the host’s word-of-mouth reputation system, profoundly influencing consumers’ trust perceptions and purchase decisions. Therefore, this study proposes the following hypotheses:
H5: The host’s word-of-mouth reputation has a significantly positive influence on consumer decisions on shared short-term rental platforms.
H5a: The host’s attainment of a “Super host” badge has a significantly positive influence on consumer purchase decisions.
H5b: The host’s booking acceptance rate has a significantly positive influence on consumer purchase decisions.
H5c: The property rating has a significantly positive influence on consumer purchase decisions.
Based on the above hypotheses, this study treats host profile photos and word-of-mouth reputation cues as heuristic information cues on shared short-term rental platforms. These are operationalized through the following measures: for profile photo cues―attractiveness score, smile intensity, stereotypes (age and gender), and blurriness; for reputation cues―“Super host” status, booking acceptance rate, and property rating. The research aims to investigate the impact of these heuristic cues on consumer purchase decisions in shared short-term rental platforms.
Regarding control variables, drawing on relevant studies in the field of shared short-term rentals (Wang & Chen, 2021; Wang et al.,2024; Xu & Liang, 2022), variables that may influence the results are incorporated, including listing price, number of listings owned by the host, host response time, host response rate, and instant booking policy. Therefore, the research model of this chapter is illustrated in
This study uses the Airbnb short-term rental platform as a case study, with data sourced from https://insideairbnb.com. The dataset comprises 75,984 property listings from March 23, 2023, to March 24, 2024, covering nine countries including the United States, France, and Germany. The provided information includes property details such as listing descriptions, housing types, and review scores, as well as host information including profile photos, “Super host” status, and response rates. Records with incomplete or meaningless attributes (e.g., garbled text) were removed. Listings with missing or broken links to the host’s profile photo were excluded. Entries where the host’s profile photo did not meet the input requirements of the Face++ facial analysis API were also discarded. Finally, records that failed facial recognition by the Face++ API were eliminated. After preprocessing, a total of 8624 valid records were retained.
This study utilizes the Face++ platform to conduct in-depth facial analysis of host profile photos. As the output of the facial recognition process is a JSON file containing face_token, it is necessary to parse this JSON file to extract facial attributes such as the host’s age, gender, smile intensity, face blurriness, and attractiveness score.